id
stringlengths 16
65
| title
stringlengths 19
278
| text
stringlengths 894
136k
| keyphrases
listlengths 1
34
| prmu
listlengths 1
34
|
|---|---|---|---|---|
springeropen-4-1-2200676
|
Transcriptome analysis of the response to chronic constant hypoxia in zebrafish hearts
|
Insufficient blood supply during acute infarction and chronic ischemia leads to tissue hypoxia which can significantly alter gene expression patterns in the heart. In contrast to most mammals, some teleost fishes are able to adapt to extremely low oxygen levels. We describe here that chronic constant hypoxia (CCH) leads to a smaller ventricular outflow tract, reduced lacunae within the central ventricular cavity and around the trabeculae and an increase in the number of cardiac myocyte nuclei per area in the hearts of two teleost species, zebrafish (Danio rerio) and cichlids (Haplochromis piceatus). In order to identify the molecular basis for the adaptations to CCH, we profiled the gene expression changes in the hearts of adult zebrafish. We have analyzed over 15,000 different transcripts and found 376 differentially regulated genes, of which 260 genes showed increased and 116 genes decreased expression levels. Two notch receptors (notch-2 and notch-3) as well as regulatory genes linked to cell proliferation were transcriptionally upregulated in hypoxic hearts. We observed a simultaneous increase in expression of IGF-2 and IGFbp1 and upregulation of several genes important for the protection against reactive oxygen species (ROS). We have identified here many novel genes involved in the response to CCH in the heart, which may have potential clinical implications in the future.
Introduction
Low oxygen levels (hypoxia) play important roles in clinical conditions such as stroke and heart failure. Insufficient blood supply leads to tissue hypoxia in the heart during acute infarction and chronic ischemia (Semenza 2001).
Effective protection of the heart against ischemia/reperfusion injury is one of the most important goals of experimental and clinical research in cardiology. Besides ischemic preconditioning as a powerful temporal protective phenomenon, adaptation to chronic hypoxia also increases cardiac tolerance to all major deleterious consequences of acute oxygen deprivation such as myocardial infarction, contractile dysfunction and ventricular arrhythmias (Kolar and Ostadal 2003). Although many factors have been proposed to play potential roles, the detailed mechanism of this long-term protection remains poorly understood. Some of the molecular mechanisms of cardiac protection by adaptation to chronic hypoxia and chronic high-altitude hypoxia have recently been reviewed (Kolar and Ostadal 2003; Ostadal and Kolar 2007). KATP channels, PKCδ as well as the different MAPK pathways were shown to be involved in the mechanism of increased tolerance of chronically hypoxic hearts and further the controversial role of ROS in hypoxia tolerance is discussed (Kolar and Ostadal 2003). Furthermore, a recent study has profiled the gene expression changes induced by chronic constant hypoxia (CCH) and chronic intermittent hypoxia (CIH) in newborn mice (Fan et al. 2005).
In contrast to most mammals (with the exception of some marine mammals), some teleosts, have developed the ability to withstand extreme chronic hypoxia (Stecyk et al. 2004). It is well assumed that these vertebrate species possess unique adaptations in order to survive short and long term oxygen deprivation. However, the molecular basis of these adaptations in fish has so far not been extensively investigated.
Several studies have profiled gene expression changes in teleosts exposed to hypoxia. Gracey et al. showed in adults of the euryoxic gobiid fish Gillichthys mirabilis (the long-jawed mudsucker), that 5 days of hypoxia induced a complex transcriptional response, including a shut down of energy requiring pathways like protein synthesis and locomotion, and an induction of genes needed for anaerobic ATP production in different tissues (Gracey et al. 2001). Recently, we described phenotypic and behavioral adaptations to long-term hypoxia and described the gene expression changes induced by chronic constant hypoxia in the gills of adult zebrafish (van der Meer et al. 2005).
Ton et al. identified global gene expression changes in zebrafish embryos. Zebrafish embryos at 48 h post fertilization were exposed to water with 5% oxygen content for 24 h. The authors identified 138 genes responsive to short-term hypoxia and could also show that transcriptional changes indicated metabolic depression, a switch from aerobic to anaerobic metabolism and energy conservation (Ton et al. 2003).
In this study, we have identified CCH-induced gene expression changes in the zebrafish heart by looking at over half of the zebrafish genome. We have compared several of these novel changes described in other species and tissues. We have here identified the heart-specific molecular adaptations to CCH. Future functional experiments are warranted to determine whether some of the findings can be used to better adapt mammalian hearts to CCH.
Material and methods
Animal handling
Adult wild-type zebrafish (Danio rerio) around 3 month of age, were obtained from a local pet store. Cichlids (Haplochromis piceatus) have been collected in the Mwanza Gulf of Lake Victoria in 1984 and were bred in our laboratory for about 20 generations. All animals were handled in compliance with animal care regulations. Our animal protocols were approved by the review board of Leiden University in accordance with the requirements of the Dutch government. Zebrafish were kept at 25°C in aquaria with day/night light cycles (12 h dark vs. 12 h light). Cichlids were kept at 25°C with the same day/night light cycles.
Hypoxia treatment
For gradual hypoxia treatment, oxygen levels were gradually decreased in 4 days from 100% air saturated water to 40% (day 1), 30% (day 2), 20% (day 3) and the final 10% air saturation (day 4). After day 4, the fish were kept for an additional 21 days at 10% air saturation. (at 100% air saturation and 28°C the O2 concentration is 8 mg/l and pO2 is 15 Torr). In parallel, a control group was kept at 100% air saturated water. Both groups were kept in identical aquaria of 100 l. The oxygen level on the hypoxia group was kept constant by a controller (Applikon Biotechnology, The Netherlands) connected to an O2-electrode and solenoid valve in line with an air diffuser. The oxygen level in the tank was kept constant by adding oxygen via the diffuser and thereby compensating the oxygen consumption of the fish. In case of immediate hypoxia exposure, tanks were pre-equilibrated to the respective pO2 concentration and fish were then directly set in the equilibrated aquaria.
Perfusion of cichlid hearts
In order to minimize blood clotting, a perfusion protocol was developed in which the whole blood volume of clinically dead animals was initially replaced with isotonic buffer and with a fixative solution secondarily.
Heart dissection
The fish were killed with an overdose of anesthetic (MS-222; Tricaine Methanesulfonate from Argent Chemical Laboratories, USA). Hearts were dissected from the fish immediately after the anesthetic worked. For RNA preparation the hearts were immediately shock-freezed in liquid nitrogen. For histology and microscopy, the hearts were left intact and fixed immediately in Karnovsky fixative (4% paraformaldehyde (PFA) and 2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2) for 4 h at 4°C. After three washes in 0.1 M phosphate buffer, pH 7.2, they were transferred to 70% ethanol.
Histology of adult fish hearts, statistical analysis and scanning electron microscopy
Hearts from zebrafish and cichlids raised either under normoxic or hypoxic conditions, were dissected from the fish and fixed for 24 h in 4% PFA in PBS. After fixation, hearts were washed with 1 × PBS, cut in halfs (sagittal through the midline) and then dehydrated through ethanol series starting at 70% ethanol, followed by 80, 90, 96 and 100% ethanol, each step was done once for 1 h except the last one which was done twice. After dehydration the hearts were embedded in increasing gradients of Historesin (Technovit 7100, Heraus Kulzer, Germany) (25, 50, 75 and 100% Historesin in ethanol, for 2 h at room temperature; 100% Historesin, 24 h at 4°C). Afterwards the plastic with the hearts was polymerized at 40°C (overnight). A 5 μm sagittal sections were made using the Ultramicrotome (Reichter-Jung) and a glass knife. Approximately 500 sections on each side of the midline were visually analyzed per heart (1,000 sections in total per heart). Sections were left to dry and later stained with hematoxylin–eosin staining. Pictures were taken with Axioplan 2 imaging, (Carl Zeiss, Jena). Azan staining of histological sections were done as follows. Paraffin sections of the hearts were prepared as described before (van der Meer et al. 2006). Sections were incubated with Azokarmine solution, for 30 min at 60°C. Afterwards they were washed in water and differentiated in 0.2% Anilin alcohol. They were then rinsed in 1% acetic acid in 95% alcohol, followed by 45 min incubation period, in 5% phosphotungstic acid. After that, sections were rinsed in distilled water and incubated for 45 min with aniline blue. Finally, they were rinsed with distilled water, differentiated and dehydrated in 95% alcohol followed by absolute, cleared in xylene and mounted in Entallan. The protocols used here for SEM had been described before (van der Meer et al. 2005). The statistical analysis of cardiac myocyte nuclei per section was done using Statistica by performing an independent t test. A P value of less than 0.05 was considered significant.
Immunohistochemistry and statistical analysis
Zebrafish, raised either under normoxic or hypoxic conditions, were killed with an overdose of anesthetic MS-222 and frozen in liquid nitrogen. Subsequently, transversal cross-sections (10 μm thick) of the body, were cut using a cryostat at −20°C and mounted on glass slides coated with Vectabond (Vector Laboratories, Burlingame, USA). Sections were fixed in 4% formaldehyde in Tris-buffered saline (TBS; 50 mM Tris and 150 mM NaCl, pH 7.5) for 10 min and subsequently washed in TBS with 0.05% Tween-20 (TBST) (Sigma-Aldrich, Zwijndrecht, The Netherlands). Subsequently, sections were incubated for 10 min with 10% normal swine serum (Vector laboratories) in TBST after which sections were incubated for 24 h at 4°C with anti-phospho Akt polyclonal antibody (Santa-Cruz Biotechnology, USA) diluted 1:50 in TBST. After incubation with primary antibody, the slides were washed in TBST and subsequently placed in 0.25% (v/v) acetic anhydride in 0.1 M triethanolamine buffer (pH 8) for 10 min followed by rinsing in TBST. After this, sections were incubated for 60 min at 20°C with secondary anti-rabbit immunoglobulin G (IgG) antibody covalently coupled to alkaline phosphatase (Vector Laboaratories) diluted 1:100 and washed in TBST. After this, sections were incubated for 5 min with alkaline phosphatase (AP) buffer (0.1 M NaCl, 0.1 M Tris, 50 mM MgCl2 and 0.1% Tween-20, pH 9.5) followed by incubation with BM Purple AP substrate (Roche Applied Sciences, Almere, The Netherlands) for 30, 45 or 60 min which was followed by rinsing in TBST. All sections were mounted in glycerine-gelatin and stored at 4°C in the dark until staining intensity was measured. The absorbance values of the BM Purple in the sections were determined using a Leica DMRB microscope (Wetzlar, Germany) fitted with calibrated gray filters using different interference filters. Absorbances for BM Purple were determined at 550 nm. Images were recorded with a ×20 objective and a Sony XC-77CE camera (Towada, Japan) connected to a LG-3 frame grabber (Scion; Frederick, MD) in an Apple Power Macintosh computer. Recorded images were analysed with the public domain program NIH-Image V1.61 (US National Institutes of Health, available at http://rsb.info.nih.gov/nih-image/). Gray values were converted to absorbance values per pixel using the gray filters and a third-degree polynomial fit in the calibrate option of NIH-image programme. Morphometry was calibrated using a slide micrometer and the set scale option in NIH-image, taking the pixel-aspect ration into account. An independent t test was used to test for differences in phospho-Akt levels in cardiac myocytes of normoxic and hypoxic fish. A P value of less than 0.05 was considered significant. Values are means ± S.E.M.
RNA preparation and biological sampling
After dissection hearts were homogenized in a Dounce homogenizer using 1 ml Trizol solution (GibcoBrl, Life technologies). The whole heart was used and for each biological sample hearts were pooled from five different animals. After Trizol extraction, total RNA was further purified using RNAeasy columns (Qiagen). RNA samples were analyzed for quality control by Lab-on-a-chip analysis (Agilent) and on agarose gels. For the array experiment five arrays were done for normoxic and 5 arrays for the hypoxic condition. Biological samples (BS) came from two independent experiments and one technical replicate (TR) was included (2BS + 2BS + TR for normoxia and 2BS + 2BS + TR for hypoxia). As mentioned above for each BS, hearts from five different animals were pooled.
Microarray analysis
The Affymetrix GeneChip® Zebrafish Genome Arrays containing 15,509 Danio rerio gene transcripts were used. Probe sets on the arrays were designed with 16 oligonucleotide pairs to detect each transcript and procedures were in full support of MIAME standards. Labeling and microarray hybridization were performed by ServiceXS (Leiden, The Netherlands), including prior a standard round of RNA amplification according to standard Affymetrix protocols. The criteria used for differential expression were greater or equal than 2-fold induced or reduced and P ≤ 0.02. Data analysis was done using Rosetta Resolver. All expression data was submitted to the NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo). The series entry number is GSE4989 and the following 10 accession numbers were assigned: GSM112796 and GSM112798-GSM112806. A complete list of differential regulated genes is shown in Supplemental Table 1.
Table 1Statistical analysis of histological sections (5 μm) from zebrafish (Danio rerio) and cichlid (Haplochromis piceatus) heartsNumber of cardiomyocyte nuclei per sectionZebrafish (nuclei per 900 μm2)Cichlid (nuclei per 10,000 μm2)NormoxiaHypoxiaNormoxiaHypoxiaMean9.8113.6714.724SD0.300.390.500.68P value9.9 × 10−177.6 × 10−12Midline sections of zebrafish and cichlids raised under normoxic and hypoxic conditions were chosen and subdivided in smaller areas. In the case of D. rerio, each subarea of the section was 900 μm2 whereas in the case of the H. piceatus it was 10,000 μm2. Then, subareas were randomly picked and the amount of cardiac myocyte nuclei were counted. The vast majority of cardiac myocytes were mononucleated cells. Hundred subareas per section were counted per specimen. Three sections per heart were visually analyzed (with a magnification of 20×). In total 6 different zebrafish hearts (from three independent experiments) were investigated for hypoxic conditions and 6 for normoxic condition. In total 1,800 subareas were quantified per heart and condition used. The same was done for the cichlids (for which bigger subareas were counted). A two-tailed t test was applied and a significant difference (P < 0.001) in the amount of nuclei present in the hearts of normoxia versus hypoxia groups was observed for both species
Gene ontology analysis
The Gene ontology analysis was performed using eGOn and the database of the Norwegian Microarray Consortium (http://www.genetools.microarray.ntnu.no/egon/) and the most recent updated Unigene numbers for zebrafish (May 2007). 13414 Unigene numbers were annotated to the Affymetrix gene chip used. eGOn Enrichment analysis was used for the entire Affymetrix set versus the identified 376 gene set based on the new Unigene numbers. A master target test was performed (using Fishers exact test) and the cut off for positives was set to P < 0.05. The available three categories, molecular function, biological process and cellular component were all tested and are shown in this dataset (Supplemental Data 1). The fold enrichment was based on the percentage found within the differentially expressed genes divided by the percentage on the total chip. A list of all Unigene numbers found in the different categories is given in Supplemental Data 2.
Real-time quantitative RT-PCR
For verification of gene expression data, we used quantitative real-time RT-PCR. Biological RNA samples were obtained from a third independent experiment and as mentioned above for the microarrays, for each BS hearts from five different animals were pooled. The Roche Master SYBR Green kit was used for the RT-PCR reactions. The annealing and synthesis temperature was 55°C alternating with 96°C for 45 cycles. Dissociation protocols were used to measure melting curves and control for unspecific signals from the primers. A measure of 100 ng of total RNA was used per reaction. A standard curve for β-actin using 1, 5, 10, 100 and 500 ng of total RNA was used for normalization. Samples were measured in the Roche LightCycler. The primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) was used to design primers for short amplicons between 50 and 100 bases. The primers used are shown in Supplemental Table 2.
Results
We describe here the survival rates of adult zebrafish upon immediate and gradual exposure to different pO2 concentrations. Immediate exposure to pO2 of 15 Torr (O2 concentration of 0.8 mg/l or 10% air saturated water) was lethal for the adult zebrafish and none survived for longer than 72 h (Fig. 1). If lowered gradually (see experimental procedures for regimen in Sect. “Hypoxia treatment”) zebrafish were able to grow and gain weight at O2 levels of 10% air saturation. The zebrafish were able to survive for longer than 6 months and no mortality was observed, demonstrating that they can well adapt to these conditions (data not shown). For the experiments with zebrafish and cichlids, the pO2 concentration was gradually lowered to 10% air saturation (0.8 mg/l) (see “Material and methods”) and fishes were kept under 10% air saturation for 3 weeks. Control groups were always kept in parallel under normoxic conditions.
Fig. 1Survival of zebrafish embryos after immediate exposure to hypoxia. Zebrafish were immediately exposed to hypoxia (15 Torr; 0.8 mg/ml; 10% air saturated water). Results are derived from three independent experiments with n = 30 in each experiment (adding to a total of 90 fish tested). After 24, 48 and 72 h, dead and alive fish were counted. Shown here is the percentage of dead fish at the respective time points and the standard deviation. At 24 h 53.3%(±4.4) of fish were dead, at 48 hours 85.3% (±5.1) and at 72 hours 100% (±0). Control groups (n = 30), which were in parallel exposed to normoxic conditions showed no mortality (not shown). None of the zebrafish survived the immediate exposure to an O2 concentration of 0.8 mg/l (10% air saturation) over the 3-day period. In contrast, fishes gradually exposed to hypoxia showed no induced mortality for even the 25 day time periods used in our experiments
Phenotypic changes in the heart of adult teleosts under chronic constant hypoxia (CCH)
In comparison to the normoxic control groups, we observed a significantly smaller ventricular outflow tract and reduced lacunae within the central ventricular cavity and reduced lacunae around the trabeculae in midline sections of hearts of both zebrafish (Danio rerio) (Fig. 2a) and cichlids (Haplochromis piceatus) (Fig. 2b) exposed to CCH. In addition to the midline sections, none of the sections investigated from hypoxia treated fishes showed a ventricular outflow tract in comparable size to the normoxic controls (data not shown). In addition to the midline sections, none of the lateral sections (to both sides of the midline) investigated from hypoxia treated fishes showed a ventricular outflow tract as well as lacunae in a comparable size to the normoxic controls (data not shown). The larger cichlid hearts were also perfused and midline sectioned and showed similar results with a smaller ventricular outflow tract and reduced lacunae (Fig. 2b-H). This might represent ventricular hypertrophy or hyperplasia in both walls and trabeculae, which could lead to the observed cavity obliteration in these sections. We quantified the number of cardiac myocyte nuclei per area in the midline sections of both zebrafish and cichlids under normoxic, as well as hypoxic conditions. A significant difference in both species was observed and showed that hypoxia led to a 1.4- and 1.6-fold increase in the number of cardiac myocyte nuclei per area in zebrafish and cichlid hearts, respectively (Table 1). Cardiac myocyte nuclei in sections were clearly distinguishable from nuclei of other cells like erythrocytes and fibroblasts and only centralized nuclei in cardiac myocytes (which are more elongated than nuclei from erythrocytes) were counted. Furthermore, scanning electron microscopy (SEM) was used to confirm these findings in the smaller zebrafish hearts (Fig. 3). Future research is warranted in order to assess further how the cardiac myocytes adapt to CCH in the fish heart.
Fig. 2Histological changes of zebrafish and cichlid hearts after exposure to chronic constant hypoxia. a shows zebrafish hearts that were dissected, sectioned and stained with a hematoxylin–eosin staining with A, B and E representing normoxic and C, D and F hypoxic conditions. Cell nuclei are seen in dark (dark blue in online version) and cell cytoplasm in light (pink in online version). Pictures A, C, E and F have the same magnification (10×). Images B and D represent a 20× magnification of cardiac muscle (D). Abbreviations used are: a atrium; v ventricle; vo ventricular outflow tract and ca conus arteriosus. b (A–F) corresponds to sections of cichlid hearts, which were treated the same way as the ones above from zebrafish and G and H show cichlid hearts which have been perfused prior to dissection and were stained with either hematoxylin–eosin (A–F) or Azan blue (G, H). In A, B, C and G pictures of normoxic conditions are shown and D, E, F and H represent the corresponding hypoxic conditions. Similar results for both the zebrafish and the cichlid hearts were observed in three independent experimentsFig. 3Morphological changes of zebrafish and cichlid hearts after exposure to chronic constant hypoxia. Scanning electron microscopy pictures of hearts from normoxia control zebrafish (a, c and d) and hypoxia-treated zebrafish (b, e and f). The bigger images show half of a heart ventricle, sectioned longitudinally. The smaller images represent a higher magnification of cardiac muscle. The scale is given at the bottom of each picture. Abbreviations used are: v ventricle and vo ventricular outflow tract
Gene expression changes in the heart of adult zebrafish under chronic constant hypoxia (CCH)
In this study, we used microarrays for the transcriptional profiling of up and downregulated genes in response to hypoxia in the zebrafish heart. We identified 376 genes that were differentially expressed under hypoxic conditions, out of which 116 genes showed a decrease in gene expression (30.9%) in comparison to 260 genes which showed increased expression levels (69.1%).
All 376 differentially expressed genes, including the ones with oligo sequences which could not be annotated so far (referred therein either as transcribed locus or zebrafish clone) are shown in the complete file (Supplemental Table 1). Functional groups are color coded and if possible, gene functions are briefly summarized and OMIM links given.
Functional groups of differentially expressed genes in the heart
We have clustered the differential expressed genes according to known functions (Table 2). Genes can have more than one particular function assigned, so some genes can appear in more than one group. In addition, a gene ontology analysis was performed using eGOn to determine gene enrichment and overrepresentation in the three categories of molecular function, biological processes and cellular components (Supplemental Data 1 and 2).
Table 2Functional groups of differentially expressed genesUniGeneGeneBankFoldGene nameUpregulated genesAngiogenesisDr.11575NM_1732442.3T-cell acute lymphocytic leukemia 1; TAL1Dr.845BG7290132.8Fibrinogen alpha/alpha-E chain Dr.4907BC0458684.2Fibrinogen, gamma polypeptideApoptosisDr.15862AF4939872.1BCL2adenovirus E1b 19 interacting protein3DrAffx.1.39AF3027892.3Death receptorDr.20106AI7222772.8Apoptosis inhibitor 5Dr.4039BQ48068821.8BAX inhibitor 1Cell adhesionDr.6007NM_1318202.9Cadherin 1Dr.25140BQ2628023.3Tumor-associated calcium signal transducer Dr.4409BC0490364.4CD9 antigenDr.25140BQ2628027.7Tumor-associated calcium signal transducer glycoproteinDevelopmentDr.11575NM_17322.3T-cell acute lymphocyte leukemia 1 (tal 1)Dr.23348BE2016532.6Bone morphogenetic protein 3b; (bmp 3)Dr25405BC0139232.8SOX2 SRY-box 2Dr.6382AW1650532.9Hedgehog-interacting proteinDr.10879U976693.0NOTCH3 Notch homolog 3 (Drosophila)Dr.15055BC0501723.6Chemokine receptor 4aDr.6787BI5334265.5NoelinDr.16720BI9808476.3notch 2Disease relatedDr.6349AW1166682.4Eparin cofactor IIDr.21064BC0460754.54hydroxyphenylpyruvate dioxygenase HPDDr.12584NM_1312115.4Gata binding protein 3 (GATA3)Dr.3530AI49754579.3Prion protein (prp) geneGrowth regulationDr.8145NM_131432.2Insulin like growth factor 2 (IGF-2)Dr.7609BI4758572.4Prolactin receptorDr.8285NM_131362.4Mad homolog 2 Dr.8947CD5947352.5Spint 2Dr.822BM1841272.5Spint 2Dr.3563CD0144882.8Tetraspan membrane protein IL-TMPDr.8587NM_173282.9Insulin-like growth factor binding protein 1 Dr.2596BM3429013.2Cyclin IDr.8587AL9108223.4Insulin-like growth factor binding protein 1Dr.26458BC0532065.6m-rasHeart relatedDr.15088BM1817494.3Lectin galactoside-binding soluble 1; (galectin10-like 3)Dr.4867AI4968405.5HaptoglobinDr.3585AY0497316.6Angiotensinogen Dr.2452BQ2848484.3Complement component C9 Dr.18453BC0445254.8Uridine phosphorylaseDr.3025BG7382042.7Alpha-2 macro-globulin; A2MGInflammationDr.12491BI6721682.1Complement C4–2Dr.4047NM_1316272.3Small inducible cytokine A (scyba)Dr.5053NM_1317232.3Kruppel-like factor 4 Dr25207X064652.5Complement component 8, gamma polypeptideDr.6845K027652.9C3 complement component 3Dr.5741BU7104823.2Complement component b fbDr.7722BI8784143.5Complement C3-H1Dr.22244AW0197813.6Complement C1sDr.22133AW0767683.7c1rs-A and clrs-BDr.5528AI4972124.2Complement component C9Dr.2452BQ2848484.3Complement component C9 Dr.1730AI7215284.8cfI-B complement control protein factor I-B Dr.2452BM7780025.8Complement component C9Dr.20291BM0363896.5Complement C3-SDr.190NM_1313387.9Complement component factor B Dr.1192AB0716012.Lipocalin-type prostaglandin D synthase-like proteinMetabolismDr.9492BI8822442.0Sulfide dehydrogenase likeDr.15574BM5714672.1Hypoxanthine _hosphor-ribosyltransferase 1Dr.3332AI9430532.2Angiopoietin 5 Dr.16130CD0148982.3Alcohol dehydrogenase 8 bDr.3959BI430012.55′-nucleotidaseDr.22205AW0194772.6OxidoreductaseDr.1699AI6672492.7Pyruvate kinaseDr.5504BI8795503.2Cystathionine-beta-synthaseDr.1202AJ2454913.9Apolipoprotein A-IDr.4111BC0532674.2Fructose-1,6-bisphosphatase 1Dr.18834AW0193214.2Urate oxidaseDr.19224BC0501674.3Aldolase bDr.4938NM_1316454.4Fatty acid desaturase 2Dr.12654BC04690114.8ELOVL family member 6,Dr.5488AI54559317.3Apolipoprotein A-IV Muscle relatedDr.3585AY0497316.6AngiotensinogenDr.2452BQ2848484.3Complement component C9ProteolysisDr.20934AF5419522.6Trypsin precursorDr.3025BG7382042.7Alpha-2-macroglobulin Dr.22139AW0189653.0Alpha-1-antitrypsinDr.25331AI6580724.1Alpha-2-macroglobulin-2Dr.12602NM_1391804.3LysozymeDr.1605BM1853884.4Protease inhibitor 1Dr.17459CD5868374.8Inter-alpha-trypsin inhibitor heavy chain H3 Dr.3073AI5850305.0Serine protease inhibitor alpha 1Dr.26371AI6676765.4ProstasinDr.3025BM5304275.6Alpha-2-macroglobulin-1Dr.3025BM3168676.5Alpha-2-macroglobulin-2Dr.2960X670553.5ITIH3 pre-alpha (globulin) inhibitor, H3 polypeptideDr.25379BI3267836.7Alpha-2-macroglobulinDr.4797AI9595347.826–29 kD-Proteinase proteinROS protectionDr.20068NM_1310752.1Metallothionein (mt)Dr.5399AI9577652.3Biliverdin I Beta ReductaseDr.14058CD0153513.5Glutathione S-transferase theta 1Dr.25160BC0494755.9Metallothionein 2Dr.3613BC0480376.0CerulopasminDr.4905.1BC0454646.5Uncoupling protein 4Signal transductionDr.9852AW8264252.1CAM kinase 1Dr.8591BM1865082.9Rho guanine nucleotide exchange factor 10Dr.6236AW1159733.1Rho guanine nucleotide exchange factor 5Dr.1267BC0511573.4Phospholipase C deltaDr.22129BC0166683.9RRAGC Rag C (Ras-related GTP binding C)Dr.7255AW1164794.4Protein phosphatase 1,Dr.4453BC0444215.8Phosphoprotein phosphataseTranslationDr.13234BM0364712.0Ribonuclease P Dr.382CB3638302.1NucleolinDr.6949AW0781162.1RNA 3′-terminal phosphate cyclase-like protein (HSPC338)Dr.13563BI8907292.3Methionyl aminopeptidase 2Dr.26328AL7236962.3Eukaryotic translation initiation factor 4A,Dr.17693BQ0782853.740 S ribosomal protein S6Dr.20270BI6740505.9Ribosomal protein L12 Dr.25224CD01533020.4Ribosomal protein L12 Dr.12439BM53384817.5Heterogeneous nuclear ribonucleoprotein K Dr.12439BM53384824.2Heterogeneous nuclear ribonucleoprotein KDr.14821BM07171433.8Heterogeneous nuclear ribonucleoprotein K Dr.12502BQ28468640.7Heterogeneous nuclear ribonucleoprotein K Dr.12439. BM53443240.9Heterogeneous nuclear ribonucleoprotein KDr.12439BQ61693045.4Heterogeneous nuclear ribonucleoprotein KTransportDr.1084BQ1097723.0Clathrin coat assembly protein AP19Dr.5562X045063.0APOB apolipoprotein B (including Ag(x) antigen)Dr.13231BM7786464.2Solute carrier family 22Dr.30444AY3296294.3Embryonic globin beta e2Dr.24250AF4891052.0Uroporphyrinogen III synthaseDr.10343NM_1316874.7Na+K+ transporting, alpha 1a.2 polypeptideDr.7634AW11575711.3HemopexinDownregulated genesAngiogenesisDr.26411BQ783571−8.9Fast muscle troponin IDr.15501BM316040−2.1Similar to CYR6 HUMAN CYR61 protein precursor, Insulin-like growth factor-binding protein 10Cell adhesionDr.251BQ285646−2.3Cadherin 11Disease relatedDr.22774AW280206−5.7ras-like GTP-binding protein RAB27ADr.1816AL720262−4.4Ataxin 2-binding proteinDr.9893BM036473−2.3FibrillarinDr.16726BI429372−2.0netrin G1Growth regulationDr.12986CA787334−5.3v-fosDr.12986BI881979−5.0v-fosDr.12986BM957279−4.5v-fosDr.1221AW510198−4.3Pmx-1b (PHOX-1)Dr.12986BI881979−4.2v-fosDr.12410NM_131826−2.4Sprouty homolog 4 Dr.6431BC049326−2.3Suppressors of cytokine signaling 3Dr.6511NM_130922−2.2B-cell translocation gene 2Dr.5365AI601685−2.2Dual specificity phosphatase 5Dr.12062BC047814−2.1Epidermal growth factor receptor kinase substrate EPS8Dr.17286BM777144−2.0Hormone-regulated proliferation-associated 20 kDa proteinDr.9448BM156058−2.0TGF-beta-inducible early growth response protein 2Heart relatedDr.20010BQ826502−7.0ATPase, Ca++ transporting, cardiac muscle (ATP2A1)Dr.1448AL717344−3.5Fast skeletal myosin light chain 1aDr.20990AY033829 AY081167−2.4–2.1TitinMetabolismDr.24950BC053305−4.1Creatine kinase CKM3Dr.9528BC045993−3.5Pyruvate dehydrogenase kinaseDr.146AI477401−2.9Carnitine O-palmitoyltransferase IIDr.21501AI667180−2.4Short-chain acyl-CoA dehydrogenase Dr.19643AL918850−2.4FabG beta-ketoacyl -reductaseDr.15059BM530407−2.2Elongation of very long chain fatty acids (Cig30)Dr.21040BC045479−2.1Glucose-6-phosphatase, transport protein 1Dr.988AW154697−2.1Dodecenoyl-coenzyme A delta isomerase Dr.11971BG727588−2.0Carnitine O-acetyl-transferaseDr.4777AW420997−2.0Succinate-CoA ligaseDr.11252BC047826−2.0Creatine kinase, mitochondrial 1 Muscle relatedDr.21800AI883923−5.0Myosin binding protein CDr.5066AF524840−3.4Alpha-actinin 3Dr.24260NM_131619−3.0Myosin, light polypeptide 3Dr.2914BC045520−2.5Myosin light polypeptide 2; mylz2Dr.20990AY033829 AY081167−2.4–2.1TitinDr.1435AI353817−2.0Caveolin 3Dr.18657BQ479700−2.1Carbonic anhydrase IIDr.26411BQ783571−8.9Troponin IProteolysisDr.3581BM101561−8.3Chymotrypsinogen B1 Dr.3581BM101561−7.5Chymotrypsinogen B1Signal transductionDr.22841AI641080−2.4Serum deprivation response protein (SDPR)TranslationDr.7939AW281840−2.7Mitochondrial elongation factor G1Dr.1286BM036808−2.2Mitochondrial ribosomal protein L48Dr.18218AL909921−2.1Mitochondrial 28 S ribosomal protein S12 TransportDr.676BC050956−4.4ADT2, ADP,ATP carrier proteinDr.25199CD014403−2.1Calcium-binding mitochondrial carrier protein Aralar2 (Citrin) Dr.2784AI942949−2.0Solute carrier family 25 Dr.11127BG306498−3.7Synaptotagmin IDr.11127AW826278−3.2Synaptotagmin I Dr.13273BI885460−2.3GTP-binding protein rab15 Dr.22748AW280026−2.7trpn1Dr.11302BG306530−2.1ATPase (Ca++ transporting plasma membrane 2)
Proteinbiosynthesis (Translation): In the group linked to proteinbiosynthesis, only three genes showed decreased expression under hypoxic conditions. All three were found to be part of the mitochondrial translational machinery. In contrast, 9 non mitochondrial genes linked to proteinbiosynthesis showed all increased expression (Table 2).
Metabolism: The group with metabolic genes contains 11 repressed and 28 genes with enhanced expression (Table 2). The repression majorly involves metabolic genes linked to β-oxidation and lipid metabolism. Among the metabolic genes with enhanced expression, are pyruvate kinase and aldolase which both are key enzymes for glycolysis indicating a shift from aerobic to anaerobic metabolism induced by hypoxia.
Protection against reactive oxygen species (ROS): The group of genes important for protection against ROS contains 6 genes which are all enhanced under hypoxia.
Apoptosis: We found four genes linked to programmed cell death to be enhanced by hypoxia. Two of these, the death receptor 5 (DR5) and the BNIP3 homologue, are considered to be pro-apoptotic, whereas apoptosis inhibitor 5 and Bax inhibitor have been shown to have anti-apoptotic properties (Tewari et al. 1997; Xu and Reed 1998).
Growth regulation: In the group of genes linked to growth regulation, we found 9 genes to be repressed and 8 genes with enhanced expression. Within the group of 8 repressed genes, we found 5 antiproliferative genes: spry4 and dual specificity phosphatase 5 both inhibit mitogen-activated kinases (MAPK), SOCS3 binds and inhibits Janus kinases (JAK) and thereby prevents STAT3 activation. BTG2 which is important in the G1/S transition and TIEG2 is a transcriptional repressor with antiproliferative functions. Although some genes involved in cell proliferation, like the transcription factor c-fos were repressed by hypoxia (see also “Discussion”), the regulation of the vast majority (13 of 17) of identified genes in this group suggests stimulation of proliferation (Table 2).
Inflammation: 14 genes involved in the inflammatory response were identified to be differentially regulated and all showed increased expression upon hypoxia treatment.
Heart-related function: Several genes linked to cardiac hypertrophy, cardiomyopathy (disease of the heart muscle) and cardiac infarction were identified in this study (see Table 2 and "Discussion").
Muscle-related function: Several sarcomeric genes linked to hypertrophy showed decreased expression under hypoxic conditions (Table 2 and “Discussion”).
Development: The genes in this group were all found to be upregulated by hypoxia among them the gene for notch-2 and notch-3. Notch receptors are transmembrane receptors with essential roles in development including heart development.
Transport (cellular and vascular): A heterologous group containing the gene for embryonic hemoglobin beta e2, important for oxygen transport and hemopexin which is important for heme and iron transport and was found to be upregulated.
Angiogenesis: It is well known that hypoxia via the hypoxia inducible factor 1α (HIF1α) pathway leads to angiogenesis. In the zebrafish heart under CCH, we observed increased expression of HIF1α as well as fibrinogen-α and -γ. Fibrinogen-α has been shown to stimulate HIF1α and VEGF expression and thereby induces angiogenesis (Shiose et al. 2004).
Expression changes of known hypoxia responsive genes
Examples for regulation of known hypoxia responsive genes are HIF1α, insulin growth factor-2 (IGF-2), insulin growth factor binding protein 1 (IGFbp1) and caveolin 3. The transcription factor Hif1α is pivotal in the cellular response to hypoxic stress (Semenza 1999). We further observed increased expression of the egl nine homolog, a gene which was shown to be induced by hypoxia through the Hif1α pathway (Pescador et al. 2005). IGF-2 gene- and protein expression had been shown to be upregulated by hypoxia (Beilharz et al. 1995). It was shown that IGFbp1 also is a hypoxia-inducible gene in zebrafish embryos and it mediates hypoxia-induced embryonic growth-inhibition and developmental-retardation (Kajimura et al. 2005a). Both IGF-2 and IGFbp1 were found upregulated in our study. We observed decreased expression of caveolin 3 in the hearts of zebrafish exposed to CCH, earlier findings showed that chronic myocardial hypoxia led to decreased caveolin-3 protein expression in rabbit hearts (Shi et al. 2000). These findings indicate that the hypoxic conditions used lead to hypoxic stress in the fish heart.
Evaluation of microarray results by quantitative real-time RT-PCR
To further verify our results, we used quantitative real-time PCR for 10 of the transcripts. We confirmed the gene expression changes found in the microarray studies for these 5 up- and 5 downregulated transcripts by this independent method (Fig. 4). The downregulated zebrafish genes tested were: c-fos, phox1, creatine kinase (ckm3), nebulin, titin, and the upregulated genes tested were: metallothionein, pyruvate kinase, apoptosis inhibitor 5, igfbp1 and notch-2. The fold induction values were not always directly comparable to the array data but in all cases induction or reduction was confirmed. Quantitative differences between array data and qPCR results have been reported before (Meijer et al. 2005; Ton et al. 2003; van der Meer et al. 2005).
Fig. 4Verification of gene expression changes by quantitative real-time PCR. 10. selected genes, which were found to be differentially expressed on the microarrays, were further analyzed by quantitative real-time RT-PCR. Relative expression is given based on normalization to β-actin. A standard curve for β-actin was included in each experiment and data represents three independent experiments each done in triplicates. The primers used are given in Supplemental Table 2
Assessment of microarray results for the IGF/PI3K/Akt pathway by comparing phospho-Akt levels in cardiac myocytes of normoxic versus hypoxic zebrafish hearts
The IGF/PI3K/Akt pathway is activated by IGFs, which are antagonized by the IGFbp1. To test the effects of the upregulation of both IGF-2 and IGFbp1, we assayed phospo-Akt levels in cardiac myocytes and showed that phospho-Akt levels were not different between normoxic and hypoxic cells. Figure 5 shows cytoplasmic immunohistochemical staining of phospho-Akt. The antibody recognizes phosphorylated and detects the phospho-Akt1/2/3 forms. The incubation times for the primary and secondary antibody as well as the BM Purple were optimized to obtain a good signal to noise ratio. Absorbances were linearly related to the time of incubation with the primary and secondary antibodies as well as that of BM Purple. Figure 5d shows for normoxic fish the absorbance values of the phospho-Akt staining in cardiac myocytes as well as skeletal muscle fibers from the tail as a function of the incubation time with BM Purple AP substrate. The absorbance values for the heart are considerably higher than for the skeletal muscle fibers. However, for both cardiac and skeletal muscle the absorbances are linearly related with the incubation time with BM Purple AP substrate and increase at the same relative rate. This implicates that the absorbance of BM Purple after 45 min incubation with BM Purple AP substrate did not reach saturation and therefore provides a semi-quantitative estimate of the phospho-Akt content in the cardiac myocytes. Absorbances of staining for phospho-Akt in normoxic and hypoxic cardiac myocytes were not shown to be significantly different (Fig. 5E, P < 0.48), which indicates that hypoxia did not change the activation of the Akt pathway in the cardiac myocytes.
Fig. 5Effects of chronic constant hypoxia of the activation of Akt in zebrafish cardiac myocytes. Immunohistochemical staining of phospho-Akt in cardiac myocytes of zebrafish raised under hypoxic a or normoxic b conditions. Specificity is shown by the control sections obtained from normoxic fish which were not incubated with primary antibody against phospho-Akt c. For both cardiac myocytes and skeletal muscle fibers, the absorbance of the BM Purple is linearly increasing at the same relative rate with the incubation time with BM Purple AP substrate at the same relative rate d. Mean absorbances of phospho-Akt staining (+S.E.M.) from normoxic and hypoxic cardiac myocytes was not different e)
Discussion
In the aquatic environment, oxygen concentrations can often vary, and being able to adapt to changes in oxygen levels can be advantageous for the survival of aquatic animals. This might be in part the reason why some teleosts have developed the ability to withstand extreme hypoxic conditions.
In this study, we have focused on the long-term response to hypoxia in the fish heart. The hypothesis is that the zebrafish heart, in contrast to most mammalian hearts, which are characterized by relative intolerance to injury or the lack of oxygen, are able to adapt to extreme hypoxic conditions.
We showed that chronic hypoxia of zebrafish caused a smaller ventricular outflow tract, reduced lacunae and increased cardiac myocyte densities in the heart. These findings suggest that hypoxia induced an increase of the cardiac myocyte volume or at least did not result in a loss of cardiac myocytes. This is in contrast to mammals where tissue hypoxia in chronic heart failure leads to apoptosis and considerable losses of cardiac myocytes (see for review, Sabbah et al. 2000a). In mammals, compensation for this loss of cardiac myocytes occurs mainly by hypertrophy of the remaining cardiac myocytes (Ostadal and Kolar 2007), although regeneration of myocardium by proliferation of cardiac myocytes may occur also but to a limited extend (Beltrami et al. 2001). Our assay for phosho-Akt did not show any enhancement of Akt activity in response to the CCH, suggesting a lack of hypertrophic signaling via the phosphotadilinositol 3 kinase pathway. However, the density of cardiac myocyte nuclei increased by 50% during CCH, which indicates substantial proliferation of cardiac myocytes and/or nuclear hyperplasia. Recently, it has been shown that the zebrafish heart has the ability to regenerate from mechanical cardiac injury by proliferation of cardiac myocytes (Poss et al. 2002). If the zebrafish heart responds to chronic hypoxia in a similar way as to mechanical dissection, this will be beneficial in preventing apoptosis of cardiac myocytes as the diffusion distance for oxygen are not increased as during hypertrophy, which will help to prevent the development of anoxic cores in the cardiac myocytes (Des Tombe et al. 2002; van der Laarse et al. 2005). In the case that the increase in cardiac myocyte nuclear density during chronic hypoxia was due to nuclear hyperplasia and nuclear ploidy, the observed smaller ventricular outflow tract and reduced lacunae may have been the result of hypertrophy of the existing cardiac myocytes. The histological assay in this study does not clearly distinguish between hyperplasia and hypertrophy of cardiac myocytes. Future investigation of the mechanisms underlying the general morphological adaptations of teleost fish heart in response to CCH requires a nuclear stain in combination with a clear membrane stain, which allows determination of cell sizes in addition to counts of cardiac myocyte nuclei.
We were interested in the underlying gene expression changes of these adaptations. We found gene regulations in a transcriptional network of the serum response element (SRE), which are opposed to the ones described in mammals. In the zebrafish heart hypoxia repressed c-fos and phox1 expression. Both genes are important in the same transcriptional network, phox1 can transduce serum-responsive transcriptional activity to the c-fos (SRE) by interacting with serum response factor (SRF) (Simon et al. 1997). In mammals many studies showed increased expression of c-fos by hypoxia, e.g. in rats hypoxia induces c-fos expression in the LV and RV (Deindl et al. 2003) and the same held true in tissue culture cells (Webster et al. 1993). Interestingly, repression of c-fos by hypoxia has also been shown for the short-term response to anoxia in anoxia tolerant turtles (Greenway and Storey 2000). Is it possible that repression of c-fos and phox-1 are important adaptations in hypoxia tolerant animals?
Several novel gene expression changes induced by CCH have been identified in this study (Table 2 and Supplemental Table 1). An example is the two Notch receptors, notch-2 and notch-3, whose expression were induced by CCH. Notch receptors have been shown to be important for heart patterning and differentiation (Armstrong and Bischoff 2004), cell fate determination and self renewal of stem cells (Androutsellis-Theotokis et al. 2006; Silvia Bianchi 2006) but so far have not been linked to the hypoxic response.
Several genes with links to human heart pathology were found to be upregulated in this study. Among these were two markers for myocardial infarction, complement component C9 and haptoglobin. C9 is used as a marker for myocardial infarction (Doran et al. 1996) and a polymorphism for haptoglobin predicts 30-day mortality and heart failure in patients with diabetes and acute myocardial infarction (Levy et al. 2002). We also observed increased expression of fetuin-α, a circulating calcium-regulatory glycoprotein that inhibits vascular calcification. Low fetuin-α levels have been associated with heart failure in mice (Merx et al. 2005). Upregulation of fetuin-α in the zebrafish heart could help to better tolerate CCH. Selenoprotein P (SEPP1) is a heparin-binding protein that appears to be associated with endothelial cells and has been implicated as an oxidant defense in the extracellular space. Human populations that are selenium deficient are susceptible to the development of Keshan disease, a potentially fatal form of cardiomyopathy (Nezelof et al. 2002). SEPP1 expression was found to be increased in our study indicating a potential protective mechanism against oxidative stress in the heart. This was further supported by the increased expression observed for several genes important for the protection against reactive oxygen species (ROS) (Table 2). Among them were metallothionein and glutathione S-transferase, which are both known ROS scavengers. Metallothionein has further been described to be protective against hypoxia-induced apoptosis when overexpressed (Wang et al. 2001). Our findings suggest ROS protection as an important adaptation to CCH in the fish heart.
Furthermore, our data show for the first time that CCH simultaneously induced upregulation of IGF-2 and the insulin-like growth factor binding protein 1 (IGFbp1). In isolated cardiomyoblasts, angiotensin II stimulates IGF-2 expression which is involved in the induction of apoptotic signaling in rat hearts (Lee et al. 2006). We have observed here both angiotensin and IGF-2 upregulation (Table 2). The upregulation of IGFbp1 seems to be a general response to hypoxia in zebrafish embryos, where it mediates hypoxia-induced embryonic growth retardation and developmental delay (Kajimura et al. 2005b). IGFbp1 is a secreted protein, which binds to IGFs in the extracellular environment and prevents receptor activation (Florini et al. 1996; Stewart and Rotwein 1996). Here, IGFbp1 by binding IGF-2, may have prevented both cardioprotective as well as apoptotic effects of enhanced IGF-2 expression. To test the effects of the upregulation of both IGF-2 and IGFbp1, we assayed phospo-Akt levels in cardiac myocytes and showed that phospho-Akt levels were not different between normoxic and hypoxic cells. This suggests that either hypoxia-induced changes in mRNA expression did not occur at the protein level or the effects of increased IGF-2 expression on the IGF-1 receptor (which can be activated by IGF-2) were blunted by the upregulation of the IGFbp1. The fact that we did not observe increased phospho-Akt levels suggests a lack of hypertrophic signaling as enhanced phospho-Akt is required for cardiac hypertrophy (DeBosch et al. 2006). This may be beneficial for the heart as hypertrophy of cardiac myocytes implicates an increase in the diffusion distance for oxygen which may cause the development of anoxic cores (Des Tombe et al. 2002), release of cytochrome c (van Beek-Harmsen and van der Laarse 2005) and production of ROS (Lee et al. 2006; Powers et al. 2007) and eventually causing apoptosis of cardiac myocytes. The importance of the IGF-2/IGFbp1 signaling in the protection of the zebrafish heart and its underlying mechanisms remain to be determined.
The gene expression changes we observed for the specific response to CCH in the fish heart were very different from the responses we observed in an earlier study in the gills (van der Meer et al. 2005). Under the same criteria as used in this study, we found that the majority of differentially regulated genes in the gills showed a decrease in gene expression (68.1% or 250 genes in total) in comparison to genes, which showed increased expression levels (31.9% or 117 genes in total). This is opposed to the heart where 69.1% of differentially regulated genes (260 genes in total) showed increased and 30.9% (116 in total) decreased expression levels. Many genes linked to protein synthesis showed a similar trend, and were downregulated in the gills (van der Meer et al. 2005) and upregulated in the heart (Table 2). The major differences observed in gene regulation between heart and gills point to very tissue specific responses to CCH. A list of genes identified in both studies, as identified by a direct comparison based on Accession numbers and old Unigenes numbers are shown in Supplemental Data 3.
Teleosts have developed specific phenotypic adaptations to low oxygen, due to the natural occurrence of hypoxia in the water environment. Here, we identified the changes in gene expression as well as associated morphological changes of the fish heart to hypoxia. We observed repression of c-fos, which differs compared to the described increase in expression in mammals (Deindl et al. 2003). Other changes found like upregulation of the two Notch receptors have not been described before to the best of our knowledge. Similarly, the simultaneous increase in expression of IGF-2 and IGFbp1 has not been shown. The changes identified here can contribute to the ability of teleosts to adapt to severe hypoxia (for example, the upregulation of fetuin-α and sepp1 levels). Future functional studies are warranted to validate the role of the identified genes in cardiac protection to hypoxia.
Electronic supplementary material
Below is the link to the electronic supplementary material.
(XLS 76 kb)
(DOC 33 kb)
(DOC 57 kb)
(XLS 52 kb)
|
[
"hypoxia",
"zebrafish",
"heart",
"gene expression",
"hyperplasia"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Arch_Microbiol-3-1-2111041
|
Structural characterization of diabolic acid-based tetraester, tetraether and mixed ether/ester, membrane-spanning lipids of bacteria from the order Thermotogales
|
The distribution of core lipids in the membranes of nine different species of the order Thermotogales, one of the early and deep branching lineages in the Bacteria, were examined by HPLC/MS and demonstrated to consist of membrane-spanning diglycerol lipids comprised of diabolic acid-derived alkyl moieties. In the Thermotoga species the core membrane lipids are characterized by the presence of both ester and ether bonds, whereas in the phylogenetically more distinct Thermosipho and Fervidobacterium spp. only ester bonds occur. A tentative biosynthetic route for the biosynthesis of these membrane-spanning lipids is proposed. Since species of the order Thermotogales are assumed to have occurred early during the evolution of life on Earth, as suggested by its position in the phylogenetic tree of life, these data suggest that the ability to produce both ether and ester glycerol membrane lipids developed relatively early during microbial evolution.
Introduction
The group of anaerobic bacteria falling in the order Thermotogales represents a very deep phylogenetic branch within the phylogenetic tree of life based upon the ribosomal RNA gene (Woese 1987). Within this order, members of the genus Thermotoga have an upper growth limit of 90°C and represent, together with species in the order Aquificales, cultivated bacteria with the highest known growth temperatures. Bacteria belonging to the Thermotogales occur widespread in the environment: they thrive within continental solfatara springs of low salinity, shallow and deep-sea marine hydrothermal systems and high-temperature marine and continental oil fields (see Huber and Hannig 2005 for a recent review). Their strict organotrophic way of life makes them consumers of microbial organic matter within high temperature ecosystems. Within the order Thermotogales, the genera Thermotoga, Thermosipho, Fervidobacterium, Geotoga, Petrotoga and Marinitoga have been described, all belonging to the single family Thermotogaceae (Huber and Hannig 2005). The type species T. maritima for the genus Thermotoga, also the first identified member of the Thermotogales, was isolated from a geothermally heated, shallow marine sediment in Italy (Huber et al. 1986). Subsequently, quite a number of phylogenetically closely related (T. neapolitana, T. petrophila, T. naphtophila) and slightly more distant (T. thermarum, T. elfii, T. subterranea, T. hypogea, T. lettingae) species were isolated, mostly from marine settings. Thermotoga-related organisms have been detected in oil reservoirs (Huber and Hannig 2005), whilst members of the genus Fervidobacterium seem to be restricted to hydrothermal settings with low salinity such as hot springs. For example, Fervidobacterium pennivorans was firstly isolated from a hot spring on San Miguel (the Azores) (Friedrich and Antranikian 1996). Members of the genus Thermosipho have been isolated from a wide variety of high temperature environments such as marine hydrothermal springs (e.g. T. africanus) and deep sea hydrothermal vents (e.g. T. melanesiensis) (Antoine et al. 1997; Huber et al. 1989).
To be able to grow at high temperatures, the members of the Thermotogales possess a variety of thermostable enzymes and proteins, which are also of biotechnological interest. In addition, the cell membranes of these organisms must be able to cope with high temperatures. Indeed, a variety of unusual membrane lipids has been reported for members of the Thermotogales. Unusual, long-chain dicarboxylic (diabolic) acids are present in the core lipids of the genus Thermotoga (Jeanthon et al. 1995; Manca et al. 1992; Windberger et al. 1989; Carballeira et al. 1997; DeRosa et al. 1988; Huber et al. 1986). About 50% of the total polar lipids of T. maritima are two amphipathic monopolar glycolipids with a very rare α-(1–4) diglucosyl structure (Manca et al. 1992).
In the present paper, we describe in detail the structures of these and other lipids in a suite of bacteria belonging to the Thermotogales and discuss their distributions. We applied an HPLC/MS method, which for the first time enabled to characterize the full membrane-spanning core lipids.
Materials and methods
Cultures
Cells were grown in 500-ml batch cultures in 1.2-l-bottles at temperatures of 55–80°C and with either medium A or B (see Table 1). Medium A was composed (per liter demineralized water) of KCl 0.335 g; MgCl2·6H2O 4.0 g; MgSO4·7H2O 3.45 g; NH4Cl 0.25 g; K2HPO4 0.14 g; NaCl 10.0 g; glucose 3.8 g; yeast extract 0.5 g; peptone 0.5 g; NaHCO3,·2.0 g; Na2S.7-9H2O, 0.40 g; trace element solution, 10 ml and vitamin solution, 10 ml, which were based on medium 141 of DSM (http://www.dsmz.de); CaCl2·2H2O, 0.140 g; resazurin, 0.5 mg. The medium was anaerobically dispensed into serum bottles and a gas phase of 180 kPa N2/CO2 (80/20 v/v) was applied. NaHCO3, Na2S.7-9H2O, glucose, CaCl2·2H2O and vitamin solution were added after sterilization. The pH was 7.0–7.2. Medium B is a modified DSM medium 640 and composed (per liter demineralized water) of NH4Cl 0.90 g; NaCl 0.90 g; MgCl2·6H2O 0.40 g; KH2PO4 0.75 g; K2HPO4 1.50 g; peptone 0.50 g; yeast extract 0.50 g; trace element solution SL-10 (see DSM medium 320), 1.00 ml; FeCl3·6H2O 2.50 mg; glucose 3.8 g; cysteine–HCl· H2O, 0.75 g; resazurin 0.5 mg. The medium was anaerobically dispensed into serum bottles and a gas phase of 180 kPa N2 (100%) was applied. Glucose and cysteine–HCl·H2O were added after sterilization. The pH was adjusted to pH 7.2. The cultures were inoculated with cells from the DSM collection of microorganisms (see Table 1 for strain number) and were harvested in their exponential growth phase. Two different batches of T. lettingae were grown and analysed separately.
Table 1Bacteria from the order Thermotogales with some background information and the culture conditions used in this studyBacterial speciesSource for the isolateOptimal growth temperature (°C)ReferenceDSM strain numberCulture conditionsMediumaT (°C) Thermotoga maritimaAnaerobic marine sediment, Italy 80Huber et al. (1986)3109A80, 55T. neapolitanaSubmarine hot spring, Italy 85Jannasch et al. (1988)4359A80T. elfiiOil-producing well, Africa 65Ravot et al. (1995)9442A65T. lettingaeSulfate-reducing bioreactor, The Netherlands 65Balk et al. (2002)14385A65T. hypogeaOil-producing well, Africa 70Fardeau et al. (1997)11164B70Thermosipho sp.Spring, New Zealand 656568B65T. africanusMarine hydrothermal spring, Obock 75Huber et al. (1989)5309B75T. melanesiensisDeep sea hydrothermal vent, Pacific Ocean 70Antoine et al. (1997)12029B70Fervidobacterium pennivoransHot mud of spa, Italy 65Koch et al. (1997)9078B65a See Materials and method sections for a description of the media used
Lipid analysis
Lyophilized cells were repeatedly (×5) ultrasonically extracted with dichloromethane (DCM)/ methanol (MeOH) (2:1 v/v). The solvent of the combined extracts was removed by rotary evaporation at reduced pressure. The extract was taken up in DCM and weighed in a small vial by removing the solvent under a gentle stream of nitrogen. An aliquot (ca. 1 mg) of the extract to which a known amount of the internal standard, 2-methylheneicosane, was added, was methylated with BF3/MeOH, filtered over a small pipette filled with silica using ethyl acetate as the eluent, and subsequently silylated with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA) in pyridine at 60°C for 15 min. These so-called total lipid fractions (TLF-1) were analysed with gas chromatography (GC) and gas chromatography–mass spectrometry (GC–MS). In selected cases, acetylation instead of silylation of the eluate obtained after chromatography over silica was performed following a procedure described elsewhere (Cox et al. 1972).
The cell residue after solvent extraction was, after addition of a known amount of the internal standard, 2-methylheneicosane, hydrolysed with 1 N KOH/MeOH (96%) by refluxing for 1 h. The hydrolysate was neutralized with 2 N HCl/MeOH and after addition of water extracted with DCM. This fraction was methylated, chromatographed over silica and silylated as described above and the resulting TLF-2 fraction was analyzed by GC and GC–MS. In specific cases alkaline hydrolysis was also performed with an aliquot of the solvent extract following the same procedure as described above. This resulted in a fraction named TLF-1-OH which was also analyzed by GC and GC–MS.
For analysis of core lipids by high performance liquid chromatography/atmospheric pressure chemical ionization mass spectrometry (HPLC/APCI–MS) aliquots of the solvent extracts (ca. 2 mg) were separated over a small Al2O3 column using hexane/DCM (1:1 v/v) and DCM/MeOH as eluent to obtain apolar and polar fractions, respectively. The polar fraction was dissolved in hexane/propanol (99:1 v/v), filtered over an 0.4 μm PTFE filter, and analysed by HPLC/APCI–MS.
GC and GC–MS
GC was performed using a Fisons GC8000 instrument, equipped with an on-column injector and a flame ionization detector. A fused silica capillary column (25 m × 0.32 mm) coated with CP Sil-5 (df 0.12 um) was used with helium as carrier gas. The samples were injected at 70°C and the oven temperature was programmed to 130°C at 20°C/min and then at 4°C/min to 320°C, at which it was held for 10 min. GC–MS was performed on a HP 5890 gas chromatograph interfaced with a VG Autospec Ultima mass spectrometer operated at 70 eV with a mass range of m/z 40–800 and a cycle time of 1.7 s (resolution 1,000). The gas chromatograph was equipped with a fused silica capillary column as described for GC. The carrier gas was helium. The same temperature program as for GC was used.
HPLC/APCI–MS
Membrane-spanning core lipids were analysed by HPLC/APCI–MS according to Hopmans et al. (2000) with minor modifications. Analyses were performed on an Agilent 1100 series HPLC/MS instrument, equipped with an auto-injector and HP Chemstation software. Separation was achieved on an Alltech Prevail Cyano column (150 mm × 2.1 mm; 3 μm). Flow rate of the hexane:propanol 99:1 (v/v) eluent was 0.2 ml min−1, isocratically for the first 5 min, thereafter with a linear gradient to 1.8% propanol in 45 min, subsequently to 10% propanol in 20 min, and subsequently isocratically for the last 20 min. Injection volume of the samples was usually 10 μl.
Results
Nine species of bacteria belonging to the order Thermotogales were analysed for their lipid composition; five of them are Thermotoga species (T. maritima, T. lettingae, T. elfii, T. neapolitana, T. hypogea), three are Thermosipho species (T. africanus, T. melanesiensis, T. sp.) and one is a Fervidobacterium species (F. pennivorans) (Table 1).
Composition of total lipid fractions
Figure 1a shows a typical gas chromatogram of a total lipid fraction (TLF-1) for T. maritima. In addition to straight-chain fatty acids (predominantly n-C16 and n-C14), a number of more unusual lipids were detected. The first group comprises long-chain dicarboxylic acids; 15,16-dimethyltriacontanedioic acid (or diabolic acid 1; see Fig. 2 for structures) and 13,14-dimethyloctacosanedioic acid (2), identified on basis of mass spectral data reported previously (Carballeira et al. 1997; Klein et al. 1979). The mass spectra of 1 and 2 as methyl ester derivatives are shown in Fig. 3. The relative abundance of the C30 and C32 diabolic acids varied substantially between different species (Table 2). F. pennivorans also contained small amounts of a tentatively identified C34 diabolic acid. In some species, isomers of these diabolic acids, with virtually an identical mass spectrum but with a different retention time, were present in varying quantities. These probably represent stereoisomers, possibly related to the presence of two chiral centres in the middle of the carbon chain of the diabolic acids. The second group of unusual components comprises glycerol monoethers. 1-O-hexadecyl glycerol (3) and 1-O-tetradecyl glycerol (4) were identified by mass spectral data (m/z 205 base peak and appropriate molecular ion) and a co-elution experiment with an authentic standard. The higher molecular weight monoethers are 15,16-dimethyl-30-glyceryloxytriacontanoic acid (5) and 13,14-dimethyl-28-glyceryloxyoctadecanoic acid (6). Glycerol monoether 5 has been identified previously and the mass spectrum of 5 as acetyl derivative was reported (DeRosa et al. 1988). For identification purposes we used acetylation in addition to silylation to be able to compare the mass spectrum of the derivative of 5 with that reported in the literature (DeRosa et al. 1988). The mass spectrum of 5 as TMS derivative is shown in Fig. 3. Component 6 was tentatively identified by interpretation of its mass spectrum in comparison with that of 5 (Fig. 3). The relative amounts of monoethers 5 and 6 in the TLF-1 vary widely between the examined species of the Thermotogales (Table 2). As observed for the diabolic acids 1 and 2, two additional minor isomers with similar mass spectrometric characteristics but different retention times are observed. The third group of lipids are glycerol dialkyl diesters and glycerol mixed ester/ethers. They were predominantly found in the TLF-1 of T. lettingae (Fig. 4a) and are composed of 1,2- and 1,3-diacyl glycerols (e.g. 7 and 8) and 1-alkyl, 2-acyl (e.g. 9) and 1-alkyl, 3-acyl (e.g. 10) glycerols. They were identified by comparison with mass spectra reported in the literature (Wood et al. 1980). Some Thermotoga species contained small amounts of 1,2-di-O-alkyl glycerols.
Fig. 1Gas chromatograms of the total lipid fractions (TLF-1) (a–b) and these fractions after alkaline hydrolysis of the residue after extraction (TLF-2) (c–d) of Thermotoga maritima (grown at 55°C) and F. pennivorans pennivorans, respectively. Fatty acids and alcohols were derivatized to the corresponding methyl esters and trimethylsilyl ethers prior to gas chromatographic analysis. Numbers refer to structures indicated in Fig. 2. FA Fatty acid, std internal standardFig. 2Structures of lipids encountered in species from the order ThermotogalesFig. 3Mass spectra (subtracted for background) of derivatised diabolic acids and diabolic acid glycerol ethersTable 2Relative distribution (in % of total quantified) of lipids in different species of the order ThermotogalesBacterial speciesFractionaLipidsbC14:0 FAC16:0 FAC18:0 FA123456Thermotoga maritima (80°C)TLF-14.244.21.121.31.79.15.110.62.7TLF-21.725.90.938.74.22.21.017.67.8T. maritima (55°C)TLF-17.554.70.514.32.311.12.56.61.2TLF-20.715.00.448.44.60.70.223.86.2T. neapolitanaTLF-14.661.32.817.31.33.70.77.11.2TLF-21.946.52.826.42.10.615.83.9T. elfiiTLF-13.561.04.028.22.60.50.10.2TLF-21.220.01.270.76.40.20.3T. lettingae (1st batch)TLF-13.037.25.924.65.26.93.98.35.0TLF-20.811.61.048.37.82.20.918.78.7T. lettingae (2nd batch)TLF-1-OH6.326.41.132.68.84.85.88.55.6TLF-20.57.40.357.614.10.40.211.87.7T. hypogeaTLF-14.854.52.314.12.37.43.58.13.0TLF-20.54.51.047.41.50.60.133.311.1Thermosipho sp.TLF-17.951.528.012.6TLF-22.616.05.066.89.6T. africanusTLF-14.457.632.52.7TLF-22.561.88.125.42.2T. melanesiensisTLF-15.562.130.81.6TLF-25.684.77.12.6Fervidobacterium. pennivoransTLF-11.054.19.535.10.3TLF-20.213.53.082.31.0a TLF-1 Total lipid fraction, TLF-2 total lipid fraction obtained after alkaline hydrolysis of the residue after extraction;TLF-OH Total lipid fraction after alkaline hydrolysis of the extractb Numbers refer to Fig. 2. Blank entries indicate that the lipid was not detectedFig. 4Gas chromatograms of the total lipid fraction (TLF-1) (a), this fraction after alkaline hydrolysis (TLF-1-OH) (b) and this fraction obtained after alkaline hydrolysis of the residue after extraction (TLF-2) (c) of Thermotoga lettingae (2nd batch). Fatty acids and alcohols were derivatized to the corresponding methyl esters and trimethylsilyl ethers prior to gas chromatographic analysis. Numbers refer to structures indicated in Fig. 2. FA Fatty acid, std internal standard
The TLF after alkaline hydrolysis of the cell residue after extraction (TLF-2) contained a similar suite of lipids as the TLF-1 (Table 2), albeit that some other components (e.g. unsaturated C16:1 and C18:1 fatty acids) were also present but only in the examined Thermosipho species. Most striking was that the characteristic long-chain diacids 1 and 2 and the complex monoethers 5 and 6 were relatively much more abundant in TLF-2 (Figs. 1c,d and 4b; Table 2), suggesting that they are predominantly formed by alkaline hydrolysis from larger lipid structures. A similar distribution was observed when the extract of T. lettingae (2nd batch) was subjected to alkaline hydrolysis (TLF-1-OH): the long-chain diacids 1 and 2 and the monoethers 5 and 6, which were not observed in the TLF-1 (Fig. 4a), are one of the most abundant lipids in the TLF-1-OH fraction (Fig. 4b). This indicates that these lipids are building blocks of larger, more complex lipids not amenable to GC.
The absolute amount of the most abundant lipids (in mg g−1 dry weight) was for a number of species measured in both the TLF-1 and TLF-2 (Table 3). This showed that the amounts released by alkaline hydrolysis from the residue after extraction are still substantial. It also showed that the highest concentrations occur in the Thermotoga species and in F. pennivorans, while concentrations in the Thermosipho species are substantially lower. Quantification of the lipids in the TLF-1-OH fraction of T. lettingae (2nd batch) showed that the concentration increases substantially after alkaline hydrolysis of the extract.
Table 3Absolute concentrations (in mg g−1 dry weight) of diabolic acids and glycerol monoether acids in different species of the order ThermotogalesBacterial speciesFractionaLipidsb1256Thermotoga maritima (55°C)TLF-15.60.512.70.50TLF-24.10.402.00.52T. lettingae (2nd batch)TLF-1TLF-1-OH7.11.91.81.2TLF-21.90.460.380.25T. hypogeaTLF-11.00.170.600.22TLF-20.700.020.490.16Thermosipho sp.TLF-10.05TLF-20.260.04T. africanusTLF-10.01TLF-20.180.02T. melanesiensisTLF-1TLF-20.01FervidobacteriumpennivoransTLF-17.90.07TLF-23.50.04a TLF-1 Total lipid fraction, TLF-2 total lipid fraction obtained after alkaline hydrolysis of the residue after extraction, TLF-1-OH total lipid fraction after base hydrolysis of the extractb Numbers refer to Fig. 2. Blank entries indicate that the lipid was not detected
Distribution of membrane-spanning core lipids in T. maritima
To characterize the membrane-spanning core lipids of the species belonging to the Thermotogales, the polar fraction of the solvent extract was analysed by an HPLC/APCI–MS method (Hopmans et al. 2000) slightly adjusted to enable the elution of more polar components than glycerol dialkyl glycerol tetraethers (GDGTs) from the LC column (see Materials and methods for details). Analysis of the polar lipids of Thermotogamaritima revealed a complex composition exemplified by the base peak chromatogram (Fig. 5a). The first eluting, small peak with a retention time of 29.0 min (labelled 11a in Fig. 5b) showed an APCI mass spectrum (Fig. 6a) characterised by peaks at [M + H]+, [M + H]+-18 (loss of water) and [M + H]+-74 (loss of glycerol). This characteristic pattern is well known from archaeal GDGTs (Hopmans et al. 2000) and non-isoprenoidal GDGTs (Sinninghe Damsté et al. 2000), suggesting a tetraether structure. Indeed, upon alkaline hydrolysis of this fraction and re-analysis by HPLC/APCI–MS, this minor peak was now the most abundant peak in the base peak chromatogram. These data indicate a GDGT structure for the major peak comprised of two glycerol units connected by two C32 alkyl chains, resulting in a molecular mass of 1077 (C70H140O6). Since 15,16-dimethyltriacontanedioic acid (1) and 15,16-dimethyl-30-glyceryloxytriacontanoic acid (5) are abundant lipids in the TLF-1 of T. maritima, it seems likely that the alkyl chains in this GDGT are the same, resulting in structure 11a. At the flank of peak 11a, at slightly higher retention time, a similar mass spectrum was observed with all ions shifted by 28 Da to lower m/z values. This compound is likely a GDGT with a C32 and C30 dimethyl alkyl chains (12a), consistent with the presence of small amounts of 13,14-dimethyloctacosanedioic acid (2) and 13,14-dimethyl-28-glyceryloxyoctadecanoic acid (6) in the TLF-1 (Table 2).
Fig. 5HPLC/APCI–MS base peak chromatogram (a) and specific mass chromatograms (b–i) of polar lipids of Thermotoga maritima grown at 55°C. The mass chromatograms, indicative for specific components, are m/z 1,078 (b [M + H]+ C70 tetraether), m/z 1,074 (c [M + H]+-18 C70 triether/monoester), m/z 1,046 (d [M + H]+-18 C68 triether/monoester), m/z 1,088 (e [M + H]+-18 C70 diether/diester), m/z 1,060 (f [M + H]+-18 C68 diether/diester), m/z 1,102 (g [M + H]+-18 C70 monoether/triester), m/z 1,074 (h [M + H]+-18 C68 monoether/triester), m/z 1,116 (i [M + H]+-18 C70 tetraester), m/z 1,088 (j [M + H]+-18 C68 tetraester). Filled and blocked peaks indicate membrane-spanning lipids based on C32/C32 and on C32/C30 branched alkyl chains, respectively. Note that some mass chromatograms are shown twice as they reveal the presence of different componentsFig. 6Partial APCI mass spectra (corrected for background) of (a) the C70 tetraether 11a, (b) the C70 triether/monoester 11b or 11c, (c) the C70 diether/diester 11d, (d) the C70 ether/triester 11e, (e) the C70 diester 11f, (f) the C70 diether/diester 13d
The first larger peak in the base peak chromatogram at a retention time of 33.0 min (labelled 11b or 11c in Fig. 5c) shows an APCI mass spectrum (Fig. 6b) in which all ions are shifted by 14 Da to higher m/z values relative to those in the mass spectrum of GDGT 11a (Fig. 6a). In addition, the [M + H]+-18 peak is substantially increased relative to the [M + H]+ peak and an [M + H]+-74 peak is also observed. Since this peak disappears when the fraction is subjected to alkaline hydrolysis, this component is likely to contain an ester moiety. This information, together with its molecular mass, indicates that this component is structurally related to GDGT 11a but with one of the ether bonds replaced by an ester bond: a glycerol dialkyl glycerol triether/monoester with two C32 dimethyl alkyl chains (11b), resulting in a molecular mass of 1091 (C70H138O7). The presence of the ester moiety likely results in the enhanced [M + H]+-18 peak in its APCI mass spectrum. Mass chromatography of m/z 1,074 (Fig. 5c), the abundant [M + H]+-18 peak, reveals that there is an additional, much later eluting isomer with an identical mass spectrum. This could relate to the position of the ester moiety in this type of components: with the parallel configuration of the glycerol groups, as indicated in 11, there are two possible isomers (11b and 11c). However, it is at present unclear why these positional isomers would differ to a large degree in polarity (i.e. the second isomer elutes much later and is, thus, much more polar than the first one). At the flank of the peak of the triether/monoester a component elutes with a mass spectrum similar to that of 11b but with all the major ions shifted by 28 Da to lower m/z values. This component is the triether/monoester with C32 and C30 dimethyl alkyl chains (12b). Mass chromatography of m/z 1,046 (Fig. 5d), the abundant [M + H]+-18 peak of this component, reveals a second, much later eluting isomer for this component as well.
The most abundant peak in the base peak ion chromatogram at a retention time of 37.0 min (Fig. 5a) shows a mass spectrum similar to that of the triether/monoester 11b but in which again all major ions are shifted by 14 Da to higher m/z values (cf. Fig. 6b and c). This component likely represents a glycerol dialkyl glycerol diether/diester with two C32 dimethyl alkyl chains (11d), resulting in a molecular mass of m/z 1,105 (C70H136O8) and an [M + H]+-18 of m/z 1,088. This component is composed of two esterified 15,16-dimethyl-30-glyceryloxytriacontanoic acid (5) units. Moreover, mass chromatography of m/z 1,060 (Fig. 5f) reveals the presence of a diester/diether with C32 and C30 building blocks (12d). Both the mass chromatograms of m/z 1,088 (Fig. 5e) and m/z 1,060 (Fig. 5f) reveal the presence of a much later eluting isomer with similar mass spectra. This latter cluster of peaks contains co-eluting components with a mass spectrum with peaks at m/z 1,120, 1,106, 1,046 and 1,028, presumably the corresponding glycerol dialkyl glycerol monoether/triester with two C32 dimethyl alkyl chains (11e). Mass chromatograms of m/z 1,074 (Fig. 5h) and m/z 1,102 (Fig. 5g) show the presence of a monoether/triester with two C32 dimethyl alkyl moieties (11e) and C32 and C30 building blocks (12e) and much later eluting isomers for both components. These isomers possess identical mass spectra (e.g. Fig. 6d). Finally, small amounts of tetraesters with two C32 dimethyl alkyl moieties (11f) and a C32 and a C30 dimethyl alkyl moiety were detected by mass chromatography of m/z 1,116 (Fig. 5i) and m/z 1,088 (Fig. 5j), respectively. A typical mass spectrum is shown in Fig. 6e. In these traces three distinct isomers were observed.
The distribution of membrane-spanning lipids in T. maritima is not depending on the growth temperature; cells grown at 80 and 55°C showed almost the same distribution.
Distribution of membrane-spanning core lipids in T. lettingae
The distribution of the membrane-spanning lipids in T. lettingae was quite different (Fig. 7) from that observed in T. maritima: it was dominated by the diether/diesters with the components with two C32 building blocks (11d; m/z 1,088; Fig. 7b) and those with a C32 and a C30 dimethyl alkyl chain (12d; m/z 1,060; Fig. 7c) in approximately equal abundance. Diether/diesters with two C30 building blocks are present in relatively small amounts as revealed by the m/z 1,032 mass chromatogram (Fig. 7d). This is consistent with the substantially higher abundance of the C30 diabolic acid (2) relative to the C32 diabolic acid (1) and the C33 glycerol monoether 6 relative to the C35 glycerol monoether 5 in the TLF-2 fraction of T. lettingae when compared to T. maritima (Table 2). The relatively high abundance of lipids with a C32 and a C30 dimethyl and, to a lesser extent, with two C30 dimethyl alkyl chains is also observed for the monoether/triesters (Fig. 7e–g) and tetraesters (Figs. 7h–j). Tetraethers (e.g. 11a), which were present in small quantities in T. maritima, were not encountered in T. lettingae.
Fig. 7HPLC/APCI–MS base peak chromatogram (a) and specific mass chromatograms (b–j) of polar lipids of Thermotoga lettingae. The mass chromatograms, indicative for specific components, are m/z 1,088 (b [M + H]+-18 C70 diether/diester), m/z 1,060 (c [M + H]+-18 C68 diether/diester), m/z 1,032 (d [M + H]+-18 C66 diether/diester), m/z 1,102 (e [M + H]+-18 C70 monoether/triester), m/z 1,074 (f [M + H]+-18 C68 monoether/triester), m/z 1,046 (g [M + H]+-18 C66 monoether/triester), m/z 1,116 (h [M + H]+-18 C70 tetraester), m/z 1,088 (i [M + H]+-18 C68 tetraester) and m/z 1,060 (j [M + H]+-18 C66 tetraester). Filled, blocked and grey peaks indicate membrane-spanning lipids based on C32/C32, C32/C30 and C30/C30 branched alkyl chains, respectively. Note that some mass chromatograms are shown twice as they reveal the presence of different components
An additional cluster of early eluting peaks not observed in the HPLC/MS analysis of the extract of T. maritima is present in the base peak chromatogram of the polar fraction of T. lettingae (Fig. 8a). The mass spectra of the peaks comprising this cluster are characterized by one dominant fragment ion which is 2 Da higher than that of the [M + H]+-18 peak of the diether/diesters (cf. Fig. 6f, c). Mass chromatography of m/z 1,090, 1,062 and 1,034 (Fig. 8b–d) reveals that this cluster of peaks is build up in a similar way as the clusters reflecting the diether/diesters, ether/triesters and tetraesters (cf. Fig. 7). These mass chromatograms also reveal that these compounds are accompanied by a later eluting isomer which possess similar mass spectra. These components likely represent glycerol trialkyl glycerol diether/diesters (13d) comprising one long-chain C32 or a C30 dimethyl alkyl chain and two shorter linear (predominantly C16) alkyl chains. It is well known from archaeal GDGT analyses that so-called “open” or acyclic (in comparison to the macrocyclic components described so far) GDGTs elute substantially earlier under these LC conditions (and are, thus, less polar) and possess APCI mass spectra that do not show an [M + H]+-74 ion resulting from the loss of a glycerol moiety (Hopmans et al. 2000). Instead, they easily loose one of the shorter alkyl chains upon APCI. Indeed, when the polar fraction of T. lettingae was analysed using an extended mass range (m/z 100–1,400), an abundant ion at m/z 792 was observed, which is likely formed by loss of a C16 moiety. Mass chromatography of ions 2 Da higher than used to detect the monoether/triesters and tetraesters reveals the presence of acyclic monoether/triesters (Fig. 8e–g) and tetraesters (Fig. 8h–j).
Fig. 8HPLC/APCI–MS base peak chromatogram (a) and specific mass chromatograms (b–j) of polar lipids of Thermotoga lettingae. The mass chromatograms, indicative for specific components, are m/z 1,090 (b [M + H]+-18 C70 “open” diether/diester), m/z 1,062 (c [M + H]+-18 C68 “open” diether/diester), m/z 1,034 (d [M + H]+-18 C66 “open” diether/diester), m/z 1,104 (e [M + H]+-18 C70 “open” monoether/triester), m/z 1,076 (f [M + H]+-18 C68 “open” monoether/triester), m/z 1,048 (g [M + H]+-18 C66 monoether/triester), m/z 1,118 (h [M + H]+-18 C70 “open” tetraester), m/z 1,090 (i [M + H]+-18 C68 “open” tetraester) and m/z 1,062 (j [M + H]+-18 C66 “open” tetraester) for the polar fraction of Thermotoga lettingae. Black, blocked and grey peaks indicate membrane-spanning lipids based on C32/C32, C32/C30 and C30/C30 branched alkyl chains, respectively. Note that some mass chromatograms are shown twice as they reveal the presence of different components
Distribution of membrane-spanning core lipids in F. pennivorans
Membrane-spanning lipids were also detected in F. pennivorans (data not shown). In this case the distribution was relatively simple: only tetraesters with predominantly two C32 dimethyl alkyl chains (11f) were detected in a distribution similar to those observed in T. maritima (Fig. 5i) and T. lettingae (Fig. 7h), i.e. dominated by three isomers. Smaller amounts of tetraesters with a C32 and a C34 dimethyl alkyl chain were also detected. In addition, the TLF-1 also contained a glycerol esterified with two 15,16-dimethyltriacontanedioic acid moieties or 15,16-dimethyltriacontanedioic acid and 15,16-dimethyldotriacontanedioic acid.
Discussion
Our results indicate that the diabolic acids and their ether derivatives previously identified in members of the Thermotogales (DeRosa et al. 1988; Windberger et al. 1989; Jeanthon et al. 1995; Huber et al. 1986) are synthesized to produce membrane-spanning lipids and probably represent biosynthetic intermediates. The relative concentration of the diabolic acids and their ether derivatives is substantially higher in the extract obtained after alkaline hydrolysis of the extracted cell residue (TLF-2) (Table 2). The HPLC/APCI–MS results clearly revealed the presence and, in some cases, the structural complexity of the membrane-spanning core lipids in species of the order Thermotogales for the first time. Previous investigators (Clarke et al. 1980) have suggested that diabolic acids may act as linkers between two glycerol moieties in polar membrane lipids of Butyrivibrio spp. It is generally thought that the production of membrane-spanning lipids in prokaryotes is an adaptation to high growth temperature, although Butyrivibrio is a mesophile. The ubiquitous presence of diabolic acid-based membrane-spanning lipids in members of the Thermotogales, which are all (hyper)thermophiles, is in line with this general idea (Table 2). However, the distribution of membrane-spanning lipids in T. maritima is not directly affected by the temperature of growth: at both 80 and 55°C diabolic acid-based membrane-spanning lipids were produced.
Despite the overall similarity in the lipid composition, there is a clear dichotomy: all members of the genus Thermotoga produce ether lipids (i.e. 3–6, 9–11e, 12a–12e, 13a–13e) in addition to ester lipids, whereas ether lipids are absent in the genera Thermosipho and Fervidobacterium (Table 2). For example, the HPLC/APCI–MS results showed only the presence of tetraesters in F. pennivorans, whereas the Thermotoga species investigated showed a wide variety of mixed ester/ether membrane spanning core lipids (Figs. 5, 7, 8). This suggests that only Thermotoga species possess enzymes to produce ether linkages. This classification based on lipid biochemistry is in good agreement with the phylogenetic classification based on the sequence of the 16S rRNA gene (Fig. 9), which reveals that the genera Thermosipho and Fervidobacterium are quite distinct from the genus Thermotoga, although the phylogenetic diversity within the Thermotoga cluster is also substantial.
Fig. 9Phylogenetic tree based on the 16S rRNA gene outlining the relationship between molecular phylogeny and lipid composition for the studied species of the Thermotogales. The neighbor-joining tree was reconstructed from distance matrices and Escherichia coli served as outgroup. The bar indicates an evolutionary distance of 0.10. Since no 16S rRNA gene sequence is available for F. penniforans, two other Fervidobacterium species were used in the construction of the tree. Gene Bank accession numbers are given in parentheses
The composition of the low-molecular-weight lipids (i.e. fatty acids, glycerol monoethers 3 and 4), higher molecular-weight lipids (1–2, 5–6) and the membrane-spanning lipids suggest a biosynthetic relationship between these different lipid classes although the relatively high concentrations of lipids 1–6 also suggest a direct functional role. It has been demonstrated that diabolic acids are produced from condensation reactions of fatty acids at the ω-1 positions (Fitz and Arigoni 1992). These authors showed by the use of [16-2H3]palmitic acid and [14-2H2]palmitic acid that unsaturated fatty acids are involved in this condensation reaction and that it may be mediated by a free radical reaction. In good agreement with the proposed biosynthetic pathway for diabolic acids is that in all species investigated in our study the C16:0 fatty acid is the major fatty acid and 15,16-dimethyltriacontanedioic acid (1), composed of two condensed C16:0 fatty acid moieties, is the major diabolic acid (Table 2). In cases where the C34 and C30 (2) diabolic acids are higher in concentration (e.g. F. pennivorans, and T. lettingae and Thermosipho sp., respectively) also a concomitant increase of the required monomer, i.e. the C18:0 and C14:0 fatty acids, respectively, is observed (Table 2). This is also seen in the presumed biosynthetic end-products, i.e. the membrane-spanning core lipids. For example, T. lettingae contains relatively more of these lipids based on C30 (2) diabolic acids, whereas in F. pennivorans the C34 diabolic acid is a more important building block. This is consistent with the published observation that during growth of T. maritima in batch culture the concentration of the C16:0 fatty acid decreased with a concomitant increase of tentatively identified diabolic acid (Zhang et al. 2002). This was interpreted to reflect the result of synthesis of diabolic acid using the C16:0 fatty acid. The same observations can be made for the ether lipids only encountered in the Thermotoga species. The presence of mono alkyl glycerol ethers in all Thermotoga species (Table 2) and mixed glycerol ether/esters in T. lettingae (2nd batch) (Fig. 4a) suggests that formation of the ether bond may also occur before the condensation reaction, resulting in the synthesis of membrane-spanning core lipids, takes place. In archaea, GDGTs are also thought to be produced from (partial) condensation of two glycerol diethers (Nishihara et al. 1989; Kon et al. 2002; Koga et al. 1993).
Generally, the distribution of the membrane-spanning lipids is quite complex in the species from the Thermotogales. It is noteworthy that several isomers with a quite different polarity were detected in the HPLC traces. The structural composition of the diacyl glycerols and the mixed glycerol ester/ethers in T. lettingae perhaps provides a clue to the observed complex composition of the membrane-spanning core lipids. Our analyses revealed that the ester moiety in these lipids can be at both at position 2 and 3 of the glycerol unit when the ether or ester moiety is at position 1 (Fig. 4a). It is quite likely (although this remains to be proven) that this also holds for membrane-spanning core lipids. This introduces a wide variety of structural isomers, which may explain why more than one peak was detected in the various HPLC mass chromatograms (Figs. 5, 6 and 8).
Our data show that some members of the Thermotogales (i.e. the Thermotoga genus) produce membrane-spanning lipids which are comprised of mixed glycerol ether/ester derivatives, whereas all examined species produce membrane-spanning diabolic acids. Mixed glycerol ether/ester lipids have been increasingly recognized in the Bacterial Kingdom, e.g. in Planctomycetes capable of anaerobic ammonium oxidation (Sinninghe Damsté et al. 2002; Sinninghe Damsté et al. 2005), in deep-branching thermophilic bacteria such as Aquifex (Huber et al. 1992) and in two mesophilic sulfate-reducing bacteria (Rütters et al. 2001). The tetraethers 11a and 12a, which occur in low amounts in T. maritima, are, to the best of our knowledge, the first example of this kind of lipid in bacteria. Non-isoprenoidal GDGTs have been detected before in peats (Sinninghe Damsté et al. 2000) but their biological origin is as yet unknown. Membrane spanning lipids are quite common in the Kingdom Archaea (Koga et al. 1993) but have also been identified in a very limited number of bacteria, e.g. in Butyrivibrio spp. (Clarke et al. 1980), and in Thermoanaerobacter species (Jung et al. 1994). It is tempting to associate the peculiar membrane biochemistry of the order Thermotogales to its position in the phylogenetic tree of life, i.e. close to the root of the tree (Woese et al. 1990). The unusually numerous archaeal-like genes (almost 25%) found in the Thermotoga maritima genome (Nelson et al. 1999) reflect that these are deep-branching bacteria. If species from the order Thermotogales occurred early during the evolution of life on Earth, as suggested by its position in the phylogenetic tree of life, our results indicate that the ability to produce both ether and ester glycerol membrane lipids developed relatively early during evolution. All Archaea produce ether membrane lipids, whereas Bacteria and Eukaryotes typically produce glycerol esters. Some members of the Thermotogales, thus, represent a hybrid in this way.
|
[
"diabolic acid",
"thermotoga",
"thermosipho",
"fervidobacterium",
"ether lipids",
"ester lipids"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Exp_Brain_Res-3-1-2039806
|
Judging surface slant for placing objects: a role for motion parallax
|
People have a variety of sources of information (cues) about surface slant at their disposal. We used a simple placing task to evaluate the relative importance of three such cues (motion parallax, binocular disparity and texture) within the space in which people normally manipulate objects. To do so, we projected a stimulus onto a rotatable screen. This allowed us to manipulate texture cues independently of binocular disparity and motion parallax. We asked people to stand in front of the screen and place a cylinder on the screen. We analysed the cylinder’s orientation just before contact. Participants mainly relied on binocular cues (weight between 50 and 90%), in accordance with binocular cues being known to be reliable when the stimulus surface is nearby and almost frontal. Texture cues contributed between 2 and 18% to the estimated slant. Motion parallax was given a weight between 1 and 9%, despite the fact that it only provided information when the head began to move, which was just before the arm did. Thus motion parallax is used to judge surface slant, even when one is under the impression of standing still.
Introduction
It is often important to accurately judge the slant of surfaces in our nearby environment. Whether placing our foot on the ground when we walk or climb stairs, or our fingers on an object when we grasp it and place it elsewhere, the interaction always involves making contact with surfaces. In order to interact successfully, we need to know the orientation of these surfaces. We have many ways to judge a surface’s orientation, including ones based on texture gradients, binocular disparity and motion parallax.
One important cue that contributes to most people’s slant perception is binocular disparity (see Howard and Rogers 1995 for an extensive review of the literature on binocular vision). The small differences between the images in the two eyes suffice to obtain information about the slant in depth. From the literature on grasping it could be inferred that binocular cues normally dominate our actions. For example, Servos and Goodale (1994) claim that binocular vision is the principal source of information for reaching and grasping movements. However, binocular information does not guarantee correct grasping (Hibbard and Bradshaw 2003), so there is reason to expect other cues to also play a role in guiding our actions.
A second cue that contributes to slant perception is the deformation of any surface texture and of shapes’ outlines as a result of perspective. We will refer to the combined information from all such sources as the texture cue. This is the cue that allows us to have a powerful and striking impression of surface slant from a flat image, as exemplified in Fig. 1. It is well known that surface texture provides valuable information on slant perception (Gibson 1950; Stevens 1981; Buckley et al. 1996; Landy and Graham 2004) and motor control (Knill 1998a; Watt and Bradshaw 2003).
Fig. 1The deformation of the regular texture of a chess board provides a profound impression that the surface is slanted relative to the plane of the 2D picture
Retrieving the slant of a surface from texture cues is based on the assumption that the distribution of texture elements over the surface is more or less even and that shapes are more or less symmetrical, see Rosenholtz and Malik (1997). This assumption holds for a wide variety of natural and artificial objects.
The third cue that we will consider is motion parallax (Rogers and Collett 1989; Rogers and Graham 1979, 1982; Ono and Steinbach 1990; Gillam and Rogers 1991; Ujike and Ono 2001). For a review of the widespread use of motion parallax in the animal kingdom, see Kral (2003). Movement of the head relative to a surface generates changes in the surface’s retinal image over time. These changes depend on the motion of the head relative to the items in the surrounding, and on the items’ relative distances. The latter dependency can be used to obtain information about depth and slant. Watt and Bradshaw (2003) have shown that motion parallax can guide human movements when binocular cues are not available.
When more than one cue is available the cues are combined. It is generally accepted that the estimated slant is a weighted average of the slants indicated by various cues, and that the weight of a cue is related to its accuracy, although some details of the mechanism are still under debate (Landy et al. 1995; Hillis et al. 2002; Hogervorst and Brenner 2004; Rosas et al. 2005; Muller et al. 2007). The relative contribution of binocular cues and texture cues depends on the distance (Hillis et al. 2004) and surface orientation (Knill 1998b; Buckley and Frisby 1993; Ryan and Gillam 1994), because binocular vision is better nearby and texture gradients change least rapidly with the angle of slant for near-frontal surfaces (for purely geometrical reasons). Knill (2005) measured how binocular cues and texture cues to surface orientation are combined to guide motor behaviour. Cue weights were found to be dependent on surface slant and also on the task: more weight was given to binocular cues for controlling hand movements than for making perceptual judgements. Knill used cue-consistent and cue-conflict stimuli in a virtual reality environment.
We wanted to find out whether motion parallax contributes to judgement of slant in the presence of other cues (such as binocular and texture cues) under more or less natural conditions. To do so, we used a setup in which the physical slant of a surface could be manipulated independently of the slant indicated by a pronounced texture. Thus a conflict between the texture cue and all the other cues was created by violating the assumption that the distribution of texture elements is homogeneous. The judged surface slant was determined by asking participants to swiftly place a flat cylindrical probe on the slanted surface (as in Knill 2005). Our main interest was in the extent to which head movements and the resulting motion parallax contribute to the perceived surface slant. So, conditions in which the head could move freely were compared with one in which a head restraint was used. We compared conditions with and without binocular vision in order to be able to evaluate the importance of motion parallax in relation to this cue.
Methods
Participants
Five people, four of whom were male, participated in the experiment. All participants gave their informed consent prior to their inclusion in the study. The experiment was part of an ongoing research program that was approved by the local ethics committee. All participants were right-handed, had normal or corrected-to-normal vision and good binocular vision (stereo acuity <40 arcseconds).
Experimental setup
Participants were standing upright in front of a large rotatable screen. A sketch of the setup is given in Fig. 2. The screen was a plexiglass plate covered with projection foil. Images were projected from below by an Hitachi cp-x325 LCD projector with a resolution of 1024 × 768 pixels. The screen and the projector could be rotated as a whole. Participants wore computer-controlled PLATO shutter-glasses, with which we could alternate between monocular and binocular vision, see Milgram (1987). A chessboard pattern was projected onto the screen (see Fig. 3). Slants were defined relative to the gravity-defined horizontal. A grey ring, 104 mm in diameter, indicated the target position for the probe. Two positions of the ring, one near the participant (the target’s centre 100 mm below the stimulus centre) and one further away (100 mm above the stimulus centre) were displayed in random order to make sure that participants did not simply repeat the previous movement. Figure 3 shows the stimuli with 0° and 10° texture slant relative to the surface’s physical slant. The geometry of the stimulus was calculated by projecting the texture-defined slant on the physically rotated surface. The projection is calculated from the point of the observer’s eyes. The 0° stimulus was a 40 cm square, with 40 cm corresponding to a visual angle of about 27°. A dark grey rim was drawn around the stimulus to mask any real edges that could become visible due to reflections within the set-up. To avoid illuminating objects around the set-up, the luminance of the image at the position of the eyes was limited to 0.4 Cd m−2. This was achieved by placing filters in front of the projector.
Fig. 2Participants were standing upright, facing the screen. They moved a probe from a starting position 50 cm to the right of the surface midline to a target position on the surface (indicated by the grey ring). Participants wore PLATO glasses with which we could switch between no, monocular and binocular vision. During the experiment, only the slanted surface was visibleFig. 3Examples of consistent (left) and conflict (right) images. The left panel shows the 0° texture slant on a 0° surface slant as seen from above. The right panel shows the 10° texture slant on the 0° surface slant. The deformation of the stimulus is consistent with the actual viewing geometry but for clarity the image is presented as seen from above. The target (grey ring) indicates the position at which participants have to place the probe. The targets are shown at the ‘near’ position. The dark grey rim’s shape is in accordance with the texture cue
The probe was a flat cylinder (diameter = 104 mm, height = 22 mm, mass = 0.2 kg). Movements of the probe were registered by an optotrak 3020 system (northern digital inc., Waterloo, ON, Canada). This system tracks the position of active infrared markers with an accuracy better than 0.5 mm. The 3D-positions of five markers on the probe were tracked at a rate of 200 Hz. The position, orientation and velocity of the probe were calculated from these data.
Procedure
Experiments were performed in a completely dark room. The dark environment and the low intensity of the stimuli ensured that there was no visible external reference frame. Participants were instructed to place the probe at the indicated target position on the surface. They were to start moving as soon as the target was visible. Stimuli were shown for 2.5 s. All movements were completed well within this interval.
In order to avoid dark adaptation a bright lamp was turned on for 5 s immediately after each trial. During this period participants placed the probe at the starting position, 50 cm to the right of the midline of the screen. Then the light was turned off for about 5 s, during which time the experimenter adjusted the orientation of the screen in preparation for the next trial.
Figure 4 shows the six combinations of surface slant and texture slant that were used in the experiment. The different combinations could be viewed monocularly or binocularly. The four consistent combinations (on the diagonal in Fig. 4) were used as a standard. The physical surface slant (see Fig. 2) was −10°, 0°, 10° or 20°, and the image on the surface was as shown on the left in Fig. 3. The two conflict combinations either involved presenting the image shown on the right in Fig. 3 on a horizontal surface (surface slant 0°; texture slant 10°) or presenting a similarly transformed image (slanted in the opposite direction) on a surface with 10° slant. All six stimuli were presented under various conditions. In total there were four experimental conditions, each consisting of 192 trials, divided into 3 blocks of 64 trials. Within every condition 75% of the trials were without conflict and 25% involved a conflict between texture and the other cues present in that condition.
Fig. 4The six combinations of physical surface slant (continuous lines) and slant suggested by texture (dashed lines) that were used in the experiment. In four cases there was no conflict between the cues (solid disks). In two cases there was a conflict (open disks). The slants have been exaggerated for clarity
We used four conditions in which different combinations of the available cues were presented. The choice of conditions will become clear when we describe the data analysis. We chose three conditions with which we could calculate the five parameters of our model, and one condition to test one of our assumptions.
In the ‘binocular’ condition viewing was binocular and head-free in both conflict and consistent trials. In this condition all cues to slant perception were available. In the ‘monocular’ condition the conflict and consistent trials were both presented monocularly and head-free. Stimuli were viewed with the left or right eye in random order. No binocular cues were available, but all other cues were present. In the ‘biteboard’ condition the head was fixed in combination with monocular viewing. The biteboards were made individually with an impression of the participant’s teeth. The biteboard severely limits head movements, removing information from motion parallax.
In the consistent trials of the ‘mixed’ condition the screen was viewed monocularly (75% of all trials), but in the conflict trials (25%) it was viewed binocularly. This condition was included to evaluate whether participants adapt their strategy at the level of a session rather than per trial. In the ‘binocular’ condition binocular information was always reliable, so participants could have learnt to use this cue. In the ‘mixed’ condition, in contrast, binocular cues were absent in the majority of trials, so participants could have learnt to use texture or motion parallax. Note that the 25% binocular, conflict trials in the ‘mixed’ condition are identical to the conflict trials in the ‘binocular condition’, whereas the 75% consistent trials are identical to the consistent trials in the ‘monocular’ condition. Thus the ‘mixed’ condition serves as a control condition to test whether the weight given to the cues stays about the same under changing viewing conditions on other trials.
Analysis
During some moments of some trials the participant’s fingers or hand occluded one or more of the five markers. The position of each marker relative to the centre of the probe is known, so the position and orientation of the probe can be calculated from any set of at least three markers. Frames in which fewer than three marker positions were known were not analysed. Figure 5 shows a schematic side-view of the path followed by the probe. The end of the movement was defined as the first sample at which the centre of the probe was less than 2 mm from the screen. The probe orientation was averaged over all samples at which the centre of the probe was between 100 and 20 mm from its position at the end of the movement (the grey line segments in Fig. 5). The last 20 mm of the path were excluded to avoid considering moments at which the edge of the probe could be in contact with the real surface of the projection screen.
Fig. 5The upper left panel gives a schematic side view of the probe’s path (curved thin line) towards the slanted surface (straight thick line). The position and orientation of the probe were measured by the optotrak system. The upper right panel shows a side view of a few paths towards the ‘far’ target position. The average probe orientation is determined during the last 100 to 20 mm before the end of the movement (indicated schematically by the dark-grey line segments). The lower panels show examples of the probe orientation with time intervals of 25 ms
To understand our method for determining the cue weights, consider the two extreme hypothetical outcomes shown in Fig. 6. At the one extreme, if information from the texture cue is not used at all, probe orientations will always follow the slant of the physical surface (left panel). At the other extreme, if the observer only relies on texture cues the orientation of the probe will follow the texture-defined slant (right panel). In the latter case the line connecting the cue-conflict conditions has approximately the opposite slope of that for consistent conditions. Note that the probe orientations in the consistent trials may differ from the physical surface slant (grey line). Some flattening may arise because the hand may still be rotating towards the surface slant during the last 100 mm of its trajectory. Participants may also rely to some extent on previous slopes that they encountered during the experiment. We will model these effects as a prior for a single surface slant for all conditions.
Fig. 6A schematic depiction of possible results for two extreme cases: probe orientation is not affected by texture at all (left panel) or only depends on the texture slant (right panel)
We assume that the weights that participants give to the different cues are the same in all conditions. This would for instance be so if observers base the weights on the reliability, as in optimal cue combination (Landy et al. 1995). We can therefore model the estimated slant (S) as a weighted average of the slant estimated from each of the available cues (si), which are all assumed to give veridical estimates except for the prior for a fixed slant (as mentioned in the previous paragraph):where wi is the weight (in arbitrary units) given to each available cue, with i ∈ {b, m, t, r, p} indicating binocular vision, motion parallax, texture, a rest category containing any other valid cues, and the prior. Note that we predict that the cue weights (wi) of all available cues will be the same in all conditions, although the relative weight given to a cue (wi/∑wi) will differ between conditions because it depends on the cues that are available in that condition.
The slant of the prior is a constant. Its value is likely to be near the mean of the slants in all previous conditions, but this is not essential for our analysis:
If s denotes the simulated slant, then the slants indicated by all other available cues are: except for the conflict trials, for which the texture differs 10° from all other available cues (see Fig. 4): Combining these equations gives: and
The sums in Eqs. 5 and 6 are only over cues that were available in each condition. The slopes (β) of the regression lines in Fig. 6 are given by the first derivatives of the estimated slant: Combining Eqs. 7 and 8 yields: and
Equations 9 and 10 apply to all experimental conditions as modified for cue availability, except for the ‘mixed’ condition. In the ‘binocular’ condition all cues yield information about the surface’s slant:
Similar equations can be written for the other conditions. Binocular information is not available in the ‘monocular’ condition, so wB does not occur in the equation:
Similarly, motion parallax ceases to contribute to the estimated slant in the ‘biteboard’ condition, giving:
Since the weights are in arbitrary units, we are free to define them in such a way that the sum of all weights is one:
First we determine the weight of the prior relative to that of texture (wP/wT) for each condition, using Eq. 10. Then, the weighted average of these ratios is calculated for each participant. We found no clear evidence that the assumption that wP/wT is the same across conditions was not justified. Next, determining the ratios between the slopes in the ‘binocular’, ‘monocular’ and ‘biteboard’ conditions from our data allows us to use Eqs. 11 to 14 to determine the values of the weights (wB, wT, wM and wR).
The data of the ‘mixed’ condition was analysed by comparing the slopes (βconflict and βconsistent) with the matching slopes in the ‘binocular’ and the ‘monocular’ conditions. Data for ‘near’ and ‘far’ target positions were pooled before calculating the slopes. The weights of the cues were calculated for individual participants. In addition, we also pooled the data of all participants before calculating the slopes, which yields the weights for ‘All’ participants.
Results
We determined average probe orientations for each participant and condition. The slopes of probe orientation as a function of surface orientation for conflict and consistent conditions enable us to calculate the cue weights, as explained in the ‘methods’ section.
Conditions
The upper panel of Fig. 7 shows the probe orientations in the binocular condition. Conflict (open symbols) and consistent (closed symbols) probe orientations are almost the same. The difference between the slopes is small but significant (P < 0.05). The second panel of Fig. 7 shows the probe orientations in the monocular condition. The slopes clearly differ between the conflict and the consistent trials (P < 0.01). In the ‘biteboard’ condition (third panel of Fig. 7) the conflict trials have almost the opposite slope than the consistent trials. The difference between the slopes is significant (P < 0.01). The cue weights were calculated using Eqs. 10 to 14 and the values of the regression slopes.
Fig. 7Pooled data for all participants. Each panel is for one of the four experimental conditions. Average probe orientations are shown for surfaces with (open symbols) and without (closed symbols) conflicts between real slant and texture slant. The error bars show the overall standard deviation. Correct probe orientations for the consistent trials (grey lines), linear regression (black lines) and regression slopes, β (numbers in lower right corner) are also shown. Asterisks indicate whether the difference between the regression slopes is significant (*P < 0.05; **P < 0.01)
The ‘mixed’ condition was included as a control to ascertain that the cue weights do not depend on the viewing conditions on other trials. We compared the monocular consistent trials with the identical trials in the ‘monocular’ condition, and the binocular conflict trials with the identical trials in the ‘binocular’ condition. The regression slope for monocular, consistent trials (βconsistent = 0.72) is not significantly different (P > 0.05) from the same trials in the ‘monocular’ condition (βconsistent = 0.76). The slope of the (binocular) conflict trials in the ‘mixed’ condition (βconflict = 0.61) is slightly but significantly lower (P = 0.04) than for conflict trials in the ‘binocular’ condition (βconflict = 0.72). Thus the weights may not be completely independent of the conditions. The difference was small enough to accept the calculation of the weights on the basis of the assumption that the condition is irrelevant. Our analysis may however underestimate the weight given to the texture cue.
Cue weights
The weight of the binocular cue lies between 50 and 90%, for individual participants (see Fig. 8). The average standard error is 11%. The errors in the cue weights are calculated by the method of propagation of errors based on the errors in the regression slopes. If we determine the slopes across all participants, the binocular weight is 71 ± 6%. The weight of the texture cue lies between 2 and 18% with an average standard error of 3% for individual participants. The weight given to the texture cue is 8 ± 1% across all participants. The weight given to motion parallax lies between 1 and 9% for individual participants with an average standard error of 6%. The weight given to motion parallax is 8 ± 3% across all participants. The weight attributed to the rest category of cues was only 3 ± 2% across all participants. The prior contributed between 6 and 23% for individual participants with an average standard error of 3%. The weight of the prior varies between 7 and 13% across conditions. Across all participants the weight of the prior is 10 ± 2%.
Fig. 8The weights given to the binocular, texture, motion parallax, other cues and to the prior, for each participant (horizontal axis). The weights are obtained by substituting the slopes of conflict data and consistent data in Eqs. 10 to 14. For ‘All’, first all participant’s data in each condition was pooled and then the weights were determined, which are all significantly different from zero. For individual participants the texture cue was not significantly different from zero in one case, the motion cue in four cases and the rest category in three cases
Head movements
Head movements were measured in the monocular viewing condition for three of the participants. Participants move their head considerably when placing the cylinder: EB moved on average 103 mm, JG 53 mm and DdG 37 mm in the lateral direction. Interestingly, the head movement only started just before the arm movement. Shortly before (100 ms) the onset of arm movement (when the probe was 10 mm from the starting position), the head had only moved 10 mm (EB), 4 mm (JG) or 6 mm (DdG). Thus the information from motion parallax is mainly picked up during the arm movement. None of the participants were aware of having made head movements.
Discussion
We used a physically rotatable screen as a surface. The projected stimulus was viewed in a completely dark environment within a space in which objects are normally manipulated. In our analysis systematic deformations that affect a single cue (like depth compression resulting from an erroneous depth estimate) were not considered. Moreover, we assume that the cue weights are the same in all conditions, so that their contributions to the percept only depend on which cues are available. We determined the contributions of binocular disparity, texture cues, motion parallax, a rest category and a prior. Under these conditions and based on these assumptions we conclude that participants mainly relied on binocular information (between 50 and 90%). Texture cues contributed between 2 and 18% to the estimated slant. Motion parallax contributed up to 9%. The prior contributed between 6 and 23%. Residual cues may account for up to 9%.
Comparing conditions with and without head movement revealed that motion parallax plays a role in slant perception. This is evident from the weights (Fig. 8) but also from a comparison of performance in the ‘monocular’ and ‘biteboard’ conditions (Fig. 7). It is not unusual to move the whole body, including the head, when making large arm movements. Beside mechanical reasons for doing so we here show that it may also have perceptual advantages.
We included in our analysis a rest category of cues that might contain information about slant that was not manipulated. The results suggest that this category indeed includes cues that yield some information about slant. Accommodation, or the rate at which the image becomes blurred with distance from fixation, might provide such information (Mather 1997; Watt et al. 2005). However, artefacts of our setup such as the possible visible micro texture (fibres in the projection foil or pixels on the screen) and the angular distribution of the light scattered from the surface could play a role too. Taken together in a rest category such cues contribute only a few percent to the estimated slant.
The probe orientations in the consistent trials in Fig. 7 are not equal to the physical surface slant. We incorporated a prior in our model to take into account behaviour that is not related to the instantaneous information, like visual or haptic information from previous trials. The slant indicated by the prior cue is a constant; i.e. it does not depend on the stimulus. Flatter slopes indicate that the prior plays a relatively large role. The weight of the prior is about as large as the weight of the texture cue. One component of the prior could be that the hand orientation is still changing towards its final value at the moment that we sample, which is slightly before contact (Cuijpers et al. 2004). Any biases towards a certain orientation of the hand or towards a certain perceived slant will also contribute to the weight of the prior.
The purpose of having the ‘mixed’ viewing condition was to check whether the weights change under different conditions. Ernst et al. (2000) have shown that haptic feedback can make more weight be given to a visual slant cue that is consistent with the feedback. In our study the haptic feedback was always consistent with the physical slant of the surface, so only the texture cue was sometimes unreliable. So, with a conflict between surface slant and texture the haptic feedback may yield a bias towards other cues than texture. Our results show a small difference between the ‘mixed’ condition and the comparable trials in the ‘monocular’ and ‘binocular’ condition. Thus here too the extent to which cues are used does probably depend to some extent on experience in previous trials. This indicates that the estimated slant does not only depend on the accuracy of the presented cues, as is often assumed in theories of optimal cue combination (Landy et al. 1995; Hillis et al. 2002; Muller et al. 2007 ). It is however possible that the difference arises from less use of motion parallax when binocular information was always available, perhaps because participants move less when there is enough information from other sources than motion parallax. We do not know whether this is the case because we did not measure head movements in all conditions. However, these differences are all too small to be taken seriously without further research.
In our study binocular disparity is given most weight, which is in accordance with binocular cues being known to be reliable when the stimulus surface is nearby and almost frontal. Because experimental conditions were all in favour of binocular disparity, the role of motion parallax and texture cues is probably smaller here than in natural viewing conditions. Motion parallax and texture cues both contribute to a small but significant extent to slant perception, although marked differences between participants were observed. Motion parallax was available only shortly and began relatively late, as the head began to move only just before onset of the arm movement. From animal studies it is known that a range of animals gain depth information by moving from side to side just before the performance of an action, see Kral (2003). We have shown that humans are able to use motion parallax during an action. It was known that monocular depth information can be used to guide our actions (Marotta et al. 1998; Dijkerman et al. 1999; Watt and Bradshaw 2003). It was not known, however, that motion parallax plays a role under conditions where other cues are dominantly available and without actively moving one’s head before starting the action. We conclude that motion parallax is used as a cue to manipulate objects in our nearby environment, even when one is under the impression of holding one’s head still. Motion parallax should therefore not be ignored in a ‘static’ task unless the head is really fixated.
|
[
"motion parallax",
"binocular",
"texture",
"cue integration",
"cue conflict"
] |
[
"P",
"P",
"P",
"M",
"R"
] |
Arch_Dermatol_Res-4-1-2254657
|
Manifestation of palmoplantar pustulosis during or after infliximab therapy for plaque-type psoriasis: report on five cases
|
Infliximab is a monoclonal antibody directed against TNF-α. It has been approved for use in rheumatoid arthritis, ankylosing spondylitis, inflammatory bowel disease, psoriatic arthritis and plaque-type psoriasis. In case reports, positive effects on pustular variants of psoriasis have also been reported. However, paradoxically, manifestation of pustular psoriasis and plaque-type psoriasis has been reported in patients treated with TNF antagonists including infliximab for other indications. Here, we report on 5 patients with chronic plaque-type psoriasis who developed palmoplantar pustulosis during or after discontinuation of infliximab therapy. In two of the five cases, manifestation of palmoplantar pustulosis was not accompanied by worsening of plaque-type psoriasis. Possibly, site-specific factors or a differential contribution of immunological processes modulated by TNF inhibitors to palmoplantar pustulosis and plaque-type psoriasis may have played a role.
Introduction
Increased expression of TNF-α has been identified as an important pathophysiological mechanism in different types of chronic inflammation including psoriasis and psoriatic arthritis. The knowledge about the central role of TNF-α in certain diseases has successfully been converted to the therapeutic level; over the last years TNF antagonizing agents such as the anti-TNF directed antibodies infliximab and adalimumab and the fusion protein etanercept have substantially improved the treatment of some of the most common chronic inflammatory conditions such as rheumatoid arthritis (RA). With 80% of the patients reaching at least a 75% reduction of their baseline skin symptoms after 10 weeks of therapy [14], infliximab is regarded as one of the most potent agents available at present for the treatment of psoriasis vulgaris. Other variants of psoriasis, such as pustular psoriasis have not been formally tested in larger clinical trials. However, in individual reports, positive effects of infliximab on generalized pustular psoriasis (GPP) with dramatic improvements in some cases have been reported [3, 13, 15]. Because this type of pustular psoriasis is believed to correspond to an extreme activation of psoriatic disease mechanisms, the high bioavailability of infliximab and its rapid onset of action following intravenous infusion have been used to explain the surprisingly fast decrease of pustule formation in patients treated with the agent. Paradoxically, on the other hand, manifestation of pustular psoriasis has been reported in patients treated with TNF antagonists including infliximab for other indications.
Palmoplantar pustulosis (PPP) has long been regarded as a localized variant of pustular psoriasis, although more recent epidemiological and genetic findings argue against this concept [2, 12]. Here, we report on five patients with chronic plaque-type psoriasis who developed PPP during or after discontinuation of infliximab therapy.
Cases and discussion
Relevant aspects of the five patients with chronic plaque-type psoriasis who developed PPP during or after discontinuation of infliximab therapy are presented in the Table 1 and Fig. 1.
Table 1Description of casesCase12345SexMaleMaleFemaleFemaleMaleAge3742673041Disease duration (years)17 2830Unknown15Psoriasis typePlaquePlaquePlaquePlaquePlaquePrior pustular psoriasisNoNoNoNoNoPsoriatic arthritisYesNoNoNoYesFamiliy history of psoriasisNegativePositiveNegativeUnknownNegativePrior UV-therapyYesYesYesYesYesPrior systemic therapiesMethotrexateLeflunomideEtanerceptPrednisoloneFumaric acid estersCyclosporineFumaric acid estersFumaric acid estersCyclosporineRetinoidsMethotrexateFumaric acid estersCyclosporineMethotrexateInitial response to infliximabExcellentcPASI 75fPASI 75PASI 75PASI 75Time of manifestation of PPPWeek 38d8 weeks after end of infliximab treatmentWeek 3dWeek 22dWeek 40dConcomitant worsening of plaque psoriasisaYesYesNoNoYesDevelopment of GPP in addition to PPPYesNoNoNoNoPotential trigger factors of PPP Infection prior to PPPYesNoYesNoNo Present smokingNoNoYes (36 packyears)UnknownUnknownDiscontinuation of infliximabYesYesYesNoYesSystemic therapy with sufficient control of PV and PPPbAdalimumab 40 mg e.o.wEtanercept 25 mg BIW plus methotrexate 7.5 mg/week orallyPUVA-therapy of palms and soles(Additional topical therapy)Etanercept 50 mg BIWSystemic therapy that failed to control PV or PPPEtanercept 50 mg BIWe plus methotrexate 15 mg/week orallyEtanercept 25 mg BIWPPP Pustulosis palmoplantaris, GPP Generalized pustular psoriasis, PV Psoriasis vulgaris, e.o.w Every other weeka Loss of >50% of maximum PASI response or increase of physician’s global assessment (PGA) by ≥ 2b All patients received additional topical therapy with glucocorticosteroids and Vitamin D analoguesc Improvement rated by PGA with “almost clear” (PGA = 1)d Week of infliximab treatmente BIW = twice weeklyf PASI 75 = Reduction in the psoriasis area and severity index (PASI) by ≥75%Fig. 1Clinical picture of pustulosis palmoplantaris in patient 3 with pustules in different stages of evolution on a sharply delineated erythematous lesion on the left sole (a) and yellowish pustules on the left palm (b). Histological examination showing intraepidermal vesiculopustular dermatitis (c, H.E. stain of a biopsy from the left plantar arch) with intraepidermal accumulation of neutrophils and subcorneal pustule formation (d)
To the best of our knowledge, the development of PPP during the treatment of plaque-type psoriasis with infliximab has not yet been reported. The occurrence of pustular skin lesions usually resembling GPP or palmoplantar pustular psoriasis has occasionally been observed in patients treated with infliximab for other indications [1, 6, 11, 16–19]. Induction of pustular skin lesions seems not to be limited to infliximab therapy, but has also been described in association with the use of the TNF-antagonists etanercept and adalimumab, including the use in one patient with plaque-psoriasis treated with etanercept [4, 8–10, 16]. One patient with seropositive RA developed GPP as well as PPP during treatment with infliximab [11]. This patient later experienced a relapse of PPP when treatment with etanercept was initiated, which also suggests that a class effect of TNF-antagonists may play a role.
In two of the three cases in whom an exacerbation of plaque-psoriasis occurred parallel to the manifestation of PPP, typical trigger factors for active psoriasis could be identified such as an infection (case 3) and the abrupt termination of anti-psoriatic treatment (case 2). These two cases are compatible with the existence of common trigger factors for plaque psoriasis and PPP. What are other factors that might contribute to the development of PPP during treatment of psoriasis vulgaris? While the exact etiology of PPP remains to be established, a history of smoking is the most important known risk factor for PPP. However, only one out of the three patients in whom a smoking history had been obtained was a smoker at the time of onset of pustular psoriasis (case 3). Streptococcal infection, a known risk factor for psoriasis vulgaris, has not been established as a risk factor for PPP and probably plays a minor role there. However, in the cases presented here, one patient (case 3) suffered an upper respiratory tract infection a few days before manifestation of PPP, while another patient (case 1) had suffered from a persistent cold 6 weeks before manifestation of pustules. In the former patient, the close temporal relationship between infectious symptoms and manifestation of PPP may point to a possible contribution of the infection to triggering PPP, and a modulation of the immune response to infliximab appears possible.
It is likely that beyond the contribution of known risk factors, other, immunological mechanisms may be involved in the manifestation of PPP under infliximab therapy. Interferon (IFN)-α has been suggested as a cytokine mediating the manifestation of psoriasiform lesions in patients treated with TNF-inhibitors as a consequence of crosstalk of TNF-α and IFN-α: TNF-α is known to suppress the generation of plasmacytoid dendritic cells that are very potent producers of IFN-α. Appearance of plasmacytoid dendritic cells (and IFN-α) in ths skin is considered to be an early and crucial step in the pathogenesis of psoriasis (reviewed in [7]). Thus, in patients treated with TNF-antagonists, the inhibition of TNF-α might induce an increase of IFN-α in the skin favoring the manifestation of psoriasiform dermatitis. In fact, an increase of IFN-α signaling has been shown in biopsy specimens from psoriatic plaques induced by TNF-inhibitors compared with traditional psoriatic plaques [6]. The relevance of IFN-α for PPP and TNF-inhibitor induced PPP, however, still needs to be determined. The observation of an improvement of pre-existing psoriasis plaques parallel to the first manifestation of PPP in two of the cases described here supports the concept that immunological mechanisms and/or local factors are not identical in the pathogenesis of plaque-type psoriasis and PPP. Differences in pathogenesis between plaque psoriasis and PPP are also supported by their different genetic background, with plaque psoriasis, but not PPP being linked to PSORS1, the major susceptibility locus for plaque-type psoriasis located on 6p21 [2]. The localized nature of the pustules on palms and soles occurring in association with infliximab therapy in some patients suggests a contribution of site-specific factors. Possibly, eccrine sweat glands that are numerous in the palmoplantar location are involved in these processes. Changes secondary to binding of infliximab to TNF-α, that has been shown to be expressed in eccrine sweat glands, have been suggested as a possible mechanism [11].
Withdrawal or dose reduction of systemic cyclosporine and systemic or topical glucocorticosteroids given for treatment of psoriasis vulgaris have been associated with manifestation of GPP, but not with PPP. To the best of our knowledge, no reports on manifestation of GPP or PPP associated with treatment with methotrexate, another drug frequently used in the therapy of psoriasis, exist. However, in the clinical trials with the CD11a-antagonist efalizumab in the indication plaque psoriasis, development of PPP has also been observed in 0.2% of 1,620 patients during the first 12 weeks of treatment with efalizumab, compared to none of the 715 patients in the placebo groups [5].
Whether therapy with TNF-antagonists may induce a more severe form of plaque psoriasis or psoriatic arthritis in some patients treated for these indications is difficult to determine, as the natural course of both plaque psoriasis and psoriatic arthritis is variable and a loss of effect is not easily differentiated from an induction of a more severe form of the respective disease. However, as a manifestation of plaque psoriasis has been noted in a number of patients treated with TNF-antagonists for other indications without a personal or family history of psoriasis [6, 16], it is likely that TNF-antagonists may in some individuals and probably in conjunction with environmental factors favor manifestation or worsening of psoriatic skin and/or bone disease.
The treatment of patients with plaque-psoriasis developing PPP should be decided on an individual basis. In case plaque-psoriasis remains controlled under infliximab therapy, addition of topical therapy may be sufficient for treatment of PPP in some instances (case 4). However, in some cases, additional topical therapy may not be sufficient and UV-therapy or another systemic therapy may be necessary in addition to or as replacement of infliximab therapy (case 3). When manifestation of PPP is accompanied by worsening of plaque-psoriasis, discontinuation of infliximab therapy is advisable and initiation of other systemic antipsoriatic agents, such as cyclosporine or a different TNF-antagonist, may be necessary (cases 1 and 5).
In summary, pustular psoriasis may show a good response to treatment with TNF-antagonists such as infliximab. However, pustular psoriasis may also manifest during treatment of rheumatological diseases, and, as described here, also in patients with plaque-type psoriasis under treatment with infliximab. Manifestation of PPP under infliximab is not necessarily accompanied by worsening of pre-existing plaque psoriasis. Management of the pustular skin lesions has to be decided on an individual basis. In a subgroup of patients, therapy with the TNF-antagonist has to be discontinued and another systemic therapy (including a different TNF-antagonist) will probably be necessary to sufficiently control plaque-psoriasis and PPP.
|
[
"pustulosis",
"infliximab",
"psoriasis",
"tnf",
"antagonist"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Fam_Cancer-4-1-2253650
|
Hereditary diffuse gastric cancer: association with lobular breast cancer
|
Hereditary diffuse gastric cancer (HDGC) has been shown to be caused by germline mutations in the gene CDH1 located at 16q22.1, which encodes the cell–cell adhesion molecule, E-cadherin. Not only does loss of expression of E-cadherin account for the morphologic differences between intestinal and diffuse gastric cancer (DGC) variants, but it also appears to lead to distinct cellular features which appear to be common amongst related cancers that have been seen in the syndrome. As in most hereditary cancer syndromes, multiple organ sites may be commonly affected by cancer, in HDGC, lobular carcinoma of the breast (LBC) and possibly other organ sites have been shown to be associated with the familial cancer syndrome. Given the complexity of HDGC, not only with regard to the management of the DGC risk, but also with regard to the risk for other related cancers, such as LBC, a multi-disciplinary approach is needed for the management of individuals with known CDH1 mutations.
Introduction
Despite an overall decrease in the global incidence of gastric cancer (GC) [1], the incidence of the subtype, diffuse gastric cancer (DGC) has remained stable and may even be increasing [2]. Within the past ten years, germline mutations in CDH1, which encodes E-cadherin, have been found [3] in over 50% of hereditary diffuse gastric cancer (HDGC) families with at least two cases of GC, with one diagnosed as DGC before the age of 50 years [4]. Within these HDGC families, we and others have noted an overrepresentation of lobular breast cancer (LBC) [4–8]. This observation has led to efforts to determine whether or not CDH1 is a breast cancer susceptibility gene, distinct from its gastric cancer risk. Recently our group has reported a novel germline CDH1 truncating mutation (517insA) in an LBC family with no known history of GC [9]. Within this review we report a germline CDH1 mutation in a second family in which breast cancer is the predominant cancer diagnosis. The management of HDGC in all patients with a particular focus on the management of the breast cancer risk associated with germline CDH1 mutations will be discussed.
Methods
The described family was referred to the ongoing HDGC study at the British Columbia Cancer Agency from a cancer genetics clinic in Seattle, WA, USA. Informed consent was obtained from the proband by the referring genetic counselor following ascertainment of a detailed cancer family history and appropriate genetic counseling prior to germline mutation testing. Our laboratory carried out the molecular genetic testing for the CDH1 mutation on a research basis. Approval for the HDGC study is by the clinical research ethics board of the University of British Columbia.
The proband (IV-4) was diagnosed with widely metastatic lobular breast cancer at age 53 years (Fig. 1a). Her family, of European ancestry, had a history of breast cancer diagnoses occurring in an autosomal dominant fashion on the maternal side of the family where her mother, aunt, and first cousin developed breast cancer in their 50’s. Due to her high-risk pedigree BRCA1, BRCA2, and PTEN genetic testing was undertaken and all were negative. CDH1 testing was also pursued.
Fig. 1(a) Pedigree of family reported showing a predominance of breast cancer. (b) Sequence from family carrying 1565 + 1G→A mutation
Results
All 16 exons were amplified for DHPLC analysis [6]. For exon 10 of CDH1, the initial amplicon failed and was therefore analyzed by direct sequencing and thus revealed a donor splice site mutation, 1565 + 1G > A (Fig. 1b). Due to its position at a donor splice site, this mutation is regarded as pathogenic [10].
The proband’s sisters (IV-2 and IV-6) participated in all aspects of the proband’s genetic consultation. They were appropriately concerned about their risk of breast cancer, but had not thought much about the possibility of getting gastric cancer until the CDH1 mutation was found. IV-2 and IV-6 had predictive genetic testing for the CDH1 mutation testing and both were found to be negative. Other family members are being informed about the availability of predictive genetic testing.
Discussion
E-cadherin
CDH1 (OMIM *192090), located on chromosome 16q22.1 encodes, E-cadherin, an epithelial transmembrane cell–cell adhesion molecule and member of the cadherin superfamily of glycoproteins. In a zipper-like fashion, its extracellular domain forms calcium-dependent homodimers between the E-cadherin molecules of adjacent epithelial cells, to act as the primary mediator of epithelial cell adhesion at the adherens junction complex [11]. Through interactions of its cytoplasmic tail with multiple signalling and structural molecules, such as the catenins, E-cadherin, maintains cellular adhesion and epithelial architecture with this link to the cytoskeleton. The cytoplasmic tail of E-cadherin directly associates with β-catenin and γ-catenin which in turn binds to the f-actin microfilaments of the cytoskeleton, directly or through α-catenin [12]. p120-catenin also associates with E-cadherin’s cytoplasmic tail at a different site, the juxtamembrane domain, and acts to both strengthen the adhesion between cells and regulate cadherin membrane trafficking and degradation [13, 14]. E-cadherin is considered to have an invasion suppressor role, where decreased expression permits cells to dissociate from each other in order to migrate and invade [15]. In cancers, this manifests as increased infiltrative and metastatic potential [16]. E-cadherin is also thought to act as a tumour suppressor, potentially through its interaction with the multipurpose β -catenin molecule which is an effector of the WNT signalling pathway [17]. Loss of E-cadherin can result in β-catenin release from the membrane and translocation to the nucleus where it complexes with Tcf/Lef-1 transcription factors to initiate transcription of WNT responsive genes [18]. Activation of these genes have been implicated in tumourigenesis through the WNT signalling pathway as seen in adenomatous polyposis coli (APC) [19]. In support of the role of E-cadherin as a tumour suppressor is the observation of abnormal or absent E-cadherin expression in precursor lesions of DGC and LBC, where the phenomenon is seen in in situ signet ring cell carcinomas found in prophylactic gastrectomy specimens from germline CDH1 mutation carriers [20] and the lobular carcinoma in situ lesions seen adjacent to invasive lobular breast cancers [21]. These examples suggest that loss of E-cadherin is an early or even tumour-initiating event, however the actual molecular basis of such a potential role of E-cadherin in such cases is unknown.
Inactivating CDH1 mutations are found in 50% of sporadic DGCs [22, 23] and cluster between exons seven and nine [11], in contrast with the low percentage of mutations seen in sporadic intestinal type GCs [23]. Decreased expression of E-cadherin in DGCs may account for morphologic differences between intestinal and DGC variants [24]. Unlike somatic CDH1 mutations, germline mutations associated with DGC are distributed throughout the gene [7] (Fig. 2). In the cancers from individuals with CDH1 mutations, CDH1 acts as a classic tumour suppressor gene with loss of expression of the wildtype allele [25, 26]. In a single study of 6 hereditary DGC cancers, inactivation of the wild-type allele could be attributed to promoter hypermethylation in 5 (83%) of cases [26]. This finding warrants verification in a larger cohort as abnormal promoter methylation in early cancers could potentially form the basis of a screening test.
Fig. 2DGC and LBC associated CDH1 germline mutations. Mutations shown above CDH1 gene schematic occur in families with DGC history and those below CDH1 occur in families with an additional or exclusive LBC history. In addition to the known CDH1 germline mutations compiled by Kaurah et al. [4], the recent mutation in an LBC family [9] and novel mutation from this paper are shown and identified below the symbol denoting mutation type. * Denotes the halfway point of the CDH1 coding sequence (1324 or the start of exon 10)
Lobular breast cancer and diffuse gastric cancer: loss of E-cadherin
Currently germline mutations in single genes account for approximately 5–10% of breast cancer [27]. High penetrance genes such as BRCA1 and 2 account for 3–8%, and TP53 and PTEN as seen in Li-Fraumeni and Cowden syndrome together only account for <0.1% of breast cancer diagnoses [28]. Other medium and low penetrance genes such as CHK2, BRIP1, PALB2 and ATM [29–32] have been identified, however, there still remains a proportion of hereditary breast cancer not yet determined. LBC accounts for approximately 10% of all breast cancers compared to the other major histologic subtype, invasive ductal carcinoma (IDC) [33]. Several factors suggest that LBC has a stronger hereditary basis relative to IDC, such as the higher frequency of bilateral disease [33], and also where excess familiality of LBC has been observed in population studies [34]. LBCs compose only 3% and 9% of the breast cancer tumour types seen in germline BRCA 1 and 2 mutation carriers, respectively [35], illustrating that the genetic risk factors for the majority of cases are unaccounted for by these genes.
The histology of LBC is characterized by infiltrative cancer cells which are isolated, highly dispersive and demonstrate a growth pattern with scattered and single files of tumor cells dispersed in stromal tissue [36]. This pathologic appearance is remarkably similar to DGCs and both LBC and DGC demonstrate characteristic mucinous, signet ring cells. This is not unexpected as E-cadherin staining is absent in 85% of sporadic invasive LBC [37] and somatic CDH1 mutations have been identified in 56% of sporadic LBCs [38]. Furthermore, in IDC, somatic CDH1 mutations are not found [38] and complete loss of E-cadherin expression is an uncommon feature. As loss of E-cadherin expression is a distinctive trait of both LBCs and DGCs, it likely contributes to the unique histopathologic features shared by the two cancers.
There are some differences with regard to the nature of the mutations seen in LBC and DGC. Generally mutations associated with sporadic LBC have been found to be nonsense or frameshift mutations [39] which encode truncated, non-functional proteins, whereas in sporadic DGC, mutations have generally been found to be splice site and in-frame mutations [11]. In sporadic LBC, mutations in CDH1 are spread throughout the gene [11] compared with the mutations seen in sporadic DGC which tend to cluster. Germline CDH1 mutations associated with DGC and/or LBC occur throughout the gene (Fig. 2). However, when the DGC and LBC associated CDH1 mutations are tabulated and compared based on their 3′ or 5′ positions relative to the halfway point of the CDH1 coding sequence (1324 or the start of exon 10), LBC associated mutations show a statistically significant trend towards clustering at the 3′ end (Fisher’s exact test, two-tailed P-value equals 0.0467) (Fig. 2). As this association is of weak statistical significance, it is unlikely to impact clinical testing strategies. Future analyses of novel germline LBC-associated CDH1 mutations should help to confirm this observation. Another difference between the molecular genetics of the two types of cancers, is that in sporadic LBC, silencing of E-cadherin expression is generally accomplished by a mutation in one allele in combination with loss of heterozygosity (LOH) or promoter hypermethylation in the remaining allele [40]. This is in contrast to sporadic DGC, where biallellic inactivation is achieved by mutations in one allele in concert with promoter hypermethylation in the other [41].
We have recently identified a truncating germline CDH1 mutation in an LBC family where analysis of the tumour was suggestive of partial LOH in the WT allele [9]. Our current case demonstrates a germline CDH1 mutation (1565 + 1G > A) in a predominantly breast cancer family, which is predicted to disrupt splicing. The mutation is in the same conserved position as a previously reported mutation (1565 + 1G > T) which was found in an Arabian HDGC family with no recorded history of breast cancer [42]. Moreover, a previous study reported a germline missense mutation in a proband with LBC but did not detail family history, or functionally characterize the missense mutation [43]. These examples demonstrate the need for further studies of germline mutations in LBC families in order to determine the mutation frequency and potential genotype-phenotype correlations.
Lobular breast cancer and HDGC
Breast cancer has been observed in HDGC kindreds to the extent where clustering of LBCs within HDGC families has led to the misclassification of families as breast cancer kindreds who test negative for BRCA1/2 mutations [4]. In 1998, Keller described the first case of histologically defined LBC in association with HDGC [5]. Since then, several more HDGC families with associated breast cancer were reported where it was observed, that these cases were LBCs when pathology was available [4, 6–8].
Prior to establishment of the association between HDGC and LBC, several efforts to determine whether CDH1 was a breast cancer susceptibility gene were attempted in view of the well-recognised phenotype of loss of E-cadherin expression displayed by the breast cancer subtype. For various reasons these studies failed to demonstrate the link. Rahman et al. examined 65 cases of lobular carcinoma in situ, however did not pre-screen the cases based on family history and included a wide age range, from 26 to 71 years, not necessarily in keeping with the usual age of onset seen in hereditary cancer syndromes [44]. Salashor examined 19 breast cancer tumours exhibiting LOH at the CDH1 locus, however of those, only 3 were confirmed to be pure LBC or mixed LBC/IDC pathology [45]. Lei examined 13 familial LBC cases and found no mutations, however did not define the extent of the family history [46].
Penetrance data based on 11 HDGC families, estimated the cumulative risk for LBC for female mutation carriers to be 39% (95% CI, 12–84%) by 80 years of age [47]. More recently we have published an estimated cumulative risk for breast cancer for females by the age of 75 years as being 52% (95% CI, 29–94%) from analysis of 4 predominantly gastric cancer pedigrees from Newfoundland with the 2398delC CDH1 founder mutation [4]. This is with the caveat that LBC risk for CDH1 mutation carriers has been assessed within high risk HDGC families, leading to a potential ascertainment bias and underestimation of the role of CDH1 mutations in LBC development. To accommodate for this we have begun analysis of CDH1 mutations within familial lobular breast cancer families or those families ascertained through a relatively young index case with confirmed LBC and have found germline CDH1 mutations in these kindreds [9].
Clinical implications of CDH1 associated LBC risk
At this time, it seems reasonable to conclude that at least four groups of women are at increased risk for LBC: women with LBC and a family history of breast cancer, women with a known CDH1 mutation, women from families with diffuse gastric cancer in whom no CDH1 mutation has yet been identified; and women with a germline BRCA2 mutation. Since there has not yet been a large population based study of the prevalence of CDH1 mutations among women with lobular breast cancer, it is premature to recommend genetic evaluation to women with a family history of breast cancer unless, at the very least, one of the breast cancers can be shown to have been lobular. Additional research can be expected to provide better guidance for these families.
Although there are not yet definitive data available on surveillance or risk reduction programs for women with known CDH1 mutations or untested women from CDH1-positive families, the high lifetime risk of LBC (39–52%) [4, 47] that these women face mandates their careful management. We suggest that they follow the recommendations for other high-risk women with hereditary breast cancer predisposition. This subgroup should be advised to practice breast self-examination; and to have annual mammograms, and semiannual clinical breast examination, beginning at least by age 30. There is certainly interest in regular bilateral breast MRI, as lobular breast cancer are known to frequently elude mammographic detection because they do not form masses or develop calcifications. These women can also be counseled to consider hormonal chemoprevention, since most LBCs are estrogen receptor positive [33], and both tamoxifen and raloxifene reduce the risk of estrogen receptor positive [48, 49] breast cancers in randomized trials. In addition, the risk reduction was greatest with both agents in women with lobular carcinoma in situ [50].
Prophylactic mastectomy may also be considered an option by some CDH1-positive women, particularly those who have been previously diagnosed with breast cancer in one breast or those who have had to undergo multiple biopsies for abnormal clinical findings. Several studies have reported a 90% reduction in breast cancer incidence with prophylactic mastectomy among women with a strong family history or with a germline BRCA1 or BRCA2 mutation [51, 52]. The published series include some lobular breast cancers, but not at numbers sufficient to permit meaningful subset analysis at this time.
Management of hereditary diffuse gastric cancer
Penetrance studies examining data from HDGC families, have estimated the lifetime risk of developing gastric cancer by age 75 and 80 respectively, to be from 40–67% in men, to 63–83% in women [4, 47]. Although identification of germline CDH1 mutations has enabled a significant proportion of HDGC families to utilise predictive testing to determine their individual risks of GC within CDH1 mutation positive pedigrees, unfortunately screening for DGC is ineffective and the current recommendation is for consideration of prophylactic gastrectomy in mutation positive individuals. Positron emission tomography [53] and chromoendoscopic-directed biopsies [54] have been proposed over basic endoscopy as more sensitive means of screening carriers, however screening methods have been consistently undermined by the recurrent discovery of multifocal DGC lesions underlying normal mucosa in prophylactic gastrectomy specimens of individuals with recent negative screening [4, 55, 56]. Regardless of the current limitations of screening, it is currently recommended that consideration for genetic testing and screening begin in at risk individuals in the late teens or early twenties [4] and that prophylactic total gastrectomy be considered in the early twenties for mutation carriers. Female mutation carriers will need specialized counseling to the potential nutritional effects on pregnancy following gastrectomy [57]. Further studies are currently underway to examine the quality of life impact of prophylactic gastrectomies. In the case report herein, although there was a GC in the maternal grandfather, the family history was more striking for the large number of breast cancer cases. This highlights the particular challenges we currently face with regard to counseling these families which appear to be mainly breast cancer, as it is unknown if the penetrance of DGC in this family is as high as it is in other HDGC pedigrees.
Conclusion
HDGC is one of a number of hereditary cancer syndromes that feature both an increased breast and gastric cancer risk (Table 1). In general, a lack of shared genetic risks for most breast and GI cancers was suggested through a recent study of 13,023 genes in 11 breast and 11 colon cancer cell lines in which the only commonly mutated gene between these two cancer types is p53 [58]. This likely reflects underlying differences in the biology of these diseases, however also highlights the unique nature of germline mutations in the CDH1 gene which are strongly associated with specific histologically defined subtypes of breast and GI cancer, namely LBC and DGC which are both part of the HDGC syndrome.
Table 1Other syndromes with familial susceptibility to breast and gastric cancersSyndromeMode of inheritanceAssociated gene(s)Sites of primary cancer(s)Evidence for association with the syndrome BRCA2 Hereditary Breast/Ovarian CancerADBRCA2BreastOvaryLarynxProstateBC is considered an integral tumor of the syndrome with an average cumulative risk in carriers by age 70 years of 45% (95% confidence interval (CI) 31–56%) [59]Familial aggregations of both BC and GC have been reported [60]In a study of 173 families, relative risk for GC was 2.6 [61]Among Ashkenazi Jewish GC patients, the frequency of 617delT mutation is five times that of the general Ashkenazi Jewish population frequency [62] Among other cancers, GC occurred in first degree relatives when mutations were located in the ovarian cancer cluster region of exon 11 of BRCA2 [63] BRCA1 Hereditary Breast/Ovarian CancerADBRCA1BreastOvary ProstateBC is considered an integral tumor of the syndrome with an average cumulative risk in carriers by age 70 years of 72.8% (95% confidence interval [CI] = 67.9% to 77.7%) [64]There is a 4 times increased risk for GC [64]Peutz-Jeghers SyndromeADSTK 11Gastrointestinal (GI) tractGC is considered an integral tumor of the syndrome with a relative risk of 213 (95% confidence interval 96–368) [65] BC is considered an integral tumor of the syndrome with a relative risk of 15.2 (95% CI 7.6–27) [65] Cowden SyndromeADPTENBreastThyroidEndometriumBC is considered an integral tumor of the syndrome with an incidence of 22–50% [66, 67]GC in situ has been reported in a patient with Cowden Syndrome [68]Li-Fraumeni SyndromeADTP53CHK2BreastAdrenal cortexConnective tissueKidneyNervous systemPancreas White blood cellsBC is frequently found in families with this cancer susceptibility syndrome [69]Chompret expanded the spectrum of cancers to include GC [70]Germline TP53 mutations have been found in GC families without CDH1 mutations [71–73]The 1100delC CHK2 allele confers a 2.2 fold risk of BC to carriers [29], however germline mutations in GC kindreds have not been identified [74]Familial Adenomatous PolyposisADAPCColon and rectumDuodenumThyroidPancreasLiterature review by Shimoyama et al. totalled 30 reported cases of GC and FAP [76]47% to 49% of primary BCs had promoter hypermethylation at the APC locus [77, 78]23% of LBCs have been shown to have LOH of APC [43]Lynch Syndrome (Hereditary Nonpolyposis Colon Cancer (HNPCC))ADhMSH2hMLH1hMSH6hPMS1hPMS2Colon and rectumEndometriumStomach Small intestineUrotheliumKidneyOvarySkinPancreasBrainWhite blood cells Biliary tractGastric cancer accounted for 5% of cancers in families harboring MLH1 or MSH2 mutations [76]hMLH1 mutations in large kindred segregated with BCs exhibiting microsatellite instability (MSI) [79]A slight increased incidence of BC was seen in hMLH1 mutation carriers [80]Germline hMSH2 mutation carrier with BC exhibited LOH for hMSH2 in tumors analyzed [81]Analysis of primary invasive BCs demonstrated that 25% of tumours were immunonegative for MSH2 staining [82] Ataxia-telangiectasia (AT)ARATMWhite blood cellsMutations causing AT in homozygotes, confer susceptibility to BC in heterozygotes, where women with ATM mutations have a ∼2-fold risk of BC and ∼15% of these women will develop the disease [32]GC has been reported in association with the syndrome [83–85]There is evidence of excess risks of GC in heterozygotes (RR = 3.39, 95% CI = 0.86 to 13.4) [86]Xeroderma pigmentosumARXPAERCC3 (XPB)XPCERCC2(XPD)DDB2(XPE)ERCC4(XPF)ERCC5(XPG)POLH(XP-V)SkinEyesBC and GC have both independently been reported with the syndrome [87,88]Werner SyndromeARWRNConnective tissue SkinThyroidGC has been reported in association with the syndrome [89] There are no reports of BC in association with Werner syndrome. Although, there is evidence supporting WRN as a low-penetrance familial BC susceptibility gene, where patients harboring both WRN Cys1367Arg or TP53 MspI variants had an increased BC risk (OR = 3.39, 95% CI 1.19–9.71) [90] AD = autosomal dominant, AR = autosomal recessive
With the recent demonstration of a CDH1 mutation in a family ascertained through an index case of LBC and in view of the additional new mutation in a predominantly breast cancer family that we have described here, the evidence for establishing LBC as part of the HDGC syndrome is strong. There now is a need for establishing the prevalence of CDH1 mutations in LBC families to avoid the ascertainment bias generated from only looking at cases from families identified because of their family history of GC. It is not currently known what the risk of GC is in these families which present predominantly as having a susceptibility to breast cancer and therefore identification of CDH1 as a true susceptibility gene for LBC could result in CDH1 screening and effective risk reduction strategies for selected breast cancer families and further studies examining their risk for gastric and other cancers.
Most hereditary cancer syndromes are associated with cancer risk involving multiple organs. Here we have discussed germline CDH1 mutations and the risks with regard to DGC and LBC, however as the recognised spectrum of related cancers broadens, more affected families will be identified and successfully managed with regard to avoidance of specific cancer risks. Longer life expectancy in individuals with penetrant mutations could potentially lead to the development of different, later onset disease as yet to be identified in these kindreds. This represents a particular challenge in hereditary cancer practice as the clinical community tends to be segregated into organ specific specialties where as the cancer risks and the risk reduction strategies for germline mutation carriers require a variety of expertise. The medical needs of the HDGC families are therefore best served through an engaged multidisciplinary team.
|
[
"diffuse gastric cancer",
"lobular breast cancer",
"hereditary diffuse gastric cancer (hdgc)",
"cdh1 mutation",
"screening",
"prophylactic total gastrectomy",
"e-cadherin mutation"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"R"
] |
J_Abnorm_Child_Psychol-3-1-1915590
|
Associations of Maternal Prenatal Smoking with Early Childhood Physical Aggression, Hyperactivity-Impulsivity, and Their Co-Occurrence
|
This study investigated associations between maternal prenatal smoking and physical aggression (PA), hyperactivity-impulsivity (HI) and co-occurring PA and HI between ages 17 and 42 months in a population sample of children born in Québec (Canada) in 1997/1998 (N=1745). Trajectory model estimation showed three distinct developmental patterns for PA and four for HI. Multinomial regression analyses showed that prenatal smoking significantly predicted children’s likelihood to follow different PA trajectories beyond the effects of other perinatal factors, parental psychopathology, family functioning and parenting, and socio-economic factors. However, prenatal smoking was not a significant predictor of HI in a model with the same control variables. Further multinomial regression analyses showed that, together with gender, presence of siblings and maternal hostile reactive parenting, prenatal smoking independently predicted co-occurring high PA and high HI compared to low levels of both behaviors, to high PA alone, and to high HI alone. These results show that maternal prenatal smoking predicts multiple behavior regulation problems in early childhood.
The literature on maternal prenatal smoking and externalizing behavior problems is clearly separated in two developmental periods. The bulk of the current literature on maternal prenatal smoking and externalizing behavior problems covers either childhood and adolescence or the preschool years. However, the rates of common childhood psychiatric disorders such as conduct disorder (CD) and attention deficit hyperactivity disorder (ADHD) and the patterns of comorbidity among them in early childhood are similar to those seen in later childhood (Egger & Angold, 2006). Limitations in current diagnostic criteria for early childhood psychopathology have resulted in a research focus on specific behaviors such as physical aggression (PA) and hyperactivity-impulsivity (HI) rather than clinical disorders in this age group. Both physical aggression and hyperactivity during early childhood appear to be typical precursors of full-blown CD and ADHD during the school years and beyond (Séguin, Nagin, Assaad, & Tremblay, 2004). A substantial number of studies show continuity between early and late childhood externalizing behavior problems (e.g. Campbell, Breaux, Ewing, & Szumowski, 1986; Keenan & Wakschlag, 2000) or have identified consistent developmental trajectories for PA and HI that start as early as age 1 1/2 years (e.g. Côté, Vaillancourt, LeBlanc, Nagin, & Tremblay, 2006; NICHD Early Child Care Research Network, 2004; Romano, Tremblay, Farhat, & Côté, 2006; Shaw, Lacourse, & Nagin, 2005).
In reviewing childhood and adolescence studies, maternal prenatal smoking has consistently been associated with CD- and ADHD- symptoms (for reviews, see Linnet et al., 2003; Wakschlag, Pickett, Cook, Benowitz, & Leventhal, 2002). The associations have been observed in clinical samples (e.g. Mick, Biederman, Faraone, Sayer, & Kleinman, 2002; Milberger, Biederman, Faraone, Chen, & Jones, 1996), ‘at-risk’ samples (e.g. Wakschlag & Hans, 2002) and large population-based samples (e.g. Braun, Kahn, Froehlich, Auinger, & Lanphaer, 2006; Kotimaa et al., 2003). Whereas these studies generally controlled for familial psychopathology and numerous environmental risk factors, there are also indications from studies using (large) twin samples that maternal prenatal smoking predicts children’s CD- and ADHD-symptoms beyond the effects of heritable risk (e.g. Button, Thapar, & McGuffin, 2005; Maughan, Taylor, Caspi, & Moffitt, 2004; Silberg et al., 2003; Thapar et al., 2003). Studies showing associations between maternal prenatal smoking and CD-symptoms and delinquency during adolescence and adulthood (Brennan, Grekin, Mortensen, & Mednick, 2002; Fergusson, Woodward, & Horwood, 1998; Räsänen et al., 1999) and studies showing behavior problems in children of mothers who smoked during pregnancy at different time points (e.g. Maughan et al., 2004; Wakschlag & Hans, 2002; Wakschlag, Pickett, Kasza, & Loeber, 2006) suggest long-lasting effects.
Although there are indications that maternal prenatal smoking is more strongly or even exclusively related to CD symptoms (Wakschlag & Hans, 2002; Wakschlag, Leventhal, Pine, Pickett, & Carter, 2006; Wakschlag, Pickett et al., 2006), only few studies tested for specificity of the relations between maternal prenatal smoking and CD- or ADHD-symptoms (Button et al., 2005; Mick et al., 2002; Thapar et al., 2003). This is an important matter considering the high comorbidity of these behavior problems (e.g. Jensen, Martin, & Cantwell, 1997). Mick et al. (2002) found a robust link with ADHD-symptoms after controlling for CD-symptoms in a clinical population. In that study, children aged 6 to 17 years with ADHD and non-ADHD controls were compared after statistical control for CD. In contrast, Thapar et al.’s (2003) and Button et al.’s (2005) population-based twin study using the Greater Manchester Twin Register did not select on either CD- or ADHD-symptoms. Whereas Thapar et al. showed that prenatal maternal smoking predicted a unique proportion of the variance in ADHD-symptoms after control for CD-symptoms in children between 5 and 16 years, Button et al. (2005) tested a more complex model showing unique variance for both CD- and ADHD-symptoms in children 5 to 18 years of age. Although the aforementioned studies used some form of statistical control for the symptoms associated with the ‘other’ disorder, only one study to date examined a relation between prenatal maternal smoking and co-occurring behavior problems (Wakschlag, Pickett et al., 2006). In that study maternal prenatal smoking predicted the co-occurrence of Oppositional Defiant Disorder (ODD) and ADHD when high levels of both were contrasted with absence of both. This suggests that maternal prenatal smoking, to the extent that it is a causal factor, may affect behavior much more seriously than initially shown in specificity studies. However, such a conclusion may be premature because that analysis did not inform us if maternal prenatal smoking was related specifically to the co-occurrence of both behavior problems or whether it was driven by its association with either behavior problem because we do not know if the combined group differed from the ODD-only or ADHD-only groups. Thus whether maternal prenatal smoking is related to the co-occurrence of behavior problems remains an open question in this literature.
In contrast to childhood and adolescence studies, there is a more limited number of early childhood studies and they tend to focus on symptoms rather than diagnoses. Tremblay et al. (2004) reported an association with early childhood trajectories of PA across time, and Romano et al. (2006) found an association between maternal prenatal smoking and hyperactive symptoms from age 2 to 7, but both studies focused on one externalizing behavior problem. Other studies focused on multiple behaviors. Orlebeke, Knol, and Verhulst (1997) and Williams et al. (1998), using the Child Behavior Checklist (CBCL) in large cohort studies of 3-year-old and 4 to 6-year old children, respectively, found associations with externalizing behavior problems (aggressive, overactive, oppositional) but not with internalizing behavior problems (withdrawn, anxious, depressed). Day, Richardson, Goldschmidt, and Cornelius (2000) reported significant associations of third trimester exposure with scores on each of the subscales of the Toddler Behavior Checklist (Physical Aggression, Oppositional Behavior, Immaturity, and Emotionality) and Activity level assessed with the Routh Activity Scale. Wakschlag, Leventhal et al. (2006) noted a persistent association between maternal prenatal smoking and disruptive behaviors during early childhood. None of these multiple behavior studies examined co-occurrence. Thus there also remains a need in the early childhood literature to examine whether maternal prenatal smoking is specifically associated with the co-occurrence of problem behaviors.
From a developmental perspective, early onset of externalizing behavior has been associated with the poorest outcomes (Brame, Nagin, & Tremblay, 2001; Lacourse et al., 2006; Moffitt, 1993; Woodward, Fergusson, & Horwood, 2002). In addition, co-occurring externalizing behaviors during middle childhood have been associated with the poorest behavioral outcomes during adolescence and adulthood (Lacourse et al., 2006; Lahey, McBurnett, & Loeber, 2000; Séguin et al., 2004). However, co-occurring behavior problems may be etiologically different from individual behavior problems, with their own social, sociodemographic and biological precursors (Waschbusch, 2002). Yet, we know very little about the prenatal markers of co-occurring behavior problems in general. Given that maternal prenatal smoking has been clearly associated with several single externalizing problems across development, in the present study, we test the hypothesis that maternal prenatal smoking is associated with co-occurring externalizing problems. Specifically we predict that maternal prenatal smoking will be associated to the co-occurrence of PA and HI in contrast with the absence of both, PA-only, and HI-only in a large early childhood sample.
Methods
Participants
The children of this study were born in 1997/1998 in the province of Québec, Canada and participate in the Québec Longitudinal Study of Children’s Development. This sample excluded very remote regions of the province populated mainly by aboriginal people (2.1% of live births), babies for whom gestational age could not be computed (1.3%), babies born in a different territory but whose parents reside in Québec (4.5%), and very premature babies (<24 weeks) and babies for whom there were delays in filing the birth records in the Master Birth Registry on time for the first assessment, i.e. babies born after 42 weeks gestation (0.1%; for full details, see Jetté & Des Groseilliers, 2000). A total of 2940 infants met inclusion criteria and were selected through a region-based stratification procedure (Jetté & Des Groseilliers, 2000). Of this original selection, a number of families could not be included in the initial 5 month-assessment for the following reasons: (1) Families not found on time (incorrect address/tel no.) (n=172, 5.9%); (2) Families excluded (total n=93, 3.2%) because of death of the baby (n=5), because of participation in other longitudinal studies (n=5), because they had no command of either English or French language (n=81), or because the instruments of the study were not designed to adequately measure development of children with severe physical or mental handicaps (n=2); (3) Families that could not be reached (n=14, 0.5%); (4) Families who declined participating (n=438, 16.4%). 2223 Families (75.6%) took part in the first assessment, which took place when the infants were 5 months old. Demographic characteristics of this Québec sample were comparable to those of a large Canadian sample consisting of 13,439 households which contains sufficient samples from each of the 10 Canadian provinces (Human Resources Development Canada, 1996: National Longitudinal Survey of Children and Youth, 1994–1995). Assessments relevant to this study took place at 17, 30 and 42 months. Parental informed consent was obtained before every assessment. 478 participants had either dropped out since the initial assessment at 5 months or had missing values for PA, HI or one or more predictor variables. Thus, the final study sample consisted of 1745 children (78.5% of families enrolled at 5 months). Table 1 indicates that, despite the fact there was a slight tendency for less advantaged families to drop out or have missing values on relevant variables, demographic characteristics were largely similar for in- and excluded families.Table 1Demographic characteristics of families enrolled in the Québec Longitudinal Study of Children’s Development (QLSCD 1) (n max=2223) and families included in the present study (QLSCD 2: n=1745)QLSCD 1QLSCD 2
n
%
n
%Sex of child Male1,12351.288450.7 Female1,06948.886149.3Age of mother at birth of first child 21 years or less47122.237221.3 Older than 21 years164877.8137378.7Mother graduated from high school Yes186584.0150486.2 No35516.024113.8Family status Two parents161779.2140780.6 One parent42420.833819.4Number of siblings in the home None83040.671340.9 One or more121559.4103259.1Family income Less than 30,000$63929.346826.8 30,000$–50,000$62528.752730.2 50,000$–80,000$61128.049928.6 More than 80,000$30714.125114.4Data courtesy of the Institut de la Statistique du Québec
Measurements
Maternal prenatal smoking
When the child was 5 months of age, mothers filled out a number of questionnaires. One set of questions concerned substance use (alcohol, tobacco, and illegal drugs) during pregnancy. The questions assessing smoking behavior during pregnancy were straightforward, ‘Did you smoke during pregnancy?’ and ‘How many cigarettes/day did you smoke whilst pregnant?’ We also asked when during pregnancy the mother had smoked, i.e. (only) in the first, second or third trimester or throughout pregnancy. These questions are similar to those found in most other assessment instruments (e.g., the Centers for Disease Control and Prevention (CDC) Pregnancy Risk Assessment Monitoring System (PRAMS) Questionnaire, see Beck et al., 2002) and like those found in most other studies (particularly those assessing large samples, e.g. Button et al., 2005; Fergusson et al., 1998; Maughan et al., 2004; Thapar et al., 2003; Wakschlag et al., 1997; Wakschlag & Hans, 2002). Because amount of cigarettes reportedly smoked tends to be a ‘rounded’ number, e.g. 5, 10, 15, etc, mothers were classified into 1 of 4 groups (0, 1–9, 10–19, or ≥20 cigarettes/day). This or similar classifications have been used in most other studies investigating smoking during pregnancy (e.g. Button et al., 2005; Fergusson et al., 1998; Maughan et al., 2004; Thapar et al., 2003; Wakschlag et al., 1997; Wakschlag & Hans, 2002). Although there is a risk for a social desirability bias, several studies have indicated a relatively strong association between retrospective self-report and blood/urine cotinine-levels (i.e. the main nicotine metabolite) (e.g. Law et al., 2003; Pickett, Rathouz, Kasza, Wakschlag, & Wright, 2005). Further reliance on self-report measures may be inferred from the strong relation between the amount reportedly smoked during pregnancy and birth weight (e.g. Huijbregts et al., 2006; Kramer et al., 2001). Other studies have shown that whereas for some prenatal exposures (e.g. alcohol, drugs) higher levels were reported postnatally than antenatally, this was not the case for tobacco exposure, which might indicate that smoking is considered a more habitual behavior that is more reliably recalled (e.g. Jacobsen, Chiodo, Sokol, & Jacobson, 2002). For the current analyses, 1307 (74.9%) mothers reported not to have smoked during pregnancy, 202 (11.6%) reported to have smoked 1–9 cigarettes/day, 174 (10%) between 10 and 19, and 62 (3.6%) reported to have smoked 20 or more cigarettes/day during pregnancy.
Physical aggression and hyperactivity-impulsivity
Maternal ratings of child behavior were obtained with the use of an early childhood behavior scale from the Canadian National Longitudinal Study of Children and Youth (Statistics Canada, 1995), which incorporates items from the Child Behavior Checklist for Ages 2–3 (Achenbach, Edelbrock, & Howell, 1987), the Ontario Child Health Study Scales (Offord, Boyle, & Racine, 1989), and the Preschool Behavior Questionnaire (Tremblay, Desmarais-Gervais, Gagnon, & Charlebois, 1987). In order to assess PA mothers were asked at 17, 30 and 42 months to indicate whether the child: (1) hits, bites, kicks; (2) fights; and (3) bullies others. The items for HI were: (1) can’t sit still, is restless (or hyperactive), (2) fidgets, (3) cannot settle down to do anything for more than a few moments, (4) is impulsive, acts without thinking, and (5) has difficulty waiting for turn in games. The first three representing hyperactivity, and the latter two representing impulsivity. The items were scored as follows: never or not true (score = 0), sometimes or somewhat true (score = 1), or often or very true (score = 2). The items were summed to obtain PA (range = 0 to 6) and HI scores (range = 0 to 10). The internal consistency values (Cronbach’s α) for PA were 0.80 at 17 months, 0.82 at 30 months, and 0.72 at 42 months. For HI they were 0.75, 0.75, and 0.71, respectively. The scales were shown to have good discriminant validity in the prediction of different types of adolescent criminal behaviors (Broidy et al., 2003; Nagin & Tremblay, 1999), and although correlated, the PA and HI scales were related in predictable ways with mother and child reports of CD and ADHD, respectively (Séguin et al., 2004). Their validity in early childhood is also well established (Tremblay et al., 2004; Romano et al., 2006).
Control variables
Potential confounders of associations between maternal prenatal smoking and early behavior problems were selected from four key domains: demographic factors (age at birth of first child, family status (separation/divorce: yes/no), presence of siblings, family income, maternal education); perinatal factors (alcohol and drug exposure during pregnancy, birth weight); family functioning and parenting (hostile reactive parenting, responsiveness, involvement), and parental background and mental health (mother’s and father’s history of antisocial behavior, maternal depression). Mothers provided information on most of these factors when their child was 5 months of age and again when their child was 17 months of age.
Family income was indicated on a 7-point scale (1 = less than $10,000 (Canadian) to 8 = more than $80,000, as was maternal education (1 = no high school diploma to 7 = university degree). The poverty line in Canada is situated at around $15,000 per capita (Statistics Canada, 2006).
Perinatal factors included exposure to illegal drugs (yes/no) and alcohol (7-point scale ranging from ‘never’ to ‘daily’) during pregnancy and birth weight. Birth weight and gestational age were derived from birth records. Birth weight for gestational age was standardized within gender for each week of gestation using Canadian norms (Kramer et al., 2001).
Parenting measures were obtained by observing the mother-child dyad during home visits at 5 and 17 months using the Home Observation for Measurement of the Environment (HOME)-Infant version (Caldwell & Bradley, 1984). The observations were carried out by trained observers who spent about 3 hours in the home to complete a battery of questionnaires and tests. These observers received a one-week training session and were closely supervised during the data collection. Three scales were obtained: hostile reactive parenting (e.g. “talks negatively about her child,” “shouts at her child,” “hits or physically punishes her child”; α=0.43 at 5 months and 0.73 at 17 months), responsiveness (e.g. “responds verbally to child’s vocalizations or verbalizations,” “tells child name of object or person during the visit,” “spontaneously praises the child at least twice”; α=0.85 at 5 months and 0.83 at 17 months), and involvement (e.g. “provides toys that challenge child to develop new skills,” “structures child’s play periods”; α=0.85 at 5 months and 0.88 at 17 months). Scores for each item on each scale ranged from 1 (never) to 5 (all the time). Only the scale score for hostile reactive parenting at 17 months was used because of the low internal reliability of this scale at 5 months. Mean scores across the 5 and 17 months assessments were used for the responsiveness and involvement scales.
Family functioning (at 5 and 17 months) was assessed with a scale containing 12 items measuring communication, problem resolution, control of disruptive behavior, and showing and receiving affection (Statistics Canada, 1995). Scores per item could be 0 (‘never’), 1 (‘sometimes’), or 2 (‘often’), thus ranging from 0 to 36 on the scale. The Cronbach α’s were 0.86 (5 months) and 0.98 (17 months).
In order to assess history of antisocial behavior, both parents completed a questionnaire at the 5-month assessment. The questionnaire included items related to childhood/adolescence (i.e. the period before the end of high school) and items related to adulthood (Zoccolillo, 2000; see also Tremblay et al., 2004), and was largely derived from the NIMH-Diagnostic Interview Schedule (Robins, Helzer, Croughan, & Radcliff, 1981). Childhood/adolescence items included ‘starting fights,’ theft, involvement with youth protection or police, expulsion or suspension from school, truancy, and running away from home. Adult items included arrests (other than for traffic violations), being fired from a job (excluding layoffs for lack of work), trouble at work, with family, or with the police due to drug or alcohol abuse, ‘starting fights’ (fathers), and ‘hitting or throwing things at the spouse or partner’ (mothers). Adolescent and adult scores (0 = no, 1 = yes) were summed for mothers and for fathers. Internal reliability for mother’s items was 0.54 (Cronbach’s α) and 0.59 for father’s items. Latent class analysis identified 3-class models for both mothers ((1) not antisocial, (2) moderately antisocial, (3) antisocial) and fathers ((1) not antisocial, (2) antisocial as an adolescent but not as an adult, (3) moderately antisocial as an adolescent, antisocial as an adult). The Center for Epidemiological Studies-Depression Scale (CES-D; Radloff, 1977) was used for report of symptoms associated with depression (at 5 and 17 months).
Studies have shown that all these factors are associated with increased risk for physical aggression, hyperactivity, CD- and ADHD-symptoms and with prenatal maternal smoking (e.g. Biederman, Milberger, & Faraone, 1995; Campbell, Shaw, & Gilliom, 2000; Huijbregts et al., 2006; Linnet et al., 2003; Maughan et al., 2004; NICHD ECCRN, 2004; Tremblay et al., 2004; Nagin & Tremblay, 2001; Romano et al., 2006; Shaw et al., 2005). Apart from the dichotomous variables (i.e. gender, presence of siblings, drug use during pregnancy, family status), all scores were standardized for statistical analyses.
Data analyses
Assignment to trajectories
Scores from the three assessment points were analyzed to identify distinctive behavioral trajectories across time (Nagin, 1999, 2005; Nagin & Tremblay, 2001). Rather than to assume that all children follow the same developmental pattern, this methodology identifies different groups of individuals who tend follow similar patterns over time. For example, some children may never show a given problem behavior (intercept model or zero order polynomial), others may show constant high levels (also intercept model), and others may increase or decrease over time (e.g., linear – 1st, quadratic – 2nd, or cubic – 3rd order polynomials). The methodology can also be adapted to accommodate various data distributions (i.e., binary, censored normal, zero-inflated Poisson, and count data). The trajectory methodology uses all available developmental data points and assigns individuals to trajectories on the basis of a posterior probability rule. Resulting groups are meant to represent approximations of an underlying continuous process. In order to identify the model that best represents development of a specific behavior during a given time frame, models with a varying number of trajectories are estimated. Model selection is dependent on a combination of statistical and investigator-guided concerns. Besides a need to determine the best model for the data distribution, key decisions are also based on Bayesian fit indices for model selection in accordance to procedures described by Nagin (2005), e.g. the higher the Bayesian Information Criterion (BIC), the better. The optimal model is thus also determined by adding trajectories to the model until the BIC ceases to improve. The investigator would then have enough information to determine the best model.
A key output of model estimation is the posterior probability of group membership. For each trajectory group this probability measures the likelihood of an individual of belonging to that trajectory group based on observations across assessments. In other words, 100% accuracy in classification is not assumed nor required. For example, in the case of an individual who scores high on hyperactivity at all assessment periods, the posterior probability of membership to the chronic group would be high whereas the probability of membership to the low trajectory group would be near 0. Participants can be assigned to the trajectory group for which they show the highest probability of belonging. Ideally, the posterior membership probability should be near 1 for this trajectory group. Further, when trajectories are joined, conditional and joint probabilities can be used to further describe the relationship between the joint factors. Posterior probabilities will then be applied to weight PA, HI and their combination when they enter further analyses (see Séguin et al., 2004).
Prediction of physical aggression, hyperactivity/impulsivity, and their co-occurrence
Weighted multinomial regressions were conducted with PROC CATMOD in SAS v.8.2. (SAS Institute Inc., 2001). First, weighted multinomial regressions were conducted separately on PA and HI between ages 17 and 42 months for descriptive purposes. The influence of maternal prenatal smoking on these behaviors was investigated in analyses with and without control variables.
For the second set of analyses predicting PA and HI between ages 17 and 42 months, we contrasted, in weighted multinomial regression, those children who were high on both PA and HI (combined group), to those who were only high on PA (PA only group), to those only high on HI (HI only group) and to the remainder of the sample.
Results
Assignment to trajectories
Models with between 2 to 5 trajectories and varied shapes for each trajectory were compared using BIC for both PA and HI. Three trajectories were modeled for PA between ages 17 and 42 months using a zero-inflated Poisson distribution: a consistently low, a moderate and rising, and a high and rising trajectory, each representing respectively 25, 50, and 25% of the sample, all of these were best modeled using a linear trend except for the low group which was best represented by a constant term (Fig. 1, left panel). The shape and level of the PA trajectories were very similar to those we identified in Tremblay et al., 2004, in a smaller sample using the same measures. It reveals that most of the children show an increase of physical aggression over time, which is consistent with other longitudinal and cross sectional studies of early childhood.
Four trajectories were modeled for HI using a censored normal distribution: consistently low, low to moderate, moderate to high and chronic HI, representing respectively, 12, 45, 37, and 6% of the sample. All were best modeled by a constant term except for the low to moderate group that was best modeled with a quadratic trend (Fig. 1, right panel).Thus in contrast to PA, HI in early childhood appears to be more stable in level across time, with very few children being atypically high.Fig. 1Observed and predicted means for developmental trajectories of physical aggression (PA) and Hyperactivity-impulsivity (HI) between ages 17 and 42 months. *Physical aggression: H-R = high-rising; M-R = moderate-rising; C-L = consistently low; Hyperactivity-Impulsivity: H-C = high-chronic; M-H = moderate-high; L-M = low-moderate; C-L = consistently low. Data courtesy of the Institut de la Statistique du QuébecTable 2Maternal prenatal smoking and physical aggression (PA), hyperactivity-impulsivity (HI) and joint PA+HI trajectories between ages 17 and 42 monthsMLE 1 (SE)MLE 2 (SE)MLE 3 (SE)
χ
2
P
Physical aggression−.40 (.09)−.24 (.07)NA23.6<.001Physical aggression adjusted for control variables−.29 (.10)−.15 (.08)NA8.4.015Hyperactivity−.56 (.15)−.41 (.11)−.22 (.11)22.8<.001Hyperactivity adjusted for control variables−.32 (.17)−.24 (.13)−.13 (.13)5.8.121Physical aggression and hyperactivity−.65 (.14)−.57 (.22)−.41 (.14)31.2<.001Physical aggression and hyperactivity adjusted for control variables−.46 (.16)−.55 (.25)−.33 (.17)10.5.015Note. MLE = Maximum Likelihood Estimate, Reference categories for PA, HI and PA+HI are high PA, high HI and high PA + high HI, respectively. NA = Not Applicable. (MLE) 1 = consistently-low PA; consistently-low HI; and low PA + low HI; (MLE) 2 = moderate-rising PA; low-moderate HI; and low PA + high HI; (MLE) 3 = moderate-high HI; and high PA + low HI. DF = 2 for PA, DF = 3 for HI and for PA+HI. Data courtesy of the Institut de la Statistique du Québec
In the current models for PA and HI the average posterior probability for the assigned trajectory group ranged between .75 and .83, thereby indicating good fit (Nagin, 1999, 2005). Further, a close match of predicted and observed means shown in Fig. 1 also illustrates this good fit. Models with three trajectories for PA and four trajectories for HI have also produced the best fit in studies using other samples that included children in early childhood (e.g. Tremblay et al., 2004; Côté et al., 2006; Romano et al., 2006; Shaw et al., 2005).
Physical aggression
A first multinomial regression revealed that maternal prenatal smoking significantly predicted PA. In an analysis without control variables, both the odds ratio (OR) contrasts between high and low PA [OR=1.49 (95% CI: 1.25 to 1.75), , p<.001] and between high and moderate PA [OR=1.27 (95% CI: 1.11 to 1.45), , p<.001] were significantly predicted by maternal prenatal smoking (, p<.001; see also Table 2). The ORs represent one categorical increase in maternal prenatal smoking, i.e. from ‘0 cigarettes/day’ to ‘1–9,’ ‘1–9’ to ‘10–19,’ and from ‘10–19’ to ≥20 cigarettes/day.
Maternal prenatal smoking was also a significant predictor of PA in a multinomial regression with control factors [, p=.015]. The ORs were reduced to 1.33 (95% CI: 1.10 to 1.61) [, p=.004] for the contrast between high and low PA, and to 1.16 (95% CI: 0.99 to 1.35) [, p=.057] for the contrast between high and moderate PA. Other significant predictors of PA between ages 17 and 42 months were gender [, p<.001], presence of siblings [, p<.001], and hostile reactive parenting (by the mother) [, p<.001].
Hyperactivity-impulsivity
A first multinomial regression predicting HI without control variables revealed significant contrasts between high and low HI [OR=1.75 (95% CI: 1.32 to 2.33), , p<.001], between high and low-moderate HI [OR=1.49 (95% CI: 1.20 to 1.85), , p<.001] and high and moderate-high HI [OR=1.25 (95% CI: 1.01 to 1.56), , p=.043], resulting in an overall significant effect (, p<.001). In a second multinomial regression with control variables the association of maternal prenatal smoking with HI was strongly attenuated, although the overall effect was still not too far from significance [, p=.12] (see also Table 2). This attenuation can also be seen in the contrasts between high and low HI [OR=1.38 (95% CI: 0.99 to 1.91), , p=.052], between high and low-moderate HI [OR=1.27 (95% CI: 0.98 to 1.63),, p=.070] and high and moderate-high HI [OR=1.13 (95% CI: 0.88 to 1.46), , p=.327]. The remaining significant predictors of HI between ages 17 and 42 months were gender [, p<.001], hostile reactive parenting (by the mother) [, p<.001], maternal depression [, p=.006], and age of the mother when she had her first child [, p=.012].
Combining physical aggression and hyperactivity-impulsivity
Although distinct, the trajectories of PA and HI were moderately but significantly related (, p<.001; Spearman r=.245, p<.001). The conditional probabilities for PA and HI in Table 3 show that, with higher PA or HI, there is also a higher probability of HI and PA, respectively. The conditional probabilities further reveal an asymmetry in the association in that most high HI children were likely to be high PA, but that most high PA children are not likely to be high HI. Table 3 also shows the joint probabilities of displaying particular levels of PA and HI simultaneously. There are 12 cells (3 PA × 4 HI). In order to test the contrasts relevant to our research question, i.e. that maternal prenatal smoking might specifically predict co-occurring PA and HI as opposed to PA only or HI only, these 12 groups were collapsed into 4 groups representing the combined group (PA3HI4), the high PA only group (PA3HI1 + PA3HI2 + PA3HI3), the high HI only group (PA1HI4 + PA2HI4), and the remainder of the sample (PA1HI1 + PA1HI2 + PA1HI3 + PA2HI1 + PA2HI2 +PA2HI3) (Table 4). For the next analysis, these four groups will constitute our dependent variable.Table 3Co-occurrence of physical aggression (PA) and hyperactivity-impulsivity (HI)Probability of PA conditional on HISumLow PAModerate PAHigh PALow HI.40.49.111.0Low-Moderate (L-M) HI.28.54.181.0Moderate-High (M-H) HI.18.49.331.0Chronic-High (C-H) HI.12.39.491.0Probability of HI conditional on PALow HIL-M HIM-H HIC-H HILow PA.19.51.27.031.0Moderate PA.12.48.36.041.0High PA.05.34.50.111.0Joint probability of PA and HILow HIL-M HIM-H HIC-H HISum/PALow PA.05.13.07.01.25Moderate PA.06.24.18.02.51High PA.01.08.12.03.24Sum/HI.12.45.37.061.0Data courtesy of the Institut de la Statistique du QuébecTable 4Co-occurrence of physical aggression (PA) (2 levels) and hyperactivity-impulsivity (HI) (2 levels)Probability of PA conditional on HISumLow PAHigh PALow HI.77.231.0High HI.51.491.0Probability of HI conditional on PALow HIHigh HILow PA.96.041.0High PA.89.111.0Joint probability of PA and HILow HIHigh HISum/PALow PA.73.03.76High PA.22.03.24Sum/HI.95.051.0Data courtesy of the Institut de la Statistique du Québec
In order to be able to conclude that prenatal smoking specifically predicts co-occurring PA and HI the a priori contrasts of all groups to the combined group should be significant. A multinomial regression with the control variables that had featured in the previous separate analyses of PA and HI showed that this was indeed the case. The overall effect of maternal prenatal smoking [, p=.015], and the predictions by prenatal smoking of the contrasts of the combined group with the PA only group [OR=1.39, 95% CI to 1.01 to 1.92, , p=.047], with HI only group [OR=1.72, 95% CI 1.06 to 2.86, , p=.029], and with the remainder of the sample [OR=1.59, 95% CI 1.16 to 2.17, , p=.004], were all significant (see also Table 2). Although additional contrasts were not significant, the contrast between the PA-only group and ‘the remainder of the sample’- group approached significance [, p=.08]. Other significant and independent predictors of the combined group were gender [, p<.001], presence of siblings [, p<.001], and hostile reactive parenting (by the mother) [, p<.001].
Discussion
The results of this study of early childhood show that maternal prenatal smoking was associated with PA but not with HI in covariate regression analyses when PA and HI were examined separately. When PA and HI were combined, a contrast between the PA-only group and the remainder of the sample, which was neither high on PA or HI, fell short of significance. However, maternal prenatal smoking predicted co-occurring elevated levels of PA and HI. Therefore the key finding of this study is that maternal prenatal smoking predicted high PA, but largely when it is co-occurring with high HI. Had we just examined PA and HI separately, as has often been done in studies of specificity, we would have missed the fact that HI is also sensitive to maternal prenatal smoking, even after controlling for several confounding variables, but only when it is combined with PA.
The first implication of these findings is that maternal prenatal smoking may be associated with the most severe forms of PA and HI, i.e. their combination, in early childhood. Second, the fact that maternal prenatal smoking predicts co-occurring PA and HI may be considered an important finding in light of evidence indicating that multiple co-occurring behavior problems at an early age are associated with a higher risk of persistent antisocial problems (Brame et al., 2001; Broidy et al., 2003; Nagin & Tremblay, 1999; Lacourse et al., 2006; Lahey et al., 2000; Moffitt, 1993; Moffitt, Caspi, Dickson, Silva, & Stanton, 1996), as well as with several other functional impairments (Waschbusch, 2002; Séguin et al., 2004). Finally, the risk factors identified for these combined behavior problems, are largely modifiable.
Our study replicates the reports from Orlebeke et al. (1997), Williams et al. (1998), Day et al. (2000), and Tremblay et al. (2004), based largely on early childhood behavior problems. However, we extend these findings because none of those early childhood studies had tested for the prediction of specific behavior problems or for the specificity of their combination. Tremblay et al. (2004) covered the same developmental period and used the exact same PA scale as we did here on another sample. We both found that the association between maternal prenatal smoking and PA was not explained by control variables. However, our results for HI appear to contrast slightly with those of Romano et al. (2006) who reported that the association between maternal prenatal smoking and hyperactivity was not explained by control variables. There are a number of key differences between the current study and Romano et al. (2006). Besides using a different sample, we included impulsivity along with hyperactivity items, and focused exclusively on behavior during early childhood. Thus, in the current study the effects of maternal prenatal smoking on HI were largely explained by control variables. It is possible that control variables have a greater effect on the impulsivity component of HI than on the hyperactivity component. We note that our results appear to be consistent with those of Wakschlag, Leventhal et al. (2006), who also failed to find a relation of maternal prenatal smoking but with ADHD, which includes impulsivity. Because of our goal for a strict test for specificity of effects on PA, it was more conservative to have a combination of impulsivity and hyperactivity items in our scale. Alternately, by having information extending beyond the early childhood period to estimate trajectories Romano et al. (2006) may have more accurately identified children for whom maternal prenatal smoking matters beyond the effects of control variables. Despite this limitation to early childhood, and despite the fact that we cannot claim to have accounted for all possible control variables (examples include exposure to environmental smoking postnatally and forms of parental psychopathology other than maternal depression and parental history of antisocial behavior such as parental ADHD), this study replicates and extends all previous studies by providing an empirical basis for how the association of prenatal smoking to both PA and HI may manifest itself. It is now much more clear that studies need to examine co-occurrence of behavior problems.
There is a leap in drawing parallels between early childhood symptoms to DSM-based disorder, i.e., from PA to CD and from HI to ADHD. CD is characterized by aggression to people and/or animals, deceitfulness or theft, vandalism and serious rule violations whereas ADHD is characterized by hyperactivity, behavior disinhibition (or impulsivity), and inattention and distractibility (American Psychiatric Association, 1994). It is not yet clear to what extent behavior problems during the early childhood years map onto ADHD and CD symptoms during later childhood and (antisocial) behavior problems during adolescence and adulthood, although we have shown that childhood PA trajectories and hyperactivity trajectories were respectively related to CD and ADHD measured in adolescence (Séguin et al., 2004). But manifestation of PA and HI symptoms are often not sufficient to warrant diagnoses. We also note that behavior problems during the early childhood years tend to be more common and typically decline over time (Côté et al., 2006; Romano et al., 2006; Bongers, Koot, Van Der Ende, & Verhulst, 2003).
A number of pioneering studies showed that behavior problems later in life (ranging from school-age to adulthood) could already be identified during early childhood (e.g. Campbell et al., 1986; Keenan & Wakschlag, 2000), but only recently a number of studies have started to show, particularly with respect to HI, a significant consistency in the likelihood to display (very) high levels from age 1 1/2 years onwards (e.g. Romano et al., 2006; Shaw et al., 2005). With respect to PA, there is a larger group of very young children showing rather high levels, but most of them will use less PA from early school age onwards (e.g. Côté et al., 2006; NICHD Early Child Care Research Network, 2004; Shaw et al., 2005). Such general trends might also underlie differences in number of groups identified by the trajectory methodology in samples with different age ranges. Whereas our three PA-groups were consistent with the three groups identified by some authors (e.g. Côté et al., 2006) covering early and middle childhood, other authors identified more than three groups for this period (NICHD Early Child Care Research Network, 2004; Shaw et al., 2005). However, most children will not yet have finished learning alternative strategies (i.e. have undergone ‘socialization of aggression,’ Tremblay, 2003) between 17 and 42 months, so the expected ‘desisting’ patterns of PA will not yet be evident. One thing the abovementioned studies have in common is that they show a degree of consistency between early and middle childhood conduct problems, i.e. children with chronic PA during middle childhood generally had high PA levels during early childhood as well. This, in turn, emphasizes the importance of searching for predictors of co-occurring early childhood behavior problems: children most at risk of subsequent behavior problems might be those who display both high PA and high HI at this stage.
The combination of conduct problems and hyperactivity-impulsivity-attention problems has been distinguished from their pure forms at a number of levels (Waschbusch, 2002). Although research has again mainly focused on older children with diagnosed disorders rather than the combination of behaviors such as PA and HI (but see Séguin, Arseneault, Boulerice, Harden, & Tremblay, 2002; Séguin et al., 2004), evidence has been provided for a number of etiological and developmental differences between the comorbid and individual conditions. For example, greater behavioral and autonomic nervous system reactivity in response to provocation has been shown in the comorbid compared to the separate conditions (Waschbusch, 2002). Children with CD+ADHD were also shown to have lower baseline sympathetic arousal than children with the separate disorders (specifically those with ADHD only; Herpertz et al., 2001). There is also evidence for differences in patterns of brain activity between CD+ADHD and the separate conditions (Banaschewski et al., 2003) and for specific heritable risk of CD+ADHD (Dick, Viken, Kaprio, Pulkkinen, & Rose, 2005; Thapar, Harrington, & McGuffin, 2001). Several studies have shown that specific family psychosocial characteristics (e.g. parent-child conflict; parental psychopathology) specifically predict CD+ADHD or predict CD+ADHD more strongly than the separate conditions (e.g. Burt, Krueger, McGue, & Iacono, 2003; Pfiffner, McBurnett, Rathouz, & Judice, 2005). Thus, although extrapolation to CD and ADHD must be done with caution, the results of the present study support an etiological basis for the combined behavior problems in an early childhood sample.
Finally, in addition to prenatal maternal smoking, we found that maternal hostile- reactive parenting was also associated with co-occurring PA and HI beyond its effects on the separate behaviors. Hostile-reactive parenting and similar ‘negative’ parenting behaviors have been associated with children’s PA and HI levels in earlier studies (e.g. Côté et al., 2006; Romano et al., 2006; Shaw et al., 2005; Tremblay et al., 2004), Negative parenting has also been related to the failure of training programs aimed at improving child conduct problems (Webster-Stratton, Reid, & Hammond, 2001). This association could be seen as resulting from the impact of maternal behavior on their child’s behavior. However, they may also reflect, at least in part, parent’s reactions to their child’s behavior. For instance, when faced with conduct problems, mothers of preschoolers at risk for ADHD and Oppositional Defiant Disorder tend to resort to more negative parenting strategies (Cunningham & Boyle, 2002). Furthermore, mothers hostile-reactive parenting toward their 5 month-old infants has been found to be partly driven by their infant’s difficult temperament (Boivin et al., 2005).
Our finding that maternal hostile-reactive parenting is associated with an increased risk of co-occurring PA and HI may thus reflect the increased challenge of parenting when a child displays multiple behavior problems (cf. Seipp & Johnston, 2005). This pattern could lead to the establishment of a family coercive process and the further learning of anti-social behaviors by the child. However, the state of evidence precludes any definite conclusions at this point, and there is a clear need to better document the early dynamics of negative parenting and externalizing problems in early childhood.
To the extent that they may be involved in causation, the good news is that prenatal maternal smoking and hostile-reactive parenting could be significantly attenuated in prevention and intervention programs. The fact that the effects of smoking during pregnancy on both PA and PA+HI were robust, added to the evidence that PA+HI increases the risk for continued serious behavior problems, emphasizes the importance of smoking cessation during pregnancy (for a review on effectiveness of cessation programs, see Lumley, Oliver, Chamberlain, & Oakley, 2004). Considering the high prevalence of prenatal smoking (around 25% in Western countries), it will be important to further refine the identification of families most at risk of having children with a prepotency to develop co-occurring PA and HI. Our results further confirm that hostile-reactive parenting should be targeted in preventive intervention aimed at children with early multiple behavior problems. However, because parenting interventions often are the least effective in multiple-risk families (for a review, see Hutchings & Lane, 2005), including those characterized by prenatal smoking (Vuijk, van Lier, Huizink, Verhulst, & Crijnen, 2006), cessation programs targeting the most vulnerable families should be emphasized, intensified, and given appropriate means to succeed.
|
[
"prenatal smoking",
"physical aggression",
"hyperactivity",
"impulsivity",
"comorbidity"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Breast_Cancer_Res_Treat-3-1-2001220
|
Femara® and the future: tailoring treatment and combination therapies with Femara
|
Long-term estrogen deprivation treatment for breast cancer can, in some patients, lead to the activation of alternate cellular pathways, resulting in the re-emergence of the disease. This is a distressing scenario for oncologists and patients, but recent intensive molecular and biochemical studies are beginning to unravel these pathways, revealing opportunities for new targeted treatments. Far from making present therapies redundant, these new discoveries open the door to novel combination therapies that promise to provide enhanced efficacy or overcome treatment resistance. Letrozole, one of the most potent aromatase inhibitors, is the ideal candidate for combination therapy; indeed, it is one of the most intensively studied aromatase inhibitors in the evolving combinatorial setting. Complementary to the use of combination therapy is the development of molecular tools to identify patients who will benefit the most from these new treatments. Microarray gene profiling studies, designed to detect letrozole-responsive targets, are currently under way to understand how the use of the drug can be tailored more efficiently to specific patient needs.
Introduction
The proliferative, invasive, and metastatic potential of breast tumors may be largely predetermined at an early stage in the course of disease, whereas genetic alterations that accumulate during progression from in situ to metastatic disease are unpredictable and result in specific phenotypic changes and loss of sensitivity to treatments [1]; for example, although the estrogen receptor-positive (ER+) phenotype may be largely stable over time [2], hormone therapy-sensitive breast tumors may develop resistance and progress to a hormone-independent state [3]. In addition, progression to hormone independence may be associated with alterations in the expression of other regulatory genes, such as human epidermal growth factor receptor 2 (HER2) [4]. In the clinical setting, a quantitative decrease in ER expression was found in over 900 patients with primary breast cancer when HER2 was amplified [5].
Recent advances in translational research studies have highlighted the complexity of ER signaling, including differential roles for the ERα and ERβ subtypes [6], and multiple regulatory interactions between steroid hormone, growth factor, and other tyrosine kinase signaling pathways [7–10]. Greater understanding of tumor biology is beginning to help physicians to individualize treatment selection based on clinical, pathologic, molecular, and genetic profiling, and to rationally design novel combinations to improve efficacy and safety.
This article reviews novel approaches with the more potent third-generation aromatase inhibitor (AI) letrozole (Femara®; Novartis Pharmaceuticals) in combination strategies with agents targeting other growth factor pathways. The review explores the hypothesis that combining letrozole with specifically targeted therapies may delay or overcome endocrine therapy resistance in postmenopausal women with hormone receptor-positive (HR+) breast cancer.
Endocrine therapy resistance
It is well established that breast cancer cells can adapt to low concentrations of estrogen by becoming hypersensitive to estradiol [11]. Long-term exposure to tamoxifen induces hypersensitivity to estradiol [12] and this adaptive change can result in resistance to endocrine therapy [13]. It has been postulated that tamoxifen is more susceptible than AIs to this phenomenon because of its intrinsic agonist properties [12]. Moreover, it has been suggested that highly potent AIs are required to block estrogen synthesis when breast cancer tumors are hypersensitive to small amounts of estradiol [13].
Recent research has provided a compelling explanation for the development of resistance [13, 14]. Experiments using MCF-7 and other breast cancer model systems have identified alternative intracellular signaling pathways used by breast tumors to enhance and activate ER signaling, thus allowing cells to escape from the inhibitory effects of endocrine therapies [15, 16]. It has been shown that long-term estrogen deprivation upregulates ERα and growth factor signaling pathways such as mitogen-activated protein kinase (MAPK), phosphatidylinositol 3-kinase (PI3K), and the mammalian target of rapamycin (mTOR) pathways [14–19]. Of note, Jeng et al. [15] reported that a specific inhibitor of MAPK (PD98059) could block the elevation of activated MAPK observed in MCF-7 cells exposed to long-term estrogen deprivation. Furthermore, studies in both wild-type and long-term estrogen-deprived MCF-7 cells suggested that mTOR has a key role in breast cancer cell proliferation and showed that mTOR inhibition by farnesylthiosalicylic acid (FTS) can reduce proliferation and induce apoptosis [19].
Growth factor pathways
The role of nongenomic pathways has been highlighted in resistance to antiestrogen therapy [20]. Classically, estrogens bind to nuclear ER to enhance transcription of genes important in breast cancer proliferation and survival (genomic pathway) [21]; however, estrogen may also act through ER located in or near the cell membrane [22]. Nongenomic actions include activation of various growth signaling pathways, including MAPK [15]. In addition, ER may indirectly activate epidermal growth factor receptors (EGFR) via coactivators, including src, leading to activation of EGFR [23–25]. Subsequently, dimerization of activated EGFR with other HER family receptors, particularly HER2, activates intracellular signaling pathways, which in turn may enhance nuclear ER signaling [4], thus completing a vicious cycle of events. Cross-talk between ER and HER2 pathways has been implicated in clinical resistance to tamoxifen [4, 7]. Shou et al. [4] reported that tamoxifen behaves as an agonist in MCF-7 breast cancer cells that express high levels of the coactivator AIB1 (src3) and HER2, resulting in de novo resistance. Interestingly, addition of an anti-EGFR tyrsosine kinase inhibitor eliminated cross-talk and restored tamoxifen’s antitumor activity [4].
It has been postulated that AIs may be more effective than selective estrogen-receptor modulators (SERMs) [26] because they can block genomic and nongenomic activities of ER [27]. Elucidation of ER biology and interactions with growth factor signaling pathways will help to identify potential therapeutic targets for HR+ breast cancer [4, 28].
Combination therapy
Several strategies to inhibit growth factor signaling and signal transduction in breast cancer have been tested in the preclinical setting (see Fig. 1). The humanized monoclonal antibody trastuzumab specifically targets the extracellular domain of HER2 [29, 30]. Amplification of HER2 occurs in approximately 25% of breast tumors and is associated with more aggressive disease and a poor prognosis [31]. Trastuzumab has been shown to restore sensitivity to estrogen and tamoxifen in estrogen-independent HER2-transduced MCF-7 cells [32]. Furthermore, trastuzumab blocked HER2 heterodimer formation and phosphorylation, reduced ERK1/2 activity, and strongly inhibited cell growth in MCF-7 cells overexpressing EGFR and HER2 and resistant to tamoxifen [18]. Of note, synergistic inhibition of the BT474 breast cancer cell line was observed with trastuzumab and the HER dimerization inhibitor pertuzumab [33], which targets a different domain of HER2 [34]. The combination of tamoxifen and trastuzumab in ER+, HER2+ BT-474 cells resulted in synergistic growth inhibition due to the enhancement of cell accumulation in the G0–G1 phase of the cell cycle, and a decreased number of cells in S phase [35].
Fig. 1Cross-talk between signal transduction pathways and ER signaling in endocrine resistant breast cancer, with opportunities for targeted intervention. Estrogen (E2)-liganded ER activates E2-regulated genes in classical pathway (thick black arrow), but following long-term tamoxifen therapy resistance can develop with bidirectional cross-talk (gray arrows) between ER and growth factor receptors, with association of membrane bound ER with growth factor receptors, and/or IGFR or EGFR/HER2 activation of ER phosphorylation. Stars show various targeted therapies. AI aromatase inhibitor, SERD selective estrogen receptor down-regulator, MoAb monoclonal antibodies, TKI tyrosine kinase inhibitor, FTI farnesyltransferase inhibitor, CCI cell cycle inhibitor. Reprinted from [28] with permission from the American Association for Cancer Research
HER signaling pathways can also be targeted by inhibiting receptor phosphorylation by intracellular tyrosine kinases [36]. Gefitinib and erlotinib both specifically inhibit the EGFR tyrosine kinase and have demonstrated inhibitory activity in models of hormone-resistant breast cancer [37, 38]. Interestingly, pretreatment of HER2+, hormone-resistant MCF-7 cells with gefitinib eliminated the agonist effects of tamoxifen and restored its antitumor activity [4]. In ER+ breast tumors, targeted therapy with single-agent gefitinib was found to be less effective than endocrine therapy [39]; however, tamoxifen plus gefitinib had greater activity than tamoxifen alone in vivo in hormone-sensitive cells [39]. Data also support the use of EGFR tyrosine kinase inhibitors in combination with HER2 antibodies, such as trastuzumab, against breast tumors that express EGFR and high levels of HER2 [37]. Alternatively, the tyrosine kinase inhibitor lapatinib provides single-agent targeting of both EGFR and HER2 [40, 41]. Using a panel of 31 human breast tumors, Konecny et al. [41] showed that response to lapatinib was significantly correlated with HER2 expression and its ability to inhibit HER2, Raf, Akt, and ERK phosphorylation. In vivo studies showed that lapatinib had a sustained antitumor effect that was further prolonged by combination with trastuzumab [41]. Another study showed that the combination of AEE788 (an EGFR/HER2 tyrosine kinase inhibitor) with letrozole enhanced antiproliferative effects of these agents by 20–30% in MCF-7 and ZR75.1 cell lines and by 60–70% in the BT474 cell line [42]. In a model system of acquired resistance to letrozole, AEE788 partially restored sensitivity to letrozole, whereas rapamycin was not effective, suggesting that letrozole resistance and mTOR activation may not be connected in this model [43]. The authors concluded that inhibition of both HER2-mediated signaling and mTOR-dependent translation may restore responsiveness to letrozole in breast cancer refractory to this AI [43].
Strategies to inhibit downstream signal pathways have also been developed. Farnesyl transferase inhibitors block the first and most important step in the activation of Ras signaling pathways [44]. Aberrant function of the Ras signal transduction pathway is common in breast cancer as a result of upstream activation via HER2 or EGFR [45]. The farnesyl transferase inhibitor R11577 (tipifarnib) was found to have antitumor activity against MCF-7 xenografts [44]. Another interesting therapeutic target is mTOR, a central regulator of G1 cell cycle protein synthesis, that precedes commitment to normal cellular replication [46]. Treatment of MCF-7 Arom-1 cells with letrozole and the mTOR inhibitor RAD001 resulted in a further 50% reduction in proliferation compared with letrozole alone [47]. Another set of experiments, developed to test the hypothesis that Akt kinase confers resistance to endocrine therapy through suppression of ASK1/JNK pathway, showed that combining RAD001 with letrozole restored activation of the ASK/JNK pathway and increased the sensitivity of MCF-7 cells with constitutively active Akt to endocrine therapy [48]. Studies have also targeted the selective estrogen receptor down-regulator (SERD) fulvestrant, which indirectly inhibits growth factor pathways by down-regulating ER [3]. In MCF-7Ca xenografts, combined treatment with fulvestrant and letrozole prevented increases in HER2 and activation of MAPK and inhibited tumor growth [49].
These preclinical models (see Table 1) suggest that treatments to reduce growth factor signaling pathways may be useful in the treatment of human breast cancer [52]. Specifically, novel combination strategies may be developed to prevent or delay the development of endocrine therapy resistance [3, 39], to restore sensitivity to endocrine therapy [53], and to treat hormone-resistant tumors [18, 46]. Anti-vascular endothelial growth factor (VEGF) therapy with bevacizumab may be able to overcome resistance to endocrine therapy and improve efficacy in HR+ metastatic breast cancer [54], and preclinical models have shown that the estrogen-induced increase in VEGF expression may be counteracted by aromatase inhibition. Inhibition of growth factor signaling and angiogenesis pathways may be rationally combined with conventional endocrine strategies for breast cancer [4, 35, 50, 55, 56].
Table 1Summary of letrozole in combination with growth factor signaling inhibitors in preclinical modelsTarget for growth factor inhibitorCombination regimenSummary of key findingsReferencesEGFR/HER2Letrozole + AEE78Combination enhanced antiproliferative effects in MCF-7 (ER+ HER2−), ZR75.1 (ER+ HER2+), and BT474 (ER+ HER2+) cell lines[42]Partial restoration of growth inhibitory effects of letrozole in refractory cell lines (LTLT-Ca; long-term letrozole treated)[43]mTORLetrozole + RAD001 (everolimus)Letrozole + RAD001 significantly increased apoptosis compared with either agent alone[50]Co-treatment increased sensitivity to letrozole in resistant MCF-7 cells with constitutively active Akt[48]RAD001 increased antiproliferative effects of letrozole in MCF-7 Arom 1 cell line[47]IGFBPLetrozole + rhIGFBP-3rhIGFBP-3 Enhanced letrozole activity in MCF-7-Ca cells in vitro and in vivo[51]rhIGFBP recombinant human insulin-like growth factor binding protein
The activity of inhibitors of growth factor signaling depends on the presence of specific cellular aberrations, such as overexpression of HER2 [57] or mutations of EGFR [58]. Consequently, targeted therapies may have limitations as single agents because the target is active in a restricted subset of patients, and breast tumors may undergo adaptive changes to render the target redundant. Since tamoxifen exhibits agonist effects on breast cancer cells exposed to long-term estrogen deprivation [12], it may be better to combine AIs with inhibitors of growth factor signaling. Letrozole is one of the most potent AIs and is one of the most extensively studied AIs in combination with new agents (see Table 1).
Clinical trials of letrozole in combination with inhibitors of growth factor signaling pathways
Based on results from preclinical studies, several clinical trials of novel combinations are under way, with the aim of improving efficacy and safety of endocrine therapy with letrozole (see Table 2). Many of these trials are being conducted in patients with locally advanced or metastatic breast cancer who have failed prior tamoxifen or have a suboptimal response to letrozole. This represents a high-risk, difficult-to-treat population who are candidates for cytotoxic chemotherapy. Preliminary results have shown that letrozole can be safely combined with trastuzumab, lapatinib, everolimus, tipifarnib, bevacizumab, and imatinib. It is too early, however, to make definitive conclusions about efficacy and clinical benefits with these novel combinations.
Table 2Clinical studies of letrozole and inhibitors of growth factor signaling pathwaysTarget for growth factor inhibitorCombination regimenStudy type and patient populationSummary of key findingsReferencesHER2Letrozole + trastuzumabPhase IIMetastatic BC, postmenopausal, ER+ and/or PR+, HER2+ (n = 31)ORR 26%; median TTP 5.8 months[59]EGFR/HER2Letrozole + lapatinibPhase IAdvanced BC (ER+ or PR+) or other tumors (n = 36)Letrozole + lapatinib safely combined at recommended single agent doses[60]Phase IIIAdvanced/metastatic BC (n = 1,200 target accrual)Ongoing trial; primary end point TTP[61]mTORLetrozole + RAD-001 (everolimus)Phase IbAdvanced BC pts with suboptimal response to letrozole (n = 6)RAD001 pharmacokinetics not altered by letrozole[62]Phase IIPresurgical therapy in patients with newly diagnosed ER+ BC (n = 255 planned)Ongoing trial of efficacy and biomarkers[63]Letrozole + CCI-779 (temsirolimus)Phase IIAdvanced or metastatic BC (n = 92)No difference in ORR, but trend to longer PFS with letrozole + temsirolimus (30 mg)[64]Phase IIIAdvanced or metastatic breast cancer (n = 1,236 planned)Terminated[65]Farnesyl transferaseLetrozole + tipifarnib Randomized phase IIAdvanced or metastatic BC that has progressed on tamoxifen (n = 121)ORR 38% for letrozole and 26% for letrozole + tipifarnib (NS)[66]Randomized, placebo-controlled phase IIAdvanced or metastatic BC that has progressed on antiestrogen therapy (n = 120)No longer recruiting[67]VEGFLetrozole + bevacizumab (anti-VEGF monoclonal antibody)Phase IIMetastatic BC, postmenopausal, candidates for AI (n = 28)Letrozole + bevacizumab is well-tolerated[54]Endocrine therapy (tamoxifen or aromatase inhibitor) + bevacizumabPhase III placebo-controlledFirst-line therapy in ER+/PR+ Metastatic BC (n = TBC)Planned trialPlanned CALGB trial [54]Bcr-ablLetrozole + imatinibPhase IIMetastatic BC, postmenopausal ER+ and/or PR+ (n = 15)Letrozole + imatinib is feasible[68]BC breast cancer, ORR overall response rate, TTP time to progression, PFS progression-free survival, AI aromatase inhibitor, TBC to be confirmed
In a clinical trial [69] designed to test whether combination therapy with letrozole and bevacizumab was possible, patients with ER+ or progesterone receptor-positive (PR+) metastatic or locally advanced breast cancer were treated with letrozole (2.5 mg daily) and bevacizumab (15 mg/kg IV every 3 weeks) [54]. The majority of patients had received prior therapy with a nonsteroidal AI. The combination of bevacizumab and letrozole was found to be well-tolerated. Common drug-related toxicities reported were hypertension, fatigue, headache, and joint pain. Median progression-free survival was reported to be 10 months, and this compares favorably with the published data on median time to progression with first-line letrozole (9.4 months) [70]. However, analysis of efficacy and biomarker data was confounded by the long duration of prestudy aromatase inhibition [54, 71]. Nevertheless, when the data were corrected for duration of previous AI therapy, the study did determine that changes in circulating endothelial cell (CEC) levels may be a biomarker of response or progression on anti-angiogenic therapy [71]. Based on these findings, a randomized, double-blind, placebo-controlled trial of bevacizumab combined with endocrine therapy in patients with ER+ or PR+ metastatic breast cancer has been initiated by the Cancer and Leukemia Group B (CALGB) (see Fig. 2) [54]. The primary end point of the trial is progression-free survival.
Fig. 2Planned CALGB trial of first-line endocrine therapy (tamoxifen or aromatase inhibitor) with or without bevacizumab
Letrozole plus trastuzumab produced durable responses in about one in four patients with ER/PR+, HER2+ metastatic breast cancer, but early progression occurred in one in two patients [59]. This suggests that common resistance pathways may be responsible for relapse [59]. Targeting multiple pathways may reduce the risk of resistance. The combination of letrozole and the dual EGFR/HER2 inhibitor lapatinib was found to be feasible and well-tolerated in a phase I study [60], and this regimen is currently being compared with letrozole plus placebo in a phase III study in women with ER/PR+ advanced or metastatic breast cancer [61] (see Fig. 3). Future studies will focus on finding the right combination or sequence of agents for different patients in specific treatment settings. Furthermore, studies in the neoadjuvant setting in patients with locally advanced breast cancer will allow correlative biomarker assessment, such as the proliferation marker Ki-67, to determine the efficacy of combination therapies. For example, everolimus and letrozole are being studied as preoperative therapy of primary breast cancer in postmenopausal women [63]. In this phase II trial, patients are randomized to receive letrozole in combination with everolimus or placebo, an adaptive design strategy, so that identification of biomarkers can be used to optimize patient selection for a future phase III trial of first-line combination therapy in patients with advanced breast cancer. Even if a biomarker is not identified, the trial is adequately powered to demonstrate a statistically significant difference in treatment effect of the combination in the overall population.
Fig. 3Letrozole and lapatinib phase III trial design. Target recruitment: 1,280 patients
Combination strategies may also change the conventional approach for treatment selection based on HR status. Interestingly, it has been suggested that treatment of HER2+, ER− metastatic breast cancer with trastuzumab may transform the tumor phenotype to become hormone responsive [72]. The hypothesis-generating study found that three of ten patients demonstrated ER overexpression at 9, 12, and 37 weeks, respectively, following the initiation of trastuzumab. Two of these patients were subsequently treated with endocrine therapy alone and one received letrozole for 3 years without evidence of progression. Further trials of letrozole used in combination with trastuzumab or sequentially in patients with HER2+, ER− tumors appear warranted.
Microarray/gene profiling studies and optimization of treatment with letrozole
The growing importance of biomarkers in oncology has been reflected in the US Food and Drug Administration (FDA), the National Cancer Institute (NCI), and the Centers for Medicare and Medicaid Services (CMS) Oncology Biomarker Qualification Initiative (OBQI), designed to improve the development of cancer therapies and the outcomes for cancer patients through biomarker development and evaluation [73]. Not only are biomarkers potentially useful as prognostic and predictive factors but they also serve as surrogate end points for long-term outcomes. For example, Dowsett et al. found that changes in Ki-67 in the neoadjuvant setting may be used to predict likely benefit (improved disease-free and overall survival) of AIs in the adjuvant setting [74], potentially expediting clinical development. Therefore, identifying biomarkers that predict endocrine resistance is essential for the optimal use of letrozole. In addition, biomarker studies that help to delineate the oncogenic pathways may be particularly useful for the development of novel combination therapies with letrozole.
Gene expression analysis of tumors represents a novel approach for biomarker development that promises to increase the understanding of breast cancer heterogeneity and facilitate the development of individualized treatment strategies. Microarray analysis, such as the Affymetrix Human Genome Gene Chips, is an exciting development in breast cancer diagnostics that allows the expression of genes in tumors to be quantified using RNA retrieved from breast cancer biopsies. Using cDNA microarray and unsupervised clustering analysis, ER+ breast cancer can be subdivided into at least two subtypes, luminal A and luminal B, with distinct gene expression patterns and clinical outcomes [75–77]. Supervised analysis, comparing gene expression patterns between tumors that relapsed and those that did not, generated a “gene signature” as an independent prognostic parameter for lymph-node negative disease [78, 79]. The National Surgical Adjuvant Breast and Bowel Project (NSABP) has developed a multigene profiling assay, using reverse transcriptase polymerase chain reaction (RT-PCR) to quantify the expression of 21 genes [80]. The 21-gene assay (Oncotype DX, Genomic Health) can predict risk of recurrence in patients with HR+ early breast cancer receiving tamoxifen [80, 81] or chemotherapy [82]. Naderi et al. [83] have also reported validation of gene expression signatures that may have predictive value in the clinic. Furthermore, newly identified interactions between ER and other signaling pathways have been studied using microarray analysis [84, 85]. For example, Bex1 and Bex2 genes have been identified as novel breast cancer-related genes and identify a subtype of ER+ tumors associated with estrogen response and nuclear factor kappa B (NF-κB) pathways [84]. Advances in gene profiling will provide further insights into tumor biology and improve prediction of likely response to specific therapies in the clinic.
Microarray analysis has shown to be valuable in predicting response to neoadjuvant endocrine therapy and was used in a study of neoadjuvant letrozole 2.5 mg/day administered to postmenopausal women with large operable or locally advanced breast cancers for 3 months [86, 87]. Changes in patterns of gene expression were assessed from tissue samples taken at diagnosis, 14 days, and 3 months (N = 58). Changes in gene expression level with treatment were identified as early as 14 days and involved classical markers of estrogen action (trefoil factors 1 and 3, LIV-1, KIAA0101) as well as tumor proliferation (cyclin D1, cyclin B2, CSK2, cell division cycle 2). The objective clinical response to neoadjuvant letrozole was 71% (N = 52 assessable). Of note, the observed changes in gene expression, when clustered, were predictive of response in all cases except one, whereas classical markers of estrogen action were not predictive [87]. In another study, postmenopausal women with primary operable breast cancer were randomized to 2 weeks of presurgical treatment with letrozole 2.5 mg/day or anastrozole 1 mg/day [88]. Microarray gene expression profiling (Breakthrough Centre cDNA chips) of biopsies taken before and during treatment identified differences in gene expression patterns between the two AIs, although the clinical significance of these preliminary findings remains to be clarified. Itoh et al. [89] also found differences between letrozole, anastrozole, and tamoxifen using microarray analysis on MCF-7 cells stably transfected with the aromatase gene (MCF-7Aro). Gene expression patterns revealed a high correlation between the AIs (letrozole and anastrozole) and a clear difference between AIs and tamoxifen [89].
Emerging data therefore suggest that estrogen-responsive genes are candidate biomarkers [90] and may be useful in the clinic as predictive factors of benefit from AI treatment [89]. The value of gene expression profiling is now being evaluated in prospective studies. MINDACT (Microarray In Node negative Disease may Avoid ChemoTherapy), an ongoing three-part randomized trial of 6,000 patients with node-negative breast cancer, is comparing the efficacy of selection of breast cancer patients for adjuvant chemotherapy based on either clinical criteria or the 70-gene microarray prognosis profile [91]. In one part of the trial, 3,500 node-negative, HR+ positive patients will be randomized to receive either 7 years of letrozole or 2 years of tamoxifen followed by 5 years of letrozole. In the US, the TAILORx Trial [92] is comparing hormone therapy with or without combination chemotherapy as adjuvant therapy for node-negative, ER+ breast cancer. The objective of this randomized phase III trial is to determine the best individual therapy using Oncotype DX gene profiling.
One of the more recent advances in transcriptional profiling addresses its potential application as an identification tool for “oncogenic pathway signatures” that could be used to guide targeted therapy. Oncogenic pathway signatures, developed in cell-line models, have been shown to predict sensitivity to therapeutic agents in vitro [93]. This offers an opportunity to identify pathway-specific drugs in endocrine therapy-resistant tumors and creates an opportunity for the rational design of combination therapy with letrozole. In addition to the RNA-based gene expression profiling approach, high throughput analysis at the DNA level, such as genome-wide microarray comparative genomic hybridization (aCGH) and DNA sequencing, has made it possible to decipher the genetic anomalies that drive a particular tumor phenotype. The use of these molecular approaches for biomarker development is still in its infancy but is now feasible with recent advances in genomic technologies.
Conclusions
Hormone-sensitive breast cancer can be regarded as a chronic disease with a persistent risk of escape from effective endocrine control. Activation of growth factor signaling pathways has been implicated in progression of HR+ breast cancer to an estrogen-independent phenotype and the development of resistance to endocrine therapy, particularly tamoxifen. It has been hypothesized that combining endocrine therapy with targeted signal transduction inhibitors may circumvent hormone-independent signaling pathways, so that patients may experience prolonged disease control.
Letrozole is one of the most potent AIs. As such, it may be more effective than tamoxifen for patients with tumor profiles associated with a high risk of developing hormone resistance (e.g., tumors with HER2 gene amplification) and represents an ideal combination partner for agents that inhibit growth signaling pathways implicated in hormone resistance. The efficacy of letrozole is currently being investigated with a variety of signal transduction inhibitors with different mechanisms of action, including monoclonal antibodies against HER family receptors, receptor tyrosine kinase inhibitors, and downstream signaling pathway inhibitors.
Gene expression profiling has been validated as a useful new tool to predict risk of relapse in patients treated with hormone therapy and chemotherapy. This approach will help physicians to identify which patient will likely benefit from specific therapies, such as letrozole. Tailoring therapy to individual patient profiles (clinical, histologic, pathologic, and genetic) will become more sophisticated in the future, helping to maximize the benefits of endocrine therapy throughout the breast cancer continuum; letrozole will undoubtedly become an integral part of the next generation of tailored combination regimens for the treatment of breast cancer.
|
[
"combination therapy",
"breast cancer",
"letrozole",
"aromatase inhibitors",
"postmenopausal",
"adjuvant therapy",
"mechanism of resistance"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"R"
] |
Mol_Biochem_Parasitol-1-5-1906846
|
Characterisation of exogenous folate transport in Plasmodium falciparum
|
Folate salvage by Plasmodium falciparum is an important source of key cofactors, but little is known about the underlying mechanism. Using synchronised parasite cultures, we observed that uptake of this dianionic species against the negative-inward electrochemical gradient is highly dependent upon cell-cycle stage, temperature and pH, but not on mono- or divalent metal ions. Energy dependence was tested with different sugars; glucose was necessary for folate import, although fructose was also able to function in this role, unlike sugars that cannot be processed through the glycolytic pathway. Import into both infected erythrocytes and free parasites was strongly inhibited by the anion-channel blockers probenecid and furosemide, which are likely to be acting predominantly on specific folate transporters in both cases. Import was not affected by high concentrations of the antifolate drugs pyrimethamine and sulfadoxine, but was inhibited by the close folate analogue methotrexate. The pH optimum for folate uptake into infected erythrocytes was 6.5–7.0. Dinitrophenol and nigericin, which strongly facilitate the equilibration of H+ ions across biological membranes and thus abolish or substantially reduce the proton gradient, inhibited folate uptake profoundly. The ATPase inhibitor concanamycin A also greatly reduced folate uptake, further demonstrating a link to ATP-powered proton transport. These data strongly suggest that the principal folate uptake pathway in P. falciparum is specific, highly regulated, dependent upon the proton gradient across the parasite plasma membrane, and is likely to be mediated by one or more proton symporters.
1
Introduction
Reduced folates are essential cofactors for one-carbon transfer reactions, including the conversion of dUMP to dTMP, which is a prerequisite for DNA synthesis. Because of this, the folate pathway has long been a target for drugs deployed against rapidly reproducing cells such as cancers and a range of microbial pathogens. Whereas most microorganisms can synthesise the folates they need from the simple precursors GTP, p-aminobenzoic acid (pABA) and glutamate, higher animals (but not plants) have lost the ability to do this and depend on dietary intake of pre-formed folate as an essential nutrient. The human malaria parasite Plasmodium falciparum is able to exploit both of these routes [1–4]. Thus, it can utilise folate provided in culture medium in vitro or salvaged from the host plasma in vivo on the one hand, or convert the above precursors de novo into folate derivatives on the other, a characteristic also shared by the related apicomplexan parasite Toxoplasma gondii
[5,6]. The relative importance of the biosynthetic and salvage pathways across the complete life cycle in vivo and the interplay between them is poorly understood, although existing data support the view that both are necessary for healthy propagation of the parasite, at least in the erythrocytic stages [7].
An important aspect of exogenous folate utilisation is the machinery and mechanism(s) by which folate is imported into the parasite. The highly polar nature of folate derivatives suggests that salvage must employ some kind of mediated transport process, as diffusion alone across the membrane is likely to be far too inefficient. Moreover, folate molecules are dianionic at physiological pH and must be imported into the parasite against an inwardly negative electropotential that has been measured as ca. −95 mV in P. falciparum
[8]. Although the transport of other key molecules, such as pantothenate, lactate, glucose and choline, has been investigated [9–13], there has been no detailed study to date of this aspect of folate metabolism in P. falciparum, a better understanding of which might lead to new ways of inhibiting parasite growth. A preliminary approach to this end has been taken by demonstrating that the anti-gout drug, probenecid, which among other things inhibits folate transport in mammalian cells [14], can increase the sensitivity of P. falciparum to antifolate inhibitors [15]. Here, we demonstrate that folate uptake by this parasite is a regulated process that is critically dependent upon provision of glucose or another sugar that can proceed through the glycolytic pathway, that the existence of a pH gradient across the plasma membrane is also required for efficient transport, and that folate is transported principally by a proton-symport mechanism.
2
Materials and methods
2.1
Chemicals
[3′,5′,7,9-3H]folic acid, 24 Ci mmol−1, 1 mCi ml−1 was from Amersham, UK, [3′,5′,7,9-3H]folinic acid (26 Ci mmol−1, 1 mCi ml−1) and [3′,5′,7,9-3H]5-methyltetrahydrofolic acid (44 Ci mmol−1, 1 mCi ml−1) were both from Moravek, California. Folic acid, folinic acid, 5-methyltetrahydrofolic acid, 2,4-dinitrophenol, concanamycin A, d- and l-glucose, d-fructose, d-xylose, d-galactose, 6-deoxy-d-glucose, probenecid and furosemide were all purchased from Sigma, UK. Note that we use the term ‘folate’ generically to indicate derivatives of the folate family of molecules regardless of their oxidation state, modifications at the 5 and 10 positions or polyglutamation status.
2.2
Parasite culture
P. falciparum was routinely cultured under 1% O2, 3% CO2, 96% N2 in RPMI 1640 medium, supplemented with d-glucose (22 mM final concentration), hypoxanthine (36 mM), HEPES (25 mM), gentamicin sulfate (50 μg/ml) and 0.5% Albumax II (Invitrogen). The cultures were synchronised by haemolysis of mature, late trophozoite-stage parasitised erythrocytes by suspension in 9 volumes of a 5% sorbitol solution at room temperature for 5 min. Cells surviving the treatment were used to set up new cultures and the process repeated where necessary to achieve a tighter synchronisation [16]. Given that synchrony is never perfect, time zero for the erythrocytic cycle was taken as the point where ca. 90% of infected erythrocytes were ring form, with the remainder as very late schizonts. To obtain free parasites, synchronised cultures, normally in the late (mature) trophozoite stages (ca. 30 h into the cycle), were quickly lysed with 0.05% (w/v) saponin at room temperature, permeabilising the red cell and parasitophorous membranes and leading to haemolysis [17,18]. Any remaining unlysed red cells were further treated with a PBS wash containing the same amount of saponin. The freed parasites were then resuspended in the appropriate buffer with or without glucose, depending upon the purpose of the assays. Saponin-freed parasites that are glucose-replete maintain intracellular levels of ATP for periods of at least 30 min and do not show evidence of leakage [19].
2.3
Uptake assay of radiolabelled folates
Folate uptake/export assays were performed on free parasites, parasitised red cells or uninfected red cells, as appropriate. Cell numbers were estimated with a haemocytometer. Normally preparations containing 107–108 parasites were used for each assay point (or ca. 107–108 uninfected red cells). Prior to uptake assays by naked parasites, infected cells were lysed in 0.05% saponin and the freed parasites washed either in PBS or folate/pABA depleted RPMI 1640, depending upon the experiment. Wash steps were also performed with different buffered or non-buffered isotonic salt solutions where necessary as indicated in the relevant text.
All cells, washed extensively to remove folate present in the culture medium, were mixed with or without glucose (20 mM final concentration) and inhibitors as appropriate, together with radiolabel (normally 1 μCi/ml, equivalent to 38 nM, unless otherwise stated) to make up a total volume of 100 μl with PBS or alternative buffer. All the components were prewarmed to 37 °C before the start of the assay, the reaction mix incubated at 37 °C normally for 30 min, and stopped by addition of 1 ml ice-cold PBS.
When assaying for folate uptake by whole parasitised red cells, these were spun down after the uptake period and the pellet washed with 1 ml ice-cold PBS at least three times to bring extracellular label down to the background level. The washed parasite pellet obtained by subsequent saponin treatment was then lysed with 0.02% SDS before counting. Uptake by naked parasites was determined in the same way. Unless otherwise indicated, the uptake values represent labelled folate that has been imported into synchronised trophozoite stage parasites. All assays were routinely performed in triplicate on different batches of parasites unless otherwise indicated, and the data expressed as the mean ± S.D.
2.4
Extraction and affinity purification of folate derivatives
For characterisation of labelled folates, parasite pellets were washed three times in 1 ml PBS to remove any extracellular label and resuspended in 1 ml extraction buffer (0.1 M Tris–HCl, 2% ascorbic acid, pH 7.5), heated in a boiling water bath for 10 min, centrifuged at 10,000 × g for 10 min and supernatants stored at −20 °C until affinity purification and HPLC analysis, performed as described [20]. Unlabelled folic acid, folinic acid and PteG2 to PteG5 of the polyglutamated forms of folic acid were added into the extracted folates as internal standards. An identical aliquot of sample was also spiked with 3H-folinic acid in a second run to confirm the identity of the radioactive peak of folinic acid that had been extracted from the parasites.
2.5
pH measurements
3H-folinic acid uptake as a function of pH was measured in a combination HEPES-MES buffer system containing 20 mM of each component, and the salt concentration made up to 154 mM with NaCl. By taking the temperature coefficients of the buffers into account, pH values were adjusted at room temperature such that the apparent pH value at the assay temperature (37 °C) was exactly the pH required. Glucose was added to 20 mM as required. The internal pH of isolated parasites was measured as described [21] using the pH-sensitive fluorescent indicator BCECF-AM. Loading was achieved by incubating isolated parasites suspended at a cell density of 0.9–2.1 × 108 cells/ml in folate and pABA depleted RPMI 1640 without Albumax II, and containing 10 μM BCECF-AM, for 10 min at 37 °C. The cells were then washed by centrifugation and resuspended in culture medium and used immediately. An aliquot of cells was transferred to an appropriate buffer in a 2 ml cuvette in the temperature-controlled chamber of a spectrofluorometer, maintained at 37 °C. The sample was measured at 520 nm after successive excitations at 440 and 495 nm. The ratio of the fluorescence emissions (excitation at 495 nm/excitation at 440 nm) was used to calculate the internal pH of the cells. A standard curve was prepared with parasites resuspended in 130 mM KCl, 10 mM NaCl, 1 mM MgCl2 and 40 mM of MES-HEPES buffer from pH 5.5 to pH 8 and 30 μM nigericin [21].
3
Results
3.1
Uptake of folates over the asexual erythrocytic cycle
To monitor the ability of P. falciparum to salvage folate at different stages of the erythrocytic cycle, synchronised parasite cultures were incubated with different forms of radiolabelled folate. 5-Methyltetrahydrofolate (5-MeTHF) is the fully reduced form most abundant in human plasma [22] and thus the principal derivative available to the parasite in vivo. Folinic acid (5-formylTHF), although not known to be a normal physiological substrate for the parasite, is a more stable member of the fully reduced (active) forms that closely resembles 5-MeTHF and is used clinically (leucovorin). The capacity of ring stages to take up both of these compounds was very low but increased markedly ca. 15 h into the cycle as the parasites entered the trophozoite stages, peaking at around 30–35 h, before falling to very low levels again in mature schizonts (Fig. 1a). The numbers of parasites at the different stages were monitored throughout such time-courses and showed that uptake of folate increased as the parasites proceeded through the trophozoite stages against a near-constant number of infected cells, as expected for a synchronous culture. At around 40 h, some schizonts began to release merozoites, leading to a sharp rise in the total number of infected cells, but the uptake from this population continued to diminish, consistent with lower levels of uptake for both rings and schizonts, relative to trophozoites (Fig. 1a). To check that the marked reduction of uptake in the latter stages of the cell-cycle (beyond 40 h) was not due to parasites becoming unhealthy or inviable, the phenomenon was also monitored through a subsequent cycle, where the same pattern was repeated (data not shown). Only minor differences were seen between parasite clones of different lineages, and labelled folic acid, the stable oxidised form extensively used experimentally to study folate metabolism and a common dietary supplement, gave qualitatively similar results to the reduced folates. This pattern strongly suggested that folate uptake is a process regulated by P. falciparum, likely to involve a transport mechanism more complex than facilitated diffusion. On the basis of the data above, parasites at the late trophozoite stage of synchronous cultures were used to investigate the characteristics of uptake in subsequent experiments, using 3H-folinic acid as the label, as this is considerably more stable than 5-MeTHF, is more readily available in labelled form, and has been shown in earlier studies to provide a better source of cofactor than equimolar amounts of folic acid [23].
Uptake of folinic acid into the parasite was also markedly dependent upon temperature. Only a very low level of transport was observed between 10 and 25 °C, over which the parasites are viable, but metabolically retarded. The rate of uptake increased dramatically above ca. 30 °C, peaking quite sharply between 37 and 39 °C, but was abolished entirely above 45 °C, a temperature at which the parasites are no longer viable. Data for free parasites are shown in Fig. 1b; very similar profiles were obtained for infected erythrocytes with the same optimum temperature (data not shown). This demonstrated that active parasite metabolism in live organisms at their normal growth temperature is required for folate to be imported to a significant degree.
3.2
Kinetics of folate uptake
Assays of the time-dependence of folate uptake into trophozoites showed that it was linear (R2 ≥ 0.98) over a period of at least 80 min (Fig. 2a). In subsequent experiments, unless otherwise indicated, a standard incubation time of 30 min was therefore adopted to ensure that adequate counts for accurate measurements were obtained, whilst remaining well within the window of linearity. Uptake rates of folinic acid into parallel samples of free parasites and infected red cells in equal numbers were comparable, with somewhat higher levels measured in the latter (Fig. 2b), indicating that the rate-limiting step is not passage across the erythrocytic membrane, but rather across the parasite plasma membrane. In free parasite cells, the folate uptake process exhibited saturable and non-saturable components (Fig. 2c), indicative of more than one mechanism. In low concentrations of the substrate (in the region of 0 to 15 μM, which encompasses the normal range of human plasma folate, 30–40 nM [24]), folinic acid uptake is non-linear, consistent with the operation of a carrier-mediated process (Fig. 2c). At higher concentrations up to mM levels, uptake became linear, a phenomenon also observed in other folate transport systems [25,26], including that of Toxoplasma
[6]. Subtraction of the linear component yielded Km values for the saturable process of 3.4 ± 0.8 and 2.4 ± 0.3 μM, respectively, for FCB and K1 parasite strains (Fig. 2d). Interestingly, while their Km values were closely similar, the observed for K1 (34.7 ± 2.6 pmol/107 parasites/h) was ca. 10-fold higher than that for FCB (3.3 ± 0.3 pmol/107 parasites/h).
3.3
Export of folinic acid
Export of folinic acid from free parasites was compared with that from uninfected cells. Uptake was first allowed to proceed into either infected or uninfected erythrocytes for 1 h in PBS plus 20 mM glucose at 37 °C with labelled folinic acid. After incubation and washing, cells were resuspended in the same buffer without label and incubated for different times (Fig. 3). Cells were then spun down and the radioactivity in the supernatant measured. To measure label exported from free parasites, they were first released from erythrocytes with 0.05% saponin and then washed and treated in a similar manner as for the uninfected red cells. The level of label in the supernatant of the red cell aliquots tripled over the 80 min incubation time, indicative of facile export. However, export of the label from free parasites was much less rapid, only increasing by ∼20% over the same period.
In the case of Lactobacillus salivarius, folate internalised over a similar period cannot be released from the cells and is recovered in cell extracts primarily in polyglutamated forms that play an important role in the retention and accumulation of intracellular folate [27,28], and previous studies using P. falciparum have demonstrated the presence of polyglutamated forms of folate produced from labelled monoglutamated precursors [2,23,20]. To examine whether the slow rate of export seen above was due to trapping by polyglutamation, the folinic acid internalised after 1 h was then extracted and analysed by HPLC on a C18 column [23,20]. The elution profiles indicated that the imported folinic acid had not been significantly modified by this stage, as no labelled polyglutamated folates could be detected in the cell extracts after this relatively short period of incubation, nor was there evidence of any conversion of label to another form of folate (data not shown). These results, together with our earlier data, where overnight incubations were required to detect polyglutamated forms [23,20], suggest that although these forms of folate probably play an important role in ultimately retaining the folates in malaria parasites and modulating the affinity of different enzymes for them, conversion to such forms is not rapid and appears not to play any immediate role in the salvage process. The low rate of net export from the parasite is however consistent with the operation of a transport system that actively imports folates into the parasite, where they are efficiently retained (see below).
3.4
The dependence of folate uptake on carbon source
To investigate possible metabolic requirements for folate uptake, its dependence on glucose and other sugars was studied in parasitised red cells. In the absence of glucose or at concentrations <1 mM, only a very low level of import was measurable. As the glucose concentration was raised, a marked increase in imported folinic acid was observed, peaking around 20 mM (Fig. 4a). This corresponds closely to the level of glucose in standard P. falciparum culture medium (22 mM), empirically determined historically to give optimal levels of growth. Time course experiments also showed that the higher the glucose concentration, the longer the uptake process could be sustained by the parasites (Fig. 4b).
d-Glucose could be replaced by the alternative hexose sugar d-fructose, which can also be utilised in the glycolytic pathway of P. falciparum
[29,30], but substitution by sugars that cannot, i.e. l-glucose, d-galactose, d-xylose or 6-deoxy-d-glucose, almost completely abolished uptake (Fig. 4a). Cellular energy in P. falciparum is predominantly provided by glycolysis, with the parasite using ca. 100 times more glucose than uninfected erythrocytes to supply their considerable energy needs [31], and uptake of folate is thus clearly dependent upon the operation of normal levels of intracellular energy metabolism.
3.5
Effect of pH on folate uptake
In the presence of glucose or other sugar susceptible to glycolysis, as in the above experiments, the malaria parasite has a higher cytoplasmic pH (7.2–7.4) and a lower parasitophorous vacuole pH (6.9), produced by proton pumping activity on their plasma membrane [21,32]. Such pumping is necessary to prevent the build-up of intracellular acidity, and the higher proton concentration thus created outside the cell can be utilised for the purposes of nutrient transport [21]. This prompted us to investigate the effect of pH on the uptake of folinic acid. For this, a HEPES-MES buffer system was used that provided a wider range of buffering capacity than either buffer alone. The HEPES-MES system contained 20 mM of each component and the salt concentration was made up to 154 mM with NaCl. Initially, glucose (20 mM) was included in all wash steps and assay reactions. For parasitised red cells under these conditions, the optimal pH was 6.5–7.0, uptake then decreasing linearly with increasing pH between pH 7.0 and 8.0 (Fig. 5a), with a similar pattern being observed for free parasites over this range (Fig. 5b). However, the transmembrane pH gradient generated in this way is dynamic, subject to change by normal cellular metabolism. To avoid this potentially confounding effect, similar experiments were also conducted in the absence of glucose, thus cutting off the energy supply for the proton pump(s), with the desired buffering and osmotic requirements maintained by NaCl, HEPES and MES. A series of artificially created, constant transmembrane proton gradients were established in this way for the uptake assay. Saponin-released parasites were washed thrice with 154 mM NaCl to remove folate and other nutrients and to effect exhaustion of the glucose, thereby bringing all parasites in a given sample to the same internal pH. The results showed that under these conditions, the uptake of folinic acid is positively correlated with the transmembrane pH gradient over a wider range (Fig. 5c). The lower the pH of the resuspension buffer, the greater the pH gradient thereby provided, which in turn yielded a higher uptake of folinic acid. Above pH 6.5, the pH response curve flattened out and exhibited a relatively constant, low level of uptake. This inflection point may correlate with the internal pH of the parasites when depleted of glucose under the above conditions [21]. In some cases this was confirmed by measuring the internal pH value with BCECF-AM as described in Section 2. Attempts to directly measure a change in pH as a result of folate import in the absence of glucose and using a minimal buffer concentration were unsuccessful, possibly because of the very small quantities that are taken up and the relatively slow rates involved.
3.6
Effect of perturbing the plasma membrane proton gradient on folate uptake
2,4-Dinitrophenol (DNP) is a widely used ionophore in the uncoupling of oxidative phosphorylation from electron transport on the mitochondrial membrane, as well as in studies of transport across the plasma membrane. Mechanistically, it provides an efficient by-pass route for protons that dissipates their gradient and hence the proton motive force. Here it was found that it also inhibits folinic acid uptake into parasite cells, with an IC50 of ca. 26 μM (Fig. 6a and Table 1), again suggesting that this uptake is intimately coupled with the energy metabolism of the parasites. Similarly, the K+/H+ antiporter nigericin had a profound effect on the uptake of label, which decreased in a dose-dependent manner, with a marked drop in the 100 pM range. Interestingly, this effect levelled off at around 1 nM to about 75% inhibition and showed little further drop up to 1 μM nigericin (Fig. 6b), suggesting that the major, but possibly not sole, factor determining the uptake of folinic acid is the magnitude of the pH gradient. A similar phenomenon was observed when the V-type ATPase inhibitor concanamycin A [33] was employed. Such an enzyme has been identified as the principal source of proton pumping in P. falciparum
[21,32]. This inhibitor was effective at sub-micromolar concentrations (IC50 ca. 16 nM; Fig. 6c), but only abolished ca. 60% of folate uptake, with no further effect beyond about 100 nM.
3.7
Metal ion dependency of folate uptake
To investigate whether folate uptake was affected by other cations, in addition to its dependence on the proton gradient demonstrated above, parasitised cells were first washed with 100 volumes of buffer containing either (i) PBS (the buffer in which other uptake experiments were conducted), (ii) a near-physiological concentration of Na+ (130 mM NaCl, 20 mM Tris–HCl, pH 7.0), (iii) an equivalent concentration of K+ (130 mM KCl, 20 mM Tris–HCl), (iv) the non-ionic polyol mannitol (130 mM mannitol, 20 mM Tris–HCl), or (v) Tris alone (150 mM Tris–HCl). Washes were repeated thrice to reduce the contribution of pre-existing salts to an insignificant level compared to those in the assay buffer. Uptake rates of folinic acid from all five of these solutions were similar (<±20% from the mean). Also, the presence of EDTA up to 5 mM or Mg2+ up to 10 mM had only a minor effect (data not shown). It thus appears unlikely that sodium, potassium or divalent metal ions are critically involved in the co-transport of folate into the malaria parasite.
3.8
Effect of antifolate drugs and other inhibitors on folate uptake
A range of other potential inhibitors and competitors were tested for their ability to affect transport of folate into the parasite (Table 1). Antifolate drugs exert their effect principally by binding to target enzymes (DHFR and DHPS) in the folate pathway, lowering pools of reduced folate and thus compromising key metabolic steps, especially DNA synthesis. However, in principle, they might also exhibit an inhibitory capability by interfering with the uptake of exogenous folate [4]. To investigate this possibility, uptake of labelled folinic acid into free parasites was monitored in the presence of the DHFR inhibitors pyrimethamine (PYR) and methotrexate (MTX), and the DHPS inhibitor sulfadoxine (SDX). Over a range of 0.4 nM to 40 μM for PYR and 0.3 nM to 322 μM for SDX, no inhibitory effect on folate uptake could be detected (the highest concentrations are ca. 900 and 500 times greater than the respective IC50 values for these drugs measured for growth inhibition). However, MTX did show inhibition of folinic acid uptake, consistent with its much closer structural relationship to folate than either Pyr or SDX. This was seen clearly when MTX (IC50 of ca. 13 μM) was compared to unlabelled folinic acid (2.6 μM), 5-MeTHF (3.5 μM) and folic acid (19.8 μM), where the drug competed with uptake of the label to a similar degree. The specificity of the transport process across the parasite plasma membrane suggested by the above results is consistent with a complete lack of competition with pantothenate (Table 1), the precursor of coenzyme A and an essential parasite nutrient [21], which is also a small but unrelated organic acid with a molecular weight about half that of folate.
To further explore the nature of the transporters involved in folate uptake, we employed two known channel blockers, probenecid and furosemide. Probenecid is known to inhibit folate uptake via an anion carrier in mammalian cells at mM levels [34,35], and is able to sensitise parasites to antifolate drugs [15]. Furosemide has been extensively used in nutrient transport studies of the malaria parasite, as it efficiently inhibits the so-called new permeability pathways (NPP) [36–38], which first appear in the erythrocyte membrane of parasitised red cells about 12–15 h into their intraerythrocytic development. The NPP show characteristics of anion-selective channels, but are of broad specificity and increase erythrocyte permeability to a range of small molecular weight anions, cations and other species [39–42], including essential nutrients such as pantothenate [11]. In experiments conducted on free parasites, where the NPP are not relevant, probenecid strongly inhibited folinic acid uptake with an IC50 of ca. 46 μM, and interestingly, furosemide was just as effective (IC50 of ca. 39 μM) (Fig. 7a and Table 1). When parasitised erythrocytes were used in the assay, furosemide blocked uptake effectively with an IC50 value of ca. 21 μM, similar to that for the free parasite, whereas probenecid was now less effective (IC50 of ca. 187 μM). The two drugs were assayed individually and also in combination under identical conditions, but no synergy was apparent between them (Fig. 7b). Indeed, the apparent absence even of an additive effect showed that when both drugs are present in equimolar amounts, the overall effect is equivalent to that seen with the drug of higher potency, i.e. furosemide. This suggests that the two drugs are likely to be acting on the same target. Given this and the above results with free parasites, where the target involved is shown by our prior experiments to behave as a specific folate transporter, it seems probable that the major effect of these two drugs on infected erythrocytes is also at folate transporters on the erythrocyte membrane, rather than the NPP, although we cannot exclude transport via the latter acting to supplement the endogenous capacity of the erythrocyte to import the higher levels of folate required by the parasite.
4
Discussion
Although P. falciparum possesses the biosynthetic machinery required to produce folate derivatives de novo, folate salvage is also an important aspect of the parasite's metabolism [7]. We present several lines of evidence to demonstrate that transport of exogenous folate into P. falciparum is primarily a carrier-mediated process dependent upon cellular energy and the trans-plasma membrane proton gradient. Uptake over the erythrocytic cycle is clearly regulated, with the highest capacity occurring at the late trophozoite stage. This coincides with a period of intense metabolic activity and the onset of multiple rounds of DNA replication, for which folate cofactors are essential, and resembles the pattern seen in the folate auxotroph Leishmania, a parasite that also imports folates maximally in the logarithmic growth phase [43–45]. The exogenous folate requirement of P. falciparum can be satisfied in vitro by folic acid, 5-formylTHF (folinic acid) and 5-MeTHF, similar to that of the folate-requiring bacterium Lactobacillus casei, but not L. salivarius, which cannot use the last of these [27]. The process also exhibits a marked temperature dependence, characteristic of a membrane carrier system, with significant uptake only occurring close to the parasite's optimum growth temperature of 37 °C. Once imported, net loss in the reverse direction appears to be a slow process, but over the short periods monitored here, did not involve a detectable level of polyglutamation that would trap the folate within the cell, although a significant portion of imported folate is ultimately modified in this way [2,23,20]. This may imply that the role of polyglutamation in this case may be more related to enzyme-substrate recognition, rather than intracellular trapping.
The Km values of ca. 2–3 μM that we measure for the saturable binding component for folinic acid of two strains of P. falciparum are higher than those of 0.4 and 42 nM, respectively, for folic acid uptake into Lactobacillus
[27], Xenopus oocytes [46], and of 250–700 nM for such uptake into Leishmania parasites, depending upon the species [47]. This may reflect the absolute dependence of these three organisms on folate salvage, whereas P. falciparum also possesses an active folate biosynthesis pathway. The values we observed are of the same order as the rate of uptake of the vitamin pantothenic acid into the parasite [11] but considerably slower than that of glucose [48,49,10], presumably reflecting the great difference between reusable cofactors required in limited quantities and the enormous consumption of the latter as the principal source of energy. The more rapid uptake of folate into the antifolate-resistant strain of K1 compared to the sensitive FCB strain, as indicated by their quite different values, could result from the selection of parasites better able to cope with drug-mediated inhibition of folate metabolism by increasing their salvage efficiency, in addition to the familiar mutation events in genes encoding folate pathway enzymes that reduce drug binding to the latter [50]. Moreover, if the mutant enzymes (DHFR-TS and PPPK-DHPS) from such parasites process their substrates in vivo with a lower efficiency than in the wild-type, compensation by an increased flux through the salvage pathway might be important for healthy growth, even in the absence of drug. These observations merit extension to other parasite strains with differing responses to antifolate drugs, to further test this hypothesis.
In principle, import of folates into the infected erythrocyte could occur via the new permeation pathway(s) (NPP) induced in the host cell membrane by the growing parasite [39,41,42]. However, mammalian cells also exhibit two types of endogenous folate transport system, the reduced folate carriers (RFC), which possess multiple transmembrane domains and have a greater affinity for the reduced folates than for folic acid, and the folate receptors (FR), which are GPI-anchored external folate binding proteins that mediate transport of both folic acid and 5-MeTHF with high specificity [51–53]. In mature erythrocytes, the FRs lose functionality with age [51] and the RFC is thought then to be the major transporter type [25,53]. Whatever the exact nature of the relevant transporters on the erythrocyte membrane, they appear not to represent a bottleneck for folate uptake into the parasite, as indicated by our rate experiments. Nor would the parasitophorous vacuolar membrane, which is highly porous and freely permeable to small molecules [54]. The rate-limiting step is therefore passage through the parasite plasma membrane, a process that is strongly inhibited in free parasites both by probenecid, a proven inhibitor of folate transport in mammalian cells [34] and furosemide. Given the similarity in these anion channel blockers (both are benzoic acid derivatives of comparable size with aminosulfonyl substituents on the benzene ring) and the evidence that they target the same sites on infected erythrocytes, we conclude that their major effect with respect to folate uptake at the erythrocyte membrane is likely to be at endogenous folate transporters rather than the NPP. Consistent with the view that the transporters involved in the parasite plasma membrane also have a high degree of specificity for folate, we note that the antifolate drugs PYR and SDX, which have some structural features in common with folate, do not compete with uptake of folinic acid even at very high concentrations. In contrast, the anticancer antifolate MTX, which is a much closer structural analogue to folic acid than either PYR or SDX, does inhibit uptake to a similar extent to the unlabelled folates themselves (μM range) in competition experiments. We note that the lack of inhibition of folate uptake by PYR that we report here rules out our earlier hypothesis [4] that such inhibition might contribute to the synergy observed between PYR and SDX.
We have demonstrated that folate uptake into P. falciparum is dependent upon the presence of d-glucose or other sugar that can be metabolised via glycolysis, a process that is essential to maintain the proton and electrochemical gradients across the parasite plasma membrane [21,8]. d-Fructose can be phosphorylated by hexokinase to fructose 6-phosphate, the substrate of the third glycolytic enzyme, phosphofructokinase-1. Interestingly, although fructose is only about half as efficient as glucose in maintaining overall growth of the parasite [30], it was as efficient as glucose in supporting folate import on a molar basis in our experiments. This may reflect the fact that only a small fraction of the glycolytic flux is likely to be necessary to support import of the relatively low amounts of exogenous folate required by the parasite. Of the other sugars we tested, which were all unable to support uptake, l-glucose is stereochemically excluded, while 6-deoxy-d-glucose and d-xylose can bind to hexokinase but cannot be phosphorylated. d-Galactose, derived principally from lactose hydrolysis, can be converted to glucose 6-phosphate in mammals via a four-step mechanism involving uridine diphosphate derivatives, but its inability to support folate import here would be consistent with an apparent lack of galactokinase and the other enzymes required for this conversion in the predicted proteome of P. falciparum.
The dramatic reduction of folate transport in the absence of a metabolisable sugar mirrors the rapid decline of intracellular ATP concentration and drop in the internal pH of the parasite seen in the same circumstances [21]. This suggested that inward movement of folates depends upon maintenance of a pH (interior alkaline) or electrochemical gradient (interior negative), or both, across the plasma membrane. Such gradients can be generated by membrane-bound ATPases, and a range of organic anions move into bacterial cells using proton symport. To differentiate between movement of folate via symport with protons or another cationic species, we examined the influence of the latter, but saw only minor changes in uptake rates in the presence or absence of Na+, K+, Mg2+ or EDTA. This differs from the folate transport system in Lactobacillus, which although similarly dependent upon glucose, also has a requirement for divalent cations [55,27].
The pH optimum that we observe for folate import (6.5–7.0) corresponds well with that (pH 6.9) measured just outside of the parasite in infected erythrocytes [32], 0.4 units lower than the pH of 7.3 in the parasite cytoplasm [21,32]. The primary mechanism for exporting protons in P. falciparum is thought to be via a V-type H+ ATPase on the plasma membrane [21,32], and consistent with this, we found that concanamycin A, a specific inhibitor of such ATPases [56], inhibited folate uptake in the low nM range. Genes encoding the A and B subunits of a V-ATPase have been identified in P. falciparum
[57,58] and immunological studies indicate that this enzyme is expressed throughout the erythrocytic cycle [32]. Interestingly though, a significant acidification of the extracellular space just outside of the parasitophorous vacuole could only be measured in trophozoites [32], the stage of the life-cycle at which we observe the highest levels of exogenous folate uptake. The substantial proton gradient that is generated via the V-ATPase is the major contributor to an inwardly negative membrane potential of ca. −95 mV [8], but protons are also extruded in a symport system together with the large amounts of lactate generated by parasite glycolysis [12]. This may explain why we did not observe 100% inhibition of uptake with concanamycin, even at μM concentrations, as it blocks only the major portion of the total proton export.
The importance of the proton gradient across the plasma membrane was emphasised by the sensitivity of folate uptake to the ionophores DNP and nigericin, both of which strongly perturb this gradient. In particular, we find that folate uptake is exquisitely sensitive to nigericin, with substantial inhibition at sub-nM levels. Whereas DNP collapses the pH gradient completely by enabling free passage of protons in both directions, the K+/H+ antiporter mechanism of nigericin is presumably limited by the degree to which the K+ gradient can be perturbed, which may explain why we also did not observe 100% inhibition of uptake with this compound. Although the profiles observed for nigericin and concanamycin A can be explained in principle by assuming that neither abolishes the pH gradient completely, we cannot exclude the possibility that there is also a component of folate transport independent of this gradient. This scenario is suggested by the fact that a low level of transport persists even when the external pH is set at the same, or a higher, level than the internal pH of parasites deprived of glucose, and would also be consistent with the two-phase nature of the concentration dependence of uptake. At first sight, this picture is difficult to reconcile with the data for DNP, which by insertion into the membrane, permits rapid equilibration of proton concentrations on each side, but we note that complete abolition of uptake only occurs at ca. 1 mM, a concentration at which quite profound changes to the membrane might be expected, possibly impacting on any proton gradient-independent component.
A plausible model encompassing our observations would be that folate uptake at the parasite plasma membrane is mediated by a transport complex or multi-binding site protein possessing a binding domain specific for folate recognition and a probenecid/furosemide-sensitive proton symport channel through which the folate is internalised. To date, no folate transporters have been functionally identified in P. falciparum, but bioinformatics studies [59] predict the existence of 3 candidate proteins (gene loci MAL8P1.13, PF11_0172 and PF10_0215 in PlasmoDB; http://www.plasmodb.org) that are related to biochemically verified molecules in Leishmania
[47,45], Synechocystis and Arabidopsis
[60]. Related putative candidates have also been identified in Toxoplasma gondii
[6]. This sub-family of folate-biopterin transporters (FBT) are thought on bioinformatic grounds to function as proton symporters [59], and on this basis, the data we present here would be consistent with one or more of these proteins providing the machinery for folate uptake in P. falciparum, which might ultimately present a novel target type for parasite inhibition. We are currently investigating the genes encoding these candidate molecules in knockout studies to better understand the relationship between the phenomena reported here and the parasite proteins involved.
|
[
"transporters",
"folate salvage",
"proton symport",
"malaria parasites",
"folate metabolism",
"bcecf-am, 2′,7′-bis-(2-carboxyethyl)-5,6-carboxylfluorescein acetoxymethyl ester",
"dhfr, dihydrofolate reductase",
"dhps, dihydropteroate synthase",
"dnp, 2,4-dinitrophenol",
"paba, p-aminobenzoic acid",
"mtx, methotrexate",
"npp, new permeability pathway(s)",
"pbs, phosphate-buffered saline",
"pppk, 6-hydroxymethylpterin pyrophosphokinase",
"pyr, pyrimethamine",
"sdx, sulfadoxine",
"thf, tetrahydrofolate",
"ts, thymidylate synthase",
"metabolic inhibitors"
] |
[
"P",
"P",
"P",
"P",
"P",
"M",
"M",
"M",
"R",
"R",
"R",
"R",
"M",
"M",
"R",
"R",
"M",
"M",
"R"
] |
DNA_Repair_(Amst)-1-5-2077361
|
Mrc1 protects uncapped budding yeast telomeres from exonuclease EXO1
|
Mrc1 (Mediator of Replication Checkpoint 1) is a component of the DNA replication fork machinery and is necessary for checkpoint activation after replication stress. In this study, we addressed the role of Mrc1 at uncapped telomeres. Our experiments show that Mrc1 contributes to the vitality of both cdc13-1 and yku70Δ telomere capping mutants. Cells with telomere capping defects containing MRC1 or mrc1AQ, a checkpoint defective allele, exhibit similar growth, suggesting growth defects of cdc13-1 mrc1Δ are not due to checkpoint defects. This is in accordance with Mrc1-independent Rad53 activation after telomere uncapping. Poor growth of cdc13-1 mutants in the absence of Mrc1 is a result of enhanced single stranded DNA accumulation at uncapped telomeres. Consistent with this, deletion of EXO1, encoding a nuclease that contributes to single stranded DNA accumulation after telomere uncapping, improves growth of cdc13-1 mrc1Δ strains and decreases ssDNA production. Our observations show that Mrc1, a core component of the replication fork, plays an important role in telomere capping, protecting from nucleases and checkpoint pathways.
1
Introduction
Telomeres are specialized DNA–protein complexes at the end of eukaryotic chromosomes. Proper telomere structure is essential for chromosome integrity and genome stability because telomeres protect natural chromosome ends from degradation and end-to-end fusion and because they ensure complete genome replication. Telomeres differ from Double Strand Breaks (DSBs) in that they normally fail to activate DNA repair or DNA damage checkpoint pathways. If that was the case, then they would undergo recombination and chromosomal fusions and dicentric chromosomes would be generated [1–4].
Many proteins associate with telomeric DNA. These proteins ensure that telomeres behave differently from DSB ends and help maintain chromosomal stability. Some telomeric proteins bind specifically to dsDNA and others show higher affinity to ssDNA. In budding yeast, there are numerous proteins with affinity for telomeric dsDNA, such as Rap1, Sir2, Sir3, Sir4, Rif1 and Yku70/Yku80 [5]. The budding yeast telomeric ssDNA ends are thought to be protected by three essential proteins, Cdc13, Stn1 and Ten1 [6–11].
If telomeres become uncapped, they activate a DNA damage response pathway leading to cell cycle arrest [12–14]. Moreover, recently it has been suggested that telomeres trigger a transient DNA damage response in each S phase in order to complete DNA replication and cap chromosome ends [15]. DNA damage response pathways are complex networks which include – among other components – mediators. Mediators facilitate the transmission of the DNA damage signal from sensors to downstream effectors; activation of the latter affecting cell cycle progression [16]. Mrc1 (Mediator of Replication Checkpoint 1) appears to take the mediator role in Rad53 activation under replication stress [17]. However, parallel pathways exist because in mrc1Δ mutants Rad53 activation occurs through Rad9 (another mediator protein), presumably because the accumulation of “DNA damage” rather than “replication defects” in mrc1Δ mutants leads to activation of Rad9 and thereby activation of Rad53 [17].
Mrc1 also appears to be directly involved in DNA replication and, because of this, mrc1Δ cells display a slow S phase [18]. Mrc1 is an active component of the replication machinery, loaded onto DNA shortly after replication initiates, and moving with other components of replication forks [19–21]. In the presence of hydroxyurea, a type of replication stress, Mrc1 appears to form a stable replication-pausing complex preventing the uncoupling of the replication machinery from DNA synthesis [20–22]. According to this model, Mrc1 mediates activation of Rad53 under conditions of replication stress so that subsequent DNA repair events occur and cell replication resumes normal function [22]. However, recent experiments suggest that the role of Mrc1 at stalled replication forks is more than activating Rad53, since mrc1AQ cells, defective in Rad53 activation, are not defective in replication fork initiation or progression [18,20,21]. mrc1AQ is a mutant allele in which SQ/TQ residues have been substituted with AQ, resulting in its inability to mediate phosphorylation and activation of Rad53 [18].
Although Mrc1 is involved in the DNA replication checkpoint, it has been shown that it is not required for the DNA damage checkpoint, since cdc13-1 mrc1Δ double mutants arrest in G2 at non-permissive temperatures [17]. It has been reported that activation of Rad53 in response to telomere shortening still occurs in the absence of Rad9 and that Mrc1 is responsible for this activation in telomerase-deficient cells, in which telomeres continually shorten until they activate a checkpoint [23]. Surprisingly, though, tlc1Δ mrc1Δ double mutants arrest cell division, suggesting that Mrc1 is not required for cell cycle arrest in telomerase negative cells. In contrast, after cdc13-1 induced telomere uncapping, Rad53 activation is entirely Rad9-dependent and Mrc1-independent [23].
Here we investigated the role of Mrc1 at uncapped telomeres, using the temperature sensitive cdc13-1 and yku70Δ mutations to uncap telomeres. Our experiments indicate that Mrc1 protects telomeres from the DNA damage response and that the role of Mrc1 in DNA replication forks, rather than in checkpoint activation, is important for protection of telomeres.
2
Materials and methods
2.1
Yeast strains and plasmids
All strains in the W303 background are RAD5 and they contain an ade2-1 mutation (Supplemental Table 1); therefore yeast extract/peptone/dextrose (YEPD) was supplemented with adenine at 50 mg/l. Strains 3393–3402 are in the S288C background (Supplemental Table 1) and they were generated by mating a single gene deletion mutant array [24] with a cdc13-1 query strain [25]. To construct strains, standard genetic procedures of transformation and tetrad analysis were used [26]. pMRC1 and pmrc1AQ, also carrying the URA3 gene were a gift from Steven Elledge [17,18].
MRC1 was disrupted in two different ways. Firstly, the MRC1 ORF was substituted with KanMX6, with a PCR based method [27]. Primers 5′-tcgttattcgcttttgaacttatcaccaaatattttagtgCGGATCCCCGGGTTAATTAA-3′ (#878) and 5′-ctggagttcaatcaacttcttcggaaaagataaaaaaccaGAATTCGAGCTCGTTTAAAC-3′ (#881), which contain homology to upstream and downstream sequences of MRC1 (bases in lowercase), were used to amplify a 1559 bp KanMX6 sequence (pFA6a-kanMX6; [27]). The PCR product was transformed into yeast and candidate colonies were selected for G418 resistance. Integration of the KanMX6 marker into the MRC1 locus was confirmed by PCR, using two sets of primers: (i) forward 5′-CCAAGAACAGACAAACAACTAAGGA-3′ (#876) with reverse primer 5′-TCAGCATCCATGTTGGAATT-3′ (#81) and (ii) forward 5′-CCATCCTATGGAACTGCCTC-3′(#82) with reverse 5′-CCTAGACTCGGGTGCCATCT-3′ (#880). Disruption of MRC1 was also confirmed by Southern blot (data not shown). Alternatively, MRC1 was substituted with URA3 using a restriction enzyme digest approach. First, pMRC1 was digested with XhoI and a 5008 bp fragment containing the full MRC1 gene was cloned into XhoI digested pIC19H vector (2.7 kb) to create pAT1065. A correct clone was identified by restriction digests. pAT1065 was digested with SpeI to remove a 2.31 kb DNA fragment containing the bulk of MRC1, which was replaced with a 1.3 kb BamHI URA3 gene fragment from pDL349 (pBSB + KS containing a BamHI fragment containing the URA3 gene) by blunt cloning following treatment with DNA polymerase I Large (Klenow) fragment (New England Biolabs). Positive clones were selected by restriction enzyme digests to identify the disruption of the bulk of MRC1 with URA3 (pAT1066). Disruption of MRC1 was also confirmed by Southern blot (data not shown). To disrupt MRC1, pAT1066 was digested with XhoI prior to transformation of yeast.
2.2
Spot tests
Single colonies were inoculated into 2 ml YEPDextrose (YEPD) and incubated overnight, with aeration, at 23 °C. The following day, 200 μl of each culture was inoculated into 2 ml of fresh YEPD and returned to 23 °C. Cells were grown for three more hours, and cell numbers were determined in a haemocytometer. The cells were then centrifuged (13,000 rpm for 10 s in a microcentrifuge), washed twice with sterile water and resuspended in water to a final concentration of 1.5 × 107 cells/ml. A five-fold dilution series of each of the cultures was prepared using sterile water in a 96 well plate and 3–5 μl spotted onto plates using a 48-prong replica plating device. Plates were incubated at various temperatures for 2–3 days before being photographed. For spot tests with strains containing pMRC1, pmrc1AQ or pRS416 the steps were as described above, but strains were grown on selective medium (-URA). All strains shown as if on a single agar plate were grown on the same plate, although in some cases their positions were moved using Adobe Photoshop and Adobe Illustrator CS. Unless otherwise stated, at least two different strains of the same genotype were spot tested and representative examples are shown.
2.3
Yeast transformation
High efficiency transformations needed for gene disruptions were performed according to Gietz et al. [28]. For plasmid transformations a more rapid method was used [29].
2.4
Western blots
Protein extracts were prepared by glass bead breakage in TCA, essentially as previously described [30,31]. Bio-Rad gels (7.5% Tris–HCl), Schleicher and Schuell Protan Nitrocellulose membranes and the Pierce Supersignal West Pico Chemiluminescent Substrate detection kit were used in a standard Western blot procedure. Rabbit anti-Rad53 polyclonal antibody (AbDL50, 1:1000 dilution, a gift from Dan Durocher [32]) was a primary antibody used with an anti-rabbit-HRP (AbDL7, 1:10,000 dilution, Dako P0448) as a secondary antibody used. Mouse anti-tubulin (TAT-1, AbDL42, 1:2000 dilution, a gift from Keith Gull [33] and anti-mouse-HRP (AbDL6, 1:10,000 dilution, Dako P0447) were used for tubulin loading controls.
2.5
Telomere length measurement by Southern blot
Strains were grown to saturation in liquid YEPD at 23 °C. DNA from each strain was subjected to XhoI cut. The digested DNA was loaded on a 0.8% agarose gel, run at low voltage overnight, transferred to a Magna nylon membrane and UV cross-linked. The membrane was then hybridised with a Y′-TG probe [34]. A non-radioactive detection kit was used for the detection of the hybridisation (Amersham, Arlington Heights, IL).
2.6
Synchronous cultures
cdc13-1 cdc15-2 bar1Δ strains with additional mutations (see Supplemental Table 1) were grown in YEPD at 23 °C overnight. In the morning, cells were adjusted to a concentration of 8 × 106 buds/ml in 250 ml. Cultures were grown for three more hours, then arrested with 20 nM α-factor for a further 2.5 h. The cultures were then released from G1 arrest by centrifugation and washed twice in YEPD and placed at 36 °C, 40 min after the culture was first centrifuged. Cell cycle position was monitored as previously described [13]. DNA was prepared and the fraction of single stranded DNA (ssDNA) was measured as previously described [35].
2.7
Asynchronous cultures
cdc13-1 strains with additional mutations indicated were grown in YEPD at 23 °C overnight. In the morning, cells were adjusted to a concentration of 1 × 107 cells/ml and temperature was raised to 27.3 °C. Every 90 min samples were taken for cell cycle position and cell density was re-adjusted to 1 × 107 cells/ml. Cell numbers were determined with a haemocytometer. Samples for Western blots were collected from exponentially growing cultures 2 h after the temperature was raised from 23 °C to 36 °C.
2.8
Cell cycle position determination
A 1 ml sample of culture was centrifuged for 8–10 s at high speed, the supernatant was aspirated, and cells were fixed at 70% ethanol overnight. The fixed cells were washed twice in water before being resuspended in 4,6-diamidino-2-phenylindole (DAPI, 0.2 μg/ml). Cell cycle distribution was monitored by DAPI staining of nuclei and fluorescence microscopy. For DAPI staining, 100 cells for each sample were counted and classified as: (1) unbudded, single DAPI-stained body; (2) small budded, single DAPI-stained body, with the bud <50% of the diameter of the mother cell; (3) medial nuclear division, single DAPI-stained body, with bud >50% diameter of mother cell, the cdc13-1 arrest point; (4) late nuclear division, two buds, and two DAPI-stained bodies, the cdc15-2 arrest point [36].
3
Results
3.1
Mrc1 contributes to the vitality of cdc13-1 and yku70Δ mutants
Since Mrc1 plays a role in the checkpoint response to stalled replication, we wondered if it also plays a role at uncapped telomeres. The temperature sensitive cdc13-1 mutation causes a defect in Cdc13, a telomere binding protein, and cells containing this mutation accumulate large amounts of ssDNA at telomeres at non-permissive temperatures [12,37,38]. Interestingly, deletion of checkpoint proteins, like Chk1, Mec1, Mec3, Rad9, Rad17, Rad24 and Rad53 improves growth of cdc13-1 strains at semi-permissive temperatures [39–41]. This is presumably because checkpoint pathways inhibit cell division by responding to low levels of ssDNA that accumulates at telomeres at semi-permissive temperatures. Deletion of other checkpoint proteins, like the MRX complex, which appears to play a role in telomere capping, worsens the growth of cdc13-1 strains [42]. Therefore, we wanted to investigate whether Mrc1 plays a role at uncapped telomeres and, if so, whether it behaved like Rad9 or MRX. Fig. 1A shows that deletion of MRC1 dramatically reduces the growth of cdc13-1 mutant strains at 25 °C. The effect of Mrc1 is not as profound as that of the MRX complex, as cdc13-1 mre11Δ and cdc13-1 rad50Δ display more severe growth defects than cdc13-1 mrc1Δ even at 23 °C (Supplementary Fig. 1). Thus, Mrc1, like MRX, but unlike the majority of checkpoint proteins, contributes to the vitality of cdc13-1 strains.
We next wanted to investigate whether analogous growth defects of cdc13-1 mrc1Δ strains occur in yku70Δ strains. Yku70 is a telomere capping protein which is also involved in dsDNA damage repair and in Non-Homologous End Joining (reviewed in Ref. [43]). Deletion of YKU70 results in a temperature sensitive phenotype at 37 °C, due to telomere uncapping, which activates a Chk1-dependent cell cycle arrest [36]. Fig. 1B demonstrates that deletion of MRC1 results in a severe growth defect of yku70Δ mutant strains at 35 °C and 36 °C. Thus, Mrc1 contributes to the vitality of yku70Δ strains. We conclude that Mrc1 contributes to the vitality of two cell types defective in telomere capping.
Mrc1, Tof1 and Csm3 are three proteins that play similar, although distinct roles in DNA replication [19–21,44]. Therefore, we wished to address whether Tof1 and Csm3, like Mrc1, contributed to the vitality of cdc13-1 mutants. Spot test analysis showed that although deletion of TOF1 or CSM3 also confers some growth defects on cdc13-1 mutants, deletion of MRC1 has a stronger phenotype (Fig. 1C). Therefore, we decided to focus on understanding the role of MRC1 at uncapped telomeres.
3.2
Growth defects of cdc13-1 mrc1Δ cells are not due to checkpoint defects
In budding yeast, two independent roles have been previously reported for Mrc1. One role implicates Mrc1 as a mediator of checkpoint activation under replication stress and the other role is as part of the replication machinery [17–21]. Therefore, we investigated whether the heightened temperature sensitivity phenotype of cdc13-1 mrc1Δ mutant strains is a result of a replication defect, a checkpoint defect or both.
cdc13-1 mrc1Δ mutants were complemented with either wild type pMRC1, pmrc1AQ or an empty vector (pRS416) and strains were grown at various temperatures. Fig. 2 shows that at 26.2 °C, complementation of mrc1Δ cdc13-1 mutant strains, with either pMRC1 or pmrc1AQ allele improves growth compared to the empty vector control (compare rows 1–3). Thus, we conclude that the checkpoint role of Mrc1 is not important for the vitality of cdc13-1 strains.
At higher temperature we noticed an increased growth of cdc13-1 mrc1Δ cells carrying the mrc1AQ allele, compared to pMRC1. However, this phenotype was observed even in the presence of the wild type MRC1 allele, suggesting that this effect is due to some type of dominant effect of mrc1AQ (Fig. 2, compare rows 2 and 11).
3.3
Exo1 inhibits growth of cdc13-1 mrc1Δ and yku70Δmrc1Δ mutants
EXO1 encodes a nuclease known to contribute to ssDNA production at uncapped telomeres of cdc13-1 and yku70Δ strains [36,39]. If Exo1-dependent ssDNA production at uncapped telomeres is responsible for the poor growth of cdc13-1 mrc1Δ and yku70Δ mrc1Δ mutants, then removing Exo1 should suppress their poor growth. Fig. 3A demonstrates that at 25 °C, cdc13-1 mrc1Δ exo1Δ triple mutants exhibit better growth than cdc13-1 mrc1Δ strains, showing that Exo1 contributes to the growth defects observed in cdc13-1 mrc1Δ strains. Importantly, deleting EXO1 also reverses the growth defect of mrc1Δ yku70Δ mutants (Fig. 3B). These data suggest that Mrc1 protects uncapped telomeres from Exo1.
3.4
Effects of checkpoint mutations on cdc13-1 mrc1Δ and yku70Δ mrc1Δ growth
To understand if checkpoint pathways are activated in mrc1Δ strains after telomere uncapping, we wanted to combine cdc13-1 mrc1Δ and yku70Δ mrc1Δ strains with checkpoint mutations. A genetic screen has revealed that mrc1Δ is synthetically lethal with rad9Δ, rad17Δ or rad24Δ checkpoint mutations [24]. Consistent with these results we were unable to recover viable offspring carrying mrc1Δ in combinations of any of these checkpoint genes (data not shown). However, we were able to combine mrc1Δ with rad53Δ and chk1Δ, encoding two downstream checkpoint kinases (analogues of human Chk2 and Chk1, respectively).
When cdc13-1 mutants grow at non-permissive temperatures Rad53 and Chk1 dependent parallel pathways are activated [36,45,46]. We deleted RAD53 or CHK1 and examined the effects in cdc13-1 mrc1Δ mutants. Deletion of RAD53 requires simultaneous deletion of SML1 to obtain viable spores. Sml1 is a small protein that inhibits the activity of ribonucleotide reductase (RNR) which catalyzes the rate-limiting step of de novo dNTP synthesis [47]. Normally Sml1 is degraded in a Rad53-dependent manner during S phase [48]. We found that removal of RAD53 and SML1 improved the growth of cdc13-1 mrc1Δ strains (Fig. 4A). However, deletion of SML1 alone (cdc13-1 mrc1Δ sml1Δ) also rescued the growth defects associated with MRC1 deletion (Fig. 4A). Thus, we were unable to observe any strong role for Rad53, in maintaining vitality of cdc13-1 mrc1Δ mutants, other than in degrading Sml1. We suggest that the reason that deleting Sml1 improves the growth of cdc13-1 mrc1Δ strains is that increased ribonucleotide reductase activity may stabilise the replication forks.
Removal of Chk1, like removal of Rad53, does not rescue growth of cdc13-1 mrc1Δ strains, indicating that Chk1 does not inhibit growth of these mutants (Fig. 4B). Therefore, we find no evidence that inactivating DNA damage checkpoint pathways improves growth of cdc13-1 mrc1Δ mutants.
A CHK1 deletion strongly rescues growth of yku70Δ mrc1Δ mutants at restrictive temperatures (Fig. 4C) similarly to its effect in yku70Δ mutants [36]. Thus, mrc1Δ yku70Δ uncapped telomeres qualitatively behave like yku70Δ uncapped mutants. The effects of chk1Δ in yku70Δ mrc1Δ and cdc13-1 mrc1Δ strains are consistent with earlier findings showing that the Chk1-dependent pathway is more important in yku70Δ mutants [36,45].
3.5
Mrc1 is not required for the cell cycle arrest after cdc13-1 uncapping
To directly test whether Mrc1 plays a checkpoint role in cdc13-1 strains, cdc13-1 strains with additional mutations were first grown at the semi-permissive temperature 27.3 °C. Strains defective in telomere capping (cdc13-1) arrest at medial nuclear division before entry to anaphase after 3 h at 27.3 °C (Fig. 5A). As expected, when the checkpoint is compromised in cdc13-1 chk1Δ, cdc13-1 rad9Δ or cdc13-1 rad53Δ cells, no accumulation at medial nuclear division is observed over 9 h (Fig. 5A). In contrast, cdc13-1 and cdc13-1 mrc1Δ strains rapidly accumulate at medial nuclear division and within 3 h more than 90% of cells are arrested at this point. Consistent with these conclusions cell numbers stopped increasing in cdc13-1 and cdc13-1 mrc1Δ strains, but continued to increase in the other strains (Fig. 5B). We conclude that Mrc1 is not required for the checkpoint response to cdc13-1 dependent telomere uncapping.
Rad53 and Chk1 are components of parallel checkpoint pathways that respond to cdc13-1 induced telomere uncapping [45,46]. It appears that the Rad53 pathway is more important for arrest of cdc13-1 mrc1Δ mutants because 80% of cdc13-1 mrc1Δ chk1Δ cells have arrested at medial nuclear division by 7.5 h, whereas there is no arrest of cdc13-1 mrc1Δ rad53Δ cells.
To test the role of Mrc1 in checkpoint control in a single cell cycle we combined mrc1Δ with cdc13-1 cdc15-2 and bar1 mutations. Over many years we and others have used these mutations to determine the effects of checkpoint proteins in responding to telomere uncapping [13,39,40,45]. Bar1 encodes a protease that degrades the mating pheromone α-factor. Cells bearing the bar1 mutation can efficiently arrest in G1 phase of the cell cycle with low levels of α-factor. Cdc15 is necessary for mitotic exit. At 36 °C, cdc13-1 cdc15-2 bar1 control strains, released from alpha factor arrest, accumulate at medial nuclear division due to cdc13-1-dependent telomere uncapping. However, if cells have escaped the G2/M checkpoint, like cdc13-1 rad9Δ cdc15-2 bar1 strains, they arrest at late nuclear division due to cdc15-2 and they are unable to proceed to the next cycle. The cdc15-2 dependent cell cycle arrest helps in two ways, it ensures that DNA damage checkpoint defects are easily quantified because cells with checkpoint defects accumulate at a later stage of the cell cycle and that DNA damage caused by cdc13-1 is not amplified during new rounds of DNA replication.
cdc13-1 cdc15-2 bar1 strains with additional mutations were arrested with α-factor, then released from G1 and transferred to a non-permissive temperature to induce telomere uncapping and the cell cycle position was monitored. Fig. 5C shows that in contrast to cdc13-1 rad9Δ strains cdc13-1 and cdc13-1 mrc1Δ strains arrest at medial nuclear division with similar kinetics at 36 °C, supporting the idea that Mrc1 does not play a role in checkpoint activation after cdc13-1 telomere uncapping. Interestingly, cdc13-1 mrc1Δ exo1Δ remain arrested at medial nuclear division (Fig. 5C and D) which contrasts to the behaviour of cdc13-1 exo1Δ that begin to escape arrest during the 4-h period in analogous experiments [39]. This difference most likely reflects the fact that cdc13-1 mrc1Δ exo1Δ mutants have more severe telomere capping defect than cdc13-1 exo1Δ mutants and therefore activate a stronger checkpoint signal.
In response to both replication stress and DNA damage, activation of the checkpoint machinery induces phosphorylation and activation of Rad53 kinase. Therefore, we addressed the role of Mrc1 in Rad53 phosphorylation after telomere uncapping. cdc13-1 strains were exposed to 36 °C for 2 h and Rad53 phosphorylation was measured by Western blot. Rad53 phosphorylation is observed in cdc13-1 mrc1Δ strains but not in cdc13-1 rad9Δ mutants (Fig. 5F), confirming a previous study [23]. We conclude that activation of Rad53 after cdc13-1 dependent uncapping at non-permissive temperatures is Rad9-dependent but Mrc1-independent.
3.6
Mrc1 contributes to telomere length regulation
If Mrc1 plays a protective role at telomeres, this predicts that strains lacking Mrc1 may have short telomeres. Fig. 6A shows that absence of Mrc1 results in shorter telomeres, compared to the wild type. However, the telomere length defects of mrc1Δ mutants are not as severe as in rad50Δ or yku70Δ mutants, and this may help explain why cdc13-1 mrc1Δ cells grow better than cdc13-1 rad50Δ cells (Supplementary Fig. 1). Our experiment is consistent with replication proteins having an important role in telomere length regulation [49–52].
3.7
Mrc1 protects telomeres from extended ssDNA accumulation
All our data suggest an important role of Mrc1 in telomere capping but no role in cell cycle arrest. To directly test the role of Mrc1 in telomere capping, we measured ssDNA accumulation on the 3′ TG strand, at PDA1, a single copy locus approximately 30 kb away from the right end of chromosome V in cdc13-1 strains (Fig. 6B). cdc13-1 strains were synchronised, as in Fig. 5C and D in order to follow the effects of Mrc1 on ssDNA accumulation at non-permissive temperatures within a single cell cycle. We find that cdc13-1 mrc1Δ mutants, like cdc13-1 rad9Δ strains, accumulate more 3′ TG ssDNA at PDA1, 30 kb from the uncapped telomere, compared to cdc13-1 strains (Fig. 6C) [39]. This shows that Mrc1, like Rad9, protects cdc13-1 mutants from ssDNA production. Consistent with our conclusion, increased ssDNA accumulation, closer to the telomere, in telomere repeats, was recently reported in cdc13-1 mrc1Δ and yku70Δ mrc1Δ mutants using both in gel and dot blot analyses [53]. Importantly, ssDNA production is reduced in cdc13-1 mrc1Δ exo1Δ strains in comparison to cdc13-1 mrc1Δ strains showing that Mrc1 protects uncapped telomeres from Exo1-dependent nuclease action. This ssDNA data is consistent with our finding that Exo1 contributes to the poor growth of cdc13-1 mrc1Δ and yku70Δ mrc1Δ mutants (Fig. 3). Taken together, we conclude that Mrc1 inhibits accumulation of Exo1-dependent ssDNA accumulation after telomere uncapping and, by this criterion, contributes to telomere capping.
4
Discussion
Our experiments demonstrate that Mrc1 contributes to the vitality of budding yeast cells with uncapped telomeres. Therefore, Mrc1 behaves differently from many other known checkpoint proteins such as Chk1, Mec1, Rad9, Rad17, Rad24 or Rad53, deletion of which improves the growth of cdc13-1 mutants at semi-permissive temperatures [39,40]. The effect of Mrc1 is more similar to that of the MRX complex, another checkpoint complex with roles in telomere capping [36,42,54,55]. It seems that it is the role of Mrc1 at the replication fork that contributes to the vitality of telomere capping mutants, rather than its role in checkpoint activation. Our findings are in accordance with recent work that demonstrated a protective role of Mrc1 in cells with cdc13-1 or yku70Δ uncapped telomeres or in telomerase deficient cells [53]. Additionally, our work demonstrates that the growth defects of cdc13-1 mrc1Δ and yku70Δ mrc1Δ mutants and enhanced ssDNA levels of cdc13-1 mrc1Δ strains are suppressed when the nuclease encoded by EXO1 is deleted. Therefore, Mrc1 protects uncapped telomeres from Exo1.
Mrc1 is recruited to the replication machinery as DNA replication initiates and is required for normal rates of replication fork progression [17,18,22]. The biochemical role of Mrc1 in replication fork progression is unclear which makes it difficult to know its precise role in telomere capping. Mrc1 is also part of a replication-pausing complex formed when replication is arrested by the S phase poison hydroxyurea (HU), and required for replication fork restart after HU. However, this restart role for Mrc1 is not universal, since Mrc1 plays no role in replication restart after cells are treated with the alkylating agent MMS [22].
It is interesting that there is evidence from budding yeast, fission yeast and human cells that telomeric sequences contain DNA regions that slow or stall replication forks [15,56,57]. From this it seems plausible that telomeric DNA may be more dependent on proteins like Mrc1, which contribute to fork stability and restart, than other chromosomal regions. That is because the replication fork struggles to reach the end of the chromosome in mrc1Δ mutants where a telomere capping defect is observed.
Numerous studies on budding yeast mutants with DNA replication defects have demonstrated interactions between DNA replication and telomere structure. For example both cdc17/pol1 and cdc44/rfc1 (large subunit of replication factor C) mutants affect telomere length [49]. Here we show that budding yeast mrc1Δ mutants have short telomeres. In S. cerevisiae, cdc17/pol1 mutants, encoding temperature sensitive DNA polymerase α, exhibit very long telomeres, high levels of telomeric ssDNA and elevated recombination at telomeres [50,51]. Interestingly, the B subunit of DNA polymerase α physically interacts with Stn1, which in turn interacts with Cdc13 [10,52]. This shows there is a very direct interaction between budding yeast telomere capping proteins and the replication fork machinery, and suggests that telomere capping is intimately linked with DNA replication. In this regard it is, perhaps, relevant that the 5′–3′ exonuclease Exo1 is involved in generating single stranded DNA at uncapped telomeres [36,39] and in processing stalled replication forks [58] and highlights the similarities between uncapped telomeres and stalled replication forks.
cdc13-1 cells maintain a functional telomere cap (low levels of telomeric ssDNA), when released from G1 arrest into the S phase poison hydroxyurea. HU stalls replication forks and stops late origins of replication from firing. However, if the same cdc13-1 cells are permitted to complete DNA replication, by removing the S phase poison HU, telomere uncapping occurs and high levels of ssDNA are observed [59]. Therefore, Cdc13-dependent telomere capping may depend on a coordinated interaction between the chromosome end, Cdc13/Stn1/Ten1 and the DNA replication fork. Further studies examining the interactions between telomeric DNA, the telomere cap and the replication fork will be necessary to better understand these interactions.
|
[
"mrc1",
"telomere",
"exo1",
"cdc13",
"yku70",
"ssdna"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Pediatr_Radiol-4-1-2292494
|
Gadolinium and nephrogenic systemic fibrosis: time to tighten practice
|
Nephrogenic systemic fibrosis (NSF) is a relatively new entity, first described in 1997. Few cases have been reported, but the disease has high morbidity and mortality. To date it has been seen exclusively in patients with renal dysfunction. There is an emerging link with intravenous injection of gadolinium contrast agents, which has been suggested as a main triggering factor, with a lag time of days to weeks. Risk factors include the severity of renal impairment, major surgery, vascular events and other proinflammatory conditions. There is no reason to believe that children have an altered risk compared to the adult population. It is important that the paediatric radiologist acknowledges emerging information on NSF but at the same time considers the risk:benefit ratio prior to embarking on alternative investigations, as children with chronic kidney disease require high-quality diagnostic imaging.
Introduction
Nephrogenic systemic fibrosis (NSF; formerly named nephrogenic fibrosing dermopathy) is a systemic disease that was first described in 1997 and recognized as a separate clinical entity in 2000 [1–4]. Until now, all confirmed cases of NSF have been reported in patients with chronic kidney disease (CKD) with estimated glomerular filtration rates (eGFR) <60 ml/min per 1.73 m2 or end-stage renal failure requiring renal replacement therapy. NSF affects both sexes with a reported age range of 8–87 years and has no race predilection [5, 6]. Relatively few paediatric cases have been reported, which may reflect the lower incidence of CKD in children.
Patients present with swelling, induration and tightening of the skin, often with a peau d’orange appearance or textured plaques, papules or nodules, usually limited to the extremities (from ankles to mid-thighs and from wrists to mid-upper arms). The skin lesions sometimes involve the trunk, but spare the neck and face. The condition develops over days to weeks and may ultimately result in severe contractures and disability. Patients may experience burning, itching or sharp pains in involved areas. Deep pain has also been described in the hips and ribs. Radiography may reveal calcifications of the soft tissue [1, 5–9]. Systemic involvement may be fatal with fibrosis of the lungs, skeletal muscles, pleura, pericardium, myocardium, kidneys, muscle, bone, testes and dura [2–4, 10]. Clinical differential diagnoses include scleroderma, scleromyxoedema, systemic sclerosis/morphoea, eosinophilic fasciitis and eosinophilia-myalgia syndrome, and the condition may mimic cellulitis, panniculitis or drug reactions [6–8, 11, 12].
The severity of CKD appears to be the main determinant of risk. Also, several contributing risk factors have been suggested: hypercoagulability and thrombotic events; cardiomyopathy; hepatic disease; idiopathic pulmonary fibrosis; systemic lupus erythematosus; brain tumours; recent surgical procedures, especially those with a major vascular component; and other proinflammatory conditions, including liver transplantation [5, 6, 8, 13, 14].
A specific aetiology is still unknown. The only common factor in all patients is the underlying kidney disease, with most patients dialysis-dependent. However, NSF has also been reported in patients who have never been dialysed [1, 5, 6, 8]. The physiopathological mechanism for fibrosis is not yet known. Histopathologically, biopsies from the skin lesions show a proliferation of fibrocytes (which may extend into subcutaneous tissue), together with thick collagen bundles and surrounding clefts with a variable amount of mucin and elastic fibres. Fibrocytes have a specific immunophenotype and express CD34, CD11b, CD45, HLA-DR, CD71, CD80, CD86, which implies that they are recruited from the circulation and do not arise as a result of proliferation in situ. Inflammatory cells are not usually evident, although small clusters of perivascular mononuclear cells may be seen [11, 15, 16].
Therapeutic options in NSF are limited. Improvements have been reported with extracorporeal photopheresis, plasmapheresis, ultraviolet phototherapy, intravenous immunoglobulin and renal transplantation [17–21].
Link with gadolinium-based contrast agents
The main concern for the radiologist is a growing body of evidence that clinical use of gadolinium-based contrast agents may trigger the disease. A number of patients with NSF seem not to have had any previous gadolinium exposure, but these are still under investigation. Grobner [22] was the first to report this association in January 2006. Grobner described nine patients with end-stage renal failure on dialysis who underwent MR angiography (MRA) with gadodiamide (Omniscan; GE Healthcare, Chalfont St Giles, UK); five (mean age 57.2±10.7 years) developed NSF after 2–4 weeks. All patients who developed NSF had confirmed metabolic acidosis, whilst all patients who did not develop NSF had normal pH and bicarbonate level at the time of the MRA. The average contrast agent volume used was 35 ml.
Marckmann et al. [23] reviewed case notes from their own nephrology department from August 2005 to May 2006 and found 13 patients with confirmed NSF, all exposed to gadodiamide. All patients were adults (33–66 years of age, mean 50 years), developed NSF 2–75 days (mean 25 days) after gadolinium exposure and had an eGFR <8 ml/min. From this cohort of renal patients the authors estimated an odds ratio between 32.5 and 47.6 for developing NSF following gadolinium exposure.
The link between gadolinium exposure and NSF has been further strengthened by two reports of gadolinium deposits in tissue samples of patients with NSF. High et al. [24] detected gadolinium in 4 out of 13 skin and soft-tissue biopsies from seven patients with documented NSF. All patients with NSF received gadolinium-based contrast medium prior to disease and the interval between dosing and biopsy could be determined with certainty only for two of four patients (age range 4–11 months). The same group reported in a research letter following quantitative analysis of the four tissue samples in which gadolinium was detected that the amount of gadolinium in the affected tissue of patients with NSF was approximately 35- to 150-fold the level of gadolinium in the bone of healthy volunteers with normal renal function. In a case report, Boyd et al. [25] found gadolinium deposits in areas of calcium phosphate deposition in blood vessels in a 68-year-old patient with NSF who had been previously exposed to gadodiamide.
Sadowski et al. [26] retrospectively reviewed data from 13 patients with confirmed NSF (eight males, five females, age 17–69 years), all exposed to gadodiamide. The median time between onset of NSF symptoms and last contrast-enhanced MRI examination was 11.5 days. The dose of contrast agent administered varied from 0.1 to 0.31 mmol/kg (actual) body weight. At the time of the contrast-enhanced MRI all 13 patients had CKD (eGFR <60 ml/min per 1.73 m2) and were hospitalized for a proinflammatory event (major surgery, infection or vascular event). The group with NSF had significantly lower eGFR, more proinflammatory events, and more contrast-enhanced MR examinations per patient (P<0.002) than did the control group. The incidence of NSF in this study and institution was 4.6% per year.
A review of six further NSF cases that were also related in time to gadodiamide administration has recently been published [27]. Patients were aged 23–71 years and the onset of symptoms consistent with NSF was between 19 days and 2 months after gadodiamide exposure. All patients had severe renal failure and were exposed to a gadodiamide dose ranging from 0.11 to 0.36 mmol/kg body weight. Despite having normal serum bicarbonate, five of the six patients had an elevated anion gap metabolic acidosis.
Broome et al. [28] have recently reported a retrospective study of 12 patients with NSF, all of whom received a double dose (0.2 mmol/kg body weight) of gadodiamide 2–11 weeks before the development of skin fibrosis. This group included eight men and four women (age 26–64 years) and all had renal insufficiency at the time of gadodiamide administration. The odds ratio for development of NSF in the gadodiamide-exposed patients compared with the gadodiamide-unexposed patients was 22.3 and the prevalence of NSF among gadodiamide-exposed dialysis patients was 4.0%. The odds ratio for development of NSF with double-dose compared to single-dose gadodiamide administration was 12.1. Ten patients were dialysed within 2 days of gadodiamide administration, but this failed to prevent the development of NSF.
Although the overwhelming majority (around 90%) of patients with gadolinium-associated NSF had been exposed to gadodiamide, there are now several reports of exposure to other gadolinium contrast agents. In a recent editorial in Radiology, Kuo et al. [29] reported 57 patients with NSF investigated by the United States Food and Drug Administration (FDA). In 43 of these the NSF was linked to gadodiamide; in 6 to gadopentetate dimeglumine (Magnevist; Berlex Imaging, Montville, N.J.); in 2 to gadoversetamide (OptiMARK; Mallinckrodt, St. Louis, Mo.); and in 3 to gadodiamide plus gadoversetamide; in 3 other patients, the associated specific gadolinium-based MR contrast agent was not definitively identified. Sadowski et al. [26] also reported one patient with confirmed NSF associated with both gadodiamide and gadobenate dimeglumine (MultiHance, Bracco Diagnostics, Milan, Italy). NSF related to specific gadolinium-based contrast agents is also monitored by the manufacturers, who have recently published several reports [30–32]. In a retrospective study in Scotland of 1,826 haemodialysis patients, Collidge et al. [33] found a positive association between the cumulative dose of gadodiamide and the development of NSF. Patients who developed NSF had received a higher median cumulative dose of gadodiamide (0.39 mmol/kg vs. 0.23 mmol/kg in patients without NSF).
A case-control study by Marckmann et al. [34] conducted in 19 patients with histologically verified NSF and 19 matched controls found that increasing cumulative gadodiamide exposure, high-dose epoietin-b treatment and higher serum concentrations of ionized calcium and phosphate increase the risk of gadodiamide-related NSF in patients with renal failure.
Pathophysiology
Although not completely understood, several factors are known to contribute to the physiopathological mechanism in NSF. Gadolinium chelates are mostly eliminated through the kidneys. Renal impairment will therefore prolong the presence of these agents in the body, facilitating the release of the toxic free gadolinium ion (Gd3+) by dissociation from its chelate (chemical imbalance, acidosis, anion gap, inflammation) and its subsequent binding to endogenous ions. This chemical process, known as transmetallation, may be promoted by acidosis and anion gap, e.g. under inflammatory conditions [35–40]. It has therefore been suggested that both the chemical stability of a particular chelate (cyclic compounds are more stable than linear ones in vitro) and the dose administered are risk factors for triggering NSF [28, 41]. Once released Gd3+ forms hydroxides and phosphates (insoluble at a pH >6.2) and these are probably engulfed by phagocytic cells [42]. This might affect the reticuloendothelial system, inhibiting the activity of certain enzymes with consequent foreign body and fibrous reactions [41, 43–45]. Gadodiamide, the agent involved in NSF in most patients to date, is the least stable in terms of susceptibility to transmetallation in vitro, being an open-chain compound [46]. Free gadolinium ions may remain in the body for weeks to months, allowing more time for toxic effects [47]. Gadolinium is a well-known inorganic calcium channel blocker and its acute toxicity may also partly be explained via this mechanism [48, 49]. A review of the properties of commercial gadolinium contrast agents and their relation to NSF is presented in Table 1.
Table 1Properties and approval status of extracellular gadolinium contrast agents (CNS central nervous system, MRA magnetic resonance angiography, N/A not available)Chemical nameTrade nameManufacturerCharge and chemical structureConditional stability at pH 7.4 (log)Excess chelate (mg/ml)Thermodynamic stability constant (log)Kinetic stabilityaBody region(s) approvedApproval statusApproved doses for imaging (mmol/kg)Approved doses for children (mmol/kg)NSF related to gadoliniumGadodiamideOmniscanGE HealthcareNonionic linear14.91216.935 sCNS, whole bodyUSA, EU, JapanBody 0.1–0.3, CNS 0.1–0.3, MRA 0.1–0.3From 6 months: 0.1Yes (180 cases worldwide)Gadopentetate dimeglumineMagnevistBayer Schering PharmaIonic linear17.70.422.110 minCNS, whole bodyUSA, EU, JapanBody 0.1, CNS 0.1–0.2, MRA 0.1–0.30.1 (doses of 0.2 may be used if necessary for children older than 2 years)Yes (78 cases worldwide)Gadobenate dimeglumineMultiHanceBraccoIonic linear16.9None22.6N/ACNS, liverUSA, EULiver 0.05, CNS 0.1, MRA not approvedNot approved <18 yearsYes (1 case in a patient coadministered Omniscan)GadoversetamideOptiMARKTycoNonionic linear15.028.416.6N/ACNS, liverUSABody 0.1, CNS 0.1, MRA not approvedNot approved <18 yearsYesGadoterate meglumineDotaremGuerbetIonic cyclic18.8None25.8>1 monthCNS, whole bodyEUBody 0.1, CNS 0.1–0.3, MRA 0.20.1NoGadoteridolProHanceBraccoNonionic cyclic17.10.2323.83 hCNS, whole bodyUSA, EU, JapanBody 0.1–0.3, CNS 0.1–0.3, MRA not approved0.1 from 2 years, caution 6 months to 2 years, contraindicated <6 monthsNoGadobutrolGadovistBayer Schering PharmaNon-ionic cyclicN/AN/A21.85 minCNS, MRAEU, CanadaBody not approved, CNS 0.1–0.3, MRA (imaging 1 FOV) 0.1–0.15, MRA (imaging >1 FOV) 0.2–0.3Not approved <18 yearsNoGadoxetic acidPrimovistBayer Schering PharmaIonic linearN/AN/A23.5N/ALiverUSA, EUBody 25 μmol/kg or 0.1 ml/kg, CNS not approved, MRA not approvedNot approved <18 yearsNoGadofosvesetVasovistBayer Schering PharmaIonic linearN/AN/AN/AN/AAbdominal and limb vesselsEUMRA 0.03Not approved <18 yearsNoaKinetic stability: dissociation half-life at pH 1.0.Table modified from Bellin MF (2006) Eur J Radiol 60:314–323.
NSF in children
NSF in children has also been reported. Five patients under 18 years of age have thus far been reported (discussed below) out of a total of ten in the central database of the International Center for Nephrogenic Fibrosing Dermopathy Research (Dr. S. Cowper, personal communication). However, 13 possible patients have been reported to an online paediatric nephrology forum in June 2007 (https://listhost.uchicago.edu/mailman/private/pedneph/) and these are still being investigated in relation to previous gadolinium exposure.
In 2003 Jan et al. [50] reported two paediatric patients with nephrogenic fibrosing dermopathy: a 16-year-old female patient with a life-long history of kidney failure and an 8-year-old boy with a 9-month history of membranoproliferative glomerulonephritis type II. Both patients had the clinical and histopathological features of NSF and no contrast-enhanced MRI procedures were known to be associated with these patients. However, this was before the publication of a possible link with gadolinium.
Jain et al. [51] reported two more patients with NSF in 2004: a 9-year-old boy on peritoneal dialysis and a 19-year-old male on haemodialysis with a 10-year history of previous failed renal transplants, both diagnosed with NSF on skin biopsies. The 9-year-old boy had a solitary MRA performed in June 2002, approximately 2 months prior to development of NFD. The history of the 19-year-old male was more complicated as he underwent 13 separate MRA procedures, all with gadolinium contrast agents, between April 2000 and August 2004. At least nine of the MRAs were performed prior to the NFD diagnosis in January 2003 [52]. Also, both patients had persistent metabolic acidosis in the months preceding the development of NSF, which may have been a contributing factor [22].
DiCarlo et al. [53] described another patient with histopathologically proven NSF. This patient was a 17-year-old male renal transplant recipient (15 years following transplant) who had recently begun peritoneal dialysis due to graft failure. No relationship to MRI investigations was mentioned.
Recently, Auron et al. [54] reported two patients with NSF. One of the patients was a 13-year-old boy previously reported at the age of 8 years by Jan et al. [50] in 2003 (discussed above). The second patient was a 20-year-old male who at the age 14 years was diagnosed with acute monoblastic leukaemia and 2 years later underwent bone marrow transplant and later developed kidney failure. No references to contrast-enhanced MRI are available from these reports.
Our own hospital has a large paediatric renal unit with up to 30 renal transplants per year. We have actively used MRI both for concurrent conditions and complications to CKD, as well as for providing a presurgical vascular road-map, using contrast-enhanced MRA (Gd-DTPA, Magnevist, Schering, Germany, 0.1–0.3 mmol/kg, i.v.). A review of all contrast-enhanced scans in children referred via our nephrology service (March 2002 to March 2007) showed that 75 nephrology patients (neonate to 18 years of age; median 9.6 years) had had 93 MRI scans with a follow-up of at least 6 months. There were no patients with NSF in this high-risk cohort.
Regulatory advice
Regulatory advice can be obtained from the websites listed in Table 2. The FDA has issued a general warning for all gadolinium-based contrast agents and advise against their use in patients with eGFR <30 ml/min per 1.73 m2 and in all patients with acute renal insufficiency “unless the diagnostic information is essential and not available with non-contrast enhanced magnetic resonance imaging”. At the FDA’s request the manufacturers of gadolinium-based contrast agents have revised the product labels, which now include a new boxed warning and a new warnings section that describes the risk of NSF. The European Medicines Agency has so far not published any update of general contraindications, and is not in a position to do so since all but one gadolinium-based agent are licensed by individual member states’ regulatory bodies. However, the issue was discussed at a meeting of its pharmacovigilance group in June and May 2007, and issued specific advice as follows (http://www.mhra.gov.uk):The use of Omniscan® and Magnevist® is contraindicated in patients with severe renal impairment (i.e. GFR or eGFR <30 mL/min/1·73 m2). Omniscan is also contraindicated in patients with renal dysfunction who have had, or who are awaiting, liver transplantation. For patients with moderate renal impairment (GFR or eGFR 30–59 ml/min/1·73 m2) or neonates and infants up to 1 year of age, Omniscan and Magnevist should be used only after careful consideration. All patients should be screened for renal dysfunction by obtaining a history and/or laboratory tests, especially before Omniscan or Magnevist are used. Haemodialysis shortly after administration of a gadolinium-containing MRI contrast agent in patients currently under haemodialysis may be useful for removal of contrast agent from the body. However, there is no evidence to suggest that haemodialysis can prevent or treat development of NSF. Careful consideration should also be given to the use of the other gadolinium-containing MRI contrast agents in patients with severe renal impairment.Table 2Websites for regulatory updates and for information and registration of cases (accessed on 20 September 2007)OrganizationInformation providedURLEuropean Medicines AgencyRegulatory informationhttp://www.emea.europa.eu/US Food and Drug AdministrationRegulatory informationhttp://www.fda.gov/cder/drug/infopage/gcca/default.htmInformation for healthcare professionalshttp://www.fda.gov/cder/drug/InfoSheets/HCP/gcca_200705.htmInternational Center for Nephrogenic Fibrosing Dermopathy ResearchUpdated information and contact details for registration of caseshttp://www.icnfdr.org/
Consequences for paediatric practice
There is an associated high morbidity and mortality with CKD, and imaging aids clinical management, such as assessment of secondary vasculopathies, infections etc, and surgical planning prior to renal transplantation. Patients at risk also include children with known or suspected CKD, as well as newborns and infants with renal immaturity or congenital cardiopathy. It seems plausible that impaired excretion, dissociation of the gadolinium chelate, and prolonged tissue retention of gadolinium is a trigger for NSF in patients with CKD. This risk increases with cumulative dose of gadolinium, with decreasing GFR <60 ml/min per 1.73 m2, and in the presence of acidosis and proinflammatory conditions. The scarcity of reported children with NSF may well be due to the lower prevalence of CKD and secondary conditions in children. It is therefore imperative that paediatric practitioners are prudent and immediately tighten their local policies for use of intravenous gadolinium-based contrast agents.
It is difficult at present to suggest general guidelines, because:
The absolute risk of gadolinium exposure, as well as the risk associated with the suspected contributing clinical and biochemical factors, is unknown. Each case must therefore be assessed individually.The risk attributable to the kinetic stability of different gadolinium-based agents is not clear. The majority of cases have been reported in association with a less stable agent (gadodiamide); however, we are still in an early stage of recording cases, and also the market share and the actual use of different agents in this patient population is not fully known. Thus, it seems reasonable to discontinue the use of gadodiamide in at-risk patients.The relationship to dose is not completely known. Since there is no indication that we are dealing with a side effect not related to dose, it seems reasonable to reduce the dose of gadolinium if possible, and also to monitor the cumulative dose in at-risk patients.It is unclear whether acidosis correction before gadolinium administration and haemodialysis immediately after gadolinium administration can prevent the development of, or treat, NSF, but such actions should be considered in children already on haemodialysis. This underlines the necessity for good communication between radiologists and nephrologists, both for issuing local guidelines and for management of the individual child.
It is often the case that MRI is requested after an inconclusive US examination. In our own experience, more meticulous patient preparation and dedicated rescanning of the patient with US can often give diagnostically adequate results. In patients with CKD not yet on dialysis CT is not a desirable alternative due to ionizing radiation and the nephrotoxicity of iodinated contrast agents. However, CT may provide an option in patients who are already on dialysis. MRI without intravenous contrast agent administration has not yet been validated for vascular anatomy prior to renal transplantation, but several techniques may be valuable, such as 3-D-balanced steady-state free precession imaging and time-of-flight angiography.
Patients with renal impairment may be referred for MRI from non-nephrology units and a crucial question is how to identify them. Our approach would be to include a mandatory tick box on the referral form requiring the referring clinician’s statement about renal function, and also to include a question about any known renal problems on the patient/guardian prescan checklist. This is a pragmatic approach, which will need revision after an initial phase.
Conclusion
We would advocate prudence, and in practice suggest:
Information collection from clinicians and patients/guardians on renal dysfunction and active collaboration with the local nephrology unit.
A case-by-case assessment in terms of the necessity for gadolinium-enhanced MRI, and whether an alternative test (repeat US, noncontrast-enhanced MRI, CT) would be acceptable.
Cumulative dose recording, reduction of the dose of any gadolinium-based contrast agent, and acting accordingly to guidance published in the literature.
Longer follow-up, especially in high-risk children.
GFR measurement or estimation should be part of a systematic prescan investigation in children with known or suspected renal dysfunction; however, a general screening is probably not justified.
To minimize the risks, we should take advantage of the current knowledge, and would suggest using cyclic nonionic compounds as they are more stable in vitro and therefore theoretically less likely to undergo transmetallation.
|
[
"gadolinium",
"nephrogenic systemic fibrosis",
"children",
"chronic kidney disease"
] |
[
"P",
"P",
"P",
"P"
] |
Graefes_Arch_Clin_Exp_Ophthalmol-4-1-2413124
|
One-year results of a multicenter controlled clinical trial of triamcinolone in pars plana vitrectomy
|
Purpose To evaluate the 1-year results of using triamcinolone acetonide (TA) in pars plana vitrectomy (PPV).
Introduction
Recently, vitrectomy has been used to treat various vitreoretinal diseases. It still remains a demanding procedure requiring a skilled and experienced surgeon, despite advances in surgical instrumentation and techniques. One major difficulty with this procedure is the transparency of the vitreous. Intra-operative triamcinolone acetonide (TA) has been used to address this problem, as it can be injected into the vitreous to visualize the posterior hyaloid or the internal limiting membrane (ILM) [1, 2]. Several previous reports have demonstrated the usefulness of TA as an adjunct to pars plana vitrectomy (PPV) [3–13]. Although their findings supported the use of TA-assisted PPV, these studies involved relatively small numbers of eyes and were not randomized. In addition, more recent reports on the use of intravitreous TA in patients not undergoing vitrectomy raised concerns about the risks of elevated intra-ocular pressure (IOP) [14–19].
We therefore undertook a multicenter prospective controlled study of TA-assisted pars plana vitrectomy (PPV). Considering the enduring nature of the effect of PPV on ocular tissue, it is important to investigate the long-term results in order to evaluate TA-assisted PPV. Moreover, the crucial issue of the effects on patients’ vision must be explored before TA-assisted PPV can be recommended. The current report presents the 1-year results of the use of TA in PPV, and evaluates the procedure in light of the aforementioned issues.
Methods
Study design, randomization and sample size
The prospective controlled clinical trial took place at eight hospitals in Japan as previously described (see Acknowledgements below) [13].
The patients were assigned to the two groups using a single-blind quasi-randomization approach within the participating clinical centers. Our preliminary study showed that the most important factor influencing the surgical outcome was not the nature of the disease for which PPV was used, but rather the hospital in which the surgery was performed (data not shown). To minimize this bias, equal numbers of patients from each hospital were assigned to each of the study groups. Thus, the patients were assigned by a single-blind quasi-randomization within the participating clinical centers. Briefly, after the patients passed the inclusion/exclusion criteria and gave consent for the study, they were numbered serially at each hospital, and alternate numbers were assigned to the TA-assisted and conventional PPV groups, respectively, as previously described [13]. Data analysis was performed by an intent-to-treat analysis method. To conceal the allocations at the participating hospitals, the surgeons were not informed of the group to which a patient had been assigned until immediately before surgery.
This is the second report of controlled clinical trial of intraocular TA in PPV. Primary outcome was whether the use of TA might reduce the incidence of intraoperative complications. The sample size was determined based upon our preliminary study, and the method has been described elsewhere [13]. Briefly, the incidence of intraoperative retinal breaks was found to be approximately 15% in the participating hospitals. The sample size calculation was based on our pilot study data, and was aimed to detect a difference in short-term postoperative complications as reported previously [13]
Eligibility criteria and surgery
Institutional review board (IRB)/ethics committee approval was obtained from all of the participating clinical establishments, such that all patients undergoing PPV at these hospitals were eligible to enroll in the study. Each patient was fully informed about the nature of the treatment, and provided written consent. The following groups were excluded: (1) patients with diagnoses that included macular holes; patients with uveitis, (2) patients with phakic lenses, (3) patients who were known to be steroid responders, to have glaucoma, or to have apparent infectious endophthalmitis, and (4) patients who were undergoing a second PPV. For ethical reasons, if at the time of surgery a surgeon felt that TA was necessary to achieve a successful PPV outcome, it was included in the procedure; for the purposes of the intent-to-treat analysis, these cases were included within the conventional PPV group.
The details of the surgical protocols varied between the hospitals, but the basic procedure was similar. PPV was performed according to a previously reported methodology, either with or without the use of a TA suspension prepared as described elsewhere [2, 13]. The recording of the operations, control of surgical quality, use of antibiotics, and termination of the study have been described elsewhere [13].
During the follow-up period, additional surgeries were performed at each hospital as and when they were needed. For example, filtering surgeries were performed when the intra-ocular pressure (IOP) remained at an uncontrollably high level (25 mmHg or greater) even with full medication. Full medication included the use of all latanoprost, beta-blocker, and dorzolamide hydrochloride eyedrops and acetazolamide medicine.
Data collection and documentation were performed as previously described [13]. For the purposes of this study, the adverse events related to surgery were defined as follows: retinal breaks, macular pucker, retinal fibrous membrane formation, retinal detachment, vitreous hemorrhage (VH), rubeosis iridis, optic-disc damage, corneal diseases, and after cataract. The serious adverse events related to surgery were defined as follows: infectious endophthalmitis, retinal degeneration, unexplained optic-disc damage, and unexplained deterioration of vision.
Masking and outcome assessment
The study records were collected by the principal investigator (TS) up until November 2006. After checking the records for harmful events, the data were sent to the controllers (YN) at Kyushu University. The remaining investigators were not permitted access to any information on the outcomes before completion of the analysis. Visual acuity was measured by ophthalmic technicians using the Snellen high-contrast acuity test. IOP was also measured by the ophthalmic technicians, and the findings were reviewed by the physicians. IOP was measured with pneumotonometer. The ophthalmic technicians were not informed of the purpose of the study or the assignment schedule. Anti-glaucoma eye drops and/or acetazolamide were used when an IOP of 22 mmHg or higher persisted for 3 days.
Statistical analysis
The visual acuity was converted to the logarithm of the minimum angle of resolution (LogMAR), and the baseline value was compared with those at 1 week, 1 month, 3 months, 6 months, and 1 year. The eyes were categorized into the following three groups: (1) the improved group, in which patients showed a 0.3 logMAR or greater improvement of visual acuity at the final examination compared with baseline, (2) the deteriorated group, in which patients showed a 0.3 logMAR or worse deterioration of visual acuity at the final examination compared with baseline, and (3) the unchanged group, in which patients did not meet the criteria of either the improved or the deteriorated group.
A Kaplan-Meier survival analysis was used to estimate the probability of improved visual acuity with time after surgery. The improvement or deterioration was analyzed by the log rank test. The earliest time point after surgery at which the events in each category were reported was designated as the time point of each event.
We also estimated the multiple-adjusted odds ratio (OR) and the 95% confidence interval (95%CI) both for TA use and for the other variables, using a logistic regression model. The parameters that were adjusted for included age, gender, and disease diagnosis, which was categorized into the following five groups: proliferative diabetic retinopathy (PDR), diabetic macular edema (DME), rhegmatogenous retinal detachment (RRD), retinal vein occlusion (RVO), and others. A diagnosis of PDR was used as a reference. Even when a single eye received multiple additional surgeries, it was counted as one eye. All statistical analyses were performed using SAS® (Proprietary Software Release 8.2; SAS Institute Inc., Cary, NC, USA). The differences in diagnosis or surgical intervention between the two groups were compared using a two-tailed Chi-squared test with the Yates’ correction. P values < 0.05 were considered statistically significant.
Results
Number of eyes
In total, 920 patients were initially registered for PPV during this period. Among these, 146 eyes were not included in the study because they did not meet the eligibility criteria or because the patient declined to participate. In total, 774 eyes were randomized, with 391 eyes (from 216 male and 175 female patients) assigned to the TA-assisted PPV group and 383 eyes (201 male and 182 female patients) assigned to the conventional PPV group. The mean ages of the patients were 63.5 ± 12.1 years in the TA-assisted PPV group and 63.0 ± 11.3 years in the conventional PPV group. In total, 19 eyes that were originally assigned to the conventional PPV group received unplanned TA based on surgical judgment. The 1-year post-operative follow-up was completed for a total of 305 eyes in the TA-assisted PPV group (78.0%) and 295 eyes in the conventional PPV group (77.0%). The overall follow-up rate was 77.5% (Fig. 1).
Fig. 1Schematic diagram of the study. TA: triamcinolone acetonide, PPV: pars plana vitrectomy
Diagnoses
The diagnoses for each group are shown in Table 1. In both groups, PDR accounted for the largest number of cases, followed by RRD. In both groups, the pattern at the 1-year follow-up was similar to that at the baseline, and the different diagnoses had significantly different distributions within the two groups: the TA-assisted PPV group had a relatively larger number of patients with DME (P = 0.018 at baseline, P = 0.006 at 1-year follow-up) and a smaller number of patients with epiretinal membrane (ERM) formation (P = 0.031 at baseline, P = 0.008 at 1-year follow-up).
Table 1Diseases of each group at baseline and 1 year after surgery TA-assisted PPVConventional PPVBaseline1 yearbaseline1 yearN (%)N (%)N (%)N (%)AMD7 (1.8%)3 (1.0%)3 (0.8%)3 (1.0%)BRVO50 (12.8%)41 (13.4%)37 (9.7%)29 (9.8%)CRVO11 (2.8%)9 (3.0%)7 (1.8%)6 (2.0%)DME67 (17.1%)60 (19.7%)42 (11.0%)34 (11.5%)ERM24 (6.1%)19 (6.2%)41 (10.7%)37 (12.5%)Lens luxation16 (4.1%)11 (3.6%)16 (4.2%)12 (4.1%)Macroaneurysm10 (2.6%)8 (2.6%)9 (2.3%)4 (1.4%)PDR113 (28.9%)85 (27.9%)131 (34.2%)101 (34.2%)RRD80 (20.5%)59 (19.3%)77 (20.1%)54 (18.3%)VH6 (1.5%)6 (2.0%)14 (3.7%)11 (3.7%)Others7 (1.8%)4 (1.3%)6 (1.6%)4 (1.4%)total383 (100%)305 (100%)391 (100%)295 (100%)TA; triamcinolone acetonide: PPV; pars plana vitrectomy: AMD; age-related macular degeneration: BRVO;branch retinal vein occlusion: CRVO: central retinal vein occlusion: DME; diabetic macular edema: ERM; epiretinal membrane: PDR; proliferative diabetic retinopathy: RRD; rhegmatogenous retinal detachment: VH; vitreous hemorrhage.
There were no significant differences between the two groups in the usage of surgical methods, such as cataract surgery (phacoemulsification P = 0.724 and pars plana lensectomy P = 0.272), ILM peeling (P = 0.075) and endolaser-photocoagulation (P = 0.413), or in the use of post-operative tamponade (sulfur hexafluoride gas P = 0.810, octa-fluorepropane gas P = 0.756, room air P > 0.999, and silicone oil P = 0.831) [13].
Visual acuity
The visual acuity in both groups improved over time. In the TA-assisted PPV group, 322 eyes were categorized as improved, 21 were categorized as unchanged, and 48 were categorized as deteriorated. In the conventional PPV group, 312 eyes were categorized as improved, 45 eyes were categorized as unchanged, and 26 eyes were categorized as deteriorated. The log-rank test revealed no significant difference between the two groups regarding the improvement (P = 0.98) or deterioration (P = 0.26) of visual acuity. The logistic regression model also showed that the intra-operative use of TA was not a significant factor for the improvement of visual acuity (P = 0.91) after adjusting for age, gender, and diagnosis (Table 2). Neither gender (P = 0.54) nor age (P = 0.31) was a significant factor for the improvement of visual acuity. However, PDR diagnosis was a significant factor for the improvement of visual acuity compared with DME (OR = 0.63, 95%CI = 0.478–0.817, P = 0.0006; Table 2).
Table 2Factors affecting improvement of visual acuityVariable (no. of improved eyes/total eyes)Odds ratio (95%CI)P-valueTA-assisted PPV (322/391) vs conventional PPV (312/383)1.10 (0.860–1.183)0.91Age1.00 (0.991–1.005)0.54Gender (male vs female)0.92 (0.785–1.080)0.31DiseasesRD vs PDR1.04 (0.834–1.340)0.71DME vs PDR0.63 (0.478–0.817)0.0006*RVO vs PDR1.21 (0.938–1.561)0.14ERM vs PDR0.78 (0.568–1.061)0.11Others vs PDR1.21 (0.924–1.579)0.17CI; confidence interval: TA; triamcinolone acetonide: PPV; pars plana vitrectomy: RVO; branch retinal vein occlusion + central retinal vein occlusion: DME; diabetic macular edema: ERM; epiretinal membrane: PDR; proliferative diabetic retinopathy: RRD; rhegmatogenous retinal detachment. * statistically significant
Post-operative complications
Subgroup analysis
In terms of post-operative complications, there was not any statistically significant difference. Fibrous ERM was observed in five eyes (1.4%) in the conventional PPV group, but in only one eye (0.3%) in the TA-assisted PPV group. A subgroup analysis showed that this difference was not significant (P = 0.10). Among the other complications, an IOP increase was found more frequently in the TA-assisted PPV group (23 eyes, 5.9%) than in the conventional PPV group (13 eyes, 3.4%), although this difference was not statistically significant (P = 0.10). There were no significant differences in the other complications between the two groups (Table 3).
Table 3Post-operative complications in each group TA-assisted PPV (total 391 eyes): n (%)Conventional PPV (total 383 eyes): n (%)P-value*After cataract18 (4.6%)17 (4.4%)0.91Macular pucker23 (5.9%)22 (5.7%)0.93IOP increase23 (5.9%)13 (3.4%)0.10Retinal detachment9 (2.3%)9 (2.3%)0.96Vitreous hemorrhage22 (5.6%)29 (7.8%)0.28Fibrous membrane1 (0.3%)5 (1.4%)0.10Iris synechiae6 (1.5%)4 (1.0%)0.55Others8 (2.0%)12 (3.1%)0.34TA; triamcinolone acetonide: PPV; pars plana vitrectomy: PPL; pas plana lensectomy: IOP; intraocular pressure: * Chi-square test with the Yates’ correction.
Additional surgery
Although the filtering surgery was more performed in TA-assisted PPV and the removal of vitreous hemorrhage was more performed in conventional PPV, there was not any statistically significant difference between the two groups with respect to causes of additional surgery after PPV. In the TA-assisted PPV group, 22 surgeries were performed (excluding filtering surgery). The most common reasons for additional surgery were removal of silicone oil (11 eyes), followed by retinal detachment (eight eyes), VH (six eyes), and others (seven eyes). In the conventional PPV group, 26 surgeries were performed (excluding filtering surgery). The most common reasons were VH (12 eyes), followed by retinal detachment (nine eyes), removal of silicone oil (seven eyes), and others (six eyes). Although there was a tendency for VH to be less common in TA-assisted PPV eyes, no significant difference was found in comparison to conventional PPV (P = 0.13; Table 4). Filtering surgery was more frequently performed in the TA-assisted PPV group (15 cases, 3.8%) than the conventional PPV group (seven cases, 1.8%), although this difference was not significant (P = 0.10). Ten eyes received multiple additional surgeries in TA-assisted PPV group and seven eyes in conventional PPV group.
Table 4Causes of additional surgery after PPV TA-assisted PPV (total 391 eyes): n (%)Conventional PPV (total 383 eyes) : n (%)P-value*Silicone removal11 (2.8%)7 (1.8%)0.36Retinal detachment8 (2.0%)9 (2.3%)0.77Vitreous hemorrhage6 (1.5%)12 (3.1%)0.14Filtering surgery15 (3.8%)7 (1.8%)0.10Others7 (1.8%)6 (1.6%)0.81Total37 (9.5%)✝32 (8.6%)✝✝0.68TA; triamcinolone acetonide: PPV; pars plana vitrectomy: * Chi-square test with the Yates’ correction: ✝ Ten eyes received multiple additional surgeries: ✝✝ Seven eyes received multiple additional surgeries.
The survival curve produced by the log-rank test showed that the incidence of additional surgery was not significantly different between the two groups (P = 0.45). The logistic regression test also showed that the use of TA was not a significant factor for additional surgery after adjustments for gender, age, and diagnosis. However, a diagnosis of PDR was a significant factor for additional surgery in comparison to DME (P = 0.005), RVO (P = 0.006), and ERM (P = 0.02; Table 5). None of the variables were related to the requirement for filtering surgery (Table 6).
Table 5Factors affecting additional surgeryVariable (no. of additional surgery/total)Odds ratio (95%CI)P-valueTA-assisted PPV (37/391) vs conventional PPV (32/383)1.23 (0.797–1.911)0.35Age0.97 (0.950–0.984)0.0001*Gender (male vs female)1.15 (0.733–1.797)0.55DiseasesRD vs PDR0.60 (0.345–1.029)0.06DME vs PDR0.29 (0.122–0.681)0.005*RVO vs PDR0.23 (0.084–0.657)0.006*ERM vs PDR0.087 (0.012–0.631)0.02*Others vs PDR0.64 (0.299–1.374)0.25CI; confidence interval: RVO; branch retinal vein occlusion + central retinal vein occlusion: DME; diabetic macular edema: ERM; epiretinal membrane: PDR; proliferative diabetic retinopathy: RRD; rhegmatogenous retinal detachment. * statistically significantTable 6Factors affecting filtering surgeryVariable (no. of filtering surgery/total)Odds ratio (95%CI)P-valueTA-assisted PPV (12/391) vs conventional PPV (5/383)1.74 (0.708–4.265)0.23Age0.98 (0.941–1.018)0.28Gender (male vs female)1.81 (0.696–4.728)0.22DiseasesRRD vs PDR–0.99DME vs PDR–0.99RVO vs PDR–0.99ERM vs PDR–0.99Others vs PDR0.83 (0.266–2.573)0.74CI; confidence interval: TA; triamcinlone acetonide: PPV; pars plana vitrectomy: RVO; branch retinal vein occlusion + central retinal vein occlusion: DME; diabetic macular edema: ERM; epiretinal membrane: PDR; proliferative diabetic retinopathy: RRD; rhegmatogenous retinal detachment
Serious adverse events
None of the serious adverse events related to surgery defined in this study (such as retinal degeneration, endophthalmitis, an unexplained decrease in visual acuity, or optic-disc atrophy) were observed in either group during the 1-year follow-up.
Discussion
The first report on this study showed that the intra-operative use of TA significantly decreased the incidence of intra-operative complications associated with PPV, and no adverse events related to surgery were found over a 3-month observation period [13]. A decreased incidence of intra-operative complications is beneficial; however, the critical issue is whether the surgery improves the patients’ vision. We therefore followed up the visual acuity, which is a crucial factor influencing vision, for 1 year after surgery. Our results revealed no significant difference in visual acuity between the TA-assisted PPV group and the conventional PPV group after 1 year. As described in the earlier report, which analyzed data from the same eyes, intra-operative retinal breaks and retinal detachment were significantly less frequent in the TA-assisted PPV group than in the conventional PPV group. Nonetheless, the decreased incidence of intra-operative complications did not significantly affect the visual acuity after 1 year. This finding can be explained by the following facts. First, surgical skill is the most important factor for surgical success. As all of the operations in this study were performed by qualified and experienced surgeons, the intra-operative complications were managed so as not to cause serious post-operative complications. Thus, the influence of intra-operative complications on visual acuity might have been minimal. Second, a patient’s vision after vitrectomy might be more strongly affected by a disease diagnosis than by intra-operative complications. Indeed, eyes with PDR were significantly more likely to show improved visual acuity compared with baseline than eyes with DME in the current study. As our present case series included various diseases, the impact of intra-operative complications on post-operative visual acuity might have been masked by another strong factor, such as disease diagnosis, even after adjustments. Third, the eyes in the current study were primarily chosen for the detection of intra-operative retinal breaks, based on the hypothesis that the intra-operative use of TA might decrease their incidence from 15% to 8% [13]. Thus, the sample size might not have been large enough to detect the superiority or inferiority of TA-assisted PPV. For example, in order to detect the difference of post-operative improvement of vision in TA-assisted PPV and conventional PPV based upon the present results, approximately 29,315 cases in each group are necessary with α power of 0.05 and 1-β of 0.8. However, considering the relative stability of the visual acuity observed here, we are confident that the intra-operative use of TA does not have a positive or negative effect on postoperative visual acuity for 1 year.
Additional surgeries carry associated risks and can be costly; they are thus important when justifying the intra-operative use of TA in PPV. The incidence of additional surgery after vitrectomy was found to be equivalent using both methods. Considering the IOP increase after surgery, the opportunity to use anti-glaucoma eye drops was high in the TA-assisted PPV group [13], and the number of filtering surgeries performed during 1 year after TA-assisted PPV was also high; however, the incidence of additional surgeries was not statistically different between the groups. Several previous reports showed that intravitreous TA injection required filtering surgery, and its frequency was less than 1% of all treated eyes [18, 19]. In contrast to intravitreous TA injection, most of the TA was removed from the eye at the end of the surgery. Thus, it was plausible that the necessity of filtering surgery in TA-assisted PPV was not significant. The other causes of additional surgery occurred at similar frequencies, and there were no significant differences between the groups. A diagnosis of PDR was a significant factor for the requirement for additional surgery, as it was for the improvement of visual acuity. Furthermore, in our earlier report on the 3-month follow-up study, a diagnosis of PDR was found to be a significant risk factor for the occurrence of intra-operative complications. PDR eyes thus appear to have several poor prognostic factors, consistent with the present results.
With regard to the post-operative complications, pre-retinal fibrosis appeared to be less common in the TA-assisted PPV group than in the conventional PPV group, although this was not statistically significant. This advantage of TA-assisted PPV has been described in previous reports [3, 5–9]. The intra-operative use of TA visualizes the posterior hyaloid clearly, which facilitates its removal [6, 7]; as the residual posterior hyaloid can act as a scaffold for pre-retinal fibrous-membrane formation, its removal could reduce the incidence of post-operative pre-retinal fibrosis [5–7]. In the current study group, the removal of the residual posterior hyaloid was carried out according to the surgeons’ preference—some surgeons removed it completely, while others did not. This might have obscured any potential advantage of the intra-operative use of TA in inhibiting post-operative pre-retinal fibrosis.
There were clear limitations to our analysis. The number of cases in each study group was primarily chosen to allow the detection of intra-operative complications; thus, the significance of the intra-operative use of TA might only become clear when studying larger numbers of patients. In addition, the sample size was too small to detect the risk of endophthalmitis. The incidence of endophthalmitis related to PPV was 0.046% in the conventional PPV group and 0.053% in the TA-assisted PPV group [20, 21]. An analysis of a greater number of cases will thus be necessary to evaluate the risk of endophthalmitis related to TA-assisted PPV. However, the present sample size was large enough to detect several factors that affected post-operative visual acuity or the need for additional surgery. Another limitation was that the post-operative visual acuity was strongly influenced by the diagnosis of disease. Eyes with simple VH clearly had a greater chance of surgery improving vision above the baseline level than did eyes with complicated PDR. In order to evaluate the influence of the intra-operative use of TA on visual acuity more precisely, we are currently performing a study on eyes with the same disease diagnosis. Furthermore, the subgroup of DME or ERM in each group was not completely matched. The interpretation of the results should be done carefully. Additionally, caution must be applied when making generalizations on the basis of our current results, and their application to clinical practice must be undertaken with care. Other factors, such as invasiveness and cost, are also important in seeking justification for this treatment. The success of surgery depends upon various factors, and surgical skill is critical. When considering the likelihood of retinal breaks and their recovery without damaging vision, comparisons of the hazard ratios of TA-assisted PPV and conventional PPV would be valid only if made by individuals who qualify as PPV specialists.
In summary, the intra-operative use of TA in PPV did not affect visual acuity over 1 year. As negative or adverse events were not observed in this case series, a more detailed study is warranted to establish the value of TA in PPV.
|
[
"steroid",
"retinal detachment",
"diabetic retinopathy",
"proliferative vitreoretinopathy",
"adjuvant"
] |
[
"P",
"P",
"P",
"M",
"U"
] |
Clin_Oral_Investig-4-1-2238784
|
The evolution of tooth wear indices
|
Tooth wear—attrition, erosion and abrasion—is perceived internationally as an ever-increasing problem. Clinical and epidemiological studies, however, are difficult to interpret and compare due to differences in terminology and the large number of indices that have been developed for diagnosing, grading and monitoring dental hard tissue loss. These indices have been designed to identify increasing severity and are usually numerical. Some record lesions on an aetiological basis (e.g. erosion indices), others record lesions irrespective of aetiology (tooth wear indices); none have universal acceptance, complicating the evaluation of the true increase in prevalence reported. This article considers the ideal requirements for an erosion index. It reviews the literature to consider how current indices have evolved and discusses if these indices meet the clinical and research needs of the dental profession.
Introduction
There is both a clinical and a scientific need to be able to measure tooth wear, and the literature abounds with many methods which can be broadly divided into quantitative and qualitative in nature. Quantitative methods tend to rely on objective physical measurements, such as depth of groove, area of facet or height of crown. Qualitative methods, which rely on clinical descriptions, can be more subjective if appropriate training and calibration are not carried out but which, with correct safeguards, can be valuable epidemiological tools. In a clinical intra-oral examination, there will be an inclination towards descriptive assessment measures, such as mild, moderate or severe, rather than quantitative measurement, which is easier to perform reliably on a model or in the laboratory. Such methods tend to be more sensitive but do not lend themselves readily to clinical use—especially in epidemiology, where fieldwork data collection is often carried out in an environment lacking sophisticated equipment.
Quantitative and qualitative methods typically utilise grading or scoring systems designed to identify increasing severity or progression of a condition; these are described as indices and are usually numerical. An ideal index should be simple to understand and use, clear in its scoring criteria and be demonstrably reproducible. Its application should be useful for research into the aetiology, prevention and monitoring of a condition, essentially being an epidemiological and clinical tool.
Review of the literature reveals the fact that many different tooth wear indices have been developed for clinical and laboratory use all over the world. Unfortunately, the production of so many indices does not allow for ready comparison of results between different working groups, and this is especially important in epidemiology when trying to define the prevalence of a condition. Confusion is further generated in the literature as the majority of researchers, in their attempts to quantify the amount of tooth tissue loss due to tooth wear, have historically concentrated on one aetiology only, and these indices tend to be surface limited. Often, the wear patterns described do not appear to reflect the aetiology suggested, and this relates to lack of uniformity with tooth wear terminology and translation errors. Many diagnostic indices do not properly reflect the morphological defects, and there is little international standardisation. All of these factors complicate the comparison of data and evaluation of the efficacy of preventive and therapeutic measures.
The literature identifies separate indices for use in clinical and laboratory situations and specific indices for attrition, abrasion, erosion and multifactorial tooth wear. There are common threads to all of the indices, such as descriptive diagnostic criteria and criteria for quantifying the amount of hard tissue loss. These generally consider the size of the affected area—as a proportion of a sound surface and/or the depth of tissue loss—often expressed as a degree of dentine exposure.
The clinical measurement of erosion
The earliest indices shared common, arbitrary criteria, relying on descriptive terms such as slight, mild, moderate, severe and extensive. Restarski et al. [26] developed a six-point grading system to evaluate the severity of erosive destruction observed on the lingual surfaces of rat and puppy molars, but concerns were raised with regards to reproducibility. With vague criteria definitions, variability in recording is expected. Each animal was allocated a total score, calculated by summing the mean molar quadrant scores. Whilst producing simple data for analysis, it is acknowledged that averaging scores in this manner leads to the loss of much data. If the number of teeth severely affected is small, the erosion score will be low; but this could mask a significant, localised clinical problem [27].
Eccles [9] originally classified lesions broadly as early, small and advanced, with no strict criteria definitions, thus allowing wide interpretation. Later, the index was refined and expanded, with greater emphasis on the descriptive criteria [10]. It was presented as a comprehensive qualitative index, grading both severity and site of erosion due to non-industrial causes, and is considered as one of the cardinal indices from which others have evolved. In essence, it breaks down into three classes of erosion, denoting the type of lesion, assigned to four surfaces, representing the surface where erosion was detected (Table 1).
Table 1Eccles index for dental erosion of non-industrial origin [10]ClassSurfaceCriteriaClass I Early stages of erosion, absence of developmental ridges, smooth, glazed surface occurring mainly on labial surfaces of maxillary incisors and caninesClass IIFacialDentine involved for less than one third surface; two typesType 1 (commonest): ovoid–crescentic in outline, concave in cross section at cervical region of surface. Must differentiate from wedge shaped abrasion lesionsType 2: irregular lesion entirely within crown. Punched out appearance, where enamel is absent from floorClass IIIaFacialMore extensive destruction of dentine, affecting anterior teeth particularly. Majority of lesions affect a large part of the surface, but some are localised and hollowed outClass IIIbLingual or palatalDentine eroded for more than one third of the surface area. Gingival and proximal enamel margins have white, etched appearance. Incisal edges translucent due to loss of dentine. Dentine is smooth and anteriorly is flat or hollowed out, often extending into secondary dentineClass IIIcIncisal or occlusalSurfaces involved into dentine, appearing flattened or with cupping. Incisal edges appear translucent due to undermined enamel; restorations are raised above surrounding tooth surfaceClass IIIdAllSeverely affected teeth, where both labial and lingual surfaces are extensively involved. Proximal surfaces may be affected; teeth are shortened
Greater accuracy was introduced by Xhonga and Valdmanis [33] who divided erosions into four levels by measurement with a periodontal probe: none, minor (less than 2 mm), moderate (up to 3 mm) and severe (greater than 3 mm). They further differentiated types of erosion by morphological descriptions, such as wedge, saucer, groove and atypical. They did not address the problem of inter- or intra-examiner variability.
The clinical measurement of tooth wear
It is perhaps significant that the earliest index documented by Broca [4] and used as a foundation for the development of further indices graded horizontal or oblique patterns of occlusal wear without presupposing the aetiology. Smith and Knight [30] introduced the more general concept of measuring tooth wear per se, irrespective of the cause, and since then more recent indices have been developed or modified from Smith and Knight that do not rely on a prior diagnosis and are more clinically relevant. Most of these stress the importance of user training sessions and calibration exercises.
Smith and Knight [30] took Eccles’ ideas a stage further, producing the tooth wear index (TWI), a comprehensive system whereby all four visible surfaces (buccal, cervical, lingual and occlusal–incisal) of all teeth present are scored for wear, irrespective of how it occurred (Table 2). This avoids the confusion associated with terminology and translation or differences in opinion for diagnosis of aetiology based on clinical findings. Guidelines for using the criteria were produced in a booklet by the authors to aid training and standardisation with other investigators; in cases of doubt, the lowest score is given. Complete enamel loss (score 4) may, however, be misleading, as there is almost always a rim of enamel at the worn surface margins—the colloquial “enamel halo.”
Table 2Smith and Knight tooth wear index [30]ScoreSurfaceCriteria0B/L/O/INo loss of enamel surface characteristicsCNo loss of contour1B/L/O/ILoss of enamel surface characteristicsCMinimal loss of contour2B/L/OLoss of enamel exposing dentine for less than one third of surfaceILoss of enamel just exposing dentineCDefect less than 1 mm deep3B/L/OLoss of enamel exposing dentine for more than one third of surfaceILoss of enamel and substantial loss of dentineCDefect less than 1–2 mm deep4B/L/OComplete enamel loss–pulp exposure–secondary dentine exposureIPulp exposure or exposure of secondary dentineCDefect more than 2 mm deep–pulp exposure–secondary dentine exposure
This index was the first one designed to measure and monitor multifactorial tooth wear; a further pioneering feature was the ability to distinguish acceptable and pathological levels of wear, by comparison with threshold normal values for the age groups studied. Tooth wear was defined as pathological if the teeth became so worn that they do not function effectively or seriously mar the appearance—before they are lost through other causes—or the patient dies. Results from inter- and intra-reproducibility exercises were within a range regarded as acceptable for epidemiological purposes, and the index appears simple to use clinically—intra-orally or from models and photographs. However, some problems have been identified with the TWI, including the time necessary to apply to a whole dentition, amount of data generated and the comparisons with threshold levels for each age group; the thresholds proposed were high, erring towards understatement rather than exaggerations of pathological wear. Full use of the index as a research tool is not feasible without computer assistance.
A sign of professional confidence in this index is its adopted use by a number of different investigators in the UK looking at tooth wear prevalence and severity [3, 20, 25, 27], aetiology and risk [1, 21]. Other researchers have used it with modifications pertaining to the particular age group being studied. Millward et al. [18] made adjustments to study erosion in the primary and secondary dentitions, excluding cervical surfaces. Subjects were then grouped together to produce three classifications: no or mild erosion—no score >1, moderate—at least one tooth in dentition score 2, severe—at least one tooth in dentition score 3 or 4.
In a survey of elderly people, Steele et al. [31] combined low wear scores representing small losses of enamel and used the worst surface score per tooth as an overall tooth score. Donachie and Walls [7, 8] outlined various flaws in the tooth wear index as an epidemiological tool in the ageing population and suggested a need to increase the sensitivity of TWI at extremes of tooth wear, to take account of the capacity of the elderly to have adequate function in cases of significant wear. They suggested altering threshold values, amplifying scoring criteria and creating a sixth point to distinguish between exposure of secondary dentine and frank pulp exposure.
Development of indices
Many other indices have been proposed for measuring erosive tooth wear [2, 14, 15, 17, 23] which have their roots in the indices of Eccles [10] and Smith and Knight [30]. Linkosalo and Markkanen [15] utilised a qualitative index with listed diagnostic criteria to confirm lesions as erosive and a four-scale grading of severity, relating to involvement of dentine. Their scoring system was modified by Lussi et al. [17] to create an erosive index that has been used widely by European workers to score the facial, lingual and occlusal surfaces of all teeth except the third molars (Table 3).
Table 3Erosion index according to Lussi [16, 17]SurfaceScoreCriteriaFacial0No erosion. Surface with a smooth, silky glazed appearance, possible absence of developmental ridges1Loss of surface enamel. Intact enamel cervical to the erosive lesion; concavity on enamel where breadth clearly exceeds depth, thus distinguishing it from toothbrush abrasion. Undulating borders of the lesion are possible and dentine is not involved2Involvement of dentine for less than half of tooth surface3Involvement of dentine for more than half of tooth surfaceOcclusal/lingual0No erosion. Surface with a smooth, silky glazed appearance, possible absence of developmental ridges1Slight erosion, rounded cusps, edges of restorations rising above the level of adjacent tooth surface, grooves on occlusal aspects. Loss of surface enamel. Dentine is not involved2Severe erosions, more pronounced signs than in grade 1. Dentine is involved
O’Sullivan [23] proposed a new index for the measurement of erosion specifically in children. The index was qualitative with a broad attempt at quantification noting whether less or more than half of the surface was affected. Every tooth was examined and assigned a three-digit score relating to the site of erosion, severity (grade 0–5) and area of surface affected. O’Brien [22] reported the use of a partial recording system for measuring erosion in children in UK children’s dental health surveys, where only the facial and lingual surfaces of the primary and permanent maxillary incisor teeth were scored for erosion. Again, the criteria were mostly qualitative and descriptive, with a broad attempt made to quantify the area involved.
Bardsley et al. [2] pioneered a new, simplified version of TWI [30] when carrying out epidemiological studies on large numbers of adolescents in North West England (Table 4). Tooth wear scoring was essentially dichotomised into the presence or absence of dentine, with even cupping of dentine scoring one. A partial recording system was used, collecting data from 40 surfaces including occlusal surfaces of the four first molar teeth and the labial, incisal and lingual–palatal surfaces of the six upper and lower anterior teeth.
Table 4Simplified scoring criteria for TWI [2]ScoreCriteria0No wear into dentine1Dentine just visible (including cupping) or dentine exposed for less than 1/3 of surface2Dentine exposure greater than 1/3 of surface3Exposure of pulp or secondary dentine
However, despite calibration and training, difficulties were experienced diagnosing dentine exposure in the epidemiological field and there is some debate as to the significance of dentinal cupping when exposed dentine does not relate to significant amounts of tissue loss [12]. In a recent study of Ganss et al. [12], teeth were visually and histologically examined for presence of exposed dentine, and the correlation in accuracy between the two examinations was poor. The diagnosis of exposed dentine is obviously important but may not be accurate from visual examination alone.
Oilo et al. [24] concentrated on a different type of scoring system, with criteria based on treatment need. They criticised the use of indices that used a nonlinear scoring method, claiming calculated mean wear scores can be misleading. Their index was based on Ryge and Snyder’s [28] system for evaluating the quality of restorations and had five categories divided into two broad camps; Romeo, Sierra and Mike were satisfactory, whilst Tango and Victor indicated unacceptable levels of wear requiring treatment. All groups except Romeo were subdivided according to degree of dentine exposed and clinical findings such as pain, sensitivity and fracture of restorations, giving the impression of a cumbersome system. Dahl et al. [6] modified it with the introduction of even more categories, with an aim to establish subjective dental criteria for present and future evaluations of tooth wear and the need for treatment. In practice, these indices require experience for reliable use; individuals with differing clinical backgrounds will not get consistent, objective results.
Indices that attempt to visualise, measure and monitor the amount of worn enamel or exposed dentine by difficult direct clinical interpretation [13] and indirectly on accurate, serial study casts which must be retained [5, 11, 29, 32] have been described. Larsen et al. [14] recommended a new clinical index based on a combination of clinical examination, photographs and study casts, with complicated qualitative and quantitative criteria. Plaque-free teeth were clinically examined and photographed prior to taking silicone impressions for epoxy resin casts. They considered clinical and photographic data to be supplemental with final wear classification based on visual inspection of casts at ×10 magnification. Each tooth surface was scored, with six grades of erosion severity modelled from Smith and Knight [30]; the index and its criteria are complicated and time consuming.
Conclusion
Review of the literature on indices for tooth wear (or erosion) is confusing; there are too many indices proposed and used, with lack of standardisation in terminology. There are many epidemiological studies reported, but it is difficult to quantify the increases in prevalence reported internationally, as results are not easily comparable. It is doubtful whether any of the indices described are sensitive enough to monitor all but the most severe changes in tooth wear clinically and these cannot be used to measure a rate of wear [19]. It is a challenge to try to develop a simple index that can be used clinically to assess progression of wear.
To date, there is not one ideal index that can be used for epidemiological prevalence studies, clinical staging and monitoring, and it may be necessary to accept that one simple index does not yet exist to meet all requirements of both clinical and research teams. There should, however, be an aim for indices that can be relevant to both fields and can be used internationally in order to strengthen knowledge of dental erosion.
|
[
"tooth wear",
"tooth wear indices"
] |
[
"P",
"P"
] |
FEBS_Lett-1-5-1942071
|
Phosphoinositide-dependent protein kinase-1 (PDK1)-independent activation of the protein kinase C substrate, protein kinase D
|
Phosphoinoisitide dependent kinase l (PDK1) is proposed to phosphorylate a key threonine residue within the catalytic domain of the protein kinase C (PKC) superfamily that controls the stability and catalytic competence of these kinases. Hence, in PDK1-null embryonic stem cells intracellular levels of PKCα, PKCβ1, PKCγ, and PKCε are strikingly reduced. Although PDK1-null cells have reduced endogenous PKC levels they are not completely devoid of PKCs and the integrity of downstream PKC effector pathways in the absence of PDK1 has not been determined. In the present report, the PDK1 requirement for controlling the phosphorylation and activity of a well characterised substrate for PKCs, the serine kinase protein kinase D, has been examined. The data show that in embryonic stem cells and thymocytes loss of PDK1 does not prevent PKC-mediated phosphorylation and activation of protein kinase D. These results reveal that loss of PDK1 does not functionally inactivate all PKC-mediated signal transduction.
1
Introduction
Phosphoinoisitide dependent kinase l (PDK1) phosphorylates a key threonine residue within the catalytic domain of a number of AGC family kinases, including diacylglycerol-regulated kinases of the protein kinase C (PKC) superfamily and phosphatidyl inositol-3 kinase (PI3K)-controlled serine kinases such as Akt/PKB and the 70-kilodalton ribosomal S6 kinase 1 (S6K1) [1]. PDK1-mediated phosphorylation of many of these AGC kinases is essential for their activation. Accordingly, deletion of PDK1 by homologous recombination removes one of the rate-limiting regulators of multiple serine kinases and causes embryonic lethality [2]. To circumvent this prenatal death and explore the role of PDK1 in different tissues, mice expressing PDK1 alleles flanked with the loxP Cre excision sequence (PDK1flΔneo/flΔneo mice) have been generated. These were first used to achieve muscle specific deletion of PDK1, producing mice that die early with severe defects in cardiac function [3]. The importance of PDK1 for T lymphocyte biology has also been explored by looking at T cell development following T lineage specific deletion of PDK1. These studies showed that deletion of PDK1 in T cell progenitors in the thymus results in a block in T cell differentiation, indicating that PDK1 has essential functions in T lymphocyte development [4].
The importance of PDK1 for T cell development in the thymus focuses attention on the identification of the PDK1 substrates directly involved in mediating the unique and non-redundant functions of this kinase in vivo. In this context, studies of mouse embryonic stem cells in which both PDK1 alleles have been deleted have shown that PDK1 is necessary for T-loop phosphorylation, and catalytic activity, of the AGC kinases Akt/PKB, S6K1 and p90 ribosomal S6 kinase (p90RSK) [5]. However, PDK1 is not required for the phosphorylation and activation of other AGC kinases such as PKA or AMPK [5].
One very important subfamily of AGC kinases that can be regulated by PDK1 are the protein kinase C enzymes. There are multiple isoforms of PKC: the classical PKCs (α, βI, βII and γ) which are regulated by calcium, diacylglycerol (DAG) and phospholipids; novel PKCs (δ, ε, η and θ) which are regulated by DAG and phospholipids; and the atypical PKCs (ζ and λ) which are insensitive to both calcium and DAG. The catalytic domain of PKCs have a conserved T-loop motif that is a substrate for PDK1, and it has been proposed that PDK1 is responsible for T-loop phosphorylation of the various PKC isoforms [6–8]. PDK1-mediated T-loop phosphorylation of PKCs is proposed to stabilize protein expression by acting as a priming signal that enables PKCs to achieve catalytic maturity [6,7]. Here, PDK1 docks to a hydrophobic motif in the C-terminus of newly synthesised PKC enzymes, enabling it to phosphorylate their T-loop sites. Following T-loop phosphorylation, PDK1 disengages enabling PKCs to autophosphorylate their C-terminal hydrophobic motifs and achieve catalytic maturity [9–11]. Once mature, PKCs remain inactive due to autoinhibition from their pseudosubstrate motifs and are only activated following the generation of second messengers, such as diacylglycerol and calcium.
The release of PDK1 from PKCs is rate limiting and so PDK1 is responsible for controlling the amount of mature/catalytically competent PKCs expressed in cells [11]. Hence, in the absence of PDK1, intracellular levels of PKCα, PKCβ1, PKCγ, and PKCε are strikingly reduced [6,12]. PKCξ levels are similar in wild-type and PDK1-null embryonic stem cell (ES cells) although PKCξ is not phosphorylated on its T-loop Thr410 site in the absence of PDK1 [6,13]. PKCδ expression is also reduced in PDK1-null cells [6,12] although there was still phosphorylation of the PKCδ T-loop site Thr505 in the residual pool of PKCδ still present in PDK1-null cells [13]. Moreover, there is some evidence in the literature that T-loop phosphorylation is not absolutely critical for PKCδ catalytic activity, since a bacterially-expressed Thr505A PKCδ mutant protein retains some activity in vitro [14]. Of note, a glutamic acid residue at position 500 within the PKCδ activation loop, which is important for catalytic activity [15], could potentially compensate for the lack of PKCδ T-loop phosphorylation in PDK1-null cells.
These studies support the hypothesis that PDK1 has an important role in regulating PKC phosphorylation and stability. However, it has not been shown unequivocally that the reduced levels of PKCs in PDK1-null cells abrogate PKC-mediated cellular responses. In this context, it has been shown that in the absence of PDK1 there is loss of Akt/PKB and S6K1 T-loop phosphorylation and a corresponding inability of cells to regulate phosphorylation of Akt/PKB or S6K1 substrates [5]. There is thus confidence that Akt/PKB and S6K1 function is defective in PDK1-null cells. This has not been determined for the PKCs because although PDK1-null cells have reduced endogenous PKC levels the integrity of downstream PKC effector pathways in PDK1-null cells has not been determined.
One well characterised target for novel PKCs is the serine kinase protein kinase D (PKD). Activation of PKD requires the phosphorylation of two key serine residues within the catalytic domain [16,17]. These residues are absolutely conserved through evolution from C. elegans to man and are substrates for novel PKCs. The activation of PKCs is thus required for PKD activation in a variety of cell lineages in response to a number of different stimuli [16,17]. For example, it has been demonstrated that PKD activity in lymphocytes is regulated by PKC-mediated phosphorylation of its catalytic domain [18,19] and phosphorylation of the PKD catalytic domain is reduced in PKCβ-deficient spleenocytes [20]. It would therefore be expected that PKD would remain inactive if PKCs cannot achieve catalytic maturity in PDK1-null cells. In support of this hypothesis, previous work has established that co-expression of PDK1 and the novel PKCε isoform, together with PKD, enhances PKD activity [21]. The hypothesis that PKD activation is dependent on PDK1 has not been fully investigated and is important to address as it is one way to assess whether there is residual functional PKC activity in PDK1-null cells. Accordingly, the objective of the present study was to examine whether loss of PDK1 results in loss of the PKC signalling pathway that regulates PKD.
2
Results and discussion
2.1
Activation of protein kinase D in PDK1−/− thymocytes
To investigate the impact of PDK1 loss on protein kinase D activation in PDK1 deficient T lymphocytes, mice expressing PDK1 alleles flanked with the loxP Cre excision sequence (PDK1flΔneo/flΔneo) were bred with transgenic mice expressing Cre recombinase under the control of the proximal p56lck promoter which induces expression of Cre in T cell progenitors in the thymus. It has been shown previously that in LckCre+PDK1flΔneo/flΔneo mice PDK1 is deleted in pre-T cells, preventing normal T cell development [4]. To explore the role of PDK1 in the regulation of PKD activity, wild-type or PDK1−/− pre-T cells were isolated and left unstimulated, stimulated with a crosslinking α-CD3 antibody (to activate preTCR signalling) or were stimulated with the phorbol ester PdBu (a diacylglycerol mimetic which activates classical/novel PKCs), as indicated. PKD catalytic activity was monitored using an antisera that recognises PKD molecules that are autophosphorylated on Ser916 [22]. The data (Fig. 1) show that in quiescent wild-type thymocytes PKD is not autophosphorylated on Ser916 whereas α-CD3 or PdBu stimulation strongly induced PKD catalytic activity, as shown by increased Ser916 autophosphorylation. The results also show that both α-CD3 and PdBu treatment induce strong PKD Ser916 autophosphorylation in PDK1−/− pre-T cells (Fig. 1). PKD activation is dependent on PKC-mediated phosphorylation of two key serine residues (Ser744/S748) within the PKD catalytic domain and as shown in Fig. 1, PKD is phosphorylated on Ser744/S748 following α-CD3 or PdBu activation in both wild-type and PDK1−/− pre-T cells.
The phosphorylation and activation of PKD in PDK1−/− pre-T cells argues that these cells still contain a functional, catalytically competent PKC(s) that is able to respond to diacylglycerol/phorbol ester signals. RNAi experiments have demonstrated that PKCε in Swiss 3T3 cells or PKCδ in HeLa cells are the PKC isoforms responsible for serine phosphorylation (Ser744/S748) within the PKD activation loop [23,24]. PKCδ has been shown to be residually phosphorylated on its T-loop Thr505 site in PDK1-null ES cells [13]. We therefore examined the phosphorylation status of the PKCδ T-loop site, Thr505, in PDK1−/− pre-T cells. The data in Fig. 1 show basal phosphorylation of PKCδ Thr505 in wild-type pre-T cells and also show that significant phosphorylation of PKCδ on Thr505 in retained in PDK1−/− pre-T cells. In contrast, as previously demonstrated, PDK1−/− pre-T cells have lost activity of the PKB/S6K1 signalling pathway that controls the phosphorylation of the ribosomal S6 subunit on Ser235/236 [4,25]. The data in Fig. 1 also explore the impact of PDK1 loss on the T-loop phosphorylation of p90RSK1 and PRK1. The data show Western blot analysis of wild type and PDK1 null pre-T cells with a phospho-specific antibody against Ser227 in p90RSK1 and show that this site is basally phosphorylated in wild-type pre-T cells but is not phosphorylated in PDK1-null pre-T cells. Similarly, basal phosphorylation of PRK1 on its Thr774 T-loop site is substantially reduced in PDK1-null pre-T cells. Thus, PDK1-null pre-T cells show defective activation of some, but not all, AGC family kinases: regulation of a PKC–PKD signalling pathway remains intact in these cells.
2.2
Activation of protein kinase D in PDK1−/− ES cells
The experiments in Fig. 1 show that PDK1 is not essential for PKD activation in thymocytes. To probe further the involvement of PDK1 in the regulation of PKC–PKD signalling, experiments in wild-type and PDK1−/− mouse embryonic stem cells were performed. Western blot analysis of wild-type and PDK1−/− ES cells with a PKCδ Thr505 phosphospecific antisera confirm that PKCδ Thr505 phosphorylation is detectable, albeit slightly reduced, in PDK1-null versus wild-type ES cells (Fig. 2A–C). To probe PKD activity, cells were left unstimulated or activated by exposure to serum or phorbol ester as indicated and PKD catalytic activity monitored using the pSer916 antisera. In quiescent, wild-type ES cells, PKD was not autophosphorylated on Ser916 but both serum and PdBu stimulation strongly and rapidly induced PKD activity as measured by increased Ser916 autophosphorylation (Fig. 2A and B). Serum or PdBu stimulation also induced PKD activity in the PDK1−/− ES cells (Fig. 2A and B) whereas p90RSK1 T-loop phosphorylation (Ser227) was basally detectable in wild-type ES cells but not in the PDK1−/− ES cells (Fig. 3). S6K1 signalling was also abrogated, as expected, in the PDK1−/− ES cells, as shown by the lack of phosphorylation of its substrate, the S6 ribosomal protein subunit (Fig. 2B).
The ability of serum and phorbol esters to activate PKD in PDK1-null ES cells is an indication that there is residual, catalytically competent PKC in these cells. PKC-mediated activation of PKD can be prevented by the PKC inhibitor GF109203X [17]. Therefore, the effects of GF109203X, on PKD activation in wild-type or PDK1−/− ES cells was examined. The data (Fig. 3) show that PKD was phosphorylated on Ser916 in activated PDK1+/+ and PDK1−/− cells but that PKD autophosphorylation on Ser916 was prevented by prior treatment of both cell types with GF109203X.
3
Concluding remarks
Loss of PDK1 in ES cells causes a reduction in levels of PKCα, βI, γ, δ, and ε [6,12] but it was not resolved whether the residual PKC in these cells is functional. As discussed above, there is a large body of evidence that PKD activation is mediated by novel PKCs such as PKCδ and ε and it has previously been proposed that PDK1 is required for activation of PKD [21], reflecting the requirement of PDK1 for phosphorylation of the key priming T-loop phosphorylation sites in PKCs. However, the present study shows that the hypothesis that PKD activation is dependent on PDK1 is incorrect. The reason this model fails is that there is not a complete loss of PKC activity in the absence of PDK1. We show that in PDK1-null pre-T cells PKCδ is still phosphorylated on its T-loop site Thr505; hence activation of pre-T cells with phorbol esters can still induce phosphorylation of PKD on its PKC substrate sequences Ser744/S748, inducing PKD activation and subsequent autophosphorylation on Ser916. In addition, the present data show that PKD can also be activated by phorbol esters or serum stimulation of PDK1-null ES cells in a PKC-regulated manner. Thus, the present results showing activation of PKD in PDK1−/− cells is evidence that at least one PKC-mediated signalling pathway is intact in the absence of PDK1.
While loss of PDK1 does not prevent the activation of PKD, the issue of whether the efficiency of PKD activation is affected in PDK1-null cells requires further experimentation. While robust activation of PKD signalling is observed in PDK1-null ES cells and pre-T cells in response to phorbol esters or CD3 crosslinking we cannot exclude the possibility that PKD activation may be compromised in PDK1-null mature peripheral T cells during physiological peptide:MHC stimulation, however these experiments await the development of genetic models to test this.
In summary, PDK1 is essential for pre-T cell development and its loss is associated with loss of PKB/S6K1-mediated signal transduction [4,25]. The present results now show loss of p90RSK and PRK1 T-loop phosphorylation in PDK1−/− pre-T cells but reveal that the PKCδ T-loop site Thr505 is still phosphorylated in PDK1-null pre-T cells. It has been proposed that the Thr505 T-loop site in PKCδ (as well as the corresponding T-loop site in PKCθ, [26]) can be autophosphorylated, which may explain why PKCδ Thr505 is still phosphorylated in the PDK1-null cells. Alternatively the present results could argue for the existence of another PKCδ priming kinase in developing thymocytes. Whatever the molecular basis for the continued phosphorylation of PKCδ following the loss of PDK1, the present results show that there is still residual functional PKC activity once PDK1 is lost. Hence, the importance of PDK1 for T cell development must reflect the requirement for PDK1 mediated phosphorylation of other members of the AGC serine kinase family, such as Akt/PKB, S6K1 or p90RSK, since phosphorylation and activity of these kinases is strictly dependent on PDK1 expression.
4
Materials and methods
4.1
Cell preparation and stimulation
PDK1+/+ and PDK1−/− murine embryonic stem cells [5] were grown on gelatinised tissue culture plates in Knockout DMEM supplemented with 10% knockout serum replacement, 15% foetal bovine serum (Hyclone), 0.1 mM non-essential amino acids, antibiotics (100 U/ml penicillin G, 100 μg/ml streptomycin and 1 μg/ml ciprofloxacin [Bayer Pharmaceuticals]), 2 mM l-glutamine, 1 mM sodium pyruvate, 0.1 mM β-mercaptoethanol and 25 ng/ml murine leukaemia inhibitory factor (LIF).
4.2
Mice
All mice used were between 5 and 7 weeks of age. Mice were bred and maintained in specific pathogen-free conditions and animal experimentation was approved by Home Office project license PPL60/3116. T-PDK1−/− mice were generated by crossing mice with PDK1 floxed alleles (PDK1flΔneo/flΔneo) with mice expressing Cre recombinase under the control of the p56lck promoter to ablate expression of PDK1 in T cells [2,4]. Control mice used for analyses of T-PDK1−/− mice were age-matched phenotypically normal littermates.
4.3
Purification of thymocyte subpopulations
Antibodies conjugated to fluorescein isothiocyanate (FITC), phycoerythrin (PE), allophycocyanin (APC) and biotin were obtained from PharMingen. Cells were stained for surface expression of the following markers: CD4, CD8, CD25, CD44, CD3ε, γδ, B220, and Thy1. The DN4 subpopulation were defined by their cell surface marker expression and subsequently sorted using a Vantage cell sorter (Becton Dickinson) or AutoMacs magnetic cell sorter (Miltenyi Biotec).
4.4
Cell lysis and Western blot analysis
Cells were lysed for 10 min at 4 °C (20 × 106 cells/ml) using buffer: 100 mM HEPES, pH 7.4, 150 mM NaCl, 1% NP40, 20 mM sodium fluoride, 20 mM iodoacetamide, 2 μM EDTA, 1 mM sodium orthovanadate, 2 μg/ml pepstatin A, 2 μg/ml leupeptin, 2 μg/ml chymostatin, 2 μg/ml antipain, 40 mM β-glycerophosphate and 1 mM phenylmethylsulfonylflouride. Soluble proteins were concentrated by precipitation with 1.5 vol. of acetone and separated by 8% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE), transferred to polyvinylidene difluoride membranes and detected by Western blot analysis with the indicated antibodies (Cell Signalling Technology).
|
[
"pdk1",
"pkc",
"pkd",
"es cell, embryonic stem cell",
"p90rsk, p90 ribosomal s6 kinase",
"pdk1, 3′-phosphoinositide-dependent protein kinase-1",
"pkc, protein kinase c",
"pkd, protein kinase d",
"s6k1, 70-kilodalton ribosomal s6 kinase 1"
] |
[
"P",
"P",
"P",
"R",
"R",
"M",
"R",
"R",
"R"
] |
Environ_Health_Perspect-116-2-2235228
|
Testicular Dysgenesis Syndrome and the Estrogen Hypothesis: A Quantitative Meta-Analysis
|
Background Male reproductive tract abnormalities such as hypospadias and cryptorchidism, and testicular cancer have been proposed to comprise a common syndrome together with impaired spermatogenesis with a common etiology resulting from the disruption of gonadal development during fetal life, the testicular dysgenesis syndrome (TDS). The hypothesis that in utero exposure to estrogenic agents could induce these disorders was first proposed in 1993. The only quantitative summary estimate of the association between prenatal exposure to estrogenic agents and testicular cancer was published over 10 years ago, and other systematic reviews of the association between estrogenic compounds, other than the potent pharmaceutical estrogen diethylstilbestrol (DES), and TDS end points have remained inconclusive.
Impaired spermatogenesis, male reproductive tract abnormalities such as hypospadias and cryptorchidism, and testicular cancer have been proposed to comprise a common underlying syndrome with a common aetiology resulting from the disruption of embryonic programming and gonadal development during fetal life, termed the testicular dysgenesis syndrome (TDS) (Sharpe and Skakkebaek 2003; Skakkebaek et al. 2001). A hormonal etiology most likely underlies this syndrome, although it is believed to have more than one cause, possibly including other than endocrine disruption. Some common causes of endocrine disruption include infection, diet and body weight, lifestyle, genetics, and environmental exposure, but endocrine-disrupting chemicals (EDCs), particularly those with estrogen-like properties, have received the most scientific attention.
The synthetic estrogenic drug diethylstilbestrol (DES) was prescribed to more than 5 million pregnant women from the late 1940s to the early 1970s to prevent abortions and pregnancy-related complications (Palmlund et al. 1993). Evidence later showed that maternal ingestion of DES during early pregnancy increased the risk of vaginal clear cell adenocarcinoma in female offspring (Herbst et al. 1971) and resulted in an increased incidence of malformations of the testes, the development of epididymal cysts, and impaired sperm quality in male offspring (Bibbo et al. 1977). During pregnancy, maternal estrogen levels are significantly elevated. However, more than 90% of maternal endogenous estrogens are effectively sequestered via binding to sex hormone binding globulin (SHBG), and thus the fetus is relatively protected (Joffe 2001; Vidaeff and Sever 2005). On the other hand, DES and ethinylestradiol do not bind well to SHBG, having a higher biopotency if ingested (Sharpe and Skakkebaek 2003; Vidaeff and Sever 2005). Additionally, transgenerational exposure is also possible when lipophilic xenoestrogens are mobilized during pregnancy and lactation (Colborn et al. 1993).
Previous systematic reviews of studies in which pregnant women were exposed to estrogens other than DES have failed to find evidence of an increased risk of urogenital abnormalities in the male offspring (Raman-Wilms et al. 1995; Storgaard et al. 2006; Toppari et al. 1996; Vidaeff and Sever 2005), and have raised the possibility that nonestrogenic or atypical estrogenic effects of DES exposure in utero induce male reproductive abnormalities. However, none of the effects of DES exposure on either male or female offspring of exposed wild-type pregnant mice were induced when administered to ERKO (ER-α knockout) mice (Couse et al. 2001), strongly suggesting an ER-α–mediated mechanism. There is, however, a body of experimental data that is consistent with an effect of antiandrogenic industrial chemicals on male sexual differentiation (Gray et al. 1999, 2000). Moreover, mechanisms other than endocrine disruption may be involved in testicular toxicity; for example, the nematocide dibromochloropropane, an alkylating agent, is one of the most potent known testicular toxins in adults (Joffe 2001). In this review we focus on the estrogen hypothesis of TDS.
Although several systematic reviews of the literature on the association between estrogenic agents and the disorders thought to belong to the TDS have been published, they are predominantly qualitative and the only quantitative summary estimate of the association between prenatal exposure to estrogenic agents and testicular cancer was published over 10 years ago (Toppari et al. 1996). The primary objective of a quantitative meta-analysis is to combine the results of previous studies examining a specific research question to arrive at a summary conclusion about a body of research. It has been found particularly useful when individual studies are too small to yield a valid conclusion, but it cannot, however, correct for bias and confounding. When applied to observational studies, subset analysis can be a useful tool to explore the reasons for discrepancies among the results of different studies.
The objectives of this research were therefore to carry out a quantitative meta-analysis of the association between three of the end points related to TDS and prenatal exposure to estrogenic agents that would account for both the size and quality of the studies included and yield updated summary estimates in light of the body of research carried out since the formulation of the estrogen hypothesis. Inclusion in this analysis was based on mechanistic criteria, and the plausibility of an ER-α–mediated mode of action was specifically explored. Moreover, subset analysis has been applied to categories of compounds with estrogenic potencies differing by several orders of magnitude in an attempt to detect the existence of any potency–response trend. Most of the studies of sperm quantity or quality have been concerned with time trends rather than etiology, and this end point was not considered further here.
Material and Methods
Identification and selection of literature
A computerized search was conducted using the databases PubMed (National Center for Biotechnology Information 2007) and Web of Science (ISI Web of Knowledge 2007) for the period 1970 to April 2007. The general search keywords were “estrogen,” “risk,” “dose,” and either “hypospadias,” “cryptorchidism,” or “testicular cancer.” A preliminary identification was performed by screening the titles and, if relevant, the abstracts of retrieved literature. The next stage was to check the citations and references of selected studies. This was an iterative process, repeated until no new study could be identified. A set of both inclusion and exclusion criteria was defined, and all relevant literature was then checked for eligibility. The inclusion criteria considered were a) study design, namely, either a case–control, cohort, or clinical trial; b) written in English; c) exposure to one or a mixture of known estrogenic compounds; and d) sufficient data reported to be used in meta-analysis.
The following exclusion criteria were used:
Exposure to a group of compounds (suspected endocrine disruptors) for which mode of action was unspecified, for example, pesticides.
Studies of exposure to phytoestrogens. Some phytoestrogens have been found to have a greater binding affinity for ER-β than for ER-α and can result in agonistic or antagonistic effects (Mueller et al. 2004).
Studies of maternal endogenous hormones.
Studies of the same cohort as this would bias the results towards the particular studies.
Incomplete data.
Data extraction and quality rating
In addition to the number of exposed and nonexposed cases and controls, and risk ratios (RRs) with their confidence intervals (CIs), information regarding the study design, estrogenic agent, geographic location of the study, and year of publication were extracted from the selected literature to allow subset analysis to be carried out. When more than one RR was reported, the following priorities were set for choice:
Adjusted RRs were used, except when the study provided only unadjusted estimates.
When multiple estimates were given, the RR estimator on which the authors had relied for their assessment of causal association was used.
Overall RRs were chosen instead of those derived from further stratifications. If an overall estimate was not provided, the RRs of the maximum duration of exposure or the maximum exposure concentration were chosen.
Several aspects of the quality of each study were also recorded according to a rating scheme adapted from those previously described (Altman 1991; Rushton 2000). Every criterion was assessed on a scale of 0 to 2, 0 suggesting that it was not present, 1 when it was unclear, and 2 when that criterion was satisfied. A maximum score of 50 and 52 could be assigned for retrospective (case–control) and prospective (cohort and clinical trials) studies, respectively. This enabled a quality sensitivity analysis to be performed to check the influence of studies with low quality on the pooled estimate.
Data analysis
Graphical representation
The RRs and CIs were plotted against the year of publication to determine whether any positive or negative trends in reporting RRs had occurred over time. Similarly, quality scores were plotted against the year of publication to investigate whether the quality of studies improved over time. To assess publication bias, a funnel plot (SE vs. RR) was produced based on the assumption that smaller studies are less precise in their RRs and thus have less weight and larger SE and should scatter more widely at the lower end of the graph, whereas larger studies will tend to be closer together (Sterne et al. 2001). Forest plots present the RRs against the reference of the study and help check homogeneity visually.
Statistical pooling
Pooled estimates and 95% CIs were calculated using both a fixed-effects model (Mantel–Haenszel method) and a random-effects model (DerSimonian–Laird method), allowing evaluation of the dependence of the conclusions of the analysis on the model assumptions. A summary estimate is considered statistically significant at the 0.05 level if its CI does not include unity.
The Mantel–Haenszel pooled effect estimate was used in a chi-square statistical test of homogeneity to assess the between-study variance. The magnitude of the test statistics depends on the weight of each study. When the number of studies is low or the studies themselves are small, the test statistic Q tends to be small. Tests of heterogeneity in meta-analyses are generally low in their power to reject the null hypothesis of homogeneity. For this reason, the chi-square statistical test of homogeneity was carried out at both 0.05 and 0.1 significance levels. Additionally, pooled estimates calculated using fixed effect and random effect models differ only if there is lack of homogeneity between studies. The estimates obtained by both methods were therefore compared to better assess potential heterogeneity between studies, in which case a single summary estimate of effect may be considered inappropriate.
Subset and sensitivity analyses
To investigate potential sources of heterogeneity between studies, we performed subset analyses for the study design, estrogenic agent, and geographic location.
Some studies exploring the influence of hormonal treatment during pregnancy did not specify the type of hormone. From what is known of the hormonal treatment of common conditions occurring during pregnancy, it was deemed reasonable to assume that they would have been likely to include estrogens, and these studies were included in the analysis. The validity of this assumption was tested by applying stricter criteria and calculating a summary estimate of effect excluding any study in which the hormone used had not been specified. Further sensitivity analysis was performed by excluding low-quality studies and extremes (exclusion of the studies with the largest and smallest RR estimators and exclusion of the studies with the largest and smallest weights) to verify that either the quality of the studies or one particular study did not have an excessive influence on the pooled estimate.
Results
A total of 50 studies were identified for the association between in utero exposure to estrogenic agents and hypospadias and/or cryptorchidism, including 16 that had not been included in previous systematic reviews. Sixteen studies, of which 8 were new studies, were included in the calculation of a summary estimate of effect for either or both end points (Table 1). Studies predating the formulation of the TDS hypothesis often were designed to explore the association of in utero exposure to a range of pharmaceuticals with birth malformations. Other than 2 recent studies for which pesticide exposure was determined by chemical analysis of specific compounds, assessment of exposure to pesticides is generally derived from the occupation of the mother and specific agents are not identified.
Of the 12 studies identified for the association with testicular cancer, only 3 were excluded from the calculation of a summary estimate of effect (Table 2).
Hypospadias
The data from studies included in the meta-analysis for hypospadias are summarized in Table 3. Three extreme values, two greater than and one lower than unity, can be identified visually from the forest plot of the RRs and their CIs (Figure 1). These extremes correspond to studies with larger SEs, and the shape given to the funnel plot (Figure 2) by those smaller positive studies would be consistent with publication bias. These two extreme positive risk ratios were, however, reported after what is commonly referred to as “third-generation exposure” to DES, when the mother herself had been exposed to DES prenatally. It was recognized that the inclusion of such studies in the meta-analysis could have introduced heterogeneity, and the influence of this choice was investigated in the subset analysis. Plots of the quality score and RRs versus year of publication did not suggest any significant trends in quality of the studies or estimates of effect over time (not shown).
The pooled estimates of effect by both the Mantel–Haenszel and DerSimonian–Laird methods are very close to unity, and no relationship between in utero exposure to estrogenic agents and hypospadias could be detected (Table 4). None of the chi-square tests allowed the rejection of the null hypothesis of homogeneity between the studies at the 0.05 or 0.1 level of statistical significance. The subsets of studies in which exposure to DES and pharmaceutical estrogens were investigated yielded statistically significant risk ratios with both models, although the modest discrepancy between the fixed-effects and random-effects estimates suggests heterogeneity. Summary estimates for the latter subset were no longer significant when studies that included DES exposure were excluded. Although these results were based on four studies that all addressed in utero exposure to oral contraceptives, some heterogeneity between studies remained. Excluding the studies of third-generation exposure to DES, values for the summary estimate of effect were found to be 1.33 (95% CI, 0.63–2.83) by the Mantel–Haenszel method and 1.31 (95% CI, 0.52–3.26) by the DerSimonian–Laird method, a very modest and nonsignificant increase in risk. Excluding third-generation exposure from the DES subset yielded estimates of 2.02 (95% CI, 1.12–3.65) by the Mantel–Haenszel method and 2.00 (95% CI, 0.97–4.15) by the DerSimonian–Laird method, on the basis of two studies investigating exposure to any estrogenic drug during the first trimester of pregnancy. The difference between the results obtained by the two models for studies of third-generation exposure to DES was reduced only slightly by excluding the study by Klip et al. (2002); the Mantel–Haenszel method yielded an estimate of 2.46 (95% CI, 0.91–6.67) and the DerSimonian–Laird method of 2.18 (95% CI, 0.64–7.46). The latter study’s cohort had been recruited in a fertility clinic, and whether results obtained with subfertile women are generalizable to all women exposed to DES in utero has been questioned (Hernandez-Diaz 2002).
Although the equality of the results obtained by both methods for the environmental estrogens subset suggests those results are robust, the influence of the weight of the study by Vrijheid et al. (2003) cannot be underestimated, as shown by the sensitivity analysis. Exclusion of this study from the analysis yielded a statistically significant Mantel–Haenszel estimate but a lower and not statistically significant DerSimonian–Laird estimate, revealing heterogeneity. A statistically significant estimate was obtained for prospective studies by the Mantel–Haenszel method, but the wide difference with the estimate using the random effect model was suggestive of heterogeneity. Geographic subsets point to a higher risk in Latin America, although the pooled estimates for this location were based on only two studies and did not reach statistical significance.
In addition to the results of the sensitivity analysis presented in Table 4, a pooled estimate of effect was calculated when a stricter inclusion criterion was applied, namely, excluding results from the study by Monteleone-Neto et al. (1981). This had little influence on the overall result, generating summary estimates of 0.97 (95% CI, 0.83–1.13) for the fixed effect model or 0.93 (95% CI, 0.80–1.09) for the random effect model.
Cryptorchidism
Data for the six studies included in the meta-analysis for cryptorchidism can be found in Table 5. The results of only two studies significantly differ from unity, as illustrated by the forest plot (Figure 3). The small number of eligible studies renders analysis of the funnel plot and potential for publication bias difficult (Figure 4). The SEs do, however, illustrate well that the studies were all relatively small. No time trends for the estimate of effect or the quality of studies could be detected (not shown).
As presented in Table 6, the pooled estimates of effect by both the Mantel–Haenszel and DerSimonian–Laird methods are marginally superior to unity, and their relative divergence implies there may be sources of heterogeneity. Chi-square tests did not, however, detect that any of the subsets were significantly heterogeneous. Excluding studies in which DES exposure was examined, either exclusively or along with hormonal therapeutics, yielded summary estimates consistent with no relationship. Statistical pooling of the studies including DES exposure generated a statistically significant estimate by the Mantel–Haenszel method, suggesting a doubling of the risk of cryptorchidism after in utero exposure to DES. The same estimate by the DerSimonian–Laird method did not, however, reach statistical significance and the difference relative to the fixed effect model is indicative of heterogeneity. The heterogeneity introduced by the DES subset of studies can again be observed by comparing the results obtained for all pharmaceutical estrogens with those obtained by pooling the two studies of accidental use of oral contraceptives during pregnancy. Study design also appeared to be a source of heterogeneity. If case–control studies are prone to recall bias, this subset also included the study with the highest estimate, itself a source of heterogeneity, as shown by the sensitivity analysis. Excluding the study by Depue (1988) reduced the difference between estimates by both models, the Mantel–Haenszel estimate then calculated as 1.29 (95% CI, 0.87–1.91) and that by the DerSimonian–Laird method as 1.23 (95% CI, 0.81–1.86). This was also observed for the American subset of studies. When the Depue (1988) study is omitted, the Mantel–Haenszel method yielded a no longer statistically significant estimate of 1.34 (95% CI, 0.84–2.14) and the DerSimonian–Laird method an estimate of 1.27 (95% CI, 0.72–2.23).
Applying a stricter exclusion criterion to studies examining hormonal treatment did not affect which studies were included in the meta-analysis of cryptorchidism. The study with the highest weight appears to lower the overall estimates, whereas increasing quality seems to reduce heterogeneity and lower the estimate of effect toward unity. These variations did not, however, influence the overall conclusion that aside from the DES studies subset, summary estimates did not detect any association between in utero exposure to estrogenic substances and cryptorchidism.
Testicular cancer
Nine studies were included in the meta-analysis of testicular cancer and the data used are summarized in Table 7. Of these, 4 had not been included in the summary estimate previously calculated by Toppari et al. (1996). The lack of homogeneity between studies is evident from the forest plot (Figure 5). Further, the funnel plot (Figure 6) also illustrates the relatively small size of the included studies. Although a positive trend over time was found for the quality of the included studies (Figure 7), no significant time trend could be detected for the effect size (not shown).
Both the fixed and random effect models yield a statistically significant estimate; however, the discrepancy between the two results is suggestive of heterogeneity despite the result from the chi-square test (Table 8). Conversely, the subset analysis was limited by the similarity of the question addressed by the studies included. Eight of the nine studies were interested in hormonal exposure and were conducted in the United States. Despite this, statistically significant heterogeneity between the studies was detected at the 0.1 level. Pooling the two studies examining DES exposure specifically produced a raised but statistically nonsignificant result. Despite the unexplained heterogeneity, all estimates that were calculated point to a doubling of the risk of developing testicular cancer after exposure to estrogenic agents in utero. The work on chlorinated biphenyls (PCBs) by Hardell et al. (2004) was the only study examining environmental estrogens. Its size was relatively small, and it did not detect such an effect.
Applying a stricter exclusion criterion to studies examining hormonal treatment excluded four studies from the meta-analysis; namely, Brown et al. (1986), Gershman and Stolley (1988), Henderson et al. (1979), and Weir et al. (2000). This resulted in a slightly lower Mantel–Haenszel estimate of 1.98 (95% CI, 1.23–3.18) and if the DerSimonian–Laird estimate remained equal to 1.59, because of the wider confidence interval (95% CI, 0.93–2.69), statistical significance was no longer achieved. The sensitivity analysis is consistent with some heterogeneity between the studies, the estimates obtained being relatively sensitive to the exclusion of particular studies varying above and below a risk estimate of 2. The quality of the studies seemed to explain at least some of this heterogeneity.
Discussion
While it is clear that hypospadias, cryptorchidism, and testicular cancer are all positively associated with prenatal exposure to DES, this meta-analysis was unable to produce evidence that such effects were associated with environmental estrogens or even accidental use of oral contraceptives during pregnancy. This is consistent with the results obtained in earlier meta-analyses (Raman-Wilms et al. 1995; Toppari et al. 1996).
The main limitations of meta-analysis are a) the susceptibility of its summary results to publication bias, b) the influence of the quality of studies, c) the possibility of including multiple results from the same study, and finally, d) heterogeneity between studies that could lead to invalid conclusions. The methodology employed in this present review attempts to address these issues. Additionally, the importance of carrying out and reporting a sensitivity analysis was illustrated by the case of hypospadias where the weight attributed to one particularly large study had a nonnegligible influence on the results. In this particular case, the study by Vrijheid et al. (2003) inferred exposure to phthalates from registry data about occupation, and although such an approach can allow the analysis of a great number of cases, assessment of exposure is much more likely to be prone to confounding. The number of studies included in meta-analyses lies typically between 5 and 15, and the results presented here also fall within this range. The size of the homogeneity test statistic depends on both the number and size of individual studies. The funnel plots offer a good visual representation of the precision and size of individual studies, and it is clear that most studies published on the association between estrogenic agents and the probable end points of a TDS were found to be relatively small. The chi-square tests had, therefore, a relatively low power to detect heterogeneity. However, in the absence of statistical heterogeneity, the results of the fixed effect and random effect models should be virtually identical, and the comparison of results obtained by applying both the Mantel–Haenszel and DerSimonian–Laird models enabled the exploration of sources of heterogeneity despite this low statistical power.
If the quality of the studies was found to explain some of the heterogeneity observed, particularly in the case of testicular cancer, the remaining heterogeneity could not be explained solely by the fact that environmental, and therefore generally much weaker, estrogens were included in the analysis. The systematic review of published literature yielded relatively few studies examining the association of male urogenital abnormalities or testicular cancer with environmental estrogens specifically; a number of studies concerned with an association with broad categories of putative endocrine disruptor, most often pesticides, were excluded from the meta-analyses. This illustrates the difficulties associated with assessment of exposure, pesticide exposure often being inferred from parental occupation rather than direct measurement. Furthermore, there is increasing evidence that, in accordance with pharmacokinetic theory, the effects of xenobiotics acting via the same mechanism can be predicted fairly accurately by concentration addition (Zhu et al. 2006). Accurately accounting for combined exposure or adjusting for the confounding introduced by environmental exposures will probably require the development of mechanism-specific biomarkers of exposure.
When DES is excluded, there is no conclusive evidence of an effect of pharmaceutical estrogens. Exposure to such estrogens is related mainly to the accidental use of oral contraceptives during pregnancy or hormonal pregnancy tests. Such estrogenic pharmaceuticals often are given in combination with progestagens, and it is legitimate to question whether unopposed estrogens would have the same effects as opposed estrogens. This also highlights another difficulty associated with exposure assessment, that of critically sensitive periods of development and the ascertainment of whether exposure took place during a “window” of susceptibility to hormone disruption. Nonetheless, studies in which maternal levels of hormones were measured in the first and third trimester of pregnancy do not support an association with elevated estrogen levels but rather indicate that a lower estrogen/androgen ratio and/or higher levels of α-fetoproteins may be beneficial (McGlynn et al. 2005; Zhang et al. 2005). If in animals both estrogenic and antiandrogenic compounds have been associated with end points consistent with those of human TDS (Fisher 2004; Veeramachaneni 2000), epidemiologic evidence remains elusive. Alternatively, the doubling of the risk estimates of all three effects associated with DES exposure would be consistent with a shared etiology and the TDS hypothesis. It does not constitute conclusive evidence of an estrogenic mode of action, however, as common etiologic factors could be related to the underlying condition for which DES was prescribed. Furthermore hypospadias, cryptorchidism, and testicular cancer have all been found to be associated with low birth weight, suggesting a potential association with an underlying placental defect.
The understanding of the importance of endogenous estrogens in normal adult testicular function is becoming clearer. Their roles during fetal life, however, remain relatively unclear, but those mediated by the ER-αor ER-β have been shown to differ (Habert et al. 2006). Interestingly, DES has been found to have similar affinity for both receptors, whereas estradiol has only a slightly stronger affinity for ER-αcompared with ER-β (Mueller et al. 2004). ER-α has been detected in undifferentiated gonads as early as 10 days postconception in the mouse and found to be localized in the Leydig cells of fetal testis in rodents (Habert et al. 2006). Studies of the expression of ER-α and ER-β in human and nonhuman primates have so far yielded inconsistent results. Gaskell et al. (2003) reported that ER-α could not be detected in human fetal testes between weeks 12–19 of gestation, whereas Shapiro et al. (2005) found that ER-α was apparent by week 12, its levels peaked at 16 weeks before diminishing, and it was localized in Leydig cells. Current research focus has shifted to the role played by testosterone, anti-Müllerian hormone and insulin-like factor 3 produced by the fetal testes during masculinization. In the male rat, exposure to high levels of estrogens has been shown not only to suppress testosterone production but also to downregulate the expression of the androgen receptor protein in reproductive target tissues including the testes, Wolffian duct, and prostate (Sharpe 2006). Further research in this area may help shed light on possible mechanisms of injury or relevance of the rodent model.
The subset analyses did not generate many clues to explain the heterogeneity of the collected data. This is, however, consistent with the wide geographic variability in the incidence of the conditions of interest (Boisen et al. 2004; Richiardi et al. 2004). Interactions between genetic susceptibility and the environment have been the focus of research in this area (Martin et al. 2007), and advances in genomics have allowed the identification of polymorphisms associated with hypospadias, cryptorchidism, and testicular cancer (Beleza-Meireles et al. 2006; Kurahashi et al. 2005; Starr et al. 2005; Yoshida et al. 2005). Such discoveries may, however, give rise to as many questions as they offer to answer. This is well illustrated by the recent identification of the association of a variant of the gene for the ER-α with hypospadias and cryptorchidism in Japanese cohorts (Watanabe et al. 2007; Yoshida et al. 2005) that has now been found to be associated with a decreased incidence of hypospadias in a European cohort (Galan et al. 2007).
Conclusion
The modest increase in risk for all three end points associated with DES exposure is consistent with a shared etiology and the TDS hypothesis, whereas the results of the subset analyses suggest the existence of yet unidentified sources of heterogeneity between studies or within the study populations. Although 10 years of further research on the potential effects of endocrine disruptors on male reproductive health have provided some clues regarding the etiology and mechanism of conditions such as hypospadias, cryptorchidism, and testicular cancer, there is still no conclusive evidence of the role played by environmental estrogens.
|
[
"testicular dysgenesis",
"estrogen",
"meta-analysis",
"hypospadias",
"cryptorchidism",
"testicular cancer",
"diethylstilbestrol",
"endocrine disruption",
"environment",
"oral contraceptives"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Osteoporos_Int-4-1-2277453
|
Effects of risedronate on fracture risk in postmenopausal women with osteopenia
|
Summary This posthoc analysis of four trials studied the efficacy of risedronate to reduce fragility fractures in postmenopausal women with osteopenia (i.e., T-scores between −1 and −2.5). Risedronate reduced the fracture risk by 73% ( = 0.023) in this population of women with low femoral neck bone mass and no prevalent vertebral fractures.
Introduction
Low bone mass, defined by bone mineral density (BMD) values in the osteopenic range, can progress to osteoporosis if left untreated. Additionally, it has recently become apparent that osteopenia is a significant risk factor for fragility fractures in older women. Studies using peripheral and central dual x-ray absorptiometry (DXA) to determine BMD have demonstrated that approximately half of all fragility fractures occur in women with osteopenic BMD [1–5]. As the number of women with low bone mass is projected to increase to 30.4 million in 2020, this condition presents a significant medical problem and socioeconomic burden [6]. Prevalence projections for 2020 are significantly higher than current rates (21.8 million women in 2002) due to expected population growth and aging [7].
Based on BMD testing, the World Health Organization (WHO) clinically defines women with osteopenia as those whose BMD T-score is between 1 and 2.5 standard deviations (SD) below the mean of peak bone mass in healthy, young normal women (i.e., −1 to −2.5 SD); women with osteoporosis include all those at or below −2.5 SD [8–10]. However, these definitions provide a diagnostic categorization of low bone mass rather than an intervention threshold that could be used to determine the need for the initiation of treatment to reduce fracture risk. Physicians are often faced with making treatment decisions for postmenopausal patients with T-scores in the osteopenic range, increasing the need for evidence-based guidance to define populations at high risk for fracture [11].
The National Osteoporosis Foundation (NOF) currently recommends an algorithm for initiation of therapy to reduce the risk of fracture [12]. According to this algorithm, treatment should be initiated in osteopenic women with T-scores <−2.0 by central DXA with no risk factors and in women with T-scores <−1.5 by central DXA with one or more risk factors [9, 10, 12, 13].
Despite the existence of NOF guidelines for the treatment of postmenopausal women with low bone mass, limited clinical data exist as a foundation for supporting therapeutic intervention in this group. Although it is possible to identify osteopenic women at higher or lower short-term risk for fracture [11], there are limited data regarding the effectiveness of osteoporosis treatments to lower fracture risk in these women. This is in part due to the inherent difficulty in collecting fracture data from osteoporosis trials, which usually exclude women with higher T-scores who may fall into the osteopenic range. The assembly of large data sets is, therefore, difficult and the feasibility of prospective osteopenic fracture trials is limited due the lower incidence of fracture in this population compared with patients with osteoporosis [14]. The use of post hoc analysis is one way to overcome some of these obstacles and provide information about effectiveness of treatments in non-osteoporotic populations.
We have analyzed data from four Phase III risedronate trials: BMD Multinational (BMD-MN) [15], BMD-North America (NA) [16], Vertebral Efficacy with Risedronate Therapy-Multinational (VERT-MN) [17], and Vertebral Efficacy with Risedronate Therapy-North America (VERT-NA) [18] in which efficacy and safety of risedronate in the prevention and treatment of postmenopausal osteoporosis have been demonstrated. Using data only from osteopenic women included in these trials, this study evaluated the effect of risedronate in reducing the risk of fragility fractures in women with femoral neck T-scores in the osteopenic range and without prevalent vertebral fracture.
Methods
Study design and patients
This retrospective analysis included the subsets of postmenopausal women with osteopenia in four risedronate placebo-controlled osteoporosis trials [15–18] of 1.5 to 3 years’ duration.
Patients included in the current analysis were postmenopausal women who had no radiographic vertebral fractures at baseline and had a femoral neck T-score between −1 and −2.5 SD as measured by DXA (NHANES III). These selection criteria were similar to those in previously published studies looking at similar populations [19, 20].
Although inclusion criteria for the VERT trials required patients to have prevalent vertebral fractures, re-adjudication conducted by central radiographic assessment after enrollment showed that between 10% and 15% of enrolled patients had no prevalent vertebral fracture and could, therefore, be included in the present study. Our analysis was based on BMD assessment at the femoral neck rather than the lumbar spine, due to its greater reliability and predictive value and wide acceptance internationally for clinical diagnosis [2, 13].
Patients treated with estrogen and estrogen-related drugs within 3 months of study entry or for more than 1 month within 6 months of study entry were excluded from the original studies. All patients received 1,000 mg calcium and, if needed, 500 IU vitamin D daily.
Adverse event information was collected at the study visits made every 3 months during the trials. The investigator recorded adverse events reported by the patients, as well as adverse events observed on examination of the patient.
Identification of fragility fractures
Fragility fractures were defined as a composite of a patient’s incident morphometric vertebral and osteoporosis-related nonvertebral fractures (i.e., six fracture types including clavicle, humerus, wrist, pelvis, hip or leg fractures). This endpoint was chosen to include all radiographically confirmed fractures.
With respect to the nonvertebral fractures, the selection of the nonvertebral fracture sites (clavicle, humerus, wrist, pelvis, hip or leg) followed the pre-specified definition of osteoporosis-related nonvertebral fractures of the risedronate trial program, which has been reported in all regulatory and previous scientific reports from this program. All osteoporosis related nonvertebral fractures required radiographic confirmation. The term “leg” fracture as used in the risedronate trials is a summary term, including the treatment emergent adverse event codes for fibula, tibia, subtrochanteric fracture, and leg, but not those for ankle or distal tibia fracture, the terms by which ankle fractures are usually reported.
Statistical analyses
The primary analysis focused on investigating the effect of risedronate to reduce fragility fractures (both vertebral and nonvertebral) in postmenopausal women who had no radiographic vertebral fractures at baseline and were considered osteopenic based on a femoral neck T-score ranging between −1 SD and −2.5 SD.
For consistency with previous risedronate studies, the statistical analysis used time-to-first-event methodology. Cumulative fracture incidence was estimated using the Kaplan–Meier method, and a Cox proportional hazards regression model was used to estimate the relative fracture risk reduction, via the hazard ratio (HR), between the risedronate 5 mg and placebo groups. The regression model was stratified for each trial to allow for separate underlying hazard functions. In order to determine the acceptability of combining data from all trials, potential heterogeneity assessments were performed using the Cox regression model to assess the trial-by-treatment, age-by-treatment and baseline femoral neck BMD-by-treatment interactions. Fracture efficacy for nonvertebral osteoporosis-related fractures and new vertebral fractures were also investigated. A sensitivity analysis was conducted to investigate whether the exclusion of patients with lumbar spine T-score <−2.5 SD would influence the outcome of the primary analyses. As the inclusion criteria for the primary analysis were low bone mass at the femoral neck (BMD T-score between −1 and −2.5 SD), we felt it was important to test whether the exclusion of patients with an osteoporotic BMD at another site would significantly affect the results.
Adverse events, serious adverse events, withdrawals, deaths, and upper gastrointestinal adverse events were summarized by treatment group. Differences in the proportions of patients between the treatment groups were tested using the Fisher’s exact test.
Results
Patients
A total of 620 postmenopausal women met the criteria for low bone mass and were included in the current analysis. Patients received either placebo ( = 309) or risedronate 5 mg ( = 311). Baseline characteristics were similar for each patient group (Table 1). The mean age was 64 years for both groups and mean baseline femoral neck T-scores were −1.84 in the placebo group and −1.85 in the risedronate group, respectively.
Table 1Baseline characteristics of patients with a femoral neck T-Score ≥ −2.5 SD and ≤ −1 SD without prevalent vertebral fractures All patientsExcluding lumbar spine T-score <−2.5 SDPlacebon n = 309Risedronate 5 mg n = 311Placebon n = 146Risedronate 5 mg n = 147Age (years)a64 (7.5)64 (7.5)63 (8.3)65 (7.0)Femoral neck T-scorea, b−1.84 (0.396)−1.85 (0.406)−1.77 (0.418)−1.76 (0.398)Height (cm)a159.8 (6.92)159.8 (6.05)160.6 (6.14)160.4 (6.32)Weight (kg)a65.9 (11.92)66.8 (11.19)69.8 (12.91)69.6 (12.27)BMI (kg/m2)a25.9 (4.54)26.2 (4.36)27.2 (5.1)27.1 (4.78)Patients with prevalent OP-related nonvertebral fractures56 (18%)59 (19%)34 (23%)31 (21%)a Mean (standard deviation)b Based on National Health and Nutrition Examination Survey III reference database
Potential heterogeneity assessments revealed that there were no significant interactions for trial-by-treatment ( = 0.495), age-by-treatment ( = 0.981) or baseline femoral neck BMD-by-treatment ( = 0.188)
Effect on fractures
Risedronate treatment was found to significantly reduce the risk of combined morphometric vertebral and nonvertebral fractures by 73% in women with a femoral neck T-score between −1 and −2.5 and no prevalent vertebral fracture; HR = 0.27, 95%CI = 0.09–0.83, = 0.023 (Fig. 1). The cumulative fragility fracture incidence was 6.9% and 2.2% respectively for placebo and risedronate.
Fig. 1Reduction of fragility fracture risk in patients with femoral neck T-score between −1 and −2.5 SD and no prevalent vertebral fractures
When separated according to nonvertebral and vertebral fractures, the cumulative nonvertebral fracture incidence was 5.4% and 0.4%, respectively, for placebo and risedronate (HR = 0.09, 95%CI = 0.01–0.71 ( = 0.022), and the cumulative vertebral fracture incidence was 4.2% and 1.8%, respectively, for placebo and risedronate (HR = 0.44, 95%CI = 0.11–1.78 ( = 0.249).
Sensitivity analyses
When patients with baseline LS spine T-scores below −2.5 SD were excluded, results consistent with those from the primary analysis population were observed, although the reduced sample size widened the confidence intervals and led to a loss of statistical significance. For the 293 patients who had both femoral neck and LS spine T-scores between −2.5 and −1 SD, the hazard ratio for fragility fractures was 0.22 (95%CI = 0.03–2.02, = 0.182), a magnitude of effect similar to that in the primary analysis population.
Safety
In the subgroup investigated for this analysis, no statistically significant differences in adverse events between the risedronate treated and the placebo group were recorded during the trial (Table 2). The observation in the subgroup analysis reported here is consistent with the findings in the overall population of the original trials [15–18].
Table 2Summary of adverse events (AEs) for patients with a femoral neck T-score ≥ −2.5 SD and ≤ −1 SD and noprevalent vertebral fractures (BMD and VERT Trials)CategoryNumber of patients (%)-valuePlacebo ( = 309)Risedronate 5 mg ( = 311)AEs290 (93.9%)293 (94.2%)0.87Serious AEs40 (12.9%)56 (18.0%)0.10AEs associated with death 2 (0.6%)0 (0.0%)0.25UGI AEs 77 (24.9%)81 (26.0%)0.78Withdrawn due to AEs 38 (12.3%)26 (8.4%)0.11-values were from Fisher’s exact test.UGI: upper gastrointestinal
Discussion
Osteopenia is common in postmenopausal women and contributes to the increased risk of fracture in this population. There is a need for clinical data examining both fracture risk in women with T-scores in the osteopenic range and the evidence for an effect of treatment on reducing that risk. Data from the National Osteoporosis Risk Assessment (NORA) study reported that more than half of women experiencing self-reported fractures of the hip, vertebrae, rib and wrist/forearm at one year fell into the osteopenic category based on peripheral BMD T-scores between −1 and −2.5 [21]. Results from the present study show that risedronate treatment significantly reduced the overall risk of fragility fractures by 73% over 3 years in patients with a femoral neck T-score between −1 and −2.5 and no prevalent vertebral fracture. A similar reduction in risk — that did not reach statistical significance due to low numbers — was observed when patients with lumbar spine T-scores below −2.5 were excluded from the analysis.
Previous studies in women with osteoporosis have shown that risedronate reduces the risk of morphometric vertebral fractures by up to 65% in the first year of treatment [17, 18], with significant reductions in clinical vertebral and nonvertebral fractures as early as 6 months [22, 23]. Moreover, a study in osteoporotic women without prevalent fracture demonstrated that treatment with risedronate reduced the risk of first vertebral fracture by 75% (i.e., a relative risk of 0.25) [24].
Previously reported studies have examined the efficacy of alendronate [20, 25] and raloxifene [19] on fracture risk reduction in osteopenic populations. A retrospective re-analysis of data from the Fracture Intervention Trial (FIT) showed a 60% reduction of vertebral fractures [RR 0.40 (0.19–0.76 95% CI)] in osteopenic women defined as having a femoral neck BMD T-score between −1.6 and −2.5 [20]. The overall analysis included patients with or without prevalent vertebral fractures, however, and since prevalent vertebral fractures are an important risk factor for further fractures and a criterion for severe osteoporosis [8, 13], classifying such women as osteopenic is subject to debate. When a subset of this population, those without prevalent fractures, was analyzed separately the results were similar, but did not reach statistical significance [RR 0.46 (0.16−1.17 95% CI)]. Results in patients with both hip and spine BMD in the osteopenic range are not reported in this paper. A re-analysis of the Multiple Outcomes of Raloxifene Evaluation (MORE) study reported similar rates of vertebral fracture risk reduction in raloxifene-treated women with osteopenia – defined as a total hip T-score >−2.5 without a prevalent vertebral fracture - compared with those with osteoporosis at 3 years [19]. The relative risk reduction for vertebral fractures with raloxifene compared with placebo was 0.53 (0.32–0.88, 95% CI) in osteopenic women; the relative risk for clinical vertebral fractures in osteopenic women was 0.25 (0.04–0.63). Information about reduction of nonvertebral fractures has not been provided in either the alendronate or raloxifene analyses. Conjugated equine estrogens significantly reduced the risk of clinical vertebral, hip and total fractures in postmenopausal women in the Women’s Health Initiative, the vast majority of whom did not have bone density testing but who were not selected based on having diagnosed osteoporosis [26]. Finally, a study of women ≥75 years of age in the UK reported that oral daily clodronate 800 mg for 3 years reduced any clinical fracture by 20%, but not hip fractures, in women unselected for having osteoporosis [27]. This reduction was independent of baseline BMD, but the number needed to treat was greater in the non-osteoporotic subjects.
The fracture benefit observed with risedronate in patients with low bone mass in our analysis may result in part from the effect of this compound on the preservation of bone microarchitecture. In previous studies of both early postmenopausal women and osteoporotic patients risedronate was observed to preserve trabecular architecture, including trabecular number, thickness and separation, compared to placebo [28, 29], which is likely to contribute to a maintenance of bone strength [30]. This beneficial effect on trabecular architecture has also been observed with long-term treatment with risedronate, as shown by recent data from a study in postmenopausal osteoporotic women [31]. Bone microarchitecture, assessed by three-dimensional micro-computed tomography, was maintained at similar levels over a five-year treatment period [31]. The sustained preservation of bone microarchitecture equates to a long-term beneficial effect on bone strength and may contribute to the decrease in fracture risk observed with risedronate treatment.
Limitations of the current study include the selection of initial populations in the VERT and BMD trials based on the presence of risk factors for osteoporosis (including low BMD and prevalent vertebral fractures). Because the original trials were designed to examine the efficacy of osteoporosis treatments in an osteoporotic population, rather than to determine representative incidence of fracture in a general osteopenic population, the data provided in the present analysis may reflect an overestimation of the fracture risk in the general postmenopausal population.
It is likely that in the relatively near future fracture risk assessment using a combination of femoral neck BMD, age, and other risk fractures will be utilized to determine absolute fracture risk and also serve as a basis for determining treatment intervention thresholds. In many older postmenopausal women, osteopenia will be an important risk factor for fracture, placing some of them at high absolute risk of a fracture event over a 10 year interval [32], and treatment with a pharmacologic agent will be indicated to lower risk. The data presented here demonstrate that in addition to the established efficacy in preventing fracture in osteoporotic women with or without prevalent fractures, risedronate significantly reduces the risk of fragility fractures in this population of postmenopausal women with osteopenic BMD values at the femoral neck and no prevalent vertebral fractures.
|
[
"risedronate",
"osteopenia",
"fragility fracture",
"vertebral fracture",
"nonvertebral fracture",
"bisphosphonates"
] |
[
"P",
"P",
"P",
"P",
"P",
"U"
] |
Cancer_Causes_Control-2-2-1764594
|
Meeting the mammography screening needs of underserved women: the performance of the National Breast and Cervical Cancer Early Detection Program in 2002–2003 (United States)
|
Objective To examine the extent to which the National Breast and Cervical Cancer Early Detection Program (Program) has helped to meet the mammography screening needs of underserved women.
Introduction
Breast cancer is the leading cause of cancer-related death among Hispanic women in the United States and the second-leading cause of cancer-related death among non-Hispanic women, causing more than 41,000 deaths in 2002 [1]. Breast cancer mortality is higher in poorer areas of the country than in wealthier areas [2]. Although timely breast cancer screening with mammography reduces mortality [3], in the United States, screening rates are low among low-income women who lack insurance coverage for mammography [4, 5]. In 1992, among women aged 40 years and older, 60% of women with health insurance and 22% of women without health insurance had received a mammogram within the previous 2 years. Among those women with family incomes greater than or equal to 200% of the federal poverty level, 64% had received a mammogram within the previous 2 years, but only 41% of women from families with incomes less than 200% of the poverty level had received this service (Robert Uhler, MS, Personal Communication, April 2006). The annual income level for a family of four at 200% of poverty in 1992 was approximately $28,000 [6, 7].
In this context, to help low-income, uninsured women gain access to mammography screening services, the U.S. Congress passed the Breast and Cervical Cancer Mortality Prevention Act of 1990 (Public Law 101-354), authorizing the Centers for Disease Control and Prevention (CDC) to establish the National Breast and Cervical Cancer Early Detection Program (NBCCEDP) [8] referred to henceforth as the Program. The Program provides services through cooperative agreements, which are in place in all 50 states, the District of Columbia (DC), 4 U.S. territories, and 13 American Indian/Alaskan Native organizations; for a total of 68 grantees. Services provided by an American Indian/Alaska Native organization were consolidated with services provided by the state-based program where the American Indian/Alaska Native organization participant resides. In this work, “state(s)” refers to the 51 geopolitical units (50 states and the District of Columbia (DC)).
The Program provides both breast and cervical cancer screening services to low-income uninsured women aged 18–64. From 1991 through June 2005, the Program screened 2.6 million women for breast and cervical cancer, provided 3 million mammograms to 1.7 million women and diagnosed 5,309 cases of in situ breast cancers and 15,493 cases of invasive breast cancers. The Program provides screening mammography to women ages 40 and older. In 1996, Program policy prioritized mammography screening of older women to allow for best use of limited resources. Of women receiving their initial screening mammogram through the Program in 2000–2004, 2.4% were under 40 years of age; 28.0% were 40–49 years old; 66.4% were between the ages of 50 and 64; and 3.2% were over 64 years of age. Fifty-two percent of the women were non-Hispanic Whites, 14% were non-Hispanic Blacks, 6% were non-Hispanic Asian/Pacific Islanders, 4% were non-Hispanic American Indian/Alaska Natives, 31% were Hispanics, and 3% were of unknown race/ethnicity [9]. The Program grew from eight grantees funded at $30 million in Fiscal Year 1991 to 68 grantees that received $204 million in 2005. Sixty percent of federal funds received by the grantees are spent on clinical services. The remaining 40% are used to fund other Program components including program management, data collection, quality assurance and improvement, partnership development, professional education, recruitment and evaluation. Treatment is covered by state Medicaid funding and other non-Program sources. A detailed description of the history of the Program is provided elsewhere [8]. Earlier research [10] found that in 1994–1996 the Program provided either breast or cervical cancer screening services to about 12–15% of eligible women aged 50–64, but that study did not estimate the proportion of women who were provided mammography screening specifically or provide estimates by race/ethnicity or for states.
The objective of the current analysis was to examine the extent to which the Program has helped to meet the mammography screening needs of this underserved population. Specifically, we estimated the numbers and percentages of women aged 40–64 who in 2002–2003 were eligible for breast cancer screening through the Program at both the state and national levels and the percentage of these eligible women who received Program-funded mammography. Finally, because disparities by race and ethnicity in the provision of screening services have commonly been reported in the U.S. [5, 11–13], we examined the extent to which the Program at the national level provided services to women of different racial and ethnic backgrounds.
Materials and methods
We examined eligibility using the Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) conducted by the U.S. Census Bureau, and screening service delivery using Program data. We used data from 2002 and 2003 because Program data were complete for those years [8].
Eligibility for NBCCEDP breast cancer screening services
Women aged 40 and older who do not have health insurance or whose insurance does not cover mammography screening are eligible for free breast cancer screening through the Program if their family incomes are below 250% of the federal poverty level. Twenty-two states and Washington, DC set eligibility criteria at lower poverty level. The annual income levels for a family of four at 200% and 250% of poverty in 2003 were approximately $37,000 and $46,000, respectively [7, 14]. Because 95.8% of women age 65 and over were covered by Medicare [15] and were therefore not served by the Program, we included only women aged 40–64 in this analysis.
Data sources
CPS ASEC
Questions about age, family size, sex, race, and Hispanic origin were included in the basic CPS, which is a monthly national survey undertaken primarily to determine the characteristics of the labor force of the U.S. civilian non-institutional population. Respondents were asked to identify their race by selecting one or more options from a list. Hispanic origin was asked separately from race. About 78,000 interviewed households in the CPS were asked a set of supplementary questions (the ASEC) about health insurance coverage, income received, and place of residence during the previous year [16]. The methods used to collect and report CPS ASEC data have been described previously [17].
NBCCEDP
Data for the number of women screened during 2002–2003 were obtained from Program service records. Grantees routinely collect income, family size, and insurance information to determine eligibility and collect screening information on each woman in the Program. These data include screening location, demographic characteristics, service dates, and outcomes. Demographic data are self-reported. Reporting of race and Hispanic origin is optional. The structure of the Program and methods for collecting and reporting Program data have been described elsewhere [8].
Data analysis
CPS ASEC
We estimated the number of women eligible for the Program by race and ethnicity (Hispanic origin) at the national level by extrapolating the respondents’ answers to the general U.S. population using standard Census Bureau methods [15–19]. Given the CPS sample size, it was impossible to produce meaningful estimates by race or ethnicity for individual states.
We categorized women who reported they were of Hispanic origin as Hispanic regardless of race. We categorized the remaining women, who were non-Hispanic, into one of the following racial groups: White, Black, American Indian/Alaska Native, Asian/Native Hawaiian/other Pacific Islander, or multiracial (which is henceforth defined synonymously with Two or More Races). In the CPS, if a respondent did not report his or her race, the respondent’s race was allocated using the race of another member of the household, or failing that, the previous record on the CPS file [17]. People were considered uninsured if they were not covered by any type of private or government health insurance for the entire previous year [18]. Poverty was computed by comparing total family income (or the person’s own income if she did not live with family members) with 1 of 48 dollar amounts called poverty thresholds [18, 19]. Each person’s poverty threshold was based on family size and number of children younger than 18 years [14]. The method of computing confidence intervals (CI) for the estimates of the eligible population is provided in Appendices 1 and 2. We use two-tailed 90-percent CI to be consistent with the Census Bureau’s practices for reporting poverty and health insurance data.
NBCCEDP
In counting the total number of women who received at least one Program-funded mammogram in 2002 and 2003, women were classified into the race/ethnic categories used to estimate Program eligibility. About 2.4% of women who were screened did not claim any race or Hispanic origin. These women (12,653) were counted in the total number of women screened, but not in the specific race/ethnicity categories.
NBCCEDP screening rates
Based on the number of women screened and estimates of the numbers of women aged 40–64 in both the U.S. population and the Program-eligible population we estimated the percentage of all U.S. women aged 40–64 and the percentage of Program-eligible women who received a Program-funded mammogram at least once in 2002 or 2003. We estimated provision of mammography screening over a 2-year period because the U.S. Preventive Services Task Force recommends that women aged 40 and older be provided with screening mammography every 1–2 years [3]. We further examined the distribution of mammography screening among women from different racial and ethnic backgrounds. The method of computing CI for the estimated screening rates is provided in Appendix 3.
Results
During 2002–2003, approximately 47 million women aged 40–64 resided in the United States (Table 1). Of those women, approximately 4.0 million (90% CI, 3.8 million–4.2 million) or 8.5% (CI, 8.1–9.0%) were estimated to be eligible for a Program-funded mammogram. Non-Hispanic White women constituted the largest number of women eligible for the Program: 2.0 million (CI, 1.8–2.1 million). However, the percentage of non-Hispanic White women who were eligible for the Program (5.7%; CI, 5.3–6.1%) was smaller than that of other racial groups except for multiracial women. The percentage of multiracial women who were eligible, 8.9%, was lower than the percentage for Hispanic women and for non-Hispanic American Indian/Alaska Native women, but was not statistically distinguishable from that of the other groups. The percentage eligible among Hispanic women (23.1%) was not statistically distinguishable from non-Hispanic American Indian/Alaska Native women (19.8%) but was higher than the percentage eligible among the other racial and ethnic groups.
Table 1Number and percentage of U.S. women aged 40–64 years who were eligible for the National Breast and Cervical Cancer Early Detection Program (NBCCEDP), 2002–2003Race/ethnicityAll women age 40–64Women eligiblea for NBCCEDP mammography screeningNumber (thousand)Percent distributionbNumber (thousand)90% CIPercentc90% CITotal46,8991004,0073,806–4,2088.58.1–9.0Non-Hispanic42,50490.62,9912,817–3,1667.06.6–7.4 White34,40373.41,9721,835–2,1095.75.3–6.1 Black5,43911.6714629–79913.111.6–14.6 AI/AN2250.54523–6619.811.3–28.4 A/NH/OPI1,9774.2221173–26811.28.9–13.4 Multiracial4601.04120–618.94.6–13.2Hispanic4,3959.41,016909–1,12223.120.9–25.3AI/AN: American Indian and Alaska Native; A/NH/OPI: Asian and/or Native Hawaiian and other Pacific Islanders; Multiracial: two or more racesa Women eligible for NBBCCEDP-funded mammography screening include women aged 40–64 with family incomes below 250% of federal poverty level, who are uninsured for mammography. The number of eligible women could be an underestimate because it excludes women who have insurance but whose insurance does not cover mammography screening. See Sect. Methods for detailsb Percent of all U.S. women aged 40–64c Percent of all U.S. women aged 40–64 in a given racial or ethnic group who were eligible for NBCCEDP funded mammography servicesSource: Authors’ tabulations of data from the U.S. Census Bureau, Current Population Survey, 2003–2004 Annual Social and Economic Supplements
During 2002–2003, 528,622 women aged 40–64 received at least one mammogram through the Program, among whom 12,653 (2.4%) were of unknown race/ethnicity (Table 2). Of all women in the United States aged 40–64, the Program provided mammography screening to approximately 1.1%. The rates of screening with mammography in the past 2 years (2002–2003), referred to henceforth as the screening rates, varied substantially by race/ethnicity. Approximately 9.7% of all American Indian/Alaska Native women were screened and approximately 0.3% of all multiracial women were screened.
Table 2Number and percentage of women eligible for the National Breast and Cervical Cancer Early Detection Program (NBCCEDP) and U.S. women provided with mammography screening services, at least once, through NBCCEDP, between 2002 and 2003Race/ethnicityNumber of women screenedPercentage of U.S. women screenedaPercentage of NBCCEDP-eligible women screenedb%90% CITotal528,6221.113.212.5–13.9Non-Hispanic349,6550.811.711.0–12.4 White221,4330.611.210.4–12.0 Black74,2591.410.49.2–11.6 AI/AN21,8829.749.225.5–72.9 A/NH/OPI30,6871.613.910.9–16.9 Multiracial1,3940.33.41.7–5.2Hispanic166,3143.816.414.7–18.1Unknown Race/ethnicity12,653–––AI/AN: American Indian and Alaska Native; A/NH/OPI: Asian and/or Native Hawaiian and other Pacific Islanders; Multiracial: two or more racesNBCCEDP mammography eligible women include: uninsured women aged 40–64 with family incomes below 250% of federal poverty levela Percent of all U.S. women in a given racial and ethnic group who were provided mammograms funded by NBCCEDPb Percent of all U.S. women in a given racial and ethnic group who are eligible and who were provided with NBCCEDP funded mammogramsSource: Authors’ tabulations of data from the U.S. Census Bureau, Current Population Survey, 2003–2004 Annual Social and Economic Supplements, and from NBCCEDP April 2005 data
Among all women eligible for the Program, approximately 13.2% (CI, 12.5–13.9%) were screened one or more times (Table 2). This estimate is based on the national Program eligibility limit of 250% of poverty. Using state-specific poverty level criteria, we estimated the screening rate for all states combined to be 14.7% (CI, 13.8–15.6%), reflecting the lower eligibility limits used in many states (data not shown).
The percentage of all eligible women who were screened in the Program varied by race/ethnicity (Table 2). The screening rates for non-Hispanic White and non-Hispanic Black women were not statistically distinguishable. Hispanic women had a higher screening rate than other women except for non-Hispanic American Indian/Alaska Native women, who had the highest screening rate, with an estimated 49.2% (CI, 25.5–72.9%) screened. The screening rate for non-Hispanic multiracial women was lowest (3.4%, CI, 1.7–5.2%).
The estimated number and percentage of women in each state who were eligible for the Program and state-specific poverty levels used as eligibility criteria are shown in Table 3. The estimated numbers of eligible women were greater in heavily populated states such as California, Texas, New York and Florida than in the other states. The estimated percentage of women eligible for the Program was highest in New Mexico (15.6%; CI 10.7–20.5%), although that percentage was not statistically distinguishable from the percentage for Louisiana, Mississippi, and five other states. Similarly, while Minnesota’s estimated percent eligible appeared lowest (2.8%; CI 1.2–4.5%), it was not statistically distinguishable from Delaware, Massachusetts, and 11 other states.
Table 3Number of women aged 40–64 and estimated number of women eligible for the National Breast and Cervical Cancer Early Detection Program (NBCCEDP), breast cancer screening, by state: 2-Year averages; 2002–2003U.S. populationEligible womenTotalPoverty criterionaNo. (thousands)90% CI (thousands)% Of totalb90% CI (%)U.S.46,899200/2503,5943,403–3,7857.77.27–8.06Alabama7482005734–807.64.6–10.6Alaska10325095–128.55.5–11.6Arizona8072507241–1028.95.3–12.4Arkansas4232003923–559.25.7–12.8California5,506200476386–5658.67.1–10.2Colorado7142505535–757.75.0–10.3Connecticut5852002312–344.02.0–5.9Delaware14325052–83.71.7–5.8District of Columbia9225063–96.63.5–9.7Florida2,805200262206–3189.37.4–11.2Georgia1,4072009555–1356.84.0–9.5Hawaii19625094–144.52.1–6.9Idaho213200169–237.24.1–10.4Illinois1,93620012385–1606.34.5–8.2Indiana1,0112006036–835.93.6–8.2Iowa4802502413–355.02.7–7.3Kansas4292502413–355.63.1–8.0Kentucky6692508054–10612.08.4–15.6Louisiana71225010875–14215.210.9–19.5Maine219250148–196.23.9–8.6Maryland9332505431–765.73.4–8.0Massachusetts1,0612504021–603.82.0–5.6Michigan1,65525011076–1436.64.7–8.6Minnesota8612502410–392.81.2–4.5Mississippi4712506140–8213.08.8–17.1Missouri8942005128–735.73.3–8.2Montana159200159–219.25.7–12.8Nebraska276225157–225.32.8–7.9Nevada3412502817–388.15.2–11.0New Hampshire225250116–164.92.7–7.0New Jersey1,4582509666–1266.64.6–8.6New Mexico2952504631–6215.610.7–20.5New York3,212250262209–3158.26.6–9.7North Carolina1,33020013797–17610.37.5–13.1North Dakota10520074–96.33.7–9.0Ohio1,96820011579–1515.84.1–7.6Oklahoma5252004023–577.64.5–10.7Oregon6152504527–627.24.5–10.0Pennsylvania2,07025011177–1455.43.8–7.0Rhode Island18125095–134.92.9–7.0South Carolina6712004828–687.24.3–10.0South Dakota11620085–116.74.1–9.4Tennessee1,0252507040–1006.84.0–9.7Texas3,205200396318–47412.410.1–14.7Utah3022501910–296.43.3–9.4Vermont11525052–74.12.1–6.1Virginia1,2452006233–915.02.7–7.2Washington1,0732007041–996.53.9–9.1West Virginia3112003222–4210.27.1–13.4Wisconsin9202504526–654.92.8–7.0Wyoming8525096–1210.57.1–13.9a 28 States and DC set income eligibility at 250% of poverty, 21 states at 200% of poverty and 1 state at 225% of poverty. The estimated number of eligible women for the U.S. is based on the eligibility criteria used in each stateb Eligible women as percentage of all women aged 40–64 years in that stateSource: Authors’ tabulations of data from the U.S. Census Bureau, Current Population Survey, 2003–2004. Annual Social and Economic Supplements.
The percentages of eligible women who were screened through the Program varied greatly across states (Fig. 1). The percentage of eligible women screened by in individual state ranged from about 2.2% (CI, 1.5–2.8%) to approximately 79% (CI, 49.4–108.9%). The median percentage screened among the states was 18.2% (CI, 11.7–24.7%). The tenth highest percentage estimate was 28.5 (CI, 23.2–33.9%) and the tenth lowest percentage estimate was 10.71 (CI, 7.4–14.1%).
Fig. 1Percentages of NBCCEDP eligible women screened for breast cancer, by State and District of Columbia, compared with the National Aggregate Percentage, 2002–2003. Note: The boxes show the percentage of eligible women screened by each state. The whiskers in each box, show the variation in the screening rate (the upper and lower CI). NBCCEDP referes to National Breast and Cervical Cancer Early Detection Program. *States includes DC, sorted by percentage of eligible women screened. Source: Authors’ tabulations of data from the U.S. Census Bureau, Current Population Survey, 2003–2004. Annual Social and Economic Supplements, and from NBCCEDP April 2005 data
Discussion
We found that a large number of women—4 million women or 8.5% of all U.S. women aged 40 to 64—were uninsured during 2002–2003 and had incomes below 250% of the federal poverty level, meeting Program eligibility requirements. Of these, only about 13.2% received mammograms funded through the Program. Although many of the estimates for groups of women classified by race/ethnicity were imprecise, as indicated by wide CI, there was clearly wide variation in numbers and percentages of women eligible for the Program and in the percentages of eligible women who were screened. The percentage of eligible women screened was highest among American Indians and Alaska Natives. Although the estimates of eligible women in each state were also imprecise, the findings clearly showed wide variability. The percentage of eligible women screened in each state ranged from approximately 2% to approximately 79%.
The Program is an important source of mammography screening services for low-income, uninsured women, but neither NBCCEDP nor other providers that serve this population are able to meet the current needs. The Program has grown since 1991/1992 when 38,476 women were screened in 12 states [8] to screen over 500,000 women during 2002/2003 in all states. A large number of federally funded community health centers, hospitals, clinics, and voluntary associations provide mammography screening services to underserved women. The numbers of women screened by these programs are not readily available. However, we know from the 2003 National Health Interview Survey (NHIS) that only 42.3% of women with no health insurance and family incomes less than 250% of the poverty level reported having had a mammogram during the previous two years (Robert Uhler, MS, Personal Communication, February 2006). Of the 4 million women we have identified in this study as eligible for the Program, the NHIS results indicate that about 1.7 million women were screened, meaning that approximately 1.2 million received screening outside of the Program. However, this leaves about 2.3 million low-income, uninsured women who did not obtain those services from either the Program or other sources.
The largest numbers of women eligible for the Program were non-Hispanic White women. In contrast, the largest percentages of women eligible for the Program were from minority groups, except for women of Two or More Races. Hispanic and non-Hispanic American Indian/Alaska Native groups had the highest percentages of eligible women. We were limited in our ability to assess the extent to which the Program met the mammography screening needs of women by race/ethnicity because about 12,000 women screened in the Program were of unknown race/ethnicity. If we had been able to correctly allocate these women to their appropriate race/ethnic groups, the percentages of women screened would have differed from those shown, potentially by an important margin. However, the findings indicate that the Program was most successful in meeting the needs of American Indian/Alaska Native women, approximately 49% of whom were screened. A possible reason for this success may be that these populations are the focus of health services through which the Program operates. In 1993, Congress amended the Breast and Cervical Mortality Prevention Act, Public Law 108-183, to authorize funding for American Indian/Alaska Native organizations and provided the opportunity to direct resources to these populations, specifically four grantees in Alaska and another nine geographically distributed across the contiguous United States.
The number and percentage of women who were eligible for the Program varied greatly from state to state, because of differences in population size, age, and sex distributions, as well as differences in income and insurance coverage, including Medicaid [18]. In addition to the aforementioned factors, reasons for variations in the percentage of eligible women screened include differences in income eligibility criteria, presence of American Indian/Alaska Native grantee, CDC funding levels, other sources of funding, and organization and efficiency of the screening Programs. The upper and lower bounds of the CI indicate that some estimates were not precise, but were useful nevertheless. These estimates have been made available to each state for their use in Program planning to better understand the need for screening services in their states.
Our study is subject to a number of limitations. As already noted, many of the estimates are imprecise because the numbers of women in the CPS ASEC sample who are eligible for the Program are relatively small. In addition, health insurance coverage could be underreported in the CPS ASEC given that the survey uses annual retrospective questions and respondents may have difficulty recalling the information [18]. Also, some women are eligible for the Program even if they have health insurance, but are underinsured, meaning the insurance does not cover mammography screening services, or there is a high copayment. Since CPS ASEC insurance questions do not measure covered services, these underinsured eligible women are not included in the denominators of our screening percentages. It is uncertain how many low-income women in the United States population are underinsured. Finally, our inability to define the race or ethnicity of some women in the study could result in an underestimate of the participation rate for any given race or ethnic group.
We suggest two strategies to improve screening rates: increasing efficiencies of the Programs and improving their collaboration with other organizations. First, the Program must seek ways of increasing its efficiency to serve more women with existing resources. A study of Program costs found that the average cost of screening a woman through the Program was lower for grantees screening greater numbers of women because of economies of scale, that is, average cost decreased as number of women screened increased [20]. States with small populations in larger geographical areas may have limited opportunities to achieve such economies of scale. CDC has recently initiated a cost-effectiveness evaluation of the Program and is developing methods to better collect and analyze information on resources and how they might be used more efficiently. A variety of means to increase efficiency will need to be pursued. For example, many women in the Program are screened annually. The U.S. Preventive Services Task Force recommends screening every 1–2 years because it has found little evidence that annual screening is more effective than biennial screening [3]. Many European programs provide screening every 2–3 years [21]. The Program may need to evaluate the potential balance of health benefits from adopting a biennial schedule that could serve more women.
Second, the Program needs to improve collaboration and coordination with other providers that serve a similar client population. The Program already coordinates substantially with private and nonprofit organizations, businesses, and other groups involved in breast cancer screening, but that coordination needs to be increased to recruit the women who are not currently being served. For example, in addition to providing screening services, the Program provides diagnostic services for eligible women screened by organizations outside of the Program. Alternative sources of diagnostic services may need to be pursued to free resources for increased screening of eligible women.
Although greater efficiency and improved coordination with other screening providers could better meet the needs of underserved women, they are unlikely to be enough. Given that about 2.3 million low-income uninsured women did not obtain recommended breast cancer screening services in 2003 and that the Program provided those services to about 500,000 women; increased efficiency and coordination alone will be insufficient to meet the needs of the eligible population.
In 2000, when Healthy People 2010 first set out its objectives of eliminating health disparities and increasing the proportion of women aged 40 and older who have received a mammogram within the previous two years to 70% [12], the greatest disparities in breast cancer screening were for women who had no health insurance, those who had no usual source of care, and recent immigrants [5]. Although progress has been made since 1987 in increasing mammography screening among low-income and uninsured women, the increases for low-income women are less than those for higher-income women, and screening among the uninsured lags far behind screening among women with private or public health insurance [5]. The Program contributes substantially to the effort to provide breast cancer screening services to those women by serving 13.2% of those eligible. However, the Healthy People 2010 objectives are still far from being met.
|
[
"mammography screening",
"breast cancer",
"screening rates",
"medically underserved"
] |
[
"P",
"P",
"P",
"M"
] |
J_Comput_Aided_Mol_Des-3-1-2039848
|
Construction of 3D models of the CYP11B family as a tool to predict ligand binding characteristics
|
Aldosterone is synthesised by aldosterone synthase (CYP11B2). CYP11B2 has a highly homologous isoform, steroid 11β-hydroxylase (CYP11B1), which is responsible for the biosynthesis of aldosterone precursors and glucocorticoids. To investigate aldosterone biosynthesis and facilitate the search for selective CYP11B2 inhibitors, we constructed three-dimensional models for CYP11B1 and CYP11B2 for both human and rat. The models were constructed based on the crystal structure of Pseudomonas Putida CYP101 and Oryctolagus Cuniculus CYP2C5. Small steric active site differences between the isoforms were found to be the most important determinants for the regioselective steroid synthesis. A possible explanation for these steric differences for the selective synthesis of aldosterone by CYP11B2 is presented. The activities of the known CYP11B inhibitors metyrapone, R-etomidate, R-fadrazole and S-fadrazole were determined using assays of V79MZ cells that express human CYP11B1 and CYP11B2, respectively. By investigating the inhibitors in the human CYP11B models using molecular docking and molecular dynamics simulations we were able to predict a similar trend in potency for the inhibitors as found in the in vitro assays. Importantly, based on the docking and dynamics simulations it is possible to understand the enantioselectivity of the human enzymes for the inhibitor fadrazole, the R-enantiomer being selective for CYP11B2 and the S-enantiomer being selective for CYP11B1.
Introduction
Aldosterone is a member of the renin angiotensin aldosterone system (RAAS) that plays an important physiological role in the regulation of electrolyte homeostasis and thereby blood pressure. This endogenous mineralocorticoid exerts its function by binding to the mineralocorticoid receptor. Upon ligand binding, the protein-ligand complex is translocated to the cell nucleus, where it modulates the gene expression of proteins involved in electrolyte homeostasis [1]. Aldosterone is produced predominantly in the adrenal cortex and is derived from cholesterol through various steps involving a number of dehydrogenases and cytochrome P450 enzymes (CYPs). This includes aldosterone synthase (CYP11B2), which catalyses the final steps of aldosterone biosynthesis.
Recently, various studies on the pathophysiology of heart failure have revealed that aldosterone plays a role in the formation of myocardial hypertrophy, reactive myocardial fibrosis, vascular remodelling and electrolyte imbalance. This may contribute to the development of arrhythmias, hypertension and congestive heart failure [2, 3]. Although great therapeutic successes have been achieved by treating heart failure patients with blockers of the RAAS (CONSENSUS trial [4], SOLVD trial [5]), the mortality due to heart failure is still high. Therefore, exploring new therapeutic possibilities is highly desirable. Recently, it became clear that aldosterone is also a key player in heart failure. In the RALES study [6], blocking the action of aldosterone using the mineralocorticoid receptor antagonist spironolactone, proved to greatly reduce mortality, hospitalisation numbers and hospitalisation time in patients with severe heart failure. Later, similar encouraging results were found for eplerenone, another mineralocorticoid receptor antagonist (EPHESUS trial [7, 8]), in patients with post myocardial infarction. From this it can be derived that reducing aldosterone action seems highly beneficial in the treatment of heart-failure and that it is particularly worthwhile to find new pharmacological manners to interfere with this hormone.
An alternative manner to reduce aldosterone action would be not to block the mineralocorticoid receptor, but to prevent formation of the hormone itself by inhibiting its biosynthesis [9]. Although the net outcome of aldosterone receptor inhibition versus aldosterone synthesis inhibition apparently is identical, a number of potential advantages for the latter approach exists: reduction of side effects as particularly observed for the antagonist spironolactone [6, 10], prevention of compensatory aldosterone synthesis inherent to mineralocorticoid receptor blockade (of which long term effects are unknown) [11, 12], and possible circumvention of inter-individual variations regarding pharmacodynamics and pharmacokinetics that are as observed for mineralocorticoid receptor antagonists [13]. The clinical use of inhibitors of aromatase (CYP19, which converts androgens into estrogens) next to estrogen receptor antagonists in the treatment of breast cancer, is one of the examples that shows that the approach of ligand synthesis inhibitors clinically can be of great use, despite the existence of receptor blockers for that ligand [14, 15].
The last steps of the biosynthesis of aldosterone are mediated by the mitochondrial cytochrome P450 11B family (CYP11B) (Fig. 1). Members of this protein family contain a heme prosthetic group in the core of the active site with which they catalyse (subsequent) oxidation reactions on C11, C18 and C19 on the β-side of the steroid skeleton (for enumeration see Fig. 1). In bovine [16], pig [17] and frog [18], aldosterone synthesis is performed by only one cytochrome, CYP11B, but in human [19] and mouse [20] the synthesis involves two isoforms, CYP11B1 (steroid 11β-hydroxylase) and CYP11B2 (aldosterone synthase). Rat possesses four isoforms for which CYP11B1 and CYP11B2 are the most important ones. CYP11B3 is only expressed in neonatal rat and carries the same activity as CYP11B2, and CYP11B4 encodes a pseudogene [21].
Fig. 1Biosynthesis of Aldosterone by the CYP11B family. Indicated with arrows are the possible substrate conversions performed by human CYP11B1 and CYP11B2 [19]. Rat CYP11B1 and CYP11B2 possess the same activities as the human isoforms, except that rat CYP11B1 can also oxidise 11-deoxycorticosterone on C19 [21]
Most remarkable is the substrate specificity of the different isoforms. In both man and rat, only the CYP11B2 isoform can perform the final oxidation of C18 to produce aldosterone [19, 21]. For the CYP11B1 isoform, C19 hydroxylation has been reported for rat but not yet for man [21], and the CYP11B1 isoform in general is known to play an important role in the biosynthesis of glucocorticoids. If an inhibitor of aldosterone synthesis is to be designed, the biosynthesis of glucocorticoids should remain unaffected, indicating that the inhibition must be CYP11B2 selective.
We developed three-dimensional in silico models as a decision making tool to facilitate the selection of potential CYP11B2 inhibitors for synthesis and in vitro testing. In order to obtain specific CYP11B2 inhibitors, we also investigated its highly homologous family member CYP11B1.
The aim of this report is (1) to deduce by homology/comparative modelling the architecture of CYP11B1 and CYP11B2 around their active site, (2) to investigate protein-substrate interactions and propose a mechanism for substrate regioselectivity and (3) to validate the homology models by correlating the in vitro activity of four known inhibitors to in silico data. The inhibitors we have chosen are metyrapone [22], R-etomidate [23], R-fadrazole and S-fadrazole [24] (Fig. 2). The in silico models not only represent an important tool in modern drug discovery but will also help in elucidating molecular mechanisms and (substrate binding) preferences of the substrate conversion of the enzymes of interest.
Fig. 2Chemical structures of the known CYP11B inhibitors, metyrapone, R-etomidate, R-fadrazole and S-fadrazole
Modelling considerations
Homology modelling has been widely applied to the family of cytochrome P450 enzymes and is mainly oriented towards enzyme inhibition. Before the first mammalian cytochrome P450 structures became available, modelling attempts were classically performed with crystal structures possessing sequence identities lower than 25% because no realistic alternatives were available. Homology models were often based on a template of bacterial CYP101 [25–27], but the introduction of bacterial CYP102 allowed for modelling the functional properties of eukaryotic class II P450s [28]. Although these homology models contain a low sequence identity with their template structures and are intuitively suboptimal, it has been shown that they can describe key features of protein-ligand interactions [27, 28]. For example, features observed for inhibitor binding in aromatase models have provided important insights for the development of drugs [27, 29]. Current models often still feature bacterial P450s as template [30, 31], but methods involving the usage of multiple crystal structures for model construction may prove to be the future trend [32, 33]. A model based on the structure of several known enzymes would be more accurate since every additional segment will improve similarity or spatial coordination of protein regions. However, structural flaws can be expected at locations where the template structures are joined and if these regions are within the active site, they need to be thoroughly refined.
Some modelling work on CYP11B1 and CYP11B2 has already been performed earlier by Belkina et al. [34] and Ulmschneider et al. [35]. The models of Belkina et al. discuss the potential spatial arrangement of the amino acids in the active site and hypothesise the hydrogen-bonding network involved in heme stabilisation. Furthermore, the effect of several amino acid mutations have been detailed. The models of Ulmschneider et al. focus on describing protein-inhibitor interactions and structure activity relations of their developed inhibitors. The so far published models were thoroughly characterised for those specific purposes, however, the goal of our model construction work was to investigate the regioselectivity of the natural ligands within the enzymes and to detail potential protein-ligand interactions.
For validation of our own three-dimensional models, the in silico data are presented in the form of molecular docking and molecular dynamics simulations. These methods are regularly used to investigate protein-ligand interactions. Because the only difference in the activity of the two isoforms CYP11B1 and CYP11B2 is the formation of aldosterone by the latter, successful 3D modelling of the isoforms relies on a careful analysis of the specific substrate conversion activities that exists between these two isoforms.
Because of this we reviewed an experimental mutation study by Bottner et al. [36] on the human CYP11B1 and CYP11B2 proteins, performed in a similar manner as by Belkina and Ulmschneider for the currently published models [34, 35]. The study by Bottner et al. showed that mutation of three residues outside the active site (L301P, E302D, A320V) is sufficient to convert the catalytic activity of CYP11B2 into that of CYP11B1, suggesting that remote steric aspects play a more important role in the substrate binding and substrate conversion than the presence of different amino acids in the active sites of both isoforms. This led us to postulate that the difference in substrate conversion is caused by a difference in the relative positioning of the substrate above the heme in the active site. To be more specific, we postulate that there is a correlation between substrate selectivity and the substrate hydroxylation distance, the distance between the heme iron and the substrate carbon. In other words, the binding mode of the natural substrate dictates which carbon atom is oxidised first, with conversion taking place on the carbon atom which is in closest proximity to the iron-oxygen complex.
For human CYP11B1 this means that C11 and C18 are to be in close proximity to the catalytic iron atom, with C11 closest to the iron. Rat CYP11B1 possesses a similar binding mode, but we expect that it also presents C19 in a position allowing oxidation. Explaining the preference for C18-hydroxylation, human and rat CYP11B2 would bind with C18 closest to the iron atom and C11 at a correct distance for oxidation.
To substantiate this hypothesis, the three dimensional architectures of the human and rat CYP11B enzymes were constructed using comparative modelling. For reasons of relevance only the CYP11B1 and CYP11B2 isoforms were investigated. We intend to show how knowledge of these various hydroxylation patterns of aldosterone precursors can result in working models for the substrate selective activity of the two isoforms. From here on, the human isoforms will be noted as hCYP11B1 and hCYP11B2, whereas the rat isoforms will be noted as rCYP11B1 and rCYP11B2.
As stated above, another aim was to validate the in silico models with in vitro activity data of four known inhibitors. These inhibitors were chosen for the following reasons. Metyrapone is a known inhibitor of CYP11B1 and is clinically used in the diagnosis of Cushing Syndrome [22, 37]. R-etomidate is clinically used as anaesthetic, but it is known to be a highly potent suppressor of the CYP11B family [23, 38]. Racemic fadrazole was designed for the selective inhibition of aromatase, a cytochrome P450 enzyme which is closely related to the CYP11B family because it oxidises steroids on C19. Next to aromatase inhibition, racemic fadrazole also shows considerable inhibition of members of the CYP11B family [24, 39].
Methods
Homology modelling
The amino acid sequences of the CYP11B family were taken from Swissprot [40] (hCYP11B1 accession P15538, hCYP11B2 accession P19099, rCYP11B1 accession P15393, rCYP11B2 accession P30099, rCYP11B3 accession P30100) and the determination of the secondary structures of the CYP11B family members was performed using the secondary structure prediction program JPred [41]. The alignment used for homology modelling was obtained with MOE-Align [42] by performing a topological alignment of the cytochrome P450 enzymes for which a crystal structure is publicly available (Fig. 3). The three dimensional architecture of cytochrome P450 enzymes consists of 12 alpha-helices annotated from A to L, as well as six beta-sheets, despite having a sequence identity among P450s less than 20%. Additionally, several short helices are present in various cytochrome P450 enzymes (annotated B′, F′, G′, J′, K′) whilst being absent in others. The structural core of all cytochrome P450 enzymes consists of a four-helix bundle composed of alpha-helices D, E, I and L, and the two alpha-helices J and K. The variable structural features involved in ligand binding and substrate specificity are situated in alpha-helices A, B, B′, F, G and their adjacent loops [43–45].
Fig. 3Topology alignment of human and rat CYP11B isoforms to related cytochrome P450 enzymes of which a three dimensional structure has been elucidated. Indicated with a & are the Arg123 in alpha-helix B′ and Glu310 in alpha-helix I. Indicated with a * is the triple mutant L301P, E302D, A320V in alpha-helix I. Indicated with a ^ is the catalytic Thr318 in alpha-helix I. Indicated with a # is the conserved Glu459 in alpha-helix L
The predicted secondary structures for the CYP11B family members were aligned to these topological features of the crystal structures, as well as by preserving several characteristic CYP features [46]: W(R/K)XXR (X indicates any amino acid) in helix C connecting and stabilising the heme prosthetic group in the active site, EXXR in helix K and (W/F)XXPXX(F/Y)XPX(H/R)(W/F) following helix K′ comprising the typical meander region, and finally XXF(G/S)XGX(H/R)XCXGXX(L/F)AXXE before helix L which contains the cysteine residue by which the heme prosthetic group is bound to the enzyme. Furthermore, a glutamic acid was aligned that is thought to be involved in the formation of a catalytically important water channel [47] (Glu366 in CYP101, Glu459 in the CYP11B family), and the entire helix I was aligned for its catalytic threonine (Thr252 in CYP101, Thr318 in CYP11B family) [47]. As a result of small variations in loop regions, some insertions and deletions were introduced in segments outside the active site. These segments did not correspond to any of the key secondary structures.
Clearly, CYP11B1 and CYP11B2 portray a high degree of homology possessing a pair wise sequence identity percentage as high as 94% for human and 83% for rat (Table 1, highlighted). This emphasises the difficulty of modelling the difference between the two isoenzymes and the challenge of reaching the level of modelling accuracy that is required. For P450 enzymes for which a crystal structure is publicly available, the overall pair wise sequence identity with the CYP11B family is found to be less than 20%, whereas it is around 30% for active site residues. This low degree of homology indicates that none of the reviewed cytochromes can be used as a sole representative for modelling the CYP11B family.
Table 1Generic pair wise sequence identity (in percentages) between the human and rat CYP11B isoforms and cytochrome P450 enzymes for which a three dimensional structure has been elucidatedChains10110210710811955512B42C52C82C92D63A4h11B1h11B2r11B1r11B2r11B3CYP101–17.330.237.739.638.523.521.223.130.823.111.519.626.424.524.526.426.42CPPCYP10216.3–24.528.339.619.229.426.934.630.832.714.847.126.426.424.524.524.51BU7CYP10720.012.3–28.350.936.521.619.221.221.225.014.321.620.818.920.818.918.91JINCYP10823.215.822.1–35.828.829.425.026.928.825.017.929.426.426.428.328.328.31CPTCYP11918.816.025.620.8–34.627.528.828.832.728.825.029.426.426.428.326.426.41F4UCYP5521.211.228.524.724.3–19.617.319.221.223.118.519.618.917.017.017.017.01ROMCYP5112.318.019.117.214.716.0–23.128.825.025.011.135.320.820.817.020.820.81EA1CYP2B414.616.716.414.015.016.016.7–57.765.459.633.333.332.132.132.130.230.21SUOCYP2C516.817.816.614.816.915.014.951.0–69.278.844.433.332.132.128.332.132.11NR6CYP2C815.817.616.415.515.815.813.453.873.6–69.244.431.430.230.230.230.230.21PQ2CYP2C915.618.017.415.016.615.314.351.077.378.4–40.733.332.132.128.330.230.21OG2CYP2D613.116.914.914.314.716.016.739.640.040.638.5–14.825.025.025.021.421.42F9QCYP3A414.122.019.614.517.214.816.522.822.022.921.917.9–32.132.130.230.230.21W0Eh11B116.016.314.615.516.112.014.917.417.615.817.314.917.1–98.181.186.884.9–h11B215.316.913.614.815.812.014.517.417.616.217.715.517.593.6–83.088.786.8–r11B112.315.413.213.815.311.013.417.416.716.617.714.417.363.663.6–88.786.8–r11B215.315.814.414.516.613.014.715.916.215.817.113.516.868.268.882.6–98.1–r11B315.415.814.414.516.613.014.715.916.215.817.113.816.868.669.283.097.3––The bottom triangle indicates the pair wise sequence identity of the whole protein, the top triangle indicates the pair wise sequence identity of the residues within 5.0 Å from the surface of the active site cavity of CYP101. Species information: Pseudomonas-Putida 2CPP, Bacillus Megaterium 1BU7, Saccharopolyspora-Erythreaea 1JIN, Pseudomonas-SP 1CPT, Archaeon Sulfolobus Solfataricus 1F4U, Fusarium-Oxysporum 1ROM, Mycobacterium Tuberculosis 1EA1, Oryctolagus Cuniculus 1SUO and 1NR6, Homo Sapiens 1PQ2, 1OG2, 1W0E and 2F9Q
Because of the low sequence identity of the CYP11B family, we have chosen to create a hybrid template for hCYP11B2 using MOE-Homology [42], constructed from the crystal structures of CYP101 (pdb code: 2CPP) and CYP2C5 (pdb code: 1NR6). Our criteria for using CYP101 and CYP2C5 involve similarity in functionality of both the cytochrome P450 reduction system and ligand characteristics, but importantly also involves the spatial positioning of active site regions.
Thus far, all modelling attempts on cytochrome P450 family 11 have included the usage of microsomal P450s such as CYP102 [30, 34] and CYP2C9 [35]. However, the CYP11B family belongs to the bacterial/mitochondrial cytochrome P450 class which obtains electrons from the ferredoxin reductase family in the electron transfer chain [48]. Using CYP101 for the modelling of mitochondrial P450s is therefore more intuitive and has been successfully applied to other mitochondrial P450s [33, 49].
The natural ligands of the CYP11B family are steroids, and steroids can be substrates for hepatic cytochromes that belong to the microsomal cytochrome P450 class. In CYP2C5 and CYP2D6 steroids are oxidised on the β-side of the steroid skeleton at carbon atoms close to C11 and C18 [50], and their crystal structures may possess the necessary interaction features for model construction. However, investigation of the crystal structure of CYP2D6 (PDB code: 2F9Q) raised doubt on the appropriateness for its use as a template structure. The crystal structure is resolved at low resolution (3.0 Å) in a tetramer complex. In this complex numerous amino acids participate in an extensive interaction network between the monomers and the structure was therefore excluded from our modelling work. Of the other hepatic P450 enzymes, CYP3A4 can also oxidise steroids, but its oxidation sites are on the opposite side of the steroid skeleton or on the different steroid rings. This means that the steroids bind with a rotated or reversed orientation in the active site [51], which may provide different active site conformation than the CYP11B family. Furthermore, many CYP2D6 substrates are characterised by a basic nitrogen [52], and the CYP2C9 substrates are mostly weakly acidic [53]. These properties are found to be less desirable for modelling of the CYP11B family than the functionally similar properties of the cyclic alkane substrates of CYP101. The structures of CYP2C9 also do not provide information about the molecular basis of regioselectivity of the substrate, since a conformational change of the active site is required to allow the substrate hydroxylation sites to contact the heme [54]. Because of the different ligand properties, the structures of hepatic CYP2D6, CYP2C9 and CYP3A4 were found to be less suitable as template structure.
An important decision criterion to use CYP101 as the template is also its spatial properties. When comparing its active site cavity to that of the mammalian P450s, we found that the steroidal ligands would not be able to fit into the active site cavity of the mammalian P450s with C11 and C18 oriented to the heme, unless additional space is introduced near helix K. Because the most important interactions between protein and ligand take place near the heme, the regions lining the active site must be modelled with the highest accuracy. From the multiple sequence alignment with the CYP11B family one can see that beta-sheet 6-1 following helix K possesses a 1 amino acid insertion in the CYP101 structure and a 1 amino acid deletion in the mammalian structures (Fig. 3). Although it is possible to model this site through a deletion in the mammalian structures, the resulting active site cavity would become strained. Since the structure of CYP101 is elongated, modelling a 1 amino acid insertion and relaxing the fold will create a better definition of this active site region. Taking into account that its substrate and reduction system are similar to that of the CYP11B family, CYP101 is found to possess the best structural core for the modelling of the CYP11B isoforms.
Because the structural core of the cytochrome P450 enzyme structure is the four-helix bundle comprising the helices D, E, I and L, and the two helices J and K, they should all be taken from one and the same template structure [43–45]. Hence, when taking the helices I and K from CYP101, the other regions must be used as well. Within the topological alignment of these regions, the mammalian P450s contain insertions and deletions in the connectivity between helix C and D, which is an additional reason why we have decided to construct the CYP11B models using the CYP101 structure.
Using the main structural features of CYP101, the remaining variant regions of the P450 fold (helices B′, F, G, J′ and the meaner region) were modelled with CYP2C5. CYP2C5 was chosen because after aligning the desired structures, it was found to possess slightly better spatial alignment with CYP101 than the other mammalian crystal structures, and additionally, CYP2C5 is a steroid synthase itself. Special attention was given to the modelling of helix B′ by CYP2C5 and its connection to the helices F and G. This three helical complex makes up the entrance of the active site cavity and closes the pocket like a lid. The helix B′ is predicted to be 3 turns long in the CYP11B family and there is an apparent insertion of 3 residues compared to that of the other cytochrome structures (Fig. 3). In the CYP101 structure, this helix is positioned too high in the active site cavity, which is probably a reason why it is regularly seen as inappropriate to model on. We chose the structure of CYP2C5 where we extended its helix from a 2 turn length to a 3 turn length.
The first 50 N-terminal residues corresponding to the membrane binding region of the human cytochrome 11B family were omitted from the models because no complementary sequence is present in the crystal structures of either CYP101 or CYP2C5. Some manual adjustments were made to the hCYP11B2 model to compensate for large steric hindrances and the model was refined by energy minimisation using a MOE succession method of steepest descent, conjugate gradient and truncated newton until an RMS gradient of 0.1 kcal/(mol Å) was reached. The Charmm22 forcefield [55] was used with a dielectric constant of 4 and all backbone atoms were tethered with a force constant of 100 kcal/(mol Å2) to prevent large movements.
To investigate the influence of the triple mutant introduced in the hCYP11B2 protein by Bottner et al. [36], a hCYP11B2-triple mutant homology model (hCYP11B2-TM) was created. The three dimensional architecture of this structure should turn out to be similar to the active site of hCYP11B1. Models for hCYP11B1, rCYP11B1 and rCYP11B2 were also constructed by amino acid replacement in the model of hCYP11B2, followed by a structural relaxation of the fold with MOE.
The four ligands 11-deoxycorticosterone (DOC), corticosterone (B), 18-hydroxy-11-deoxycorticosterone (18OH-DOC) and 18-hydroxycorticosterone (18OH-B) were fit in the active sites of all CYP11B models matching all three CYP11B hydroxylation sites (C11, C18, C19) to the oxygen atom occupying the 6th ligating position of the heme iron. First the steroids were docked into the active site using automated docking to get an impression of the relative positioning of the C11, C18 and C19 above the heme. Next, alternate orientations of the steroids were investigated by visual inspection, taking into account alternate folding of amino acid side chains. A distance of 2.5 Å was chosen as a starting distance between the oxygen and carbon atoms, which is a near-optimal distance for hydrogen atom abstraction during hydroxylation [56]. A threshold of 5.0 Å was estimated to be the representative maximal distance for hydroxylation. The models were then equilibrated with MOE without tethering the backbone atoms, allowing the model active sites to diverge.
Ligand docking
The docking of all ligands was performed using GOLD v3.0 [57], in order to investigate the protein-ligand interactions and investigate the application of the models for inhibitor potency prediction. The docking parameters used were taken from the default 1 GOLD GA settings. These settings were used for all docking runs presented in this study. Each ligand was docked 5 times for 100 poses each for which the GOLD Fitness score was obtained using the standard Goldscore function. The amount of docking runs and poses was taken to ensure an appropriate sampling of the ligand conformations in the active site of the protein. For the inhibitors, the fitness score was averaged for the top 10 ranking poses if the conformation was within an RMSD of 1.0 Å from the best ranking pose (results Table 5). Subsequently, to get an impression of the correlation of the docking with the in vitro data, the Goldscore measure of binding affinity was calculated by the method reported by Verdonk et al. [57] per ΔGbinding = −0.1075 * Goldscore−2.2665 (R2 = 0.5529, N = 60, ΔGbinding expressed in kcal/mol). The average poses were also rescored with the Chemscore scoring function [58].
Docking of steroids was performed in the presence of an iron-bound oxygen atom and their conformations were afterwards checked to investigate alternate orientations of the steroid in the active site cavity. The protein structures used for docking the substrates were the unequilibrated structures, whereas for docking the inhibitors, the hCYP11B1 and hCYP11B2 models were used after they were equilibrated with the ligand 18-hydroxycorticosterone. All docking runs were performed in the absence of water molecules. For each inhibitor, the best ranked pose was used as input for the molecular dynamics study.
Molecular dynamics of inhibitors
Several molecular dynamics simulations were performed to investigate the stability of the enzyme models of hCYP11B1 and hCYP11B2, and the dynamic behaviour of the three inhibitors inside the respective active sites. The simulations were performed with the NAMD package [59] using the Charmm22 forcefield [55]. The protein models of hCYP11B1 and hCYP11B2 were solvated in a 80 × 80 × 80 Å equilibrated water box, removing any overlapping water. Counter-ions were added to ensure an overall net charge of zero. The protein-ligand complexes were slightly equilibrated for 25 ps at a temperature of 100 K and were subsequently simulated for 1ns at a temperature of 310 K in an NPT ensemble. All simulations were carried out using periodic boundary conditions. To calculate the electrostatic interactions we chose the implementation of NAMDs Particle Mesh Ewald [60]. To accommodate the filling of the active site cavity volume by the three ligands, three and four explicit water molecules were positioned in the active sites of hCYP11B1 and hCYP11B2 respectively (volumes are hCYP11B1 360 Å3, hCYP11B2 334 Å3, etomidate 269.7 Å3, fadrazole 241.9 Å3, metyrapone 247.7 Å3, water 15.5 Å3). The positioned water molecules were optimised in the equilibration step.
Cellular assay for measuring inhibitor in vitro activity
For determining inhibitor potencies for CYP11B1 and CYP11B2, an assay was used in which cells overexpressing CYP11B1 or CYP11B2 convert the used steroid substrates for these cytochromes into their products. Product formation rates were assessed by HPLC. V79 cells overexpressing CYP11B1 (stably transfected with a pcDNA3.1 vector, carrying a hygromycin resistance box) were constructed at NV Organon (Oss, the Netherlands). These cells were constructed in the following manner. The full length cDNA for human CYP11B1 was obtained by PCR from human adrenal cDNA as described by Kawamoto et al [61] and cloned into pPCR SCRIPT (Stratagene, La Jolla, USA). After digestion with Xho I and Not I, the obtained cDNA was inserted into a Xho I/Not I digested pcDNA3.1 vector (InVitrogen, Breda, Netherlands), carrying a hygromycin resistance box. Following transfection to the V79 cells, positive cells were selected based on their hygromycin resistance. The presence of the CYP11B1 gene was confirmed by PCR and the presence of 11β-hydroxylase activity (i.e., the ability to produce corticosterone or cortisol from 11-deoxycorticosterone or 11-deoxycortisol respectively). Although no electron-transporting proteins were co-transfected, the CYP11B1 expressing V79 cells showed abundant 11β-hydroxylase activity, as was previously also shown by Denner et al. for both CYP11B1 and CYP11B2 [62]. In normal (non transfected) V79 cells, no detectable 11β-hydroxylase activity or PCR signals were found. V79 cells stably (over)expressing CYP11B2 were developed in the laboratory of Prof. R. Bernhardt, Institute of Biochemistry, Saarland University, Saarbrücken, Germany [62–64].
Cells were cultured under standard conditions in DMEM/FK12 medium (Gibco, Gaitersburg, USA) supplemented with 10% foetal calf serum (Hyclone, Logan, USA), penicillin/streptomycin (100 U/mL and 100 μg/mL, respectively, Gibco), inside a humid environment of 37 °C, and 5%CO2 atmosphere. For assessing inhibitor potencies, cells were transferred to 12 well plates and grown until they were confluent. Next, cells were incubated for 1 h in serum-free medium with cumulative inhibitor concentrations, followed by addition of 500 nM (0.5 times Km) 11-deoxycorticosterone (Sigma-Aldrich, St Louis, USA) as substrate. Finally, V79 CYP11B1 cells were incubated for another 1 h and V79 CYP11B2 cells for another 3 h. Medium was removed and the (produced) steroids were extracted with 5 mL diethylether from 1 mL medium aliquots, mixed with 1 mL 1 M sodium-glycine buffer (pH 10.5) containing 500 nM methylprednisolone (Sigma-Aldrich) as internal standard. Extracts were dried under nitrogen and dissolved in mobile phase for automated HPLC analysis, using a stationary phase consisting of an MR column (4.6 × 50 mm, particle size 2.5 μm) and a mobile phase consisting of a mixture of 680/320/1 (v/v/v) milliQ water, acetonitrile and trifluoro-acetic acid. Detection of UV absorption was performed at 243 nm (Shimadzu, Tokyo, Japan).
Results and discussion
Model quality
One of the theorems often applied to the quality assessment of a protein model is the Ramachandran Plot, which is an indicative measure for the correctness of the residue torsion angles. The plot is a graphic display of torsion angle φ (Cn−1–Nn–Cα,n–Cn) versus torsion angle ψ (Nn–Cα,n–Cn–Nn+1) for each residue of the protein of which the secondary structure character of the residue can be extracted. The alpha-helix character of a protein backbone is located roughly in the region where −60° < φ < −30° and −120° < ψ < −30°, and the beta-sheet character is located roughly in the region where −180° < φ < −60° and 90° < ψ < 180° [65]. In Table 2, the results of the Ramachandran Plot are summarised for both hCYP11B1 and hCYP11B2 models after they were equilibrated with the ligand 18-hydroxycorticosterone.
Table 2Validation results for the lowest energy models of CYP11B1 and CYP11B2 and the crystal structures which were used for the template, part IRamachandran Plot (core regions) (%)Ramachandran Plot (favourable regions) (%)Errat2 (quality factor, %)aVerify 3D (total score)bTemplate (PDB, resolution)CYP101 (2CPP, 1.63 Å)92.1100.096.0197CYP2C5 (1NR6, 2.10 Å)87.899.293.6195ModelhCYP11B178.894.784.1126hCYP11B278.794.787.5125hCYP11B2-TripMut80.696.581.1117rCYP11B179.796.580.2113rCYP11B282.496.580.1114a Outliers are positioned in the loop regions before alpha-helix F where the two template structures connect and at the end of alpha-helix D where a large insertion was introducedb Additional outliers are positioned at the end of alpha-helix G
For both hCYP11B1 and hCYP11B2 around 95% of the residues are positioned in the favoured and core regions of the Ramachandran Plot, indicating that for hybrid models, the structures are of acceptable quality. Due to the high quantity of alpha-helices and beta-sheets, the majority of residues is positioned in the expected regions. The residues which are situated in disallowed and unfavoured regions of the plot, are located in loop regions outside the active site. In total, 9 residues in the hCYP11B1 model are situated in the disallowed regions and 15 residues in the unfavoured regions. For the hCYP11B2 model, 10 residues are situated in disallowed regions and 14 residues in the unfavoured regions. The causes for these disparities are several insertions or deletions introduced in the models for which the structural minimisation was not sufficiently adequate to correct the backbone dihedrals. In particular, these regions are a relatively large insertion between alpha-helix D and beta-sheet 3-1, and an insertion between helix G and H.
Additionally, the amino acid environment of the models was evaluated with Errat [66] and Verify3D [67] (Table 2). By comparing the results, it can be seen that the models all score equally well but are less accurate than the template structures. An Errat quality factor of 95% is expected for crystal structures resolved at a resolution of 2.5 Å. Using the Errat score per amino acid we were able to locate the deviations in the loop regions at the end of helix E where our two template structures connect and at the end of helix D where a large insertion was introduced. It is likely that in the minimisation protocol with tethered heavy atoms, the connectivity between template structures has not been fully optimised. Investigation of the amino acid environment with Verify3D resulted in similar conclusions as found with Errat and the Ramachandran data. The low scoring of Verify3D can be attributed to a bad folding of the regions around the helices D and F, and additional potential errors were located at the end of helix G, where again the two templates have been connected to each other.
Finally, the stereochemistry of the models was analysed with MOE (Table 3 [68]). Except for the already determined structural flaws in the regions around helices D, F and G, no additional flaws were found using these measurements, indicating that the overall fold of the protein is of acceptable quality.
Table 3Validation results for the lowest energy models of CYP11B1 and CYP11B2 and the crystal structures which were used for the template, part IIMOE-protein reportObserved CYP101 (2CPP)Observed CYP2C5 (1NR6)Observed CYP11B1 modelObserved CYP11B2 modelReference valuesaParameterMeanS.D.MeanS.D.MeanS.D.MeanS.D.MeanS.D.Trans-Omega176.62.7178.90.9172.07.7172.07.7180.05.8C-alpha chirality32.83.634.31.730.811.130.810.833.84.2Chi1-gauche minus−63.017.4−63.314.8−62.321.7−62.622.1−66.715.0Chi1-gauche plus55.420.756.316.451.526.753.028.364.115.7Chi1-trans185.313.3184.212.9186.321.9186.820.3183.616.8Helix phi−65.211.9−67.415.5−60.719.7−61.119.8−65.311.9Helix psi−41.216.5−37.617.9−42.525.1−41.925.2−39.411.3Chi1-pooled S.D.15.513.822.521.615.7Proline phi−65.811.3−61.99.5−65.319.6−67.620.5−65.411.2Dihedral outliers041517Bond angle outliers0034Bond length outliers0001Results were generated with the MOE module: protein Eval. The thresholds were chosen to be 5 for the Z-Score and 70 for the vanderWaals contactsa Reference values were published in a statistical survey of the high-resolution data in the Protein Data Bank [68]
None of the structural flaws are located at residues in the active site or at residues lining the active site. The occurrence of Ramachandran errors and problems with modelling external loop regions seems an inevitable circumstance in homology modelling [69, 70] and will probably not pose a problem for modelling protein-ligand interactions. All the abovementioned errors are expected to be alleviated during further investigation of protein-ligand interactions using molecular dynamics simulations. Therefore the current models have been selected as appropriate starting points for further analysis.
Protein-substrate interactions
After quality assessment of the created protein models encompassing the various ligands, we investigated the steroid binding mode. All docking results from GOLD favoured the β-side of the steroid oriented to the heme with the C3-carbonyl pointing towards alpha-helix B′ (Fig. 4). Visual inspection revealed that in case of reverse orientation of the steroid, unfavourable clashes of the C3-carbonyl in the opposite side of the pocket would occur. After docking the steroids into the active site, the protein–ligand complexes were subjected to MOE minimisation as described in the homology modelling section. After minimisation, the distance between the heme iron atom and the different substrate hydroxylation sites (C11, C18 and C19) was measured. The results are summarised in Table 4.
Fig. 4Hypothetical binding of 18-hydroxycorticosterone (18OH-B) inside the CYP11B2 active site for the synthesis of aldosterone. The heme prosthetic group contains a bound oxygen atom needed for catalytic function. The CYP11B2 model possesses Arg123 in alpha-helix B′ which is stabilised by Glu310 in alpha-helix I. 18OH-B possesses several hydrogen bonds: one internal hydrogen bond between the C18-hydroxyl and the C20-carbonyl, two hydrogen bonds between the C21-hydroxyl and the backbone carbonyls of Gly379 and Phe381, and finally a hydrogen bond between the C3-carbonyl and Arg123Table 4Hydroxylation distance table (iron atom–carbon atom) after minimisation with MOE (distances in Angstrom)hCYP11B2hCYP11B2-TripMutHCYP11B1rCYP11B1rCYP11B2C11C18C19C11C18C19C11C18C19C11C18C19C11C18C19DOC4.724.305.614.374.655.324.304.565.484.304.754.834.704.245.5418OH-DOC4.334.305.42a4.314.515.21b4.314.605.19b4.304.685.17b4.324.315.39aB5.394.065.465.374.405.225.434.395.285.334.494.945.284.215.2018OH-B4.864.215.50a5.424.645.29c5.384.625.26d5.474.625.28d5.294.355.29aa Ligand C18-hydroxyl group forms a hydrogen bond with the C20-ketone group of the ligandb Ligand C18-hydroxyl group forms a hydrogen bond with the iron-oxygen of the proteinc Ligand C18-hydroxyl group forms a hydrogen bond with the C11-hydroxyl group of the ligandd Ligand C11-hydroxyl group forms a hydrogen bond with the C18-hydroxyl group of the ligand
All ligands showed two very distinct interactions in the modelled active site cavities. Firstly, the ligands possess a steric fit for the C20-carbonyl and the C21-hydroxyl in a small cavity created between helix K and beta-sheet 6-1. Inside this cavity, the C21-hydroxyl group possesses two hydrogen bonding interactions with the protein backbones of Gly379 and Phe381 (Fig. 4). The presence of these amino acids in the active site cavity coincides with the models of Belkina et al., but for those models, no interactions between protein and ligand were discussed [34].
Secondly, the ligands possess an interaction between the C3-carbonyl and active site residue Arg123 in helix B′. Arg123 is stabilised by Glu310 in helix I, which has further stabilising interactions with the protein backbone. Glu310 coincides with an aspartic acid in the CYP2 family which from visual inspection of the crystal structures of CYP2C5 (pdb-code 1NR6) and CYP2C9 (pdb-code 1OG5) seems to play a specific stabilising role in the active site structure; it is expected that Glu310 has the same stabilising role. The difference in chain length between an aspartic acid and a glutamic acid also determines the flexibility of helix B′. It can move 1.5 Å further out of the active site cavity in all CYP11B models allowing the steroid to fit parallel to the heme. The ligands also possess many hydrophobic interactions in this region, particularly with Phe130 (Fig. 4). Due to the close interactions with the A and B rings of the steroid skeleton, this amino acid might play an important role in substrate stabilisation. The presence of Phe130 in our models coincides with the models of Ulmschneider et al. [35] where it is seemingly involved in ring stacking with their inhibitors.
There are also two striking differences between the CYP11B1 and CYP11B2 models. Firstly, the active site cavity near beta-sheet 6-1 is smaller in both CYP11B2 models compared to the CYP11B1 models, with rCYP11B1 being the largest. This small cavity is formed by the loop region between helix K and beta-sheet 6-1 (Fig. 5). Comparing this cavity in both the hCYP11B1 and hCYP11B2 models, we find that in hCYP11B1 this cavity has a 9–14% larger volume, dependent on amino acid flexibility as well as the probe radius chosen for calculation of the volume (differing from 1.0 to 1.5 Å). This difference in size is caused by the folding of Leu407 which induces a 1.0 Å outward shift of the loop region. A comparison of the models with the crystal structures of the CYP2 family reveals that the loop region in the CYP11B models is in closer proximity to helix I (by 1.5 Å). This contact is defined by two amino acids having relatively smaller sidechains (Pro322 and Val378 in CYP11B) than observed in the CYP2C family (generally Thr/Val and Leu/Ile). Changes in helix I such as the mutant A320V between hCYP11B1 and hCYP11B2 are likely to have a direct influence on the folding of this region.
Fig. 5Active site volume difference between the hCYP11B1 (white) and hCYP11B2 models (orange). Shown for both hCYP11B1 and hCYP11B2 are the active site volume and the backbone trace. R-fadrazole and S-fadrazole are indicated with blue and purple respectively. It is clear to see that hCYP11B1 contains a larger active site between helix I and sheet 6-1. This cavity allows S-fadrazole to fit the cavity, but not R-fadrazole (black arrow). On the other side of the active site near Helix B′ and Arg123, hCYP11B2 contains the larger cavity, which might rationalise the better fit of R-fadrazole in the cavity
Secondly, helix B′ is shifted outward by 1.1 Å in both CYP11B2 models (the backbone RMSD of helix B′ of hCYP11B1 and hCYP11B2 is 2.13 Å). The change in spatial positioning of this helix is probably caused by the mutants L301P and E302D in helix I, but because this region is on the surface of the protein, the exact cause is less clear. In the hCYP11B2 model, Glu302 contacts helix B′ at Lys127, which forms the counter charge for the C-terminal end of this helix. Hasemann et al have posed that movement of helices B′, F and G is involved in the opening of the active site cavity [71]. Changing the stabilising environment of helix B′ has a likely effect on the structural stability of the helix and may result in a loss of activity.
Introducing the triple mutant investigated by Bottner et al. [36] by means of the hCYP11B2-TripMut model showed that the hydroxylation distance pattern of the substrates shifts as expected from hCYP11B2 to that of hCYP11B1. Both hCYP11B1 and hCYP11B2-TM display similar active site cavities near beta-sheet 6-1 (backbone RMSD 1.43 Å) with only slight deviations in the hydroxylation distances for the ligands (Table 4). This confirms the behaviour of the triple mutant in enzymatic activity found by the study of Bottner et al. To speculate further, Bottner et al have shown that the A320V mutant alone or the L301P/E302D mutant alone, is not enough to significantly change aldosterone synthesis [36]. Only mutation in both regions results in an almost complete loss of the aldosterone synthesis capabilities of hCYP11B2. It is likely that the subtle changes on both sides of the active site go hand-in-hand to (nearly) completely convert the activity of hCYP11B2 into that of hCYP11B1.
For DOC, our results indicate that hCYP11B1 and rCYP11B1 preferentially catalyse C11-hydroxylation, whereas hCYP11B2 and rCYP11B2 preferentially catalyse C18-hydroxylation. For both CYP11B2 models, the shifted alpha-helix B′ causes DOC to present its C18 closest to the iron as it is repositioned by the strong interaction of the C3-carbonyl group with Arg123 in alpha-helix B′. Additionally, the larger active site of rCYP11B1 around beta-sheet 6-1 allows DOC to fit further into the niche presenting C19 into hydroxylation range (distance less than 5.0 Å). A point of argument against a preferred C18-hydroxylation of DOC by CYP11B2 is that in vitro measurements indicate higher levels of the C11-hydroxylated product B [19, 21]. However, both C11- and C18-hydroxylated products of DOC can be promptly consumed as a substrate for the production of 18OH-B and subsequently aldosterone, in in vitro [19, 21]. This apparent discrepancy between observed and predicted regioselectivity of DOC hydroxylation by CYP11B2 may possibly indicate that other factors than hydroxylation distances are involved in the formation of the actual products.
For 18OH-DOC, the C18-hydroxyl group forms a hydrogen bond with the iron–oxygen limiting hydroxylation to C11 which is in closest proximity to the iron. In the CYP11B2 models both C11 and C18 are in approximately equal distance to the iron and an internal hydrogen bond is formed by the substrate between the C18-hydroxyl group and the C20-carbonyl. Although both C11- and C18-hydroxylation are shown to be possible in such a complex, oxidation on the unsubstituted C11 is likely to form a more stable reaction intermediate.
In all models, the two C11-hydroxylated ligands B and 18OH-B only portray C18 in close proximity to the iron, as the C11-hydroxyl group blocks access of C11 to the heme iron. The positioning of the C18-hydroxyl group of 18OH-B in the active site cavity appears to determine conversion into aldosterone. In both CYP11B2 models, the natural substrate is shifted above the heme, which creates a slightly larger active site cavity near beta-sheet 6-1 for the C18-hydroxyl group to rotate in. This difference in size allows room for an internal hydrogen bond between the C18-hydroxyl group and the C20-carbonyl for both 18OH-DOC and 18OH-B. In rCYP11B1 and hCYP11B1, the C18-hydroxyl group forms a hydrogen bond with the heme oxygen atom which blocks C18 for hydroxylation (Fig. 4). Interestingly, in the hCYP11B2-TripMut, the C18-hydroxyl group forms an internal hydrogen bond with the C11-hydroxyl group, and thereby also blocks the C18 for hydroxylation.
In conclusion, we propose that the immediate folding of the active site around the substrate C18-hydroxyl group may be the key difference between the two isoforms leading to the production of aldosterone by CYP11B2 and not by CYP11B1. To speculate further on the mechanism of the enzymatic formation of aldosterone from 18OH-B, we hypothesise that a stable internal hydrogen bond for the C18-hydroxyl group of the ligand may be essential to stabilise its C18 against further oxidation. A second oxygen atom can then be inserted between C18 and one of its hydrogen atoms, forming a C18-gem-diol intermediate which forms aldosterone by elimination of water. Although the formation of a C18-gem-diol has been proposed by Johnston et al. [72], the precise mechanism of oxidation is still unknown. This proposed method of substrate stabilisation may explain the regioselectivity of the two CYP11B isoforms and can be attributed to the subtle changes observed on two sides of the active site cavity; the cavity near helix K and the positioning of helix B′.
Protein-inhibitor interactions
The non-steroidal inhibitors, metyrapone, R-etomidate, R-fadrazole and S-fadrazole were docked flexibly into the active site of the human CYP11B models where their aromatic nitrogen atom forms a strong interaction with the sixth ligating position of the heme iron atom. The accessible electron lone pair of the heterocyclic nitrogen atom is required for heme iron complexation, a well-known interaction for non-steroidal CYP inhibitors [73, 74].
The GOLD docking results showed all ligands to bind in one favourable conformation. The enantiomers R-fadrazole and S-fadrazole mainly favoured one particular orientation in the active site cavities of hCYP11B2 and hCYP11B1 respectively. The higher affinity of S-fadrazole to hCYP11B1 and R-fadrazole to hCYP11B2 can be attributed to the active site difference near beta-sheet 6-1 and helix I (Fig. 5). The difference in affinity is determined by the steric aspects of the active site cavity allowing only S-fadrazole to fit in the active site of hCYP11B1, where R-fadrazole possesses a steric clash (black arrow). In hCYP11B2, it is less clear why R-fadrazole fits the active site best, but this is mainly determined by steric effects near helix B′. Due to the presence of rotatable bonds, R-etomidate was able to dock in different orientations, but favoured one particular conformation with its phenyl ring pointing into the direction of helix B′. The main flexibility observed was caused by the ethyl-ester group and the docking score of those conformations was very similar. Metyrapone was able to dock with both pyridine rings to the iron, but mainly favoured a conformation with the carbonyl facing Arg110. Since all ligands favoured only one orientation, it was this orientation that was investigated by MD.
Our in vitro test results show a striking enantioselectivity of fadrazole binding for the CYP11B family. We find R-fadrazole to be the most active enantiomer for CYP11B2 as well as being selective (Table 5). In contrast, for CYP11B1 it is S-fadrazole which appears to be the most potent enantiomer. A similar stereoselectivity as for CYP11B1 has also been described for aromatase, wwith the S-enantiomer being the better aromatase inhibitor [75]. The docking results from both Goldscore and Chemscore predict the same enantioselectivity for fadrazole in the CYP11B family as observed in vitro, indicating that the three dimensional models contain promising accuracy for the valid prediction of enantiomer selectivity. Docking results of the two other inhibitors metyrapone and R-etomidate confirm the trend of inhibitor potency as determined by the in vitro experiments (Table 5).
Table 5Correlation of docking and molecular dynamics results to in vitro data for both human CYP11B1 and CYP11B2 modelsIC50 (nM)GoldscoreΔG Goldscore (kcal/mol)ΔG Chemscore (kcal/mol)Unon-bonded (kcal/mol)hCYP11B1Metyrapone46.4 ± 10.457.33−8.43−8.73−48.4 ± 3.7R-Etomidate0.5 ± 0.266.21−9.38−9.25−56.0 ± 2.4R-Fadrazole118.6 ± 8.954.01−8.07−8.14−38.4 ± 2.5S-Fadrazole39.5 ± 4.456.67−8.36−8.77−56.3 ± 3.4hCYP11B2Metyrapone207.8 ± 4.549.99−7.64−7.95−36.2 ± 7.5R-Etomidate1.7 ± 0.965.21−9.28−9.21−54.4 ± 2.9R-Fadrazole6.0 ± 1.963.20−9.06−9.38−55.9 ± 3.3S-Fadrazole171.2 ± 51.753.81−8.05−8.12−44.3 ± 1.8Indicated are the Goldscore and the extracted Goldscore binding free energy as well as the Chemscore binding free energy. Unon-bonded indicates the total non-bonded energies between the protein and the ligand for the molecular dynamics simulations
Our human CYP11B1 and CYP11B2 models are able to rationalise the (inverse) enantioselectivity of CYP11B1 and CYP11B2 inhibition by fadrazole as observed in the in vitro tests. The enantioselectivity can be attributed to the small cavity size differences of the models. Both enantiomers bind with a steric fit comparable to the A, B and C rings of the endogenous substrates, where the cyanobenzyl moiety forms a π–π stacking interaction with Phe130. In addition to the steric interactions, both enantiomers possess a strong polar interaction between their cyanide moiety and Arg123, mimicking the interaction that the C3-carbonyl of the natural substrates possesses with this amino acid.
Metyrapone was not found to possess any hydrogen bonding or polar interaction with the active site other than the aromatic nitrogen. Instead it possesses a pronounced steric fit with its second aromatic ring overlapping the space occupied by the A-ring in case of the steroidal substrate. R-etomidate, which was not designed as a CYP11B specific inhibitor, does possess a hydrogen bonding interaction. The hydrogen bond is made between its ester moiety and the catalytic Thr318, and is present in both hCYP11B1 and hCYP11B2. This may explain the non-selective inhibitory action of R-etomidate on both isoforms. Its aromatic phenyl ring also possesses strong steric interactions, overlapping the same space of the steroidal A-ring like metyrapone.
Molecular dynamics
Given the promising prediction of the models for substrate conversion, substrate binding, inhibitor binding and the satisfying Ramachandran Plots, we conducted molecular dynamics simulations to obtain a better understanding of the dynamic behaviour of the inhibitors in the active site of the enzyme models. To investigate the structural integrity change during the simulation, we calculated the relative root mean square deviation (RMSD) over all heavy atoms (Fig. 6).
Fig. 6RMSD for the molecular dynamics simulations of the different protein-ligand complexes of the hCYP11B1 and hCYP11B2 models. Metyrapone is coloured black, R-etomidate green, R-fadrazole red, and S-fadrazole blue
During the first 500 ps, the RMSD increased and the protein still adapted towards its optimal conformation. After this point in time hardly any change in the three dimensional structures of the proteins was observed. The largest fluctuations of the protein were found in the flexible regions with peak values located in the structures around alpha-helix D (not shown). In the random coil following alpha-helix D we introduced a large insertion of seven amino acids, which elongates the alpha-helix by one turn before it connects to the following beta-sheet. Inside the water box, this region is found to protrude into the water without any stabilising protein interactions and unfolds due to interaction with water. In all the simulations we observed an opening of the active site and the continuous flow of water molecules in and out of the active site cavity. Several water molecules retained key positions, such as the water molecules that make up the channel towards the conserved Glu459 (not shown).
All the inhibitors maintained the same interactions as observed in the docking study. In Fig. 7, the resulting poses of the inhibitors metyrapone and R-etomidate are displayed. Both possess a ring stacking with Phe130 in the active site, and R-etomidate also possesses a hydrogen bond with Thr318. Arg110 stabilises the heme in the active site. In Fig. 8, the poses of R-fadrazole in the CYP11B2 active site and S-fadrazole in the CYP11B1 active site are compared. Here it can be seen that in the CYP11B2 model, alpha-helix B′ is moved further out of the active site. In the CYP11B2 model, Arg123 possesses a polar interaction with Glu310, whereas Phe130 provides a horizontal ring stacking for R-fadrazole. In the CYP11B1 model, S-fadrazole possesses the same interactions with the protein, although the ring stacking with Phe130 is vertical rather than horizontal. Importantly, the interaction points of the fadrazole enantiomers and those of the natural substrate coincide very well.
Fig. 7CYP11B2 active site model with metyrapone (purple) and R-etomidate (blue) in the active site. Indicated are the stabilising interaction between Arg110 and the heme, as well as Phe130 which accommodates a ring stacking with the ligands. The ester group of R-etomidate possesses a hydrogen bond with Thr318Fig. 8CYP11B1 model (orange) containing S-fadrazole (purple) in the active site and CYP11B2 model (white) containing R-fadrazole (blue) in the active site. Indicated, the stabilising interaction between Arg110 and the heme, as well as the stabilising hydrogen bonds between Arg123 and Glu310. Phe130 accommodates a ring stacking with the fadrazole enantiomers; horizontally for the R-enantiomer and vertically for the S-enantiomer
R-fadrazole and S-fadrazole displayed different behaviour in the other protein models. R-fadrazole lost its direct interaction with the heme iron atom in the hCYP11B1 model, exchanged it for a water molecule and drifted to the top of the active site. S-fadrazole displayed a similar behaviour in the hCYP11B2 model. Repeated simulations did not change either scenario, indicating that unlike the eutomers, the distomers do not comfortably fit our starting models.
Because the complex was stable over the last 500 ps, we sampled the non-bonded interactions between the protein and the ligand for this time period (Table 5). These interactions contribute to the binding free energy, although the solvation effects of free ligand still need to be subtracted [76]. These energies follow the same generic trend which emphasise the strong interactions for R-etomidate in both hCYP11B models and the strong interactions of S-fadrazole in hCYP11B1 and R-fadrazole in hCYP11B2 respectively.
Conclusion
We have constructed homology models of the two isoforms of the human CYP11B family, as well as promising starting models for the rat isoforms. These models are based on the knowledge of substrate specificity, which were defined as the differences in hydroxylation distances between the active site and the C11, C18 and C19 of the steroid skeleton. As both hCYP11B1 and hCYP11B2 possess virtually similar active sites, the steric fit of the different steroidal ligands seems to be one of the strongest determinants for substrate specificity.
We have found that within the active site of the hCYP11B2 model, the endogenous ligand 18-hydroxycorticosterone forms a stabilising internal hydrogen bond. This is not observed for the hCYP11B1 model, which might rationalise why 18-hydroxycorticosterone is solely a substrate for CYP11B2 to yield aldosterone. In addition, the interactions of the C3-carbonyl with Arg123 in alpha-helix B′, the interaction of the C21-hydroxyl with the backbones of Gly379 and Phe381, and the presence of Phe130 are also important for the stabilisation of the ligand in the protein active site. Determinant for the interaction of Arg123 is the presence of Glu310 in alpha-helix I, which stabilises Arg123 and alpha-helix B′ in the active site.
The known non-steroidal CYP11B inhibitors metyrapone, R-etomidate, R-fadrazole and S-fadrazole were postulated to occupy the same space in the active site as the endogenous substrates. Conclusions from both molecular docking and molecular dynamics simulations corroborate the measured activity data from in vitro experiments, supporting the validity of the constructed models for these ligands. Importantly, these models rationalise the enantioselectivity of fadrazole, with the R-enantiomer being most potent on CYP11B2 and the S-enantiomer being most potent in CYP11B1.
The constructed models are useful tools in trying to understand some of the molecular mechanisms involved in ligand binding and substrate conversion for the CYP11B family. As such, these models might also be appropriate tools for more detailed protein-inhibitor modelling studies as well as for ligand design or database screening, following further model optimisation and model tuning.
|
[
"cyp11b2",
"cyp11b1",
"fadrazole",
"molecular docking",
"molecular dynamics",
"aldosterone synthesis",
"homology modelling"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Crit_Care-8-1-420054
|
Clinical review: Outreach – a strategy for improving the care of the acutely ill hospitalized patient
|
We examined the literature relating to the safe care of acutely ill hospitalized patients, and found that there are substantial opportunities for improvement. Recent research suggests substantial benefit may be obtained by systems of outreach care that facilitate better integration, co-ordination, collaboration and continuity of multidisciplinary care. Herein we review the various approaches that are being adopted, and suggest the need for continuing evaluation of these systems as they are introduced into different health care systems.
Introduction
Health care providers in many countries are increasingly aware that quality of care can be improved by measures to reduce errors. Evidence suggests that the greatest opportunities for improvement may be in the management of the acutely ill hospitalized patient, where uncertainty, urgency and lack of integration substantially increase the risk of errors leadinf to adverse outcomes. Different approaches to this problem have evolved, including hospitalists in the USA, medical emergency teams in Australia, and outreach care in the UK. Critical care has a central role to play in all three developments, which share the common aim of improving the safe care of acutely ill patients as they travel through the health care system. We will review the background, methods, roles and benefits of these various systems which we group under the heading of 'outreach' care.
Method
We focused on identifying research publications that examined ways of improving the integration of critical care with acute care services as a means of improving the safe care of acutely ill hospitalized patients. Publications were initially identified by an electronic search of Medline and Cinahl, and the cited references provided additional material. The initial date range searched was 1995–2003 to ensure that current research and up to date literature was reviewed. However, this produced only a limited number of references, and the time period was therefore extended to 1990. Keywords searched singularly and in combination were 'acute pain team', 'suboptimal care', 'patient at risk', 'critical illness' and 'cardiopulmonary resuscitation'. Only English language references were included in the analysis.
Background
Trends in hospital care
Hospitals are becoming increasingly complex environments because of developments in medical technology, more potent treatments, and an ageing and dependent population of patients. The proportion of emergency admissions continues to rise in most countries [1], whereas the stock of hospital beds has fallen. Combined with political demands for cost containment, this has resulted in shorter hospital stays, and increased bed occupancy and throughput [2]. At the same time there has been a trend toward greater transparency in decision making, rising public expectations, and improved public access to information about quality of care and outcomes from medical interventions. These changes have coincided with increasing difficulty with the recruitment and retention of trained nurses [3], and constraints on doctors' hours of work. The implementation of the European Working Time Directive [4], and similar trends in the USA [5], means that proportionately fewer staff are available to manage this increased workload, particularly the out-of-hours and emergency elements. Shorter training times reduce the expertise acquired through apprenticeship. The net effect is that sicker patients are receiving care from fewer and less experienced staff.
Error and quality in health care
These pressures have been accompanied by a growing awareness of the problem of error in health care and an exponential increase in litigation. The much quoted Institute of Medicine report 'To Err is Human' [6] estimated that between 44 000 and 98 000 patients die each year in the USA as a result of preventable clinical errors. Studies in Australia and the UK suggest that around 10–16% of hospitalized patients experience an adverse event related to clinical care, with a mortality rate in these patients of 5–8%, and overall financial costs in the range US $4.7–29 billion [7-9]. Clinical error is now widely recognized as a systems problem; that is, adverse events affecting individual patients are usually caused by a sequence of events in the macro- and micro-environment that involve deficiencies in the structure and organization of health care, and are not simply a consequence of human error by the responsible clinician who is the terminal link in the chain. To improve safety and quality in health care requires a systems approach in which all participants – politicians, administrators and health care professionals – assume responsibility for patient outcomes. This ambitious task requires focusing of efforts on those patients most at risk.
Identifying at-risk populations
Acutely ill patients in general
The risk of error and adverse outcomes would appear to be higher in the context of acute and emergency care, for example in emergency departments [7] or in general medical (rather than elective surgical) admissions [8]. The elderly are more susceptible [7]; they are also more likely to be admitted as emergencies and exposed to emergency surgery [10]. Inexperienced clinicians and unsupervised trainees (who often deliver first-line care out-of-hours) have a higher error rate [11,12]. The risk for an adverse event increases by approximately 6% per day for patients admitted with emergency conditions [13], and is much increased in severely ill patients who undergo life-saving invasive interventions [14]. Discontinuities in care [15] created by shift working or poor information transfer contribute to error by failing to recognize trends in a deteriorating patient's condition.
Postoperative patients
In the UK the 1993 National Confidential Enquiry into Perioperative Deaths showed that two thirds of perioperative deaths occurred 3 or more days after surgery when the patient had been returned to the ward. The majority of these deaths were from cardiorespiratory complications, and many were considered preventable by earlier identification and treatment [16]. The 1999 report [17] analyzed deaths within 30 days of a surgical procedure in elderly patients (>90 years), and found that suboptimal fluid management was a major cause of serious postoperative morbidity and mortality in this group. It recommended more accurate monitoring and recording of fluid balance and earlier recognition and correction of problems as a means of reducing the incidence of postoperative complications. The report also identified deficiencies in multidisciplinary care, despite the high incidence of comorbid disease in these elderly patients, and recommended more collaborative working between surgeons, anaesthetists and physicians with expertise in the care of the elderly.
This type of audit based on a large observational database is essential for identifying current practice and opportunities for improvement. However, recommendations were based on peer review and data from questionnaires provided by assessors who were unblinded to clinical outcomes, examining only those patients who died – there are no denominator data.
Cardiopulmonary arrest
Hospitalized patients who undergo cardiopulmonary resuscitation commonly exhibit premonitory signs and symptoms many hours before the cardiac arrest [18-21]. Schein and coworkers [18] studied 64 patients following cardiac arrest, and reported that 84% of vital signs charts showed an acute deterioration in the patient's condition prior to arrest. Franklin and Mathew [20] examined the case notes of 150 consecutive patients who had suffered a cardiac arrest on a general ward, and found documented prior clinical deterioration in 99 (66%). A common finding was the failure of the nurse to notify a physician of deterioration in the patient's condition.
The interface with critical care
Critical illness increases the opportunity for clinical error [22,23], and this is related at least in part to the complexity of diseases, the multiplicity of therapies, frequent invasive interventions and, within the intensive care unit (ICU), the intensity of monitoring and observation, which may paradoxically increase the apparent error rate in this environment simply by improved detection. Iatrogenic complications are a common cause for ICU admission [24], and suboptimal care before referral to intensive care is associated with a markedly increased mortality [25]. Premature discharge from intensive care of patients recovering from critical illness is also associated with a markedly increased hospital mortality [26], supporting the view that organizational aspects of clinical care profoundly influence patient outcomes [27]. Common errors include lack of attention to detail, poor communication, failures of organization, lack of knowledge, failure to appreciate clinical urgency, insufficient supervision and failure to seek advice.
Treatment limitation decisions, futility, and end of life care
Studies in the USA and Europe have demonstrated considerable diversity of practice and opportunities for improving autonomy, patient centred decision making, and quality of end-of-life care [28,29]. The Study to Understand Prognosis and Preferences for Outcomes and Risks of Treatment in the USA [29] demonstrated that 31% of the cohort of patients admitted to intensive care would have preferred not to be resuscitated, but that clinicians were aware of this preference in only 47% of the subset and in half it had not been documented. Aarons and Beeching [30] surveyed the use of 'do not resuscitate' orders in a community hospital in the UK and suggested that end-of-life care of dying patients and their families in hospital could be substantially improved in terms of symptom relief, communication and respect for the patients' wishes. They also concluded that poor decision making by health care workers could be improved by education. One consequence of a lack of communication between staff and a reluctance to discuss these difficult issues with patients and families is the inappropriate resuscitation and subsequent admission to the ICU of patients for whom further intervention would be futile and add to the burden of suffering.
A more appropriate approach to management of acutely ill patients at risk for critical illness or cardiac arrest would be to prevent these complications, or agree treatment goals and limitations, by earlier recognition of simple warning signs. We consider the various approaches that are being adopted below.
Systems for earlier recognition and management of patients at risk
The problems described above require a systems approach to improving the safe care of the acutely ill hospitalized patient. The USA, the UK, Australia, Canada and Denmark are instituting national systems for improving patient safety [31-33]. Within this general framework, three countries have specifically identified the need to improve the care of acutely ill patients. In the USA the Leapfrog Group [34] has recommended that ICUs be managed by intensivists – doctors specifically accredited in critical care medicine – while in a parallel development many hospitals are appointing 'specialist generalists' ('hospitalists') to provide inpatient care on the wards [35]. Australia has promoted the establishment of medical emergency teams (METs) led by doctors as an alternative to cardiac arrest teams [36,37]; and the UK has implemented the recommendation of the expert group report 'Comprehensive Critical Care' [38] to establish multidisciplinary outreach care. All these developments are based on the concept of earlier intervention by people with appropriate knowledge and skills in managing acutely ill patients. Is this concept valid, and which model is the best?
Structures and processes for early intervention
In the UK a survey of intensive care facilities found that, in 94% of the units questioned, staff regularly visited wards in response to requests for advice from medical and nursing members of the admitting team [39], demonstrating that this is clearly an important role for intensive care staff. However, there is little information describing the experiences and perspectives of ward based staff who care for acutely ill patients outside the intensive care environment. Gibson [40] found that ward staff often lacked confidence and felt ill prepared to deal with acutely ill unstable patients, and that they experienced increased stress and anxiety. Such emotions are unlikely to enhance staff retention or reduce sickness rates. It therefore makes sense to put in place systems that improve the support not only of sick patients but also of the staff responsible for their care, and that reduce discontinuities in clinical care. The key to this is empowerment through adequate resourcing and training, and simple methods of clinical monitoring. Three models are currently employed: hospitalists in the USA, the MET in Australia, and outreach care in the UK.
Hospitalists
Hospitalists are internists who specialize in acute hospital medicine. This new speciality has appeared in the USA during the past 8 years in response to perceived difficulties with primary care clinicians maintaining continuity of care for their patients admitted to hospital. Currently at around 5000, their numbers are expected to increase rapidly. Hospitalists are usually salaried employees of managed care organizations, which favour their development as a means of reducing costs and duration of hospital stay [41,42]. The training of hospitalists is rooted primarily in internal medicine, but there is no national core curriculum. They do not appear to have responsibility for surgical patients, and therefore cannot be considered generalists in the sense of dealing with all acutely ill patients. The relationship between hospitalists and intensivists has not been defined [43], although there clearly are opportunities for interaction. Hospitalists occupy a role that is close to that of the general physician in UK hospitals, except that the latter group often have a subspeciality.
This model is suited to the system of care in the USA, where primary care physicians often have continuing responsibility for inpatient care. It might also be of value to other health care systems in which increasing specialization is creating a need for generalist acute care clinicians in hospital practice.
Acute pain teams
A joint expert committee report from a Royal College of Surgeons and College of Anaesthetists working party [44] reviewed the evidence related to postoperative pain management and recommended the establishment of an acute pain service in all major hospitals. However, the Audit Commission reported in 1997 [45] that only 57% of hospitals in the UK had established a pain service. The situation had improved by 2000 when the Clinical Standards Advisory Group found that 88% of responding UK hospitals had set up an acute pain service, although in some cases this was only a token service. The 1990 working party did not describe a model for the ideal acute pain team [44], and this has led to many inconsistencies in the provision of pain management throughout the UK. Acute pain teams ensure adequate postoperative pain relief by supporting and educating ward staff [46]. Concerns have been expressed that they may de-skill ward nurses [47], although there is no research evidence to support this view.
Because more complicated surgery is performed in older patients, who have more comorbidities, we can expect to see an increase in the number of life-threatening postoperative complications [16]. Surgical patients make up 60–70% of the workload of ICUs in the UK [48,49]. It has been suggested that the role of acute pain teams should be extended to identify nonpain problems in patients, and liaise with other specialties to manage them [46]. Some acute pain services have already extended their role from pain management to include fluid balance, oxygen therapy, management of nausea and vomiting, and anticoagulant prophylaxis [50]. This role extension requires proper training in the management of acutely ill patients and an understanding of the interface between acute medicine and intensive care.
Medical emergency teams
The concept of a better integrated, multidisciplinary approach to postoperative care [51,52] has been extended to other groups of patients in the form of specific groups or teams of clinicians, usually centred on the ICU. Examples of this evolution include postoperative care teams [53], patient at risk teams [54] and METs [36,48,52]. All suggest that the introduction of such teams to identify and manage complications on the ward might prevent unnecessary admissions to the ICU and reduce morbidity and mortality.
The team based approach uses calling criteria based on abnormal clinical or physiological variables, specific conditions, or patients causing concern to ward staff. The research evidence to support the validity of these criteria is limited by the practical and ethical difficulties of conducting randomized controlled trials in the clinical environment. Although postoperative care teams provide additional support, expertise and equipment for postoperative patients [54], and similar approaches could improve suboptimal ward care in general [25], there are no a priori definitions of inadequate care that have been prospectively calibrated against outcome in this context. However, pragmatic studies of the impact of METs indicate a reduction in cardiac arrest rates and in the use of intensive care resources for cardiac arrest survivors [55,56].
Critical care outreach
'Outreach' care is a systems approach for identifying and managing patients at risk of critical illness through collaborative care and education. Rather than providing a service through an external group, it aims to empower ward staff by offering them regular support, usually led by critical care trained nurses visiting the wards, with the facility to call on more expert assistance if required. Currently, critical care is seen as occurring within a defined environment (the ICU) and patients must achieve a certain level of severity of illness to merit admission. Outreach services facilitate a more flexible approach that is based on the needs of the patients and the skills and abilities of the ward staff.
Both the Audit Commission [57] and the Department of Health working party report 'Comprehensive Critical Care' [55] supported the development of outreach care as a means of improving the care of acutely ill patients in hospital wards. The latter report identified three goals for a critical care outreach team. The first was to identify patients at risk for critical illness and either prevent their admission to ICU by timely interventions at a ward level or ensure early appropriate admission to ICU. The second was to facilitate timely and safe discharge from intensive care by following up patients discharged to the ward. The third was to share ICU skills with ward staff. Given the wide variations in hospital size, specialties, staff expertise and skills, the Department of Health report did not prescribe a standard structure for achieving these three goals. The development of outreach services should not be taken in isolation from other critical care initiatives and should be part of an integrated, multidisciplinary, hospital-wide delivery of critical care services that improves liaison between intensive, high dependency and ward care [56].
The majority of UK centres have nurse led outreach systems, supported by critical care doctors with sessional recognition for this service. However, the aim of outreach is to diffuse skills across many disciplines, and to enhance collaborative care. Many professional groups can therefore contribute to this approach, including physiotherapists and nutritionists as well as physicians. Studies are needed to evaluate the efficacy of this development, but current experience indicates a strongly favourable response from ward based staff. Standards for the development of outreach care are now available [58].
Identifying the patient at risk – scoring systems for decision support
Traditionally, the process of identifying critically ill and deteriorating patients has relied on the clinical intuition of staff. The value of experienced clinical judgement is well recognized [36,54,59,60], but given the trend toward shorter training times and reduced hours of work in the clinical environment, objective systems are becoming increasingly important because inexperienced practitioners may fail to recognize impending critical illness and the need for assistance [61].
There are several such systems that utilize combinations of physiological variables as indicators of risk [36,54,56,62-66]. A summary of component variables is given in Table 1. Hodgetts and coworkers [67] analyzed factors that predicted risk for cardiac arrest and could therefore be used as triggers to call for help ('activation criteria'). Variables with predictive capacity included chest pain, staff concern, systolic blood pressure, oximetry, pulse and respiratory rate, and temperature. The choice of vital signs is necessarily constrained to those that are easy to record in the ward environment. More finesse may be achieved with the inclusion of laboratory tests such as serum electrolytes, blood sugar and acid-base analysis, or the inclusion of diagnosis, but the problem with this approach is that recording vital signs is often omitted or inaccurate [17,54], and the greater the degree of complexity, the more likely it is that errors or omissions will occur. Scoring systems for use in ordinary wards must be simple, and should direct attention to patients who need more intensive observation; they should be a trigger for investigation, not a precise tool for predicting individual patient outcomes. Thus, although physiological abnormalities may be unreliable predictors of the need for intensive care admission, they remain important stimuli for empowering staff to call for assistance in improving simple aspects of care of acutely ill patients.
Does earlier intervention improve outcome?
It may seem unnecessarily argumentative to suggest that something that is so evidently 'good' requires evaluation. Is it not self-evident that the earlier a life-threatening disease process is identified and treated, the better? This may be so, but complex systems – of which the acutely ill patient is one example – demonstrate sensitivity to initial conditions; that is, the outcome may be more difficult to predict with earlier application of an intervention. The outcome from cardiopulmonary resuscitation is well defined – a survival rate of around 10–15%, with death usually occurring when resuscitation attempts are discontinued. Is it not possible that earlier intervention that prevents cardiac arrest might result in delayed death following prolonged organ system support in the ICU, with the attendant suffering that this may cause? What is the cost–benefit of implementing the different models of care? Who will manage the process of discussing treatment goals or limitations with patients, and what do patients themselves feel about it?
Physiological goals
The literature relating to preoperative optimization of systemic oxygen delivery in high-risk surgical patients suggests that early intervention is beneficial [68-72]. However, a recent large-scale study [73] conducted in surgical patients did not confirm this in terms of benefit from goal-directed therapy guided by pulmonary artery catheterization, perhaps because patients in the control group were already being optimally managed (and experienced a low overall mortality) as a result of improvements in clinical practice derived from earlier research.
In critically ill septic patients, early intervention to optimize oxygen delivery with fluid resuscitation, vasoactive drugs and respiratory support appears to reduce mortality [74], whereas similar interventions applied later in the course of illness do not [75-77]. It seems reasonable to conclude that using fluids and supplemental oxygen to optimize circulating volume, cardiac output and systemic oxygen delivery as early as possible in acutely ill hospitalized patients will tend to reduce the incidence and severity of organ dysfunction related to a systemic oxygen debt. These simple measures may need to be supplemented later by more complex interventions, but the ability to manage the initial phase of preventive care should be within the ability of most health care staff.
Clinical outcomes
Early intervention may reduce morbidity and mortality, but it is also possible that the earlier application of supportive treatment could contribute additional burdens. Buist and colleagues [62] demonstrated that the introduction of a MET contributed to a reduction in ward cardiac arrest rates, but this was also associated with an increase in emergency admissions to intensive care with no significant change in ICU mortality rates. This suggests that for some patients earlier intervention may have the effect of shifting the burden of mortality from the ward to the ICU, replacing a 'cheap' death with an expensive one. More recently, however, Bellomo and colleagues [63] also demonstrated a marked reduction in cardiac arrest rates associated with the introduction of a MET, and a parallel reduction in the use of intensive care resources on cardiac arrest survivors. It seems likely, therefore, that hospitals with high ward based cardiac arrest rates may well benefit from the introduction of systems aimed at identifying and managing sick patients earlier. Further multicentre studies from Australia are awaited.
Autonomy and treatment limitation decisions
In an ideal world we would all possess clear advance directives giving guidance on how we would wish to be treated given differing circumstances. However, treatment preferences will not be static over time and are likely to be influenced by circumstances and the provision of information. How do patients respond to opportunities to discuss treatment preferences? Studies conducted in elective or out-patient settings demonstrate that many patients wish to be involved in decisions about treatment goals and intensity. They also demonstrate that the accuracy of information substantially alters preferences for resuscitation [78]. In the acute care context, however, there is considerable variability in clinical practice; moreover, many patients may not wish to discuss their preferences [79], even though this results in inappropriate treatment decisions [80]. It is clear that discussions about treatment preferences must be conducted with sensitivity by staff with appropriate experience and training, who have had an opportunity to develop a relationship of trust with the patient. This mandates a collaborative approach between the various medical and nursing teams, the patient and the family. Early intervention may buy time for these complex discussions to take place and an appropriate decision to be made.
Changing behaviour through education and training
The key to improving safe care of acutely ill hospitalized patients is through team working and education, combined with improvements in resources for integrated delivery of care. In accident and emergency care, team based working reduces clinical error rates and enhances overall quality of care [81,82], and the principles of crew resource management drawn from aviation have important messages for clinical practice in this respect, by empowering all members of the team to make contributions to safety [83]. To achieve this we must incorporate team based attitudes in medical education, starting at undergraduate level and following this through into speciality training so that there is more overlap between disciplines than is currently the case. A competency based core curriculum for acute care that incorporates education in patient safety is essential. This process has started in intensive care medicine [84], and it is hoped that other disciplines will follow.
Personal responsibility and continuity of care are important features of quality care, and one of the challenges of implementing outreach is that it may encourage the attitude that someone else is responsible for the patient, and thus disem-power and de-skill ward staff [85]. Team working and continuity of care may be secured through the development of collaboratively produced guidelines or protocols that seek to support and guide the interaction between ward and outreach staff. Outreach staff should also recognize the limitations of their own expertise. Education is a key element in outreach activities, and is a two-way process that requires sharing of expertise, collaborative support, and blurring of traditional boundaries [86]. Important attitudinal attributes of outreach staff must therefore include the capacity to teach, learn from, and support other clinicians, sometimes under difficult circumstances.
Conclusion
There are substantial opportunities for improving the safety and quality of care delivered to acutely ill hospitalized patients. The methods which are adopted will vary according to local circumstance, but common elements include the need for better integration of care across disciplines and systems for earlier identification of patients at risk, and we refer to these as 'outreach' care. We do not yet know which of these various approaches will best improve patient outcomes, and there is a need for prospective studies in this area which take into account the difficulties of using randomization and controls, and which employ long-term follow-up. METs appear to reduce the incidence of cardiac arrests in ordinary wards, and consequential use of scarce intensive care resources. Outreach-based systems which support and educate ward-based staff in delivering clinical care appear to have achieved a high degree of acceptance in the UK. Health care managers need to work closely with clinicians to introduce these methods of team-working into hospital practice, while evaluating their effectiveness.
Competing interests
None declared.
Abbreviations
ICU = intensive care unit; MET = medical emergency team.
|
[
"critical care",
"acute pain team",
"suboptimal care",
"patient at risk",
"critical illness",
"cardiopulmonary resuscitation"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Neurogenetics-2-2-1513515
|
Investigation of autism and GABA receptor subunit genes in multiple ethnic groups
|
Autism is a neurodevelopmental disorder of complex genetics, characterized by impairment in social interaction and communication, as well as repetitive behavior. Multiple lines of evidence, including alterations in levels of GABA and GABA receptors in autistic patients, indicate that the GABAergic system, which is responsible for synaptic inhibition in the adult brain, may be involved in autism. Previous studies in our lab indicated association of noncoding single nucleotide polymorphisms (SNPs) within a GABA receptor subunit gene on chromosome 4, GABRA4, and interaction between SNPs in GABRA4 and GABRB1 (also on chromosome 4), within Caucasian autism patients. Studies of genetic variation in African-American autism families are rare. Analysis of 557 Caucasian and an independent population of 54 African-American families with 35 SNPs within GABRB1 and GABRA4 strengthened the evidence for involvement of GABRA4 in autism risk in Caucasians (rs17599165, p=0.0015; rs1912960, p=0.0073; and rs17599416, p=0.0040) and gave evidence of significant association in African-Americans (rs2280073, p=0.0287 and rs16859788, p=0.0253). The GABRA4 and GABRB1 interaction was also confirmed in the Caucasian dataset (most significant pair, rs1912960 and rs2351299; p=0.004). Analysis of the subset of families with a positive history of seizure activity in at least one autism patient revealed no association to GABRA4; however, three SNPs within GABRB1 showed significant allelic association; rs2351299 (p=0.0163), rs4482737 (p=0.0339), and rs3832300 (p=0.0253). These results confirmed our earlier findings, indicating GABRA4 and GABRB1 as genes contributing to autism susceptibility, extending the effect to multiple ethnic groups and suggesting seizures as a stratifying phenotype.
Introduction
Autism is a neurodevelopmental disorder characterized by severe impairment in social interaction and communication, as well as repetitive behavior. Autism is part of a spectrum of disorders denoted autism spectrum disorders (ASD), which in addition to autism, include Asperger syndrome, childhood disintegrative disorder, and pervasive developmental disorder not otherwise specified (PDD-NOS). Onset of this disorder is typically before 3 years of age. The incidence of autism is estimated at approximately one in 1,000 individuals, with an increased incidence in males [1–3]. Incidences of approximately two to three in 1,000 are reported when a broader diagnosis includes the entire spectrum of pervasive developmental disorders [2, 4].
Evidence indicates that this disease has a strong genetic component. Twin studies have shown that monozygote twins have a higher recurrence rate than dizygotic twins. While the actual percentages have varied, recurrence rates have been measured as high as 95% in monozygote twins and 23% in dizygotic twins [5–7]. Despite this strong evidence for a genetic component, the inheritance appears to be complex, with estimates that more than 15 genes may be involved in its inheritance [8, 9], with different families carrying different combinations of contributing genes. The genes may act independently or interactively, adding to the complexity of determining the genetic contribution to this disorder.
There are two approaches to identifying genetic contributors to disease. The first is a genome wide search in which linkage or association analysis is used to identify regions of the genome that may contain autism susceptibility genes. The second is the candidate gene approach, which investigates a specific gene or genes for involvement in autism risk. In the candidate gene approach, genes are chosen for study based on either what is known about the gene’s function, its location (for example in a recognized linkage peak), or a combination of both. Several candidates are hypothesized to be involved in autism; however, no single candidate gene has consistently emerged as involved in autism risk.
One candidate pathway that is hypothesized to be involved in autism is the GABAergic system. Hussman [10] suggested that autism is the result of an imbalance of the excitatory glutamatergic and inhibitory GABAergic pathways, resulting in overstimulation in the brain and inability to filter out excess stimuli from environmental and intrinsic sources. This theory is supported by multiple lines of evidence. First, histological, biochemical, and molecular approaches have demonstrated altered levels and distribution of GABA and GABA receptors in peripheral blood and plasma, as well as in the brain, including decreased GABA-A receptors and benzodiazepine binding sites in the hippocampal formation [11–13]. There are also reported alterations in GABAergic neurons, as demonstrated by the increased packing density of GABAergic interneurons in the CA3 and CA1 subfields, and by the decreased numbers and reduced size of cerebellar GABAergic Purkinje cells [14, 15]. In addition, duplications and isodicentric chromosomes in the region containing the three clustered GABA receptor subunits GABRB3, GABRA5, and GABRG3 on chromosome 15q have been associated with autism [16, 17]. As well, evidence for both linkage and allelic association have been reported for this same GABA gene cluster, although the findings have not been consistent across datasets [18–22]. Investigation of association of GABA receptor subunits outside of the chromosome 15 region has been limited [23]. Lastly, mutations have been reported in multiple GABA receptor genes in families with epilepsy [24]. Given the high comorbidity of autism with epilepsy and seizures, these data suggest that a similar molecular etiology could exist between the disorders.
Signaling in the GABAergic system is mediated by receptors for the neurotransmitter GABA. There are 19 known GABA receptor subunits arranged in clusters throughout the genome. Functional pentamers formed by various combinations of these subunits result in receptors of varying properties and sensitivities. The amounts and functional capabilities of individual receptor subunits that form a specific pentamer can affect the amount and quality of signaling in different parts of the brain.
Previously published work from our laboratory analyzed 14 autosomal GABA receptor genes on four different chromosomes, using 70 SNPs in a Caucasian dataset of 470 families. This analysis revealed allelic association to rs1912960 (p=0.01) within gamma-aminobutyric acid (GABA) A receptor, alpha-4 (GABRA4) on chromosome 4 [23] and a significant interaction with rs2351299 within the neighboring gene gamma-aminobutyric acid (GABA) A receptor, beta-1 (GABRB1).
Despite similar prevalence rates between Caucasian and African-Americans [25, 26], autism studies in African-Americans are rare. Risk alleles may be different between ethnic groups or the same risk alleles may have differential effects in each ethnic group, warranting studies in multiple groups. Evidence that phenotypic factors, including indicators of language development, may be more severe in African-Americans, compared to non-Hispanic Caucasians, [27] is consistent with these possibilities and underscores the need to investigate autism in different ethnic groups. In this study, we present a validation of our previous report in an independent dataset of 54 African-American families, as well as confirmation of our previous results in an expanded Caucasian sample of 557 autism families.
Materials and methods
Samples
All families were drawn from a large multisite study of autism genetics conducted in the southeastern United States. These families are recruited through the Center for Human Genetics (CHG) at Duke University Medical Center (DUMC), the University of South Carolina, and the Center for Human Genetic Research at Vanderbilt (N=54 African-American and 557 Non-Hispanic Caucasian families) through support groups, advertisements, and clinical and educational settings. All sites recruited, enrolled, and sampled individuals with autism and family members, per study protocols approved by their respective institutional review boards (IRBs). Written informed consent was obtained from parents and from children who were able to give informed consent. Families were enrolled based on probands meeting the following core inclusion criteria of: (1) probands ranging from three to 21 years of age; (2) a presumptive clinical diagnosis of autism; and (3) an expert clinical diagnosis of autism using DSM-IV criteria [28], supported by the Autism Diagnostic Interview-Revised (ADI-R) [29] and in some cases, the Autism Diagnostic Observation Schedule (ADOS) [30]. Our original ascertainment protocol relied on clinical expertise and the ADI-R. The ADOS was subsequently added as a required diagnostic method, and families that were missing ADOS assessments were updated as available. To assure valid ADI-R results, all participants who met current diagnostic criteria for autism were included only if they had a minimal developmental level of 18 months, as extrapolated from the Vineland Adaptive Behavior Scale score [31], or had an IQ equivalent greater than 35 based on scores from a standardized measure of cognitive ability such as an the Wechsler Scale of Intelligence for Children—4th edition [32], Differential Abilities Scale [33], Mullen Scales of Early Learning [34] or Leiter Intelligence Scale-Revised [35]. IQ data was derived from medical records or direct assessment. Exclusion criteria for participation in the larger genetics study included: severe sensory problems (e.g., visual impairment or hearing loss), significant motor impairments (e.g., failure to sit by 12 months, or walk by 24 months), or identified metabolic, genetic, or progressive neurological disorders, based on screening by clinical staff. Additional samples are from the Autism Genetic Research Exchange (AGRE). These individuals were qualified using similar methods including ADI-R, ADOS, VABS, and a standardized measure of IQ.
Thirty-nine African-American families were used in an initial GABA receptor screen. Follow-up analysis of significant findings was performed in 54 African-American families. Analysis of the extended Caucasian dataset included 557 non-Hispanic Caucasian families. One-hundred and five new non-Hispanic Caucasian families were added to the analysis (18 families previously analyzed by Ma et al. [23] were newly identified as Hispanic, and were omitted from the current study in an effort decrease heterogeneity in the Caucasian dataset).
Classification of history of seizure activity in autism patients was based on question 92 from the ADI-R, which queries for both current and lifetime presence of convulsions, seizures, and epilepsy. Caregiver responses to question 92 are coded to indicate no seizure activity, seizure activity with no definitive diagnosis of epilepsy, and seizures with a definite diagnosis of epilepsy. Using lifetime ratings, two groups of families were defined: those in which no seizure activity was reported, and those in which seizure activity was present in at least one autism patient. In addition, question 92 allows for coding of febrile seizures. Families with only febrile seizures were classified as negative for seizure activity and not included in the seizure subset analysis. Both families with positive and negative history of seizure activity were included in our overall dataset. This resulted in a dataset of 41 Caucasian families with a positive history for seizures.
Molecular analyses and genotyping
The analysis of 14 GABA receptor subunit genes was performed in 39 African-American families as previously described [23]. Briefly, between three and seven intronic, UTR and synonymous coding SNPs within each gene were identified from Applied Biosystems (ABI, Foster City, CA, USA) Assay on Demand (AoD) products, resulting in 70 SNPs within 14 GABA receptor genes on four autosomes. Genes analyzed were: GABRA1, GABRA6, GABRB2, GABRG2, and GABRP from chromosome 5; GABRA2, GABRA4, GABRB1, and GABRG1 from chromosome 4; GABRB3, GABRA5, and GABRG3 from chromosome 15; and GABRR1 and GABRR2 from chromosome 6.
Additional SNPs within GABRA4 and GABRB1 were analyzed in the extended African-American (N=54) and Caucasian (N=557) datasets to expand the coverage of variation across this region. Thirty-five SNPs, representative of different linkage disequilibrium (LD) blocks across the two genes (20 in GABRA4 and 15 in GABRB1), were genotyped (Fig. 1). SNP’s for genotyping were selected from online databases (University of California Santa Cruz http://genome.ucsc.edu and NCBI dbSNP http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=snp) and from resequencing of exons and surrounding areas of both GABRB1 and GABRA4 genes.
Fig. 1Approximate locations of SNPs for extended analysis within GABRA4 and GABRB1. Horizontal arrows represent transcriptional direction. Vertical arrows represent SNPs, as numbered and identified below the diagram. Pter indicates the P-terminal end of the chromosome and cen indicates the centromere
SNP genotyping was performed using Taqman allelic discrimination assays (Applied Biosystems). DNA was extracted from whole blood according to established protocols [36], and 3 ng of genomic DNA was used per reaction. Amplification was performed on GeneAmp PCR Systems 9700 thermocyclers, with cycling conditions as recommended by Applied Biosystems. Fluorescence was measured using Applied Biosystem’s 7900. Genotype discrimination was conducted using ABI Prism SDS 2.1 software. Quality control, to ensure accurate genotyping, involved two different CEPH DNAs in quadruplicate on each 384 well plate, as well as the presence of samples which were replicated elsewhere in the sample list. Additionally, ≥95% genotyping efficiency is required.
Statistical analysis
To ensure genotyping quality, Pedcheck was run for detection of Mendelian inheritance inconsistency. One affected and one unaffected individual from each family were selected randomly for tests of Hardy–Weinberg equilibrium (HWE), which was assessed using exact tests from the Genetic Data Analysis program [37]. Pairwise LD between markers was calculated using graphical overview of linkage disequilibrium (GOLD) [38] in the parents of autism cases for both the African-American and Caucasian samples. LD was evaluated in parents to increase the available sample size for analysis and comparison between the two ethnic groups. The pedigree disequilibrium test (PDT) and its extension the genotypic pedigree disequilibrium test (genoPDT) [39, 40] were used to test for association to autism susceptibility.
The EMDR [23, 41], an extension of the MDR [42, 43], was used to test for potential gene–gene interaction, to identify specific locus combinations of interest for further investigation and validation of previous results. EMDR analysis was performed using seven SNPs, the four in GABRA4 found to show significant allelic or genotypic association in the Caucasian sample-set, and the three in GABRB1 found to be significant in the seizure subgroup. One-, two-, and three-way analysis was performed on the Caucasian dataset. For case-control pairs used in EMDR, the proband (or most completely genotyped affected child) from each multiplex and triad family was selected (n=470 total) as a case. Controls were generated using the untransmitted alleles from parental genotypes. In this study, a cross-validation option was not utilized. The sample size of the African-American dataset is too small to provide power for EMDR analysis, not allowing us to test for validation of the interaction seen in the paper of Ma et al. within the African-American dataset.
The haplotype family-based association test (HBAT; [44]) was used for haplotype association analysis using the significant SNPs in GABRA4.
Results
Allelic association studies of 70 SNPs across the 14 GABA receptor subunit genes in the 39 African-American screen set of families, revealed association in rs2280073 (GABRA4; p=0.0053) and hcv2119841 (GABRB1; p=0.0343), the same two genes identified through allelic association and interaction analysis in the Caucasian dataset [23]. Genotypic association analysis revealed the same GABRA4 SNP, rs2280073 (p=0.0262), and marginal significance within GABRP, rs1862242 (p=0.0471). The remaining SNPs showed no significant association (data not shown).
Analysis of the screening SNPs and newly identified SNPs within GABRA4 and GABRB1 in the Caucasian population (N=557), and within the extended African-American population (N=54) (Table 1), revealed new SNPs with significant association. In the Caucasian dataset, rs1912960, also significant in the study of Ma et al. (p=0.012), showed significant allelic association (p=0.0073). Additional significant SNPs were identified in GABRA4 as well, rs17599165 (p=0.0015) and rs17599416 (p=0.0040). Genotypic association was also seen in these SNPs (p=0.0046, 0.0009, and 0.0043, respectively), as well as in a fourth SNP, also in GABRA4 (rs7660336, p=0.0368). In the African-American dataset, rs2280073 (p=0.0287), identified in the smaller African-American dataset above, and rs16859788 (p=0.0253), were found to be associated with the allele based test. Genotypic association was also identified in rs16859788 (p=0.0412). No SNPs within GABRB1 were found to be associated with autism in either ethnic group.
Table 1Analysis of GABRA4 and GABRB1 in extended Caucasian and African-American datasetsGABRA4GABRB1SNPCaucasianAfrican-AmericanSNPCaucasianAfrican-AmericanPDTaGeno PDTaPDTaGeno PDTaPDTGeno PDTPDTGeno PDTRS76783380.93500.99230.22300.4190RS18669890.30710.38600.55530.7892RS64475170.88260.70340.5050.3930RS23512990.45290.08220.57750.8614RS175991020.80550.89130.75180.8805RS100163880.15850.26600.23670.4259RS76603360.08330.0368(G/G)0.51640.7410RS13724960.23620.37400.24820.3715RS15121360.90520.98690.28880.3575RS31140840.09340.19420.20590.3281RS175991650.0015(T)0.0009(T/T)0.65470.4304RS44827370.14950.25041.00001.0000HCV15925450.77980.94270.22300.5313HCV113535240.19590.31171.00001.0000RS76855531.00000.84190.49130.6044RS37755340.18930.18310.49130.4913RS19129600.0073(C)0.0046(C/C)0.41110.5110HCV21198410.33520.08380.22780.4111RS20559430.96710.94340.49270.7607RS62870.40450.35710.61710.6984RS22800730.14040.09550.0287(G)0.1100RS62890.93490.55540.86580.9220RS168597880.31730.31730.0253(A)0.0412(A/A)RS62900.42850.19730.31730.4594RS175994160.0040(A)0.0043(A/A)0.80840.80844P04131.00001.00000.73890.7389RS37922081.00000.49800.17970.1797RS100289450.32720.45841.00000.8179RS105171740.94840.08940.71500.8903RS38323000.40940.63520.65470.6547RS76940350.43370.62660.36570.3657RS37922110.90570.83820.65470.2895RS22299400.78730.93750.54850.7866RS131517590.75290.93260.23670.5759RS131517690.47400.74360.18240.4768Bold indicates significant valuesaAssociated allele/genotype shown in parenthesis
The majority of pairwise r2 values between the significant SNPs were less than 0.3, in both ethnic groups (Table 2). However, a few SNPs have values between 0.3 and 0.35. SNPs, rs17599165, and rs17599416 have r2 values of 0.709 in African-Americans and 0.853 in Caucasians, and rs7660336 and rs2280073 have a pairwise r2 of 0.907 in Caucasians and 0.905 in African-Americans. Allele frequencies are similar, yet not identical between the two groups. One SNP, however, did have large differences in minor allele frequencies (MAFs) between the two groups, with a MAF of 0.24 in African-Americans but only 0.001 in Caucasians (Table 2). Haplotype analysis, using the four SNPs with significant allelic or genotypic association in the Caucasian families, revealed a significant global test (p=0.014) in the Caucasian population, further supporting the involvement of these SNPs or another variant on the haplotype background. The haplotype composed of all risk alleles trends towards significance with autism as a risk haplotype (p=0.0763), while the haplotype of all protective alleles resulted in being highly significantly protective (p=0.0006; Table 3)
Table 2Minor allele frequencies and linkage disequilibrium in Caucasian and African-American datasetsBold indicates significant valuesaMAF minor allele frequencyTable 3Haplotype frequency and association in Caucasian populationRS7660336RS17599165RS1912960RS17599416Haplotype frequencyP-valueGTCA0.5030.0763CTCA0.2770.9214CTGA0.1240.6094CAGG0.0820.0006Bold indicates significant values
Subsetting of the GABRA4 and GABRB1 data to analyze all families with positive history for seizures revealed no association to GABRA4. However, three SNPs within GABRB1 were found to be both allelically and genotypically associated with autism: rs2351299 (p=0.0163 and p=0.0189 for PDT and genoPDT, respectively), rs4482737 (p=0.0339 and p=0.0339), and rs3832300 (p=0.0253 and p=0.0357). These three SNPs all had pairwise r2 values less than 0.1 (data not shown).
In the Caucasian population, EMDR verified the single locus effect identified through PDT analysis in rs1912960 (p=0.024), and identified two different significant two-locus gene–gene effects between GABRA4 and GABRB1, rs1912960 with rs2351299 (p=0.004), and rs17599416 with rs2351299 (p=0.014). Several three locus effects were also significant [rs7660336, rs1912960, and rs2351299 (p=0.012); rs17599165, rs1912960, and rs2351299 (p=0.012); rs1912960, rs17599416, and rs2351299 (p=0.038); and rs7660336, rs17599416, and rs2351299 (p=0.047)] (Table 4).
Table 4EMDR results in Caucasian dataset between GABRA4 and GABRB1Input SNPsSignificant interactionsGeneSNP numberSNPSNPsP-valuesGABRA41RS7660336One-way30.0242RS17599165Two-way3, 50.0043RS19129604, 50.0144RS17599416Three-way1, 3, 50.012GABRB15RS23512992, 3, 50.0126RS44827373, 4, 50.0387RS38323001, 4, 50.047
Discussion
We have confirmed the involvement of GABRA4 in autism through identification of significantly associated SNPs within an independent African-American population. Furthermore, we have strengthened our original findings, including identification of additional associated SNPs and a significant interaction between GABRA4 and GABRB1 and in an extended dataset (N=557) of Caucasian autism families. The AA dataset contains only 54 families, and in general power can be a problem in small sample sets if an effect is not seen. However, for the GABRA4 gene, we did find significant results in the AA dataset. Furthermore, studies in small datasets such as the AA dataset can also be less robust than in larger datasets. However, the fact that we find the same gene, GABRA4, significant in two different ethnic groups is supportive of the role of GABRA4 in autism risk.
The identification of two different two-way interactions between GABRA4 and GABRB1 provides additional evidence of the complex interaction of these two genes in autism. The rs1912960 (GABRA4) with rs2351299 (GABRB1) interaction is between the same two SNPs previously reported in the study of Ma et al., and is still significant in our larger dataset. A second significant two-way interaction was found, also including rs2351299 in GABRB1 as well as rs17599416 in GABRA4, further supporting that interactions between these two genes are involved in autism risk. Although rs2351299 does not have a significant PDT result, it is found in both significant interaction pairs, and in each of the significant three-way interactions. The MDR approach used in these analyses is specifically designed to detect interaction effects both in the presence and absence of main effects.
It is possible that both of the two-way interactions are being identified due to LD between the two GABRA4 SNPs (rs1912960 and rs17599416). There is significant correlation between the two SNPs, although the r2 value (Table 2) between them is not large (0.32). Given that these do not appear to be causative variants, it is likely that the true variant, yet to be identified, is in LD with these GABRA4 SNPs. Examination of interaction in the independent dataset of African-American families was not possible due to the limited sample size.
We also identified variants within GABRB1 as associated within the autistic population with seizures. One of these variants, rs2351299, also shows a significant interaction with SNPs in GABRA4 (Table 4). While no effect was seen in GABRA4 for the seizure subset, the sample size may be too small to conclusively determine its role in seizure status in autism. However, the enhanced findings in GABRB1 in the seizure subset implicate GABRB1 as a contributor to genetic risk in these patients.
Despite the identification of GABRA4 in both ethnic groups, different SNPs were found to be associated. The identification of distinct SNPs within these populations may indicate differences in allele frequency and linkage disequilibrium within the two racial groups, differences in the haplotypic background in which identical causative variations originated, or differences in the causative variation. SNP rs16859788 for example, which is significant in the African-American group, shows little variation in the Caucasian dataset, therefore, providing no power for detection of an effect in this group. Other SNPs, however, show similar allele frequencies between the two populations.
Some differences in LD do exist between the two ethnic groups as well; however, the majority of the differences are small. The largest differences in LD are in pairwise values with rs16859788, which appear to mostly be due to the fact that the SNP is practically monoallelic in the Caucasian population. The Caucasian dataset suggests that there is a significant association of SNP haplotypes with autism risk, while the African-American set does not. However, this difference may be due to the small size of the African-American dataset. Therefore, while it appears that minor allele frequency differences can account for the lack of association of rs16859788 in the Caucasian dataset, additional studies are needed to determine whether the other differences in results between the two ethnic groups are due to sample size differences, differential LD with the causative variation in the two populations, or are population-specific risk alleles.
While we have identified several associated SNPs, we do not predict any of the ones in GABRA4 to have functional consequences; therefore, it is unlikely that these are primary variants leading to the autism susceptibility. One of the SNPs identified in GABRB1 in the seizure subset, however, is in the 3′ untranslated region (UTR). Given that multiple GABA receptor subunits combine in varying combinations to form a functional GABA receptor, even minor changes in levels of a particular subunit may alter the makeup of receptors within a particular cell type, and alter the GABAergic signaling. Therefore, variations within potential regulatory regions, such as untranslated regions and promoters, could play an important role. It will be important to look at potential changes that may result from this and other potential GABRB1 UTR variations, as well as sequence coding and potential regulatory regions to identify the primary variation, or variations leading to altered autism susceptibility.
In summary, these data show that the GABA receptors are implicated in the etiology of autism in two different ethnic populations and suggest seizures as a stratifying phenotype. Furthermore, these results support our earlier findings, indicating GABRA4 and GABRB1 as genes contributing to autism susceptibility, independently, in the case of GABRA4, and through complex interactions with each other.
|
[
"autism",
"gaba receptors",
"ethnicity",
"association",
"snps"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Eur_Arch_Otorhinolaryngol-4-1-2254466
|
Management and prognostic factors of recurrent pleomorphic adenoma of the parotid gland: personal experience and review of the literature
|
The aim of this study was to investigate the management and prognostic determinants of recurrent pleomorphic adenoma (RPA). A retrospective analysis was performed to examine the clinical features, the prevalence of surgical complications, and new recurrences of RPA. Tumor recurrence rate was estimated by the Kaplan–Meier method, and the prognostic value of some of the variables was tested by univariate analysis using the log rank test. The study focused on 33 patients, 18 female (54.5%) and 15 male (45.5%), aged 12–71 years (median 41). A total or extended total parotidectomy was performed in 16 cases (48.5%), a superficial parotidectomy in 10 cases (30.3%), and a local excision in 7 cases (21.2%). In ten patients (30.3%), a branch or the trunk of the facial nerve was deliberately sacrificed. Major complications included one unexpected definitive paralysis of the marginal mandibular branch of the facial nerve and 14 cases of Frey syndrome. Follow-up varied from 2 to 25 years (median 10.5 years), and there were 11 new recurrences (33.3%) within a period varying from 1 to 16 years (median 6 years). The estimated tumor recurrence rates were 14.1 ± 6.6% at 5 years, 31.4 ± 9.4% at 10 years, 43.0 ± 10.8% at 15 years, and 57.2 ± 14.8% at 20 years. Presence of a multinodular lesion and the type of intervention performed were significantly associated with a higher probability of recurrence. RPAs are prone to new recurrences, especially when multinodular and treated with a local excision. Surgical treatment should include facial nerve resection in selected cases. Follow-up for the patient’s lifetime is warranted.
Introduction
Pleomorphic adenoma is the most common neoplasm of the parotid gland. It is a benign tumor composed of epithelial and myoepithelial cells arranged in various morphological patterns. Thinning or absence of the pseudocapsule and the presence of fingerlike projections of the tumor have been observed in all histologic subtypes of pleomorphic adenoma, in particular the myxoid type [1]. Even though many hypotheses for recurrences of parotid gland pleomorphic adenoma have been advanced, including cell biological and genetic factors [2, 3], obvious or underestimated tumor spillage, incomplete excision, and violation of the pseudocapsule of the tumor are considered the only proven reasons contributing to recurrent disease [4].
This is why the abandonment of enucleation techniques in favor of more extended surgical procedures, which require the tumor to be excised with the surrounding normal tissue and the facial nerve to be identified and preserved, have dramatically reduced the recurrence rate of pleomorphic adenomas of the parotid gland from 20–45% to less than 4% in the last decades [5–7].
However, pure en bloc excision in parotid benign tumor disease has been considered impossible by most surgeons when the pleomorphic adenoma is lying on the facial nerve, and so partial enucleation is often the only surgical option in those cases [5, 7], thus explaining the possibility of recurrence even after proper surgery.
Most of the recent literature on recurrent pleomorphic adenoma (RPA) of the parotid gland has attempted to analyze the causes of primary failure and discusses other possible risk factors for recurrence [8–12]; however, there are very few papers focused on the prognostic factors of RPA and retreatment policies [13–18].
The present study was undertaken to address these important issues based on a follow-up period of up to 20 years. Here, we present our findings and review some of the more recent literature.
Materials and methods
A retrospective analysis of the charts of patients treated for benign parotid neoplasm between 1983 and 2004 was performed to identify those patients operated on for RPA. We collected information on patient demographics and clinical history, clinical features of the lesions, diagnostic work-up, the type of surgery, operative findings, facial nerve management, and adjuvant treatment.
The study analyzed the prevalence of surgical complications and new recurrences. Statistical analysis was performed using the SPSS statistical package. Tumor recurrence rate was estimated by the Kaplan–Meier method: the entry point was the date of surgery for a recurrent tumor seen for the first time in our department, and the end point was the date of a new recurrence or the date of last consultation for censored observations.
The prognostic value of certain categorical variables (age, gender, number of previous operations, type of previous operation, lobe of origin, type of surgery performed for RPA, maximum diameter, facial nerve resection, presence of a pseudocapsule at the surgical margins, rupture of the pseudocapsule, number of neoplastic nodules, postoperative radiotherapy) was tested by univariate analysis using the log rank test.
Results
Among 524 patients treated for a benign parotid neoplasm between 1983 and 2004, 33 underwent surgical treatment for RPA. Eighteen (54.5%) patients were female and 15 (45.5%) male; their ages ranged from 12 to 71 years (median 41). None had been previously affected by a malignant neoplasm or submitted to radiation treatment. Seven patients were previously operated on in our department (with a prevalence of recurrence after first treatment of 1.3%) and 26 patients had previous surgical treatment(s) performed elsewhere. A summary of the patient’s previous clinical history is reported in Table 1. A swelling in the parotid region was the most significant complaint for 87.9% (29/33) of patients when they came to our attention for the recurrence; pain or facial nerve paresis was associated with the swelling in two cases. In four (12.1%) cases, the relapse was recognized by routine postoperative follow-up examinations. Other complaints, present from previous surgeries, were facial nerve paralysis in four patients and Frey syndrome in one.
Table 1Summary of clinical history of patients treated for the first time at the Department of Otolaryngology of the University of Brescia for RPA of the parotid glandPrevious recurrences One63.6% (21) Two 18.2% (6) Three 12.1% (4) Four 6.1% (2)Median time to recurrence First recurrence (33) 6 years (range 1–23) Second recurrence (12) 5 years (range 2–28) Third recurrence (6) 5.5 years (range 3–19) Fourth recurrence (2) 8 years (both patients)Last operation Subtotal parotidectomy 33.3% (11) Local excision 33.3% (11) Superficial parotidectomy 21.2% (7) Not reported 12.1% (4)Number of patients in parentheses
Ultrasonography was the most frequently used diagnostic tool (23 patients, 69.7%), followed by fine needle aspiration biopsy in 21 (63.6%) patients, magnetic resonance imaging in 14 (42.4%), computed tomography in 3 (9.1%), and sialography in 2 (6.1%). The different types of diagnostic procedures employed reflect the multi-decade period analyzed. At present, after ultrasonography, which is routinely adopted for follow-up, magnetic resonance imaging is used to delineate the extension of the recurrent lesions.
In 12 (36.4%) patients, the lesion involved the superficial lobe of the parotid gland, in 9 (27.3%) both lobes, and in 12 (36.4%) only the deep lobe was involved. In 5 (15.2%) patients the lesion also extended to the parapharyngeal space. Information on multinodularity of the lesion could be obtained for 31 patients: multiple lesions were observed in 24 (77.4%) of them.
Surgical treatment was tailored to the single patient in an attempt to obtain a margin of healthy tissue (Table 2). Postoperative complications are listed in Table 3. Nine (27.3%) patients were submitted to postoperative radiotherapy. There were no standard criteria for postoperative radiotherapy, but it was performed when the excision did not guarantee sufficient free margins of healthy tissue based on the judgment of the surgeon.
Table 2Surgical treatment for RPA of the parotid gland at the Department of Otolaryngology of the University of BresciaTotal or extended total parotidectomy 48.5% (16)Superficial parotidectomy 30.3% (10)Local excision 21.2% (7)Partial or total facial nerve resection30.3% (10)Number of patients in parenthesesTable 3Postoperative complications of patients treated for RPA of the parotid gland at the Department of Otolaryngology of the University of BresciaFrey syndrome 43.7% (14/32)Temporary paresis of a branch or of the entire facial nerve30.4% (7/23)Definitive paralysis of the marginal mandibular branch of the facial nerve4.3% (1/23)Salivary fistula3.0% (1/33)Keloid3.0% (1/33)
The patients were followed by yearly ultrasonography. Follow-up varied from 2 to 25 years (median 10.5 years). Three patients died from a metachronous malignant neoplasm, one of them after a new recurrence. The other patients are still alive. Overall, 11 (33.3%) new recurrences were observed. No patient with a single node had a recurrence. Four of the new recurrences had been operated on again and one had a further recurrence. Seven patients with small (less than 1 cm) and stable new recurrences are still on follow-up. The time for new recurrence varied from 1 to 16 years (median 6 years). There were 2 cases of focal atypical cells but no malignant transformation in the new recurrences. Distribution of the recurrences in relation to the different variables under analysis is shown in Table 4. The estimated tumor recurrence rates were 14.1 ± 6.6% at 5 years, 31.4 ± 9.4% at 10 years, 43.0 ± 10.8% at 15 years, and 57.2 ± 14.8% at 20 years (Fig. 1). Among the variables evaluated with the log-rank test, only the presence of a multinodular recurrence and the type of intervention performed were significantly associated with a higher probability of a new recurrence (Table 5). The high variation in tumor recurrence rates observed after 20 years of follow-up (Table 5) is probably related to the low number of patients with longer follow-up times.
Table 4Distribution of the new recurrences after treatment for RPA of the parotid gland at the Department of Otolaryngology of the University of Brescia (No. = 11; 33.3%)Variable%P-valueAge≤30 years (3/10)30.01>30 years (8/23)34.8GenderFemale (4/18)22.20.1Male (7/15)46.7Number of previous operations1 (7/21)33.31>1 (4/12)33.3Type of previous operationsEnucleation (2/11)18.20.2Parotidectomy (8/18)44.4Lobe of originSuperficial (5/21)23.80.1Deep (6/12)50.0Intervention performedParotidectomy (6/26)23.10.02Other (5/7)71.4Maximum diameter≤2 cm (7/16)43.80.3>2 cm (3/15)20.0Facial nerve resectionNo (7/23)30.40.7Yes (4/10)40.0Involvement of surgical marginsNo (8/23)34.80.5Yes (2/8)25.0Capsular ruptureNo (6/23)26.10.2Yes (4/8)50.0Multiple nodulesNo (0/7)00.04Yes (10/24)41.7Postoperative radiotherapyNo (9/24)37.50.3Yes (2/9)22.2Fig. 1Estimated recurrence rate after treatment for recurrent pleomorphic adenoma of the parotid gland at the Department of Otolaryngology of the University of Brescia (Kaplan–Meier method)Table 5Estimated tumor recurrence rates after treatment for RPA of the parotid gland at the Department of Otolaryngology of the University of Brescia (log-rank test)VariableEstimated tumor recurrence rates (%)P-value5-year10-year15-year20-yearAge≤30 years (10)12.5 ± 11.741.7 ± 18.6––0.8> 30 years (23)14.7 ± 7.926.9 ± 10.542.4 ± 12.856.8 ± 15.7GenderFemale (18)13.4 ± 9.030.7 ± 13.130.7 ± 13.130.7 ± 13.10.4Male (15)14.3 ± 9.430.7 ± 13.056.7 ± 15.171.1 ± 16.3Number of previous operations1 (21)11.8 ± 7.925.4 ± 11.145.6 ± 14.763.7 ± 17.80.9>1 (12)18.5 ± 11.941.8 ± 16.341.8 ± 16.341.8 ± 16.3Type of last previous operationEnucleation (11)9.1 ± 8.724.2 ± 15.624.2 ± 15.624.2 ± 15.60.2Parotidectomy (18)20.0 ± 10.345.1 ± 14.058.9 ± 15.9100Lobe of originSuperficial (21)013.8 ± 9.132.1 ± 13.654.8 ± 20.60.07Deep (12)35.8 ± 14.459.9 ± 16.459.9 ± 16.459.9 ± 16.4Intervention performedParotidectomy (26)5.3 ± 5.116.4 ± 8.730.5 ± 11.747.9 ± 17.40.001Other (7)42.9 ± 18.7–––Maximum diameter≤2 cm (16)19.6 ± 10.226.9 ± 11.644.3 ± 14.072.2 ± 20.90.4>2 cm (15)10.0 ± 9.532.5 ± 15.532.5 ± 15.532.5 ± 15.5Facial nerve resectionNo (23)9.4 ± 6.327.2 ± 10.635.3 ± 12.151.4 ± 16.70.4Yes (10)25.0 ± 15.340.0 ± 18.2––Involvement of surgical marginsNo (23)14.0 ± 7.524.7 ± 9.737.9 ± 11.753.4 ± 16.00.3Yes (8)25.0 ± 21.762.5 ± 28.6––Capsular ruptureNo (23)16.4 ± 8.838.7 ± 12.838.7 ± 12.838.7 ± 12.80.8Yes (8)12.5 ± 11.712.5 ± 11.740.0 ± 18.2100Multiple nodulesNo (7)00000.02Yes (24)19.6 ± 8.936.8 ± 11.253.2 ± 13.4100Postoperative radiotherapyNo (24)14.2 ± 7.734.3 ± 11.848.9 ± 12.965.9 ± 16.40.3Yes (9)12.5 ± 11.625.0 ± 15.325.0 ± 15.325.0 ± 15.3
Discussion
It is generally accepted that the standard treatment of pleomorphic adenoma of the parotid gland is a formal parotidectomy or at least the removal of the tumor with a surrounding cuff of normal parotid tissue after identification, dissection and preservation of the facial nerve [7]. Unfortunately, if the tumor is in contact with the facial nerve, even a total parotidectomy can not guarantee the presence of a cuff of normal tissue, because the surgeon must carry on the dissection between the tumor and the branches of the nerve with the risk of leaving microscopic disease behind or causing microscopic spillage of the tumor. This accounts for 1–4% probability of recurrence [5, 7]. These low percentages, together with the increasing number of institutions that treat parotid neoplasms, have reduced the number of cases of RPA being referred to a single institution; this, in turn, has contributed to a lack of experience in managing these tumors. Moreover, with the number of reported patients usually varying from less than 15 to about 50 [6, 8, 9, 11, 12, 14–17, 19–24], and only a few studies reporting about a hundred patients [10, 13, 18, 25], there has been relatively little discussion of the prognostic factors for RPA [13–18].
Follow-up of parotidectomies is currently performed by ultrasonography, but most patients do not avail themselves of regular follow-up and only 12.1% of our patients with an RPA were diagnosed by routine ultrasonography, while the majority came to our observation for a swelling in the parotid region. In 77.4% of our patients, there was a multinodular lesion (in Stennert et al. [4], 90% were multinodular at pathologic examination), reinforcing the pathogenetic hypothesis of incomplete excision, violation of the pseudocapsule of the tumor, or obvious or underestimated tumor spillage during the first surgery due to the impossibility of obtaining healthy free margins when the lesion is adjacent to the facial nerve.
Delineation of the extension of the RPA is better depicted by MRI, which better demonstrates the multinodular nature of the disease [24] and potential deep-lobe or parapharyngeal extension. Many patients that we treated came from other institutions without clear documentation of their previous treatment(s); in such cases, MRI can also help in determining the amount of parotid gland still present after any previous surgery. The signal and enhancement characteristics of RPA of the parotid gland are nonspecific [26]. The lesions are usually round; they are of low intensity on T1-weighted images and of high intensity on T2-weighted MR images. With contrast administration, the lesions show mild enhancement [26]. However, MRI can also be inadequate for identifying all nodules, and frequently the surgeon’s microscope or the pathologist’s microscope reveals many more nodules than suspected from clinical assessment [4].
Management options for RPA include observation only, localized resection, resection with facial nerve dissection, wide-field resection with facial nerve sacrifice, and radiotherapy [2, 4, 24]. Only observation has been suggested for the elderly or medically infirm patient [2, 4, 24]. When surgical excision is considered, it should be tailored to the individual patient [19, 23]. In general, it is recommended to resect the scars of any previous surgery. Localized resection of the tumor has been employed after multiple recurrences or after previous total parotidectomy when it is the only option to preserve the facial nerve [10, 14]. Superficial or total parotidectomy, depending on the location of the recurrence, has been suggested when the previous operation was a simple local excision [6, 8–15, 17, 18, 20, 23–25]. Radical and extended parotidectomies have been considered for patients with infiltration of branches or the main trunk of the facial nerve, and in cases of multinodular recurrences [9, 11–13, 15, 18, 23–25]. In particular, facial nerve resection has been suggested for patients with a history of multiple recurrences or failed radiotherapy [24]. In 30.3% of our patients, the resection of at least a branch of the facial nerve was performed (Table 2). Facial nerve dissection and preservation were performed in all cases with still healthy parotid tissue between the nerve and the tumor, and in the case of facial nerve encasement when a patient refused facial nerve resection. The observation of many cases with unexpected multinodularity persuaded Stennert et al. [4] to routinely perform a total parotidectomy for all RPAs. However, such an aggressive policy of treatment does not prevent leaving microscopic residuals; indeed, a recent study by the same group reported a 52% probability of new recurrences [18].
The role of radiotherapy remains questionable [9, 13, 16, 17, 24, 27]. Some authors suggest that radiotherapy should be reserved only for malignant tumors since it carries, especially in young people, a risk of malignant transformation that increases over time [17]. Nonetheless, its use is not infrequent [6, 9, 13–16, 19, 20, 27]. Jackson et al. [9] used postoperative radiotherapy in patients in whom histology showed that resection margins of the specimen were not free of tumor or in those with intraoperative spillage of tumor. Renehan et al. [13] recommended postoperative radiotherapy in multinodular recurrences. Douglas et al. [16] reported their experience of fast neutron radiotherapy and recommended radiotherapy for selected cases: when complete extirpation is not possible, when sacrifice of the facial nerve is necessary, or after multiple recurrences. Glas et al. [17] suggested postoperative radiotherapy for difficult cases in which further surgical treatment of recurrent disease is not recommended for technical reasons. Kamida et al. [27] suggested the use of stereotactic radiosurgery for difficult selected cases, such as the patient they treated for skull base invasion with no regrowth after 5 years. Our experience, including only nine patients, is too limited to draw any conclusion about indications for radiation therapy.
Complications of treatment of RPA are similar to those for primary surgery for parotid pleomorphic adenoma. The main difference is the higher risk of facial nerve damage in the former, with unexpected permanent facial nerve paralysis in more than 30% of cases [24].
Even though we did not have cases of malignant transformation, a high incidence is expected, especially after multiple recurrences, reaching up to 15% in the experience of Mercante et al. [22].
The main purpose of our study was to analyze risk factors for new recurrence after treatment for RPA. The percentage of new recurrences varies from 10 to 58% [8, 10, 12–15, 17–20, 23–25]. In two surveys of the literature with a 10-year follow-up, the prevalence of a new recurrence was 43 and 45%, respectively [12, 25]. Wittekindt et al. [18] reported a recurrence rate increasing from 42% at 5 years to 75% at 15 years. We observed a prevalence of 33.3% with an estimation of recurrence rate increasing from 14.1 ± 6.6% at 5 years to 31.4 ± 9.4% at 10 years, 43.0 ± 10.8% at 15 years, and 57.2 ± 14.8% at 20 years (Fig. 1).
Among the variables tested with the log-rank test against the estimated new recurrence rate (Table 5), just two were significantly associated with the likelihood of developing a new recurrence: local excision of the RPA instead of a formal parotidectomy and the presence of multinodular disease. To date, no new recurrences have been observed for uninodular recurrences. There was a trend toward increased risk of a new recurrence for deep-lobe localization, although it was not statistically significant.
Few papers have analyzed prognostic factors for RPA [13–18]. Renehan et al. [13] studied 114 patients with first-recurrence RPA and reported significantly better, new recurrence-free survival for patients treated with postoperative radiotherapy when the RPA was multinodular. Age, gender, time to first recurrence, and the type of prior treatment were all not significant predisposing factors for a new recurrence [13]. Yugueros et al. [14] followed 39 patients treated for RPA at their institution: they found no significant difference in development of another recurrence in relation to gender, time after the previous treatment, size of the tumor, type of resection, and postoperative radiotherapy. Carew et al. [15] analyzed 31 patients treated for RPA and concluded that the extent of the first parotid operation was the only factor with a statistically significant impact on tumor control after resection of the recurrence. Local control was achieved in all patients whose initial procedures involved local excision. Superficial lobe location and postoperative radiotherapy showed better results but were not significant [15]. Fast neutron radiotherapy of 16 patients with residuals of RPA stopped progression of the tumor in patients without gross residuals, without multifocal disease and with an interval from diagnosis of less than 25 years; however, the results were not statistically significant [16]. Glas et al. [17] analyzed 52 patients with RPA, treated after one or more recurrences: no significant influence was seen in the recurrence rate with respect to gender, age at initial treatment, time of first recurrence, and intraoperative tumor spillage. Wittekindt et al. [18], in a group of 108 patients, observed a significantly higher rate of new recurrences in female patients, in younger patients, and in patients treated with a simple enucleation.
Conclusion
The management of RPA is a major challenge for the clinician. RPAs, particularly multinodular tumors, are prone to new recurrences especially when treatment of the initial tumor is performed according to currently accepted standards. MRI is considered to be the best tool to delineate the extension of the lesion, although it can miss microscopic nodules. Options for management include pure observation, not only for the elderly or infirm, but also in cases of small lesions until they begin to grow. When surgery is elected, it should be tailored to the single patient, because even if on one side a limited local excision is considered acceptable, on the other side a total or extended parotidectomy may be inadequate to control an RPA adjacent to the nerve. In these cases, facial nerve resection and reconstruction must also be considered and discussed with the patient in the preoperative counseling. Postoperative radiotherapy is an option, particularly for older patients for whom the risk of inducing other malignancies is considered to be low. Follow-up for patients treated for RPA should be done with regular ultrasonography for the lifetime of the patient and with magnetic resonance in selected cases with deep lobe or parapharyngeal involvement. All patients should be informed about the possibility of the need for multiple treatment procedures.
|
[
"pleomorphic adenoma",
"tumor recurrence",
"parotid gland neoplasms"
] |
[
"P",
"P",
"R"
] |
Clin_Oral_Investig-4-1-2238792
|
Methodological considerations concerning the development of oral dental erosion indexes: literature survey, validity and reliability
|
Within the context of preventing non-communicable diseases, the World Health Report (2002) and the WHO Global Oral Health Program (2003) put forward a new strategy of disease prevention and health promotion. Greater emphasis is placed on developing global policies in oral health promotion and oral disease prevention. The Decayed, Missing, Filled Teeth (DMFT) index does not meet new challenges in the field of oral health. Dental erosion seems to be a growing problem, and in some countries, an increase in erosion of teeth is associated with an increase in the consumption of beverages containing acids. Therefore, within a revision of the WHO Oral Health Surveys Basic Methods, new oral disease patterns, e.g. dental erosion, have to be taken into account. Within the last 20 years, many studies on dental erosion have been carried out and published. There has been a rapid growth in the number of indexes quantifying dental erosion process in different age groups. However, these indexes are not comparable. This article discusses quality criteria which an index intended for assessing tooth erosion should possess.
Introduction
Within the context of preventing non-communicable diseases, the World Health Report (2002) and the WHO Global Oral Health Program (2003) put forward a new strategy of disease prevention and health promotion [52]. The main orientation of the WHO Oral Health Program (2003) is that oral health is integral, essential and interrelated to general health and a determinant factor for quality of life. Proper oral health care reduce the burden of disease as well as premature mortality. Risk factors for oral diseases are common with cardiovascular diseases, diabetes, cancer and chronic obstructive pulmonary diseases.
The actual WHO Global Oral Health Program focuses on priority action areas as: “Diet, nutrition and oral health”, “Oral health and fluorides”, “Tobacco and oral health”, “School children”, “Elderly people” and “Oral health services”. Nutrition affects oral health in many ways, e.g. dental erosion. To minimize the occurrence of dental erosion which is particularly related to acidic beverages, the WHO/FAO (Food and Agricultural Organization of the UN) recommends (within their recently published Global Strategy on Diet, Physical Activity and Health) a reduction in the amount and frequency of intake of soft drinks and juices. WHO recommends that the Ministries of Health should ensure that the mechanisms for intersectorial collaboration are strengthened. Strategies include taxation and pricing, food labeling, school lunch policies and nutrition programs. This WHO Programme is elaborated for all countries of the World. The Regional Offices of WHO play an important role in the improvement of oral health within risk factor approach in disease prevention and health promotion. Therefore, the Global Oral Health Programme will help to achieve greater equity in oral health. But to evaluate the Global Oral Health Programme, we need suitable and validated indicators and indexes [52]. There is currently only one global oral health index to measure oral health. It is the Decayed, Missing, Filled Teeth (DMFT) index. According to the WHO Oral Health Data Bank in the year 2000 from 184 countries, 68% had a DMFT index less than 3 [52]. Up to now, there is only the DMFT index for 12-year-old children in the Health for all 21 Database (www.euro.who.int: HFA-21 DB, downloaded June 13, 2007).
The European Commission launched from 2003 to 2005 the EU-Project: European Global Oral Health Indicators coordinated by the Université Claude Bernard de Lyon, to support European Member States in their efforts to reduce the toll of morbidity, disability related to oral health diseases and especially [18]:
To identify indicators of oral health, of critical oral health care, its quality of care and of essential health resourcesTo strengthen the ability at the local, national and regional levels to measure, compare and determine the effects of oral health services and use of resourcesTo identify indicators of oral health (problems, determinants and risk factors related to lifestyle) of critical oral health careTo identify the types of data generation and management problems within the Health Information SystemTo identify principles for guiding the selection and use of oral health indicatorsTo identify a set of core indicators for oral healthTo review the recent oral health-indicator selection efforts
In total, 66 indicators are proposed, but no indicator for dental erosion is considered [18]. Therefore up to now, the only international reference in the area of oral health is the DMFT index. The European Commission assesses the quality of oral health data as inadequate for planning, implementation, management and evaluation. The number of internationally recommended indicators complicates the national selection of indicators and the comparability of indicators and indexes [18]. The statements of the European Global Oral Health Indicators Report coincide with the experience in the field of development and using of different erosion indexes. Therefore, it would be necessary to strengthen the international alliance of dentists, epidemiologists, statisticians and other scientists and politicians, to develop an adequate oral health indicators set, i.e. as a subset or user-window of the ECHI-Database (European Community Health Indicators-Database) and within the Oral Databank of the WHO and to develop an internationally agreed and accepted erosion index.
Societies around the world today are increasingly witnessing significant changes in diets and lifestyles which have an impact on oral health. Dental erosion seems to be a growing problem, and in some countries, an increase in erosion of teeth is associated with a higher consumption of beverages containing acids [52]. Within a revision of the WHO Oral Health Surveys Basic Methods, new oral disease patterns, e.g. dental erosion, have to be taken into account [52]. The WHO urges the development of methodologies and approaches for evaluating the effectiveness of community oral health programs focusing on health promotion and disease prevention. Additionally, the formulation of new WHO oral health goals up to the year 2020 have been initiated [52].
Most indexes use different clinical examination standards for measuring tooth erosion especially in preschool- and schoolchildren [25]. Such examination standards could be:
–The full mouth or partial recording–The examination of primary and/or of only permanent teeth–The examination of all surfaces or partial recording of surfaces
On this basis, a lot of indexes for the clinical diagnosis of erosive tooth wear have been proposed, which are more or less modifications of combinations of the index published by Eccles [17] or Smith and Knight [48]. The most cited examples of Erosion Indexes developed during the last 20 years are [adopted from 20]:
–The Smith and Knight Tooth Wear Index (TWI) (1984) [48]–The Eccle’s Index (1979) [17]–UK National Survey of Children’s Dental Health Index (1999/2003) [35]–Erosion Index according to Lussi (1996) [33]–Modified scoring system of Linkosalo and Markkanen (1985) [19]–Aine Index 1993 [1]–The Larsen and Westergaard Index (2000) [30]–The O’Sullivan Index (2000) [39]
Recently, the question has arisen how reliable and valid current diagnostic criteria and data on erosion are. The indexes developed and used during the last 20 years are not comparable; a gold standard does not exist, and validation studies have not had the effect of identifying an index that could be used as a standard for assessing tooth erosion. The following questions should be answered to advance the definition and assessment of tooth erosion:
–Which erosion indexes are mostly used in the scientific literature?–Which quality criteria should indexes possess?–What differences exist between individual- and population-based erosion indexes?
Finally in the discussion, we will debate whether it is possible to involve erosion indexes into Health Indicator Sets and into International Classification of Diseases (ICD; e.g. ECHI, HFA21, country databases on oral Health, ICD-11).
Erosion index in the scientific literature
A literature analysis with regard to erosion indexes for the period 2000–2006 was carried out. For this purpose, a Medline research was done considering mesh terms and keywords “tooth erosion” and “dental erosion” in connection with “index”. All human studies published in English found by this search strategy were analyzed for the erosion indexes used. In the literature analyses, a total of 1,380 articles were found with regard to the mesh terms and keywords “tooth erosion” or “dental erosion”. Approximately 326,112 articles were found with regard to the mesh term and keyword “index”. The conjunction of both revealed a total of 91 articles. From these articles, 40 were excluded as they were published before the year 2000, three were excluded because they were not English-language publications, eight studies were not included as no humans were involved, and one review was excluded. In total, 40 publications were considered in the literature review. Of 22 publications, 14 cross sectional studies were found in children, and 18 publications/cross sectional studies were found in adults. The erosion indexes used are presented in Tables 1 and 2 for children and adults, respectively. In children, the UK children dental health survey index [35] was used in four studies, the O’Sullivan Index [39] in three, an index developed in 1993 by Aine et al. [1] was used in two studies, and the Smith and Knight TWI [48] in one study. However, in most studies (n = 4), differently designed erosion indexes were considered. In adults, the Smith and Knight TWI [48] was used in most studies (n = 10) followed by the Eccle’s Index [17] (n = 2), Lussi Index [33] (n = 1), UK adult dental health survey index (n = 1) and an index developed by Schweizer-Hirt et al. 1978 [47]. In one study, the method for estimating the tooth erosion was not mentioned.
Table 1Used dental erosion indexes in human cross-sectional studies for children found in Medline from the years 2000–2006Aim of the studynStudy populationReferred indexSourceErosive tooth wear463Kindergarten, 3- to 5-year-old childrenO’Sullivan IndexWiegand et al. [54]Dental erosion15311-year-oldsO’Sullivan IndexCaglar et al. [10]Prevalence of dental erosion499School setting; 12-year-old childrenO’Sullivan IndexPeres et al. [43]Prevalence of dental erosion1,949Preschool children; 3- to 5-year-old childrenUK Children dental health survey indexLuo et al. [32]Prevalence of dental erosion8326- and 12-year-old school childrenOwn erosion indexTruin et al. [50]SES and ethnicity and oral health1,753Random sample of 12- and 14-year-old childrenUK Children dental health survey indexDugmore and Rock [16]Accuracy and reproducibility of school dental screening570Primary-school childrenNew index was developedHetherington and White [24]Prevalence of and risk factors for dental erosion95Children and adultsOwn erosion indexJohansson et al. [28]Oral health and gastro-oesophageal reflux52Children with gastro-oesophageal reflux diseaseAine IndexLinnett et al. [31]Asthma and dental erosion418Random Sample of 14-year-old childrenSmith and Knight TWIAl-Dlaigan et al. [2]Prevalence of dental erosionOral health and gastro-oesophageal reflux37Children with gastro-oesophageal reflux diseaseAine IndexDahshan et al. [15]Prevalence of dental erosions987Preschool children, 2- to 5-year-old childrenUK Children dental health survey indexAl-Malik et al. [4]Oral health of children with clefts914-, 8-, and 12-year-old childrenUK Children dental health survey indexChapple and Nunn [11]Dental erosion and consumption of oranges1,01012-year-old childrenOwn erosion indexKünzel et al. [29]Table 2Used dental erosion indexes in human cross-sectional studies for adults found in Medline from the years 2000–2006Aim of the studynStudy populationRefered indexSourceTooth survey (surface?) check! loss155Patients attending a dental hospitalUK Adult dental health survey indexRafeek et al. [44]Tooth wear among psychiatric patients143Psychiatric patientsSmith and Knight TWIAl-Hiyasat et al. [3]Tooth wear in elderly690Local survey, elderlyEccles IndexTaiwo et al. [49]Tooth surface in winemakers36WinemakersSchweizer-Hirt et al. 1978Chikte et al. [12]Workplace and dental erosion20Silicon workersOwn erosion indexJohansson et al. [27]Oral and dental health34Inpatients in treatment of alcohol disordersNo information givenAraujo et al. [6]Dental erosion and gastro-oesophageal reflux?Patients with gastro-oesophageal reflux diseaseSmith and Knight TWIMoazzez et al. [36]Prevalence of dental erosion18,555Permanent dentitionSmith and Knight TWIBorcic et al. [9]Dental erosion and gastro-oesophageal reflux253Patients with gastro-oesophageal reflux diseaseEccles IndexMunoz et al. [37]Monitoring of tooth wear500Patients referred for a variety of restorative procecduresSmith and Knight TWIBartlett [7]Risk factors of tooth wear506Patients attending a dental hospitalSmith and Knight TWIChuajedong et al. [13]Sport drinks and dental erosion304AthletesLussi IndexMathew et al. [34]Prevalence of tooth wear126Patients attending a dental hospitalSmith and Knight TWIOginni and Olusile [41]Recreational drug and tooth surface loss13Undergraduate studentsSmith and Knight TWINixon et al. [38]Risk factors for dental erosion10Male military Saudi inducteesOwn erosion indexJohansson et al. [28]Oral health status of workers68Workers exposed to acid fumesSmith and Knight TWIAmin et al. [5]Methamphetamine and tooth wear43Methamphetamine usersSmith and Knight TWIRichards and Brofeldt [45]Dental erosion and gastro-oesophageal reflux20Patients with gastro-oesophageal reflux diseaseSmith and Knight TWIGregory-Head et al. [21]
Quality criteria of indexes for measuring tooth erosion
Instruments or indexes for measuring a construct of interest—quality of life or intelligence, mental health status or tooth erosion—should possess certain quality characteristics. Otherwise, the scientific value of the obtained results is questionable [8]. Important quality characteristics are validity and reliability as well as sensitivity and specificity, which will be explained in the following paragraphs.
The validity of an instrument indicates to what extent it measures what it is supposed to measure. There are different—partly overlapping—types of validity which emphasize different aspects [42]. The most important types are content, construct, and criterion validity which will be explained in the following paragraphs.
Content validity describes whether all aspects, which are relevant to grasp the construct of interest, have been considered adequately. For example, a test that is supposed to measure the ability to calculate should not be restricted only to addition and subtraction but include all basic arithmetic operations. In our context, e. g. the erosion of surfaces of all teeth has to be considered, not only the erosion of incisors. The extent to which an instrument possesses content validity cannot be determined numerically. It is solely based upon the subjective meaning of established experts [8, 23]. Therefore, a major step in developing a new instrument is a thorough discussion of which aspects should be included.
The optimal method to validate a newly developed instrument is the comparison with a so-called gold standard, which measures the same construct. A gold standard is “the method, procedure or measurement that is widely accepted as being the best available” [46]. The so-called criterion validity is high if the results of the new instrument and the gold standard are highly correlated. For example, a thermometer might serve as a gold standard for self-reported temperature [23].
However, because a gold standard is seldom available, construct validity is of major importance [8]. Construct validity is subdivided into convergent and discriminant validity. There is convergent validity if the results obtained by the new instrument are correlated with the results of an established instruments that measures similar aspects. If, for example, results of a new instrument for measuring physical health are correlated with the results of an established instrument which measures activity of daily living, convergent validity is indicated [8]. Or in our context: if the results of a tooth-wear index is correlated with tooth-erosion index, convergent validity is indicated. On the other hand, an instrument possesses discriminant validity if the results of this instrument are not too highly correlated with the results of an established instrument that measures a different construct. For example, results of an instrument for measuring mental health should not be too highly correlated with an instrument for measuring physical function [23]. With regard to teeth, a high correlation between tooth erosion and wedge-shaped defects might indicate insufficient discriminant validity. Both convergent and discriminant validity have to be given for a complete confirmation of construct validity.
The reliability of an instrument indicates how precise it is able to measure, independent of whether it really measures what it is supposed to (this is a question of validity) [42]. There are different aspects of reliability. For our purposes, inter-examiner reliability and intra-examiner reliability are important.
Inter-examiner reliability can be determined if two or more dentists (the “examiner”) assess dental erosion of a number of patients independently of each other [14]. Roughly spoken, the more often the examiners agree in their assessment, the higher the inter-examiner reliability. Intra-examiner reliability or test–retest reliability can be determined if each dentist rates dental erosion of a group of patients twice [23]. The period of time between the two assessments should be fairly long, otherwise the examiners may be overly consistent because they remember their former ratings. On the other hand, if the time interval between the two assessments is too long, changes in the erosion status might bias the reliability estimates. The more often the assessments at the two points in time concerning each examiner are identical, the higher the intra-examiner reliability. A widespread index to calculate the degree of agreement is Cohen’s Kappa [14], which “is now fairly well disseminated as one of the standard summary statistics used in the medical literature” [22]. The calculation of Kappa is based on the percentage of agreement which is adjusted for agreement expected by chance.
The sensitivity of an instrument indicates its ability to detect—in our case—dental erosion. In contrast, an instrument with high specificity is able to indicate no dental erosion if dental erosion is not present. Both assessment of sensitivity and specificity require the comparison with a gold standard. Comparing a new instrument that is supposed to detect dental erosion with an instrument that is known to be able to detect dental erosion, i.e. a gold-standard instrument, can yield one of the four following outcomes: (1) dental erosion is indicated by the new instrument if there is in fact dental erosion (correct positive), (2) dental erosion is indicated if there is no dental erosion (false positive), (3) dental erosion is not indicated if there is dental erosion (false negative) and (4) dental erosion is not indicated if there is no dental erosion (correct negative). The degree of sensitivity is calculated by dividing the frequency of correct positive results by the sum of correct positive and false negative results, while the degree of specificity is determined by dividing the frequency of correct negative results by the sum of correct negative and false positive results [46].
Individual versus population-based erosion indexes
An individual-based index should fulfill other criteria than a population-based index. The first should allow assessment of the full truth of the construct “tooth erosion”. Therefore, a very good individual-based erosion index should nearly be a gold-standard instrument. Using a population-based erosion index, the individual assessment should take less time to allow an assessment in big population samples. Therefore, for practical reasons, a population-based erosion index is often a short form of an individual-based index. However, a good population-based erosion index should possess very high validity as well as high reliability.
Discussion
There is no common sense in the usage of the above-mentioned erosion indexes. A wide range of different forms and indexes was used in the last 6 years. Therefore, comparability between different studies was not given, and meta-analyses were not possible. If anything, it can be stated that for adults, the Smith and Knight TWI [48] has been widely used in most Medline-cited cross-sectional studies. For children, there is no index that is most used. However, the literature analysis is limited. Only one search strategy was used. Some articles which do not use the mentioned key words or mesh headings might not be included in the analysis. However, the results presented help to get a first view on the different indexes on tooth erosion cited in the last 6 years.
In all these indexes, the criteria to record the grading for erosion differed [26]. Most of the indexes used are based on the clinical severity of erosion, focusing on accessible teeth but not all [29]. For example, the structure of the Smith and Knight TWI is shown in Table 3. However, this index is made for scoring tooth wear in general irrespective of its predominant aetiology. It is only a system for quantifying and grading the amount of tissue loss but does not include diagnostic criteria for erosion. The UK National Survey of Children’s Dental Health Index was a modified version of the Smith and Knight TWI assessing only buccal or palatal surfaces of maxillary incisors and added erosion criteria. However, while using the first index, the conclusion was drawn that the figures might not be completely accurate [40]. The next version focused on erosion of the palatal surface of at least one maxillary anterior tooth with the majority of surfaces having dentin involvement [29]. To date, literature assessing the prevalence of erosion has been published from different countries. Erosion was recorded using various indexes. In addition, selection criteria of study populations, sampling techniques and considered age groups differed. The prevalence of dental erosion of different studies is therefore difficult and nearly impossible to compare [29].
Table 3Smith and Knight Tooth Wear Index [48]ScoreSurfaceCriterion0B/L/O/I/CNo loss of surface characteristic, no loss of contour1B/L/O/I/CLoss of enamel surface characteristics, minimal loss of contour2B/L/OLoss of enamel exposing dentine for less than one third of the surfaceILoss of enamel just exposing dentineCDefect less then 1 mm deep3B/L/OLoss of enamel exposing dentine for more than one third of surfaceILoss of enamel and substantial loss of dentine not exposing secondary dentine or pulpCDefect 1–2 mm deep4B/L/OComplete loss of enamel, or pulp exposure, or exposure of secondary dentineIPulp exposure or exposure of secondary dentineCDefect more than 2 mm deep, or pulp exposure, or exposure of secondary dentineB Buccal or labial, L lingual or palatal, O occlusal, I incisal, C cervical
A good index is characterized by a conducted validation study. For those indexes for which a validation study has not been conducted so far, that should be made up to check whether they possess the quality criteria as described above. In a validation study, it is not only the validity of an index that is examined but also aspects of reliability as well. Intra-examiner and inter-examiner reliability should be checked. That is done quite easily as described in “Quality criteria of indexes for measuring tooth erosion”. At the beginning of the study, it has to be considered how many patients and how many examiners respectively have to be involved. This is because the number of patients and examiners should be as small as possible for practical and economic reasons. On the other hand, the number of patients and examiners has to be sufficiently high because otherwise a certain degree of reliability, which actually exists, might be overlooked. Therefore, solid sample-size calculations should be conducted with the help of an experienced statistician.
So far, there is no consensus concerning a gold-standard instrument. However, a gold standard is important for comparison. Otherwise, criterion validity cannot be checked. If one attempts to develop a gold-standard instrument, the first step would be a thorough discussion of which aspects of the construct “tooth erosion” have to be included. This should be conducted by established experts to ensure content validity. When the instrument is constructed, inter-examiner reliability and intra-examiner reliability can be examined as described above. However, this should be done not only in a national but in an international context.
A gold-standard instrument would explain the full truth of the construct “tooth erosion”. Nevertheless, the measurement procedure in general takes a long time. It is not possible to deal with such a comprehensive assessment instrument in a population-based study design. Therefore, more simple assessment instruments should be available. These assessment instruments should fulfill the following qualifications: (a) a validation study should be available where all quality criteria of the assessment instrument were tested and (b) the instrument should be used in other scientific studies, too. If the population-based short version of an assessment instrument and the gold-standard instrument are measuring the same construct, and when both conditions are fulfilled, calibration studies might be possible. Calibration refers to a process in which values from one method are quantitatively related to values from a superior, gold-standard method. The aim of such an analysis is to ensure that the measurement uncertainty is known and is consistent with the required measurement capability. In a so-called calibration study, known data on the observed relationship between an independent variable of the population-based short version of the assessment instrument and the dependent variable of the gold-standard instrument are used to make estimates of other “true” values of the independent variable from new observations of the gold standard.
One of the prospective targets ought to be the inclusion of indexes on dental health prevention especially of an internationally agreed “erosion index” into the existing Health Indicator Sets. These would be the Health for All Database of the WHO (HFA21), The European Community Health Indicator Set (ECHI), Country Databases on Oral Health and the inclusion of Dental Erosion into the International Classification of Diseases and Health-Related Problems.
In April 2007, the World Health Organization announced plans to revise the ICD-10, in other words, to prepare for ICD-11. Within the actual ICD-10, there exists two codes for Dental Diseases: “K02” for Caries and “K03” for “Other diseases of hard tissues of teeth” including with “K03.2—Erosion of teeth.” On the four-digit level, we find the following different diagnoses presented in Table 4.
Table 4Code description for K03.2—Erosion of teeth (ICD 10 code) [51]ICD 10 codeDescriptionK03.2Erosion of teethK03.20Occupational erosion of teethK03.21Erosion of teeth due to persistent regurgitating or vomitingK03.22Erosion of teeth due to dietK03.23Erosion of teeth due to drugs and medicamentsK03.24Idiopathic erosion of teethK03.28Other specified erosion of teethK03.29Erosion of teeth, unspecified
The Updating and Revision Committee proposes major changes like addition of new codes or deletion of codes and developed submission guidelines. Therefore, it would be possible to check critically if the codes for dentistry within the Application of the International Classification of Diseases to Dentistry and Stomatology are adequate to future developments or have to be replaced or completed [53].
Conclusion
Further efforts have to be made in the development of an internationally agreed index which is able to assess dental erosion with as much reliability and validity as possible. Current recommendations of the WHO and the European Union and health task forces within countries and Dental Associations should be used to develop and to discuss the concept of tooth erosion and the development of a unified erosion index.
|
[
"index",
"validity",
"reliability",
"tooth erosion",
"gold standard"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Appl_Microbiol_Biotechnol-3-1-1914249
|
Co-existence of physiologically similar sulfate-reducing bacteria in a full-scale sulfidogenic bioreactor fed with a single organic electron donor
|
A combination of culture-dependent and independent methods was used to study the co-existence of different sulfate-reducing bacteria (SRB) in an upflow anaerobic sludge bed reactor treating sulfate-rich wastewater. The wastewater was fed with ethanol as an external electron donor. Twenty six strains of SRB were randomly picked and isolated from the highest serial dilution that showed growth (i.e. 108). Repetitive enterobacterial palindromic polymerase chain reaction and whole cell protein profiling revealed a low genetic diversity, with only two genotypes among the 26 strains obtained in the pure culture. The low genetic diversity suggests the absence of micro-niches within the reactor, which might be due to a low spatial and temporal micro-heterogeneity. The total 16S rDNA sequencing of two representative strains L3 and L7 indicated a close relatedness to the genus Desulfovibrio. The two strains differed in as many as five physiological traits, which might allow them to occupy distinct niches and thus co-exist within the same habitat. Whole cell hybridisation with fluorescently labeled oligonucleotide probes was performed to characterise the SRB community in the reactor. The isolated strains Desulfovibrio L3 and Desulfovibrio L7 were the most dominant SRB, representing 30–35% and 25–35%, respectively, of the total SRB community. Desulfobulbus-like bacteria contributed for 20–25%, and the Desulfobacca acetoxidans-specific probe targeted approximately 15–20% of the total SRB. The whole cell hybridisation results thus revealed a consortium of four different species of SRB that can be enriched and maintained on a single energy source in a full-scale sulfidogenic reactor.
Introduction
Many industrial processes, such as metal smelting, flue gas scrubbing and mining generate sulfate-rich wastewater (Lens et al. 1998). These wastewaters usually do not contain any organic carbon, and the addition of external electron donor is required for their treatment. These waste streams cause a range of problems associated with the now well-established anaerobic treatment of these wastewaters, such as a decrease in methane production (Colleran et al. 1995), sulfide toxicity (O’Flaherty and Colleran 2000), foul smell (Lens et al. 2001) and corrosion (Vincke et al. 2001). Of late, bioreactor processes have been developed for treating these sulfate and metal-rich wastewaters. This technology developed by the company Paques BV in The Netherlands (Pol et al. 2001) makes use of the dissimilatory sulfate-reducing capacity of sulfate-reducing bacteria (SRB) to simultaneously remove sulfate and metals in the form of metal sulfides. A typical application is a two-step process that includes dissimilatory sulfate reduction to sulfide (Gibson 1990) as the first step. Subsequently, the sulfide will be bound by heavy metals, such as cobalt, lead, nickel and zinc and precipitated as metal sulfide (Colleran et al. 1995). The second step is the biological oxidation of the remaining sulfide to insoluble elemental sulfur, which is either recovered by sedimentation or a small portion converted to H2SO4 and re-used to neutralise the alkalinity generated because of sulfide production in the first step.
Several full-scale sulfate removal bioreactors are currently in operation. The dominant sulfidogenic communities in these bioreactors are heterotrophic SRB belonging to the family Desulfovibrionaceae (Kaksonen et al. 2004; van Houten et al. 2006). The genus Desulfovibrio represents a phylogenetically coherent group; all species incompletely oxidise lactate to acetate but can utilise hydrogen, formate and ethanol as well (Widdel and Bak 1992).
In an engineered system fed with a single nutritional source and kept under constant operational parameters such as pH, temperature and salinity, competition among species might tend to reduce the species and sub-species diversity according to the principle of competitive exclusion (Gause 1934). However, physiologically competing species can co-exist if they occupy different niches, whereby each species uses distinct parts of the resource base. In the present study, we investigated the co-existence of physiologically similar hydrogenotrophic SRB in a full-scale sulfidogenic bioreactor treating sulfate-rich wastewater using a combination of cultivation and molecular techniques. The reactor was fed with ethanol as carbon and energy source. SRB isolated in pure culture were characterised genetically and physiologically. Micro-diversity among the dominant culturable isolates was assessed by repetitive enterobacterial palindromic polymerase chain reaction (rep-PCR) (Versalovic et al. 1994). Whole-cell hybridisation with fluorescently labelled general and specific probes was used for SRB community characterisation and for the estimation of the relative abundance of the different SRB populations in the reactor.
Materials and methods
Sludge source
Granular sludge was obtained from an upflow anaerobic sludge bed (UASB) reactor, treating sulfate-rich wastewater from a chemical plant located in Emmen, The Netherlands. The wastewater fed to the reactor did not contain any organic compound, so ethanol was added as an external electron donor and carbon source. The reactor had a volume of 300 m3 and a feed rate of 60–75 m3 h−1 The ratio between the amount of electron donor added and the amount of sulfate in the reactor was around 0.4 kg/kg. Sulfide produced in the reactor was either converted to elemental sulfur through biological sulfide oxidation or precipitated with toxic metals. The reactor was operated at a temperature of 30°C and a pH of 7.0–7.5. The concentration of sulfate in the wastewater was approximately 1,500 mg/l and ca. 100 mg/l in the effluent.
Culture media and isolation of SRB
A basal bicarbonate-buffered and sulfide-reduced medium was used for the enumeration and isolation procedures. The mineral medium contained per litre of distilled water: 0.2 g KH2PO4, 0.25 g NH4Cl, 0.5 g KCl, 0.1 g CaCl2·2H2O, 0.4 g MgCl2·6H2O, 1.0 g NaCl and 0.5 ml of a resazurin solution (0.5 mg ml−1) as a redox indicator. The medium was supplemented with (per litre) the following: 30 ml 1 M NaHCO3 solution, 1 ml of a vitamin solution, 1 ml of trace element solution with ethylenediamine tetraacetic acid (EDTA; Widdel and Bak 1992) and 0.1 g of yeast extract. As a reducing agent, 7.5 ml l−1 of 0.2 M Na2S·9H20 was added. Either lactate or ethanol (20 mM) was used as an electron donor and sulfate (10 mM) as electron acceptor. Enumeration of the potentially dominant heterotrophic sulfate reducers was performed by serial dilutions in Hungate tubes. The Hungate tubes were incubated at 30°C in the dark for 7 weeks. The highest positive dilution tubes that showed growth were selected for further isolation. Growth was assayed by measuring sulfide production photometrically (Cord-Ruwisch 1985).
Pure cultures were obtained by repeated transfer in agar shake tubes (Widdel and Pfennig 1984). Purity of the isolates was checked by microscopic observation and further confirmed by denaturing gradient gel electrophoresis (DGGE) analysis of PCR-amplified 16S rRNA gene fragments (Teske et al. 1996).
Phenotypic characterisation
Substrate utilisation by the isolated strains was determined in duplicate in the same bicarbonate-buffered medium as used for enumeration and isolation but without yeast extract. Growth was tested in 100-ml serum bottles closed with butyl rubber stoppers and aluminium crimp seals. The inoculum size was 1% (v/v). The cultures were incubated at 30°C for 5 weeks in the dark. Growth was determined by substrate consumption or product formation as well as by checking for increase in optical density at 660 nm (OD660). The following substrates were tested as electron donors in the presence of 10 mM sulfate: 10 mM each of pyruvate, fumarate, butyrate, formate, propanol, ethanol, methanol, serine and cysteine; 5 mM each of malate, glycolate and glycerol; 2.5 mM of benzoate; and 0.5% (w/v) of casamino acids. Sulfite, thiosulfate and 2% (w/v) elemental sulfur were tested as electron acceptors in the presence of lactate as electron donors.
For measuring catalase activity, the fully grown isolates were centrifuged at 13,000 × g. The cell pellets were re-suspended on glass slides with a drop of 3% (v/v) H2O2, bubbles indicated the presence of catalase. Detection of desulfoviridin was performed according to Postgate (1959). Gram staining was performed as previously described (Bartholomew 1962).
Analytical methods
Acetate and other volatile fatty acids were analysed either by gas chromatography (GC) or high performance liquid chromatography (HPLC). For GC, a Chromopack 9001, equipped with a flame ionisation detector and a fused-silica capillary column 15 × 0.53 mm HP-Innowax, was used. The column temperature was 120°C. The temperature of the injector and detector were 180 and 200°C, respectively. Helium was used as a carrier gas. An Aminex HPX-87H column from Bio-Rad (T = 60°C) coupled to a UV and a RI detector was used for HPLC; phosphoric acid (0.05 M) was used as an eluent. Sulfide was measured quantitatively by a colorimetric assay (Cline 1969).
DNA extraction
Genomic DNA was isolated from the bacterial cultures using the Ultra Clean Soil DNA extraction kit (MOBIO Laboratories, California) according to the manufacturer’s protocol. The quality of the extracted DNA was examined on a 1% (w/v) agarose gel and the amount quantified by absorption spectrophotometry using the Nanodrop ND-1000 TM (NanoDrop Technologies, Delaware). Extracted DNA was stored at −20°C until subsequent use in different PCR reactions.
PCR amplification and DGGE of 16S rRNA genes
For DGGE analysis, amplification of partial 16S rRNA gene was carried out using the primers 341F-GC and 907R as described by (Schäfer and Muyzer 2001), while primers GM3 and GM4 (Muyzer et al. 1995) were used to amplify the nearly complete 16S rRNA gene for sequencing and subsequent phylogenetic analysis. PCR amplification and DGGE was performed as described previously (Schäfer and Muyzer 2001). The quality of the PCR products was examined on 1% (w/v) agarose gel, and the yield was quantified by absorption spectrophotometry using the Nanodrop ND-1000 TM (NanoDrop Technologies).
DNA sequencing and phylogenetic analysis
The nearly complete 16S rRNA gene fragments, obtained from the strains L3 and L7, were purified using the Qiaquick Gel Extraction Kit (Qiagen, Hilden, Germany). Purified PCR products were sequenced by the company BaseClear (Leiden, The Netherlands). The DNA sequences of about 1,400 bp were first compared to the sequences deposited in public databases using the NCBI BLAST search tool (http://www.ncbi.nlm.nih.gov/BLAST; McGinnis and Madden 2004). Subsequently, the sequences were imported into the ARB software programme (Ludwig et al. 2004) and aligned using the automatic aligner function. The alignment was further corrected manually, and a phylogenetic tree was constructed using the neighbour-joining algorithm with Felsenstein correction.
rep-PCR fingerprinting
The genetic diversity of the isolates was analysed by rep-PCR (Versalovic et al. 1994) using the primer GTG5 (5O′-gTggTggTggTggTg-3′). The amplification reaction was performed as previously described (Foti et al. 2006). A 1-kb size marker and 600 ng of the PCR product were loaded onto a 1.5% (w/v) agarose gel containing 0.5 × TAE-buffer (200 mM Tris-acetate, 0.5 mM EDTA, pH 8). The electrophoresis was performed for 14 h in a cold room at a constant voltage of 65 V. The gel was subsequently stained with ethidium bromide (0.5 μg/ml) and photographed under UV illumination using the GelDoc UV Transilluminator (Bio-Rad, Hercules, CA).
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis
Denaturing sodium dodecyl sulfate-polyacrylamide gel electrophoresis of whole-cell protein (in cell-free extract obtained by sonication) was performed on 12% polyacrylamide gels according to Laemmli (1970). The gels were stained with Coomassie Brilliant Blue and photographed using visible light.
Design of oligonucleotide probes
Specific oligonucleotide probes for the 16S rRNA of the two strains were designed using the Probe Design tool of the ARB software package (Ludwig et al. 2004). The requirements for designing the probes included high specificity, with no organism outside the intended target group having 100% similarity within the target sequence, and the target sequence being located within the high accessibility region of 16S rRNA molecule as suggested previously (Behrens et al. 2003). The probes were named with a number that indicates the position of the first base in the target sequence (by Escherichia coli numbering). The oligonucleotides used for in situ hybridisation are given in Table 1.
Table 116S rRNA-targeted oligonucleotide probes used in this studyProbe nameaTarget organismsProbe sequence (5′–3′)ReferenceEUB338_IMost bacteriaGCT GCC TCC CGT AGG AGTAmann et al. (1990)EUB338_IIPhylum PlanctomycetesGCA GCC ACC CGT AGG TGTDaims et al. (1999)EUB338_IIIPhylum VerrucomicrobiaGCT GCC ACC CGT AGG TGTDaims et al. (1999)ARCH915ArchaeaGTG CTC CCC CGC CAA TTCStahl and Amann (1991)SRB385Most DeltaproteobacteriaCGG CGT CGC TGC GTC AGGAmann et al. (1990)SRB385DbSome DeltaproteobacteriaCGG CGT TGC TGC GTC AGGRabus et al. (1996)DSR660Genus DesulfobulbusGAA TTC CAC TTT CCC CTC TGDevereux et al. (1992)DSBA1017Desulfobacca acetoxidansGTT GCC AGG CAC CCC CATDar et al. (2007)DSV119Desulfovibrio strain L3GGC AGA TCA TCC ACG CGTThis studyDSV139Desulfovibrio strain L7CGC TGT TAT CCC GAT CACThis studyaEUB338 is a combination of EUB338_I, EUB338_II and EUB338_III.
Whole cell hybridisation
Cells from strains L3 and L7 and from the original reactor were fixed, washed and spotted onto Teflon-coated multi-well microscopic slides as described previously (Dar et al. 2007). Hybridisation was carried out according to the protocol described by Manz et al. (1992) using a formamide concentration of 35% (v/v). Quantification of the hybridised cells was performed as described previously (Neef et al. 1996). The hybridised cells were analysed by two independent observers for determining the fraction of positive signal from each probe relative to the signal visualised with general probes for bacteria (EUB338 I, II and III), SRB (SRB385 and SRB385Db) or with the general DNA stain DAPI (4′, 6′ -diamidino-2-phenylindole). In addition, a general probe specific for members of the domain Archaea (ARCH915) was used. The hybridisation experiments were done in duplicate using different fluorochromes for each probe. Different microscopic fields on each slide were analysed to confirm the results. Hybridisation stringencies of the newly designed probes were determined by performing hybridisations with increasing formamide concentrations as described previously (Manz et al. 1992) using target organism(s) and non-target organism displaying three mismatches within the target region.
Sequence accession numbers
The nearly complete rRNA gene sequences of strains L3 and L7 have been deposited in GenBank under accession nos. EF055876 and EF055877, respectively.
Results
Reactor performance
The amount of ethanol dosed to the reactor was continuous and limited to the amount required for sulfate reduction according to the net stoichiometric reaction for sulfidogenic oxidation of ethanol and its major degradation intermediate, acetate. The equations are as follows:
The sulfate removal efficiency at the time of sampling was more than 93%. Less than 1 mg of acetate was estimated in the effluent and no significant biogas (CH4) production from the oxidation of organic source was observed at the time of sampling.
Enumeration and isolation of SRB
The maximum number of culturable SRB observed with lactate or ethanol as substrate in the serial dilution tubes was in the order of 108 cells ml−1. The tubes with growth in the highest dilution were used for isolation in agar shake tubes. In total, 26 strains were obtained in pure culture, 17 on lactate and nine on ethanol as substrate. Sub-cultivation of the isolated strains was carried out on lactate. Apart from microscopic observation, the purity of the isolated strains was confirmed by single bands obtained from the DGGE of partial 16S rRNA genes amplified from the strains. The DGGE results of the 26 isolated strains identified two melting types (results not shown).
Genomic fingerprinting
The isolated strains were subjected to genomic fingerprinting, i.e. rep-PCR, to resolve any higher degree of genetic diversity among the strains. The rep-PCR profiles of the strains (Fig. 1) revealed the presence of two distinct genotypes with no micro-diversity among the strains. Furthermore, the whole cell protein profile of representative strains of the two groups, grown on the same substrate (i.e. lactate and sulfate) and under similar conditions of pH and temperature, gave only two distinct profiles (Fig. 2), confirming the presence of two genotypes. Two isolates, strain L3 and L7, were chosen as representatives of the two genotypes and were subjected to a more detailed phylogenetic and phenotypic analysis.
Fig. 1rep-PCR patterns of different strains isolated from the full-scale sulfidogenic bioreactor. M is the molecular weight markerFig. 2Whole cell protein profile of different strains isolated from the full-scale sulfidogenic bioreactor. M is the molecular weight marker. The arrows depict protein bands that are expressed in one or the other genotype
Phylogenetic analysis
A similarity check using the NCBI BLAST search tool of the nearly complete 16S rRNA gene sequences obtained from strain L3 and L7 indicated 98% sequence similarity to Desulfovibrio strain SB1 and 99% sequence similarity to Desulfovibrio mexicoense, respectively. The phylogenetic affiliation of the obtained sequences is presented in Fig. 3. A neighbour-joining tree based on nearly complete 16S rRNA gene sequences was generated, confirming the close affiliation of the isolated strains L3 and L7 to Desulfovibrio strain SB1 and Desulfovibrio mexicoense, respectively.
Fig. 3Neighbour-joining tree based on nearly complete 16S rRNA gene sequences showing the phylogenetic affiliation of the two isolated SRB, strain L3 and L7. The sequence of Desulfobacter postgatei was used as an outgroup but was pruned from the tree. Dots on the nodes indicate bootstrap values of 90% or higher (1,000 replicates).The bar indicates 1% sequence difference
Phenotypic characterisation of the strains
All isolates had a vibrio to spiral cell morphology. Two distinct groups could be identified based on size and motility. One group of cells related to Desulfovibrio strain SB1 was motile, and their size ranged from 4–7 μm in length and 1 μm in width. The other group was non-motile, and the size was 1–2 μm in length and 0.5 μm in width.
Table 2 shows the substrate utilisation patterns of the two isolates. Both strains used lactate, ethanol, pyruvate, glycerol and casamino acids as electron donors. Hydrogen and formate could only be used as substrates in the presence of acetate as carbon source. The organic substrates were incompletely oxidised to acetate. Acetate, propionate, butyrate, glycolate and methanol could not be used as electron donors by both strains. Fumarate and malate were used by strain L3 but not by strain L7, while serine and cysteine were utilised by strain L7 but not by strain L3. Sulfate, thiosulfate and sulfite were used as electron acceptors by both strains. In addition, strain L7 could also use elemental sulfur as an electron acceptor.
Table 2Phenotypic characterisation of strains L3 and L7CharacteristicsStrain L3Strain L7Cell morphologyVibrioVibrioMotility+−Desulfoviridin++Catalase++Gram staining−−Electron donors H2 plus acetate++ Pyruvate++ Lactate++ Acetate−− Propionate−− Fumarate+− Butyrate−− Formate++ Propanol−− Ethanol++ Methanol−− Serine−+ Cysteine−+ Malate+− Glycolate−− Glycerol++ Benzoate−− Casamino acids++Electron acceptors Thiosulfate++ Sulfite++ Elemental sulfur−+
Whole cell hybridisation
After isolation of the most abundant culturable SRB, whole cell hybridisation with specific oligonucleotide probes was performed to estimate the abundance of strains L3 and L7 in the original sludge sample. The specificity of the designed oligonucleotide probes (DSV119 and DSV139) was verified using growing cells of strain L3 and Desulfovibrio gigas for probe DSV119 and cells of strain L7 and Desulfovibrio mexicoense for probe DSV139. Strain L3 served as a non-target species for L7 and vice versa. A formamide concentration of 35% (v/v) was found stringent enough to discriminate between the two strains (Fig. 4a). In addition, probes specific for cells of Desulfobulbus (i.e. DSR660) and Desulfobacca acetoxidans (i.e. DSBA1017), which we had detected previously in a similar sulfidogenic reactor (Dar et al. 2007), were also used.
Fig. 4a Mixture of cells of strain L3 and L7 hybridised with probe Dsv139 labelled with Cy5 (blue), probe Dsv119 labelled with Cy3 (red) and probe SRB385 labelled with Fluos (green). b Sludge sample from a full-scale sulfidogenic bioreactor hybridised with probe Dsv139 labelled with Cy3 (red), probe Dsv119 labelled with Cy5 (blue) and SRB385 labelled with Fluos (green). c Sludge sample hybridised with probe Dsv139 labelled with Cy5 (blue), probe Dsv119 labelled with Cy3 (red) and EUB338 labelled with Fluos (green). d Sludge sample hybridised with probe DSR660 labelled with Fluos (green), probe SRB385 labelled with Cy3 (red) and EUB338 labelled with Cy5 (blue). e Sludge sample hybridised with probe SRB385 labelled with Fluos (green), probe DSBA1017 labelled with Cy3 (red) and probe EUB338 labelled with Cy5 (blue). f Sludge sample hybridised with probe Arch915 labelled with Fluos (green), probe SRB385 labelled with Cy3 (red) and EUB338 labelled with Cy5 (blue). Bar is 20μm
The relative percentage of cells that hybridised with probe DSV119, which is specific for strain L3, was 30–35% of the total SRB385-positive cells and 15–20% of the EUB338-positive cells (Table 3; Fig. 4b,c). The probe specific for strain L7, i.e. DSV139, was detected between 25–35% of the SRB385-positive cell and 10–20% of the EUB338-positive cells (Table 3; Fig. 4b,c). The Desulfobulbus and Desulfobacca-specific probes also gave a positive signal, but these SRB were somewhat less abundant than Desulfovibrio positive cells. The Desulfobulbus-specific probe, DSR660, targeted approximately 20–25% of the total SRB cells and 10–15% of the total bacterial cells (Table 3; Fig. 4d). The Desulfobacca acetoxidans-specific probe DSBA1017 targeted approximately 15–20% of the total SRB cells and 8–10% of bacterial cells (Table 3; Fig. 4e). The percentage of cells that hybridised with probe specific for Archaea (probe ARCH915) was less than 0.1% of the total DAPI stained cells (Table 3; Fig. 4f).
Table 3Relative abundance of SRB and ArchaeaaProbeSRB385EUB338DAPIDSV11930–3515–2012–16DSV13925–3510–208–16DSR66020–2510–158–12DSBA101715–208–106–8ARCH915––<0.1aPercentage of positive cells relative to those detected by a mixture of probes SRB385 and SRB385Db (SRB385), a mixture of probes EUB338 I, II and III (EUB338) and to cells stained with the DNA stain DAPI.
Discussion
Based on the observed sulfate removal efficiency (>93%) and the amount of ethanol dosed to the reactor, it may be assumed that a major part of electron flow from the substrate is scavenged by the SRB. The presence of less than a milligram of acetate in the effluent stream with no significant production of CH4 points to the oxidation of ethanol mainly through sulfate reduction, although fermentation of ethanol through acetogenesis cannot be ruled out. The metabolic products of fermentation and acetogenesis, mainly acetate and hydrogen, may serve as substrates for methanogens and sulfate reducers; however, under high sulfate concentrations in the reactor, hydrogen and acetate would be more readily used by hydrogenotrophic and acetate-utilising sulfate reducers, respectively, because of more favourable substrate affinity (Ks) values of the SRB for these substrates (Stams et al. 2005). Fermentation of ethanol with propionate as the major reduced end product has also been described in the literature (Schink 1984; Tholozan et al. 1992); and under such fermentation, propionate will serve as an energy source for the members of the genus Desulfobulbus among the SRB.
Genetic diversity among the dominant culturable isolates
Compared to high micro-diversity among phylogenetically similar strains detected in sediments (Sass et al. 1998; Wieringa et al. 2000; Klepac-Ceraj et al. 2004), no such micro-diversity was observed within the 26 strains isolated from the bioreactor. The chemical complexity of sediments with steep gradients of substrate concentrations, redox potential and pH may give rise to a number of physico-chemical and depth-defined micro-niches, resulting in the evolution of co-existing but genetically distinct sub-populations (Gray et al. 1999). In addition, Torsvik et al. (2002) suggested structural complexity of sediments being an important factor that allows nutritional partitioning, creating numerous niches that, in turn, results in genetic diversification of populations. From the 26 strains isolated, the ones that had identical 16S rRNA gene sequences were also identical by genomic fingerprinting (Fig. 1) and whole cell protein electrophoresis (Fig. 2). Based on the rep- and protein profiles, two genotypes could be identified among the 26 strains isolated. The absence of micro-diversity among the dominant culturable Desulfovibrio populations suggests the existence of a few ecological niches in the reactor. The long-term operation of the reactor under stable operational parameters, like constant temperature, pH, salinity etc., combined with the selection pressure because of nutritional limitation (i.e. the use of a single energy source), might be the reason for absence of micro-diversity. In addition, the upward flow of wastewater through the reactor results in frequent mixing, which, in turn, might prevent the creation of micro-habitats that are assumed to be important for the evolution of genetically distinct sub-populations (Torsvik et al. 2002).
Although the viable cell count of SRB observed (108 cell ml−1) was comparable to the previous studies (Vester and Ingvorsen 1998; Oude Elferink et al. 1999; Roest et al. 2005; van Houten 2006), it cannot be ruled out that potential biases associated with culture-based enumeration techniques might have underestimated the overall SRB population diversity.
Phenotypic and phylogenetic characterisation of the dominant culturable isolates
The physiological characteristics (Table 2) of strain L3 and L7 indicated that they are sulfate reducers. The two isolates oxidised lactate and ethanol incompletely to acetate in the presence of sulfate as electron acceptor. Both used H2 and formate in the presence of acetate as carbon source. These metabolic traits, in addition to the presence of desulfoviridin and a typical vibrio-shaped morphology, suggested that they are the members of the genus Desulfovibrio (Widdel and Bak 1992). Desulfovibrio are Gram-negative sulfate reducers, most of which oxidise their substrates incompletely to acetate (Widdel and Bak 1992). Previous studies have demonstrated the dominance of Desulfovibrio species in freshwater sediments (Sass et al. 1998), in oil wells (Voordouw et al. 1996) and in several wastewater treatment plants (Santegoeds et al. 1998; Dar et al. 2005). 16S rRNA sequence analysis of the two strains, L3 and L7, confirmed their affiliation to other members of the genus Desulfovibrio (Fig. 3),with Desulfovibrio strain SB1 and Desulfovibrio mexicoense as closest relatives, respectively. Desulfovibrio strain SB1, a mesophilic, Gram-negative SRB was isolated from anaerobic sludge of a gas lift reactor-treating sulfate and zinc-rich wastewater (van Houten 2006); while Desulfovibrio mexicoense was isolated from a UASB digester-treating wastewater from a cheese-manufacturing factory in Mexico (Hernandez-Eugenio et al. 2000).
Co-existing SRB populations in the reactor
After isolation of the most abundant culturable SRB, i.e. the motile strain L3 and non-motile strain L7, whole cell hybridisation using fluorescently labelled oligonucleotide probes was performed to search for their specific signals in the fixed sludge sample. Hybridisation results not only confirmed the presence of the two isolates but also gave an estimate of their abundance relative to the total SRB population and to the overall bacterial community present (Table 3; Fig. 4). Fluorescence in situ hybridisation (FISH) results indicated that the two isolates indeed made up a major part of the hydrogenotrophic SRB community present in the reactor. Cells detected by probe Dsv119, specific for the motile strain L3, appeared as single cells or as chains of three to four cells, while cells detected by probe Dsv139, specific for the non-motile strain L7, appeared as individual cells or as loose aggregates.
According to the ecological principle of competitive exclusion, co-existence of physiologically related populations in the same habitat can be understood if they occupy distinct ecological niches (Gause 1934). Although a number of similarities can be drawn between the strains, L3 and L7, from their substrate utilisation profiles (Table 2), the phenotypic differences together with differences in the range of substrates used by the two isolates may allow them to adapt to slightly different niches within the reactor. The two isolates differed in as many as five physiological properties, which are the use of malate, formate, serine and cysteine as energy sources and sulfur as an electron acceptor (Table 2). The use of elemental sulfur in particular by strain L7 will give it a selective advantage over strain L3 in those places in the reactor where elemental sulfur might be available because of the re-circulation fluid from the aerobic reactor; whereas the motility of strain L3 will confer a competitive advantage to it over strain L7. Potential benefits of motility may include increased efficiency of nutrient acquisition and avoidance of toxic substances (An et al. 2006). The presence of Desulfobulbus-like SRB besides the members of Desulfovibrio in the same habitat could be explained because of the ability of Desulfobulbus to use the fermentation product propionate as an energy source as well; furthermore, under limiting sulfate concentrations, Desulfobulbus competes more successfully for ethanol than other sulfate reducers by its ability to ferment ethanol (Laanbroek et al. 1982). The simultaneous presence of Desulfovibrio and Desulfobulbus in wastewater treatment systems has often been reported in the literature (Nanninga and Gottschal 1987; Raskin et al. 1995; Okabe et al. 2003).
Among the complete oxidisers, the probe specific to Desulfobacca acetoxidans gave a positive signal, suggesting their dominance among the acetotrophic sulfate reducers. Desulfobacca acetoxidans, first isolated from a sulfidogenic bioreactor (Oude Elferink et al. 1999), is a Gram-negative SRB that can utilise acetate as the only source of organic carbon and electron donor. Kaksonen et al. (2004) also found Desulfobacca acetoxidans-like SRB in their lab-scale fluidised-bed reactors that were fed with a single electron donor, i.e. lactate or ethanol.
In summary, this study demonstrated the presence of a consortium of four sulfate-reducing populations in the reactor maintained on a single substrate (ethanol). This is in contrast to the findings of Kaksonen et al. (2004) who reported a relatively more diverse consortium of SRB in a fluidised-bed reactor fed with ethanol or lactate as the only energy source, using clone libraries and DGGE as molecular methods for microbial characterisation. The observed difference in the extent of diversity might be due to the difference in the molecular methods employed. The general probes like DSR660, used during FISH analysis, cannot discriminate among the different species, and the probes like DSV119, DSV139 and DSBA1017 target only the specific populations. A likely complete oxidation of ethanol through sulfate reduction might thus be assumed to be taking place through a combined effort of hydrogenotrophic sulfate reducers (i.e. the motile strain L3, the non-motile strain L7 and Desulfobulbus spp.) oxidising ethanol incompletely to acetate and Desulfobacca acetoxidans oxidising acetate completely to CO2. Genetic diversity analysis of the most dominant culturable Desulfovibrio populations suggests that long-term stable operation of a reactor with constant operational parameters and frequent mixing might result in the absence of micro-diversity.
|
[
"sulfate-reducing bacteria",
"ecological niches",
"microbial ecology"
] |
[
"P",
"P",
"R"
] |
Neurochem_Res-3-1-1794624
|
The Role of Citrullinated Proteins Suggests a Novel Mechanism in the Pathogenesis of Multiple Sclerosis
|
The pathogenesis of MS is unknown. In our studies, we have demonstrated an important role for citrullinated myelin basic protein (MBP). The accompanying loss of positive charge compromises the ability of MBP to interact with the lipid bilayer. The conversion of arginine to citrulline in brain is carried out by an enzyme peptidyl arginine deiminase (PAD) 2. The amount of PAD 2 in brain was increased in MS normal-appearing white matter. The mechanism responsible for this increase involved hypomethylation of the promoter region in the PAD 2 gene in MS, but no change (compared to normal) was found in thymus tissue DNA from the same MS patients. In addition, no change was observed in other neurological diseases, including Alzheimer’s, Parkinson’s, and Huntington’s. We propose that citrullinated MBP, resulting from elevated levels of PAD 2 represents an important biochemical pathway in the pathogenesis of MS.
Introduction
Demyelinating diseases, such as multiple sclerosis (MS), the major demyelinating disease of humans affects not only young adults, but also children. It is the leading cause of neurological deficit in North America and Western Europe [1]. It is characterized by the patchy destruction of the myelin sheath around the axon and is, therefore, classified as a primary demyelinating disease [2]. The reason for this patchy destruction is not known. The MS plaque is characterized by a focal area of myelin destruction associated with astroglial scar formation. These lesions are scattered throughout the CNS with a predilection for optic nerves, brain stem, spinal cord, and periventricular white matter [3]. Perivascular lymphocytic infiltration with macrophages, microglial cells and astrocytes is a common feature. More recently, it has been recognized that extensive oligodendrocyte apoptosis and myelin macrophages are found in MS tissue, suggesting a primary neurodegenerative mechanism, followed by the autoimmune response as a secondary event [4]. In this type of apoptosis, the cardinal feature is damage to the most distal part of the oligodendiocyte process in the periaxonal region [5]. The chemical changes in myelin that precede this destruction are under intensive investigation. In the present overview, we suggest that enzymatic changes induced in myelin basic protein (MBP) may represent an important factor.
The cause of MS is unknown. Two major theories have emerged: an autoimmune or a neurodegenerative cause. The autoimmune etiology is considered to arise from sensitization of T-cells in the periphery by a mechanism of molecular mimicry, usually involving a viral antigen with peptide sequences similar to those found in myelin proteins, especially MBP, a candidate autoantigen in MS. The neurodegenerative hypothesis involves metabolic changes in the myelin constituents, which destabilize the membrane architecture, resulting in myelin degradation. Since the myelin proteins are primarily responsible for maintaining the structure of the membrane we focused our attention on myelin proteins, in particular, modifications that are found on MBP.
Myelin basic protein is an unusual protein. It is a single protein chain consisting of 170 amino acids. Since it does not contain cysteine, it cannot form disulfide bonds. It is a flexible structure, which adapts to environmental conditions, by changing conformation. It is a member of a group of proteins considered to be intrinsically disordered and, therefore, have considerable conformational flexibility (reviewed extensively in the contribution of Harauz and Musse in this issue). A further unusual feature of MBP is the large number and variety of post-translational modifications, such as N-terminal acylation, GTP- and ADP-ribose binding sites, deamidation, methylated arginine, methionine sulphoxide, phosphorylation, and deimination of arginyl residues. Citrullination occurs on six sites in the 18.5 kDa MBP from humans, while phosphorylation is found on ten sites at least. These are summarized in Fig. 1.
Fig. 1Post-translational modification of MBP. All post-translational modifications, except methylation of Arg107 (human sequence), decrease the positive charge. The known modifications are acylation of the N-terminus, GTP binding, ADP-ribosylation (if R25 is not citrullinated), deamidation, methionine sulphoxide, mono, and dimethylation of Arg107, phosphorylation at Ser/Thr residues and Tyr and deimination of arginyl residues (six sites in the 18.5 kDa MBP from normal brain). The structure, shown as a hairpin, occurs at the tri-proline region at the site of the bend. The figure was adapted and modified from Moscarello (1990, Myelin basic protein: a dynamically changing structure. In: Dynamic interactions of myelin proteins, George Hashim and Mario Moscarello (eds) Prog Clin Bio Res, vol 336)
One of the modifications, in which we are particularly interested, involves the conversion of peptide-bound arginine to peptide bound citrulline, an enzymatic reaction called deimination. The conversion of arginine to citrulline involves the release of ammonia. For each arginine converted to citrulline, one positive charge is lost from the protein. We first demonstrated that citrulline was present in a crude MBP fraction from myelin [6]. We then purified the citrulline containing MBP on CM52 cation exchange columns essentially as described earlier [7–9]. From these columns, we obtained a number of fractions, all of which were MBP. Fractionation of the various MBPs was shown to be dependent on the net charge of the highly modified MBP, resulting from phosphorylation, deamidation, C-terminal arginine loss, and others (Fig. 1). The citrullinated MBP was recovered in the flow-through volume and required further purification by high-performance liquid chromatography (HPLC). The purified protein was sequenced and the arginyl residues that were deiminated were identified. These were R25, R31, R122, R130, R159, and R169 [10]. This decreased positive charge affected the ability of the protein to interact with and organize lipid vesicles. Using liquid X-ray diffraction, we were able to show that the citrullinated MBP (termed “C-8” for convenience) was unable to organize lipid bilayers into compact multilayers, which was readily achieved by the non-citrullinated charge isomer (termed C-1, so-called because it bound most tightly to the CM52 column) [11]. At this time, we speculated that this failure to organize bilayers into compact multilayers, may contribute to myelin instability as seen in MS. These data demonstrated that the citrullinated MBP had a limited ability to organize and compact the lipid bilayers.
The citrullinated MBP was isolated from white matter of MS brain and compared to that isolated from normal brain. We reported that the relative amount of the citrullinated MBP was increased in MBP isolated from MS brain. Whereas, it accounted for 20% of the MBP isolated from normal brain, a threefold increase was found in chronic MS. Thus, the ratio of citrullinated MBP/non-citrullinated MBP was 0.82 in the normal and it was 2.45 in the MS MBP [12]. In a single case of Marburg’s disease (a fulminating form of MS), 80–90% of the MBP was citrullinated, with a citrullinated MBP/non-citrullinated MBP ratio of 6.7 [13, 14]. Therefore, MBP in Marburg’s type of MS was almost totally deiminated. Citrullination of MBP has been reported to diminish charge density, which diminishes its interaction with lipids, making it more vulnerable to proteolytic attack [15].
The effects of deimination on protein structure and protein-lipid interactions in myelin are diverse and have been thoroughly reviewed [16]. Briefly, deimination of MBP alters the 3D structure of MBP, producing a more open conformation. The more open conformation of MBP increases its susceptibility to proteases such as cathepsin D, a myelin associated protease. MBP containing six citrullinyl residues is digested four times faster than non-deiminated MBP. The MBP containing 18 citrullinyl residues was digested 45 times faster than MBP from normal brain. These data support the view that deimination of MBP not only decreases the positive charge on the molecule, but also creates a more open structure [17].
These perturbations in structure affect the interactions of the protein with the lipid bilayer and decrease the compaction of the multilayer structure of myelin. The deiminated protein had a reduced ability to bind and aggregate lipid vesicles and the highly deiminated form of MBP caused fragmentation of lipid vesicles [18], a process which may have implications for myelin breakdown in MS.
Electrostatic interactions between the negatively charged phospholipids and the positively charged amino acids (Arg and Lys) are the origin of attraction of MBP to the lipid. MBP is postulated to work as a glue for adjacent bilayers [19, 20] by binding to the polar head groups. This mechanism of interaction is interfered with by deimination of arginyl residues, since the resulting citrulline carries a neutral charge.
The conversion of arginine→citrulline in proteins is carried out by the enzyme peptidyl arginine deiminase (PAD). Of the five known PAD enzymes, PAD 2 is the one involved in the deimination of MBP. Since our studies on the deiminated MBP in humans was obtained from autopsy material, the question of the origins of the citrullinated protein, as possibly artefactual, was addressed. To resolve this issue, we explored the temporal relationship between the appearance of the citrullinated protein and the up-regulation of the PAD 2 enzyme, which could only be done in a relevant animal model. Such a model was available to us. It is a transgenic model, containing 70 copies of the transgene for the myelin proteolipid protein DM20. This mouse is normal for the first 3 months of its life and then demyelinates spontaneously. We measured PAD enzyme and citrullinated MBP. We found that PAD 2 was up-regulated at 2 months of age and citrullinated MBP at 3 months of age, suggesting a precursor-product type relationship [21]. We postulated that up-regulation of PAD 2 was the primary response. To substantiate this hypothesis, a thorough study was carried out on PAD 2 regulation.
In this regard, we have reported that paclitaxel (Taxol) inhibited the activity of purified PAD 2 isolated from bovine brain in an in vitro assay [22]. In in vivo studies, we demonstrated that paclitaxel attenuated clinical disease in the spontaneously demyelinating mouse model described above [23]. In a recently submitted manuscript, we extended our Taxol studies further by demonstrating that the addition of the methyl donor, vitamin B12, suppressed both immune and non-immune demyelinating disease in mice (F. G. Mastronardi et al. 2006, submitted data), better than paclitaxel alone. These data suggest that drugs, which target PAD 2, in the presence of a methyl donor, such as vitamin B12, may be beneficial in demyelinating disease.
Another PAD enzyme, PAD 4, has been implicated in the pathogenesis of psoriasis [24], rheumatoid arthritis [25, 26] and MS (F. G. Mastronardi et al. 2006, submitted data). PAD 4 is different from other PAD enzymes, because the gene contains a nuclear localization signal (PPAKKKST) [27], not present in any of the other PAD enzymes. Translocation of the enzyme from the cytoplasm to the nucleus, has been shown to involve the cytokine tumor necrosis factor α (TNFα). The substrate for PAD 4 in the nucleus is histone (F. G. Mastronardi et al. 2006, submitted data). Deimination of arginyl residues of histones decreases the positive charge on histone, which compromises its ability to interact with DNA, possibly resulting in apoptosis of oligodendrocytes. On the other hand, deacetylation of histones would increase the positive charge on histone. Therefore, these two processes, the deimination of arginyl residues and the deacetylation of lysyl residues of histones affect the positive charge on histone in opposite ways, and these two processes may have a regulatory function.
Since we determined that up-regulation of PAD 2, preceded demyelination in the animal model, we searched for the mechanism responsible. An examination of the nucleotide sequence of the PAD 2 promoter revealed that it was high in CpG sequences. CpG sequences are important, because cytosine is a target for methylation by DNA methyltransferase, which produces 5-methylcytosine. Methylation of the promoter region of a gene silences the gene, whereas demethylation increases transcription.
Accordingly, we isolated DNA from white matter obtained from normal individuals and patients who died with MS. The DNA was treated with bisulphite to convert cytosine into uracil. The methylcytosine is not affected by bisulphite. We found that methylcytosine accounted for less than 4% of the cytosines in the PAD 2 promoter in MS (n = 17), whereas in the normal, the methylcytosine in the PAD 2 promoter was 12% (n = 8). Therefore, the PAD 2 promoter was hypomethylated in the DNA from MS white matter samples. DNA was also obtained from thymus of the same MS patients. The amount of methylcytosine was the same in both MS and normal samples. Therefore, the hypomethylation found in MS appeared to be tissue specific. When we examined other neurological diseases, including Alzheimer’s, Parkinson’s, and Huntington’s, the amounts of methylcytosine were the same as normal (Mastronardi et al. 2006 in preparation). Therefore, the hypomethylation of the PAD 2 promoter was unique to MS brain. This hypomethylation of the promoter of PAD 2 may account for the increase in PAD 2 in MS brain (summarized in Fig. 2).
Fig. 2Deimination pathway in the pathogenesis of demyelinating disease. Normal myelin is shown on the left. The promoter of the PAD 2 gene contains methylated cytosines (closed lollipops) along its CpG island, which regulates its transcription. Some MBPcit is made under normal conditions by the deimination of arginyl residues “C” on MBP. The amount of MBPcit is not enough to destabilize the bilayer. During demyelination, the promoter of PAD 2 is demethylated (open lollipops) at cytosines in the CpG island. This demethylated promoter increases transcription of PAD 2, which increases the amount of MBPcit in myelin with subsequent destabilization of the myelin. Degradation of myelin protein fragments, in particular MBPcit fragments, become available to sensitize T-cells in the periphery, initiating the immune response
The removal of methyl groups from methylcytosine in CpG sequences is carried out by a DNA demethylase. We proposed that up-regulation of the demethylase may account for the hypomethylation state of the PAD 2 promoter in MS. When we measured the total demethylase activity in brain homogenates from normal and MS white matter, using a synthetic methylcytosine-guanosine (CpG) substrate [28] we demonstrated that the enzyme in NAWM from MS brain was increased twofold over normal, suggesting the DNA demethylase may have an important role in the hypomethylation of the PAD 2 promoter. Therefore, the hypomethylated state of the PAD 2 promoter may involve up-regulation of the DNA demethylase at least in part.
As a result of the work summarized in this overview, we postulate the following sequence of events during the pathogenesis of MS. The citrullinated MBP, because of its decreased positive charge, prevents proper compaction of the bilayer, so that the multilamellar structure is less stable. Since PAD 2 has been shown to be present in myelin by immunogold electron microscopy (unpublished results), the conversion of non-citrullinated to citrullinated MBP can occur locally in myelin. Preliminary results have shown that immunogold labeling of PAD 2 in mouse optic nerve was not random, but was found in clusters in the axon itself and in the myelin sheath. In the latter, the clusters were found mainly at the junction of the myelin sheath and the axon (periaxonal) with some gold particles distributed in the myelin sheath. The presence of the PAD 2 enzyme at these sites suggests that the citrullinated MBP would be formed in these clusters in myelin, eventually resulting in plaques.
When we examined the extent of labeling in the optic nerve of the spontaneously demyelinating transgenic mouse, we found a threefold increase in the number of gold particles. This increase in PAD 2 labeling demonstrated that increased amounts of PAD 2 were present in the transgenic mouse, accounting for an increase in the citrullinated MBP. Since citrullinated MBP is present in the normal myelin sheath (20% of the total MBP), citrullinated MBP can be tolerated to this extent. However, increased amounts as found in chronic MS (threefold), and in fulminating MS (six to sevenfold), cannot be accommodated in a compact myelin sheath, resulting in destabilization.
Peptidyl arginine deiminase 2 has been reported recently to be elevated in primary open-angle glaucoma, where it was considered responsible for the deimination of several optic nerve proteins, in addition to MBP. Concomitant with this protein deimination, a decrease in arginine methylation was also observed [29]. Therefore, the PAD 2 enzymes have important roles in several nervous system diseases.
Oligodendrocyte apoptosis has been reported in MS white matter in the absence of lymphocytes or myelin phagocytes [4]. The major feature of this type of injury is damage to the most distal part of the oligodendrocyte process (periaxonal), which results in clumping of proteins [5], followed by apoptosis. The possible role of the PAD 2 enzyme in this process has not yet been demonstrated. If it can be established that protein deimination at sites of PAD 2 clusters does occur, it may suggest that demyelination begins in the periaxonal myelin sheath and subsequently moves outward. Although this hypothesis is speculative, it warrants investigation. The increased amount of PAD 2 in MS white matter probably reflects a greater production of the enzyme. Since the PAD 2 promoter is rich in CpG islands, we showed that the promoter was hypomethylated. These studies are ongoing in our laboratory. They suggest immediately that the reversal of the hypomethylated state may be a therapeutic target in MS.
The importance of methylation in a therapeutic strategy of MS has been suggested by us recently [(F. G. Mastronardi et al. 2006, submitted data), 30]. In these studies, we demonstrated that vitamin B12 in combination with either interferon β or paclitaxel, effectively ameliorated demyelinating disease in both EAE and non-immune demyelinating disease. In a detailed mass spectrophotometric study of the various components of MBP, we showed arginine methylation was decreased in most of the components of the MBPs from MS material [31], supporting the view that defects in methylation may involve protein methylation as well. The methylation process requires Vitamin B12, which transfers its methyl group to homocyteine synthesizing methionine, which is then converted to S-adenosytmethionine, the methyl donor in all biological methylation reactions, including methylation of cytosine in DNA to methylcytosine. These studies with the preliminary studies reported above suggest an important role for methylation/demethylation reactions in the pathogenesis of MS.
|
[
"citrulline",
"multiple sclerosis",
"myelin",
"peptidyl arginine deiminase",
"demyelination",
"apoptosis"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Matern_Child_Health_J-2-2-1592156
|
Toxoplasmosis, Cytomegalovirus, Listeriosis, and Preconception Care
|
Toxoplasma gondii (T. gondii), cytomegalovirus (CMV), and Listeria monocytogenes (L. monocytogenes) can all negatively affect pregnancy outcomes. Preconception counseling about such effects can reduce the risks posed by these pathogens. Informing women of childbearing age about these pathogens and how to prevent their negative effects can help women make informed decisions about prevention. This brief summarizes some basic information about these infections and provides some web sites and articles for further information about how to prevent them within the context of preconception care.
Toxoplasmosis is a disease caused by Toxoplasma gondii (T. gondii), a protozoan parasite mainly transmitted to humans via three routes: a) ingestion of raw or undercooked contaminated meat; b) exposure to T. gondii oocysts (a form of the organism passed in cat feces), through cat litter or soil (e.g., from gardening or unwashed fruits or vegetables), or contaminated water; and c) congenital in which maternal infection is passed transplacentally via blood to the fetus [1]. Congenital infection leads to stillbirth and severe neurological illness in some instances, although the majority of infected newborns are asymptomatic at birth and some develop sequelae such as mental retardation, blindness, and epilepsy later in life [2]. Extrapolation from regional studies suggests that ∼400–4,000 cases of congenital toxoplasmosis occur each year in the United States [2].
Adults with normal immune function who are infected with T. gondii are usually asymptomatic or have self-limited symptoms (e.g., fever, malaise, and lymphadenopathy) [1]. Once infected, these individuals usually develop an immune response against toxoplasmosis [3, 4]. A recent study based on the National Health and Nutrition Survey conducted from 1988–1994 (NHANES III) reported that, among women aged 15–44 years, seroprevalence of T. gondii antibodies was 15%, suggesting that ∼85% of women of childbearing age are susceptible to T. gondii infection [5].
Three principal interventions are presently used to reduce morbidity and mortality from congenital toxoplasmosis: a) education about how to prevent infection (especially during pregnancy); b) prenatal and newborn screening to identify and treat congenital infection; and c) animal rearing and production methods designed to reduce T. gondii contamination of meat. Of the three, education about how to prevent infection is applicable to adolescents and women in the preconception period. Education programs during pregnancy have been associated with improved knowledge and behavior and a reduction in infection rates [6–9]. Toxoplasma infection can be prevented by one or more of the following: a) cooking meat to a sufficient temperature to kill Toxoplasma; b) peeling or thoroughly washing fruits and vegetables before eating; c) cleaning cooking surfaces and utensils after they have contacted raw meat, poultry, or unwashed fruits or vegetables; d) pregnant women avoiding changing cat litter or using gloves, then washing hands thoroughly; and e) not feeding raw or undercooked meat to cats and keeping cats inside to prevent acquisition of Toxoplasma by eating infected prey [2].
Although prenatal and newborn screening programs have been evaluated, they are controversial because of the lack of proven efficacy of treatment, side effects of treatment, and potential complications of invasive procedures such as amniocentesis to evaluate fetal infection [10–23]. Animal rearing and production methods to reduce T. gondii contamination of meat are effective and have been associated with a reduction in the prevalence of T. gondii in important meat sources such as pork [24]; however the need for improvement in producers’ knowledge and production practices is acknowledged [25]. Education about toxoplasmosis is an important component of preconception care that can be integrated with information about other diseases that affect women.
Human cytomegalovirus (CMV) is the largest DNA virus belonging to the herpesvirus family. Humans are the only reservoirs for the human herpesviruses, and they can transmit these agents through direct contact with infected blood, tissues, bodily fluids, feces, and fomites [26]. A pregnant woman infected with CMV can transmit this virus to her unborn fetus, which can cause damage to the central nervous system, hematopoietic system, kidneys, endocrine glands, gastrointestinal tract, lungs, and liver. Long-term sequelae include cerebral palsy, mental retardation, and hearing loss [26, 27]. The birth prevalence rate of congenital CMV infection varies between ∼0.6–1.5% [28–30] in the United States, making it the most commonly transmitted virus in utero [26, 31] and a major cause of cerebral palsy, mental retardation, and hearing loss among children [27, 31].
CMV infects almost all humans at some point in their lives. Adults with normal immune function infected with CMV are usually asymptomatic or might experience mild flu-like symptoms, or even mononucleosis with symptoms such as malaise, persistent fever, myalgia, and cervical lympadenopathy [26, 31]. Once a human is infected, the virus passes into a latent state [26]. Although the virus can be reactivated, it is usually kept under control, because adults with normal immune function usually retain lifelong immunity against CMV.
Primary CMV infection of women during pregnancy or periconceptionally results in transmission to the fetus transplacentally in ∼30–40% of maternal infections [26, 32–34]. Preexisting maternal immunity strongly reduces the risk of transmission to the fetus [34]. However, approximately 10–15% of newborns with congenital CMV will be symptomatic and from ∼6–25% of those born without symptoms will develop late sequelae [35].
The most common means for women to be infected with CMV is by exposure to toddlers who shed large amounts of the virus in their saliva and urine for many months following their first (usually asymptomatic) infection [26, 36–40]. Daycare providers and pregnant women who have a toddler of their own are at high risk for infection [41–46]. Sexual transmission, blood transfusion, and organ transplantation are other means by which CMV is transmitted [26]. It is important to counsel all women about safe sex practices.
Currently, there are no vaccines available for preventing CMV infection, although some promising advances have been made [47, 48]. It is, therefore, extremely important to provide women of childbearing age with information about how they can prevent CMV infection before conception. The most effective means for preventing CMV infection is handwashing [26, 49–53]. Education about careful hygiene and frequent handwashing, especially after contact with the saliva and urine of young children, and careful disposal of diapers, tissues, and other contaminated items can reduce the transmission of CMV. Avoiding sharing drinking glasses and eating utensils with young children can also prevent transmission. This is especially true for women who work in daycare settings and for those who are pregnant [26, 49–54].
As some studies suggest that CMV infection prior to conception can result in congenital CMV infection [26, 30, 55], counseling all women of childbearing age about how to prevent transmission preconceptionally could significantly reduce the incidence of congenital CMV infection.
Listeriosis is an invasive foodborne infection caused by the motile, gram-positive bacterium Listeria monocytogenes (L. monocytogenes). The disease affects primarily pregnant women, newborns, and adults with weakened immune systems [56]. Maternal infection during pregnancy is usually a self-limited, nonspecific acute febrile illness of the third trimester. In contrast to the maternal illness, fetal and neonatal infection is severe and frequently fatal. Infection in the fetus and newborn occurs by transplacental transmission or possibly from exposure to L. monocytogenes in the perinatal period. The effects of intrauterine infection on the fetus and neonate include preterm labor, amnionitis, spontaneous abortion, still birth, and early-onset of the neonatal sepsis syndrome, evident at birth or shortly thereafter. Late-onset neonatal disease is likely due to infection at or around the time of birth, and presents as meningitis at one to several weeks of age [57].
Listeriosis is a rare disease; the incidence rate in 10 states participating in the Foodborne Diseases Active Surveillance Network (FoodNet) was 2.7 cases per 1,000,000 population in the year 2004 [58]. However, the incidence rate in susceptible subgroups is much higher. The rate among neonates younger than 28 days of age in FoodNet sites was 52.8 per 100,000 population in the year 2000 (CDC unpublished data). More importantly, listeriosis has a very high case fatality rate (20–30% in neonates) and is responsible for an estimated 500 deaths each year in the United States [59].
The food items implicated in outbreaks of listeriosis include ready-to-eat meats, such as turkey delicatessen meat [60], meat paté [61], pork tongue in jelly [62], hot dogs [63], and dairy products made from unpasteurized milk, in particular fresh soft cheeses [64].
Primary prevention for listeriosis focuses on improvements in food processing and on consumer education. Substantial efforts by the food industry and food regulatory agencies have been directed toward reducing the likelihood that high risk foods will be contaminated with L. monocytogenes [65]. Despite this, pregnant women, immunocompromised persons, and the elderly should be advised to avoid paté, fresh soft cheeses made from unpasteurized milk and to cook ready-to-eat foods such as hotdogs, delicatessen meats, and left over foods until steaming [66]. Information regarding these foods at high risk of contamination with Listeria can be incorporated into preconception care dietary recommendations.
For more information
CDC web sites
Toxoplasmosis: http://www.cdc.gov/ncidod/dpd/parasites/ toxoplasmosis/default.htm
CMV: http://www.cdc.gov/cmv
Listeria: http://www.cdc.gov/ncidod/dbmd/diseaseinfo/ listeriosis_g.htm
Publications for practitioners
Disclaimer: These references are included for information only. CDC has no control over the information in these articles. Views and opinions of these organizations are not necessarily those of CDC, the Department of Health and Human Services (HHS), or the U.S. Public Health Service (PHS).
American College of Obstetricians and Gynecologists. Perinatal viral and parasitic infections. ACOG Practice Bulletin 2000;20.
Brundage, SC. Preconception health care. American Family Physician. 2002 June; 65(12):2507–14.
Hanlin RB. Congenital infections and preconception counseling. J S C Med Assoc. 2002 Oct;98(6):277–80.
|
[
"toxoplasmosis",
"cytomegalovirus",
"listeriosis",
"preconception care",
"listeria",
"infection"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
J_Interv_Card_Electrophysiol-4-1-2292475
|
Acute success and short-term follow-up of catheter ablation of isthmus-dependent atrial flutter; a comparison of 8 mm tip radiofrequency and cryothermy catheters
|
Objectives To compare the acute success and short-term follow-up of ablation of atrial flutter using 8 mm tip radiofrequency (RF) and cryocatheters.
Introduction
Atrial flutter is a common arrhythmia [1], difficult to suppress with medication [2], and is associated with significant symptoms. Since it was first proposed [3–5], ablation for atrial flutter has increasingly been used for its therapy, especially since induction of bidirectional cavo-tricuspid isthmus block was shown to be associated with better immediate outcomes and lower recurrence rates [6–11]. Failure to successfully ablate atrial flutter in the long-term may be due to particular anatomic problems [12–14], poor catheter stability in this region, and incorrect interpretation of isthmus block. Important developments in the field of atrial flutter ablation have been a better understanding of the anatomy of the isthmus, refinements of the definition of bidirectional isthmus block, and the arrival of new catheter technology.
Numerous studies have compared different energy types, different catheter tip sizes and different energy settings [15–19], as well as the use of advanced cardiac mapping systems [20]. The preferred tools now are 8 mm, or irrigated tip catheters [21]. From a recent meta-analysis it would appear that these two technologies are equally effective with acute success quoted at 84% and 85% for the primary catheter chosen before changeover, whereafter final success rates of up to 99% can be achieved [22]. Not all studies have this high success rate and the preponderance of studies on ablation of atrial flutter suggest that these high success rates are not repeated in all centers [23]. Also of importance is a clinical recurrence rate, in the face of acute success, of at least 5 to 12% when using radiofrequency [19, 24, 25]. Disadvantages of radiofrequency (RF) energy include pain, overheating with “popping”, char formation and risk to coronary arteries.
Cryoablation is a relatively recent addition to the transvenous ablation armamentarium and has been shown to be comparable to RF for some arrhythmias. It may have some advantages over RF, especially as regards discomfort during ablations [17, 26]. The development of 6 and 8 mm tip cryocatheters has increased the interest in this technology for atrial flutter ablation. As far as we are aware there have been no randomized comparative studies published comparing 8 mm tip cryocatheters with 8 mm tip RF catheters.
Methods
Patient population
Consecutive patients with ongoing symptoms and documented atrial flutter with or without fibrillation were included. At least one recent episode of atrial flutter (within the last 6 months) was documented on a 12 lead ECG and was suggestive for isthmus dependency. Patients with drug-induced flutter (and prior AF) could be included, and had their drug therapy continued after ablation. Patients were excluded if they had undergone a previous flutter ablation, if thrombus was present in the atria, or after previous cardiac surgery for valvular or congenital heart disease. Baseline investigations included a standard echocardiogram, a simple questionnaire asking about clinical wellbeing and a subjective assessment of the arrhythmia burden both in terms of duration and frequency. The study was approved by the ethics committee of our institution. All patients signed written informed consent.
Assessment before ablation
Antiarrhythmic drugs, except for AV nodal slowing agents, and amiodarone, were discontinued at least five half-lives prior to ablation. Patients were studied in the fasting, post-absorptive state. A coronary sinus (CS) catheter was inserted through the left subclavian vein, and a multipolar circular right atrial catheter with alternating 2–10–2 mm interelectrode distance was positioned from the right groin, with the tip positioned immediately lateral to the planned position of the isthmus line and anterior to the crista terminalis. Heparin 100 U/kg was given and a further 5 U/kg given if the procedure lasted longer than 180 min.
If the patient was in sinus rhythm, isthmus conduction was confirmed by pacing. If flutter was present, entrainment was performed to confirm isthmus dependence, and the patient left in flutter. If AF was present, the patient was cardioverted after a transoesophageal echocardiogram, and then isthmus conduction confirmed. Absence of isthmus conduction or non-isthmus dependence was not seen in the selected patients. No induction of arrhythmia was attempted if patients had sinus rhythm.
Radiological assessment of the right atrial isthmus was made in a right anterior oblique (RAO) 30°, and a left anterior oblique (LAO) 45° view (each with 40 cc at 18 cc/s). Angiograms were acquired digitally to allow for post-hoc analysis. The treating physician was able to view the angiographic findings to optimize the planned ablation line. The isthmus length was assessed from the inferior hinge point of the tricuspid valve to the IVC at the end of atrial diastole (the frame before opening of the tricuspid valve) [14]. Morphology was assessed visually as to the presence of a Eustachian valve or a recess, as well as to the general shape i.e. flat or concave.
Ablation procedure
The catheters were a 9Fr 8 mm tip catheter (FreezorMax, Cryocath Technologies Inc, Kirkland, Canada) with a cryoconsole for the cryoablation group, and a 7Fr 8 mm tip single sensor catheter (EPT Blazer II, Boston Scientific, Natick, MA, USA) with an EPT-1000XP generator for the radiofrequency group. A large curve was initially selected in both groups, with change out of catheter curve during the study only as necessary. Applications of −75°, for 4 min were given with cryothermy, and applications of 60 Watt, for 60 s, targeted at 60°C for RF. Lines were made with discrete applications between the tricuspid valve and the inferior vena cava at an approximately 6 o’clock position in LAO 45°, unless otherwise dictated. If termination of atrial flutter occurred, or if the patient was in sinus rhythm, continuous pacing from the proximal coronary sinus was employed to continually assess isthmus conduction. After the first line, a new assessment of conduction was performed. If conduction over the isthmus remained present, gaps were sought. If there was still isthmus conduction, a slightly more medial or lateral line was made. In no patient was an attempt made to perform a septal ablation line. Final assessment of acute block was confirmed after 30 min waiting.
The end point for successful ablation was induction of complete bidirectional isthmus block, defined as the presence of reversal of activation on the lateral and septal wall when pacing the CS os and low lateral RA, the presence of widely split potentials along the isthmus line, by activation mapping across the isthmus, and by differential pacing. All 4 were required before calling the ablation successful. In the case where bidirectional block was not achieved, ablation was stopped when no large, sharp signals could be identified over a broad area of the isthmus.
As pain perception was assessed, sedation was standardized. Before venous puncture 5 mg of diazepam was given intravenously, and repeated at the patient’s request. Fentanyl 50 μg intravenously was given when the patient requested pain control and the physician considered this necessary. This was repeated as needed. Dosages of both diazepam and fentanyl were recorded.
In the initial 40 patients creatine kinase (CK) and CK-MB were taken before the procedure, 2 and 24 h after the start of the procedure. For the final 22 patients the laboratory had changed the measurement to CK mass. We then modified the protocol to measure CK-MB mass, Troponin T, and Myoglobin at 4 and 24 h after the start of the procedure.
No crossover, other than in catheter curve, was allowed in an attempt to remove any possible bias. Change over to an irrigated tip ablation catheter was also not allowed. In patients in whom no block could be induced, a repeat procedure was scheduled not earlier than 6 weeks after the initial ablation, at the physicians’ discretion. The choice of energy source at that time was at the physicians’ discretion.
Patients were all questioned with regard to pain perception using a visual analogue score, where patients are shown a line from 0 to 10, where 0 is no pain, and 10 is the highest pain level imaginable, and were asked to point to the position on the line where their pain level during ablation was.
Follow-up
All patients received an event-recorder for the first 6 weeks after the procedure and were requested to send at least daily strips as well as strips made during symptoms. Patients visited the outpatient clinic 6 weeks after the procedure. After this period all patients were asked to report symptoms and if these were present were given a further event monitor until documentation of symptoms was obtained. If at all possible a 12 lead ECG was also obtained. A second assessment with a questionnaire was performed after 3 months, again asking a question about general clinical well being and also symptom burden in regard to duration and frequency. Clinical files were followed up after 9 months.
Statistical analysis
For patients in whom another ablation was performed (AVNRT in two, and pulmonary vein ablation in six others) procedure and fluoroscopy times were limited to the flutter approach, which was done first, including 30 min waiting time. Biomarker assessment was not done in these patients. Continuous variables are expressed as mean ± standard deviation, with medians as necessary. Parametric and non-parametric tests were used where appropriate. A p-value of < 0.05 was considered significant.
Results
Patient data
62 patients were included as planned, with clinical characteristics as outlined in Table 1. There were no significant differences in any of the parameters between the group assigned to radiofrequency (RF group) versus that assigned to cryotherapy (cryo group). The large number of patients with prior AF is due to the fact that we had initially taken a decision to perform isthmus ablation first in all patients with AF who had shown typical atrial flutter on any 12-lead ECG, prior to performing a left atrial procedure, initially in a separate procedure and only later in the same session.
Table 1Demographics AllCryoRFp valueNumber623230Age (years ± SD)56 ± 1055 ± 1156 ± 9NSMale/female27/527/528/2NSAtrial fibrillation history47 (76%)25 (78%)22 (73%)NSCryo Cryoablation; NS not significant; RF radiofrequency; SD standard deviation
Angiographic data
Right atrial angiography was not performed in 4 patients because of mild renal dysfunction or allergy to contrast material. The angiogram was of insufficient quality in eight others.
The mean isthmus length was 35.2 ± 14.6 mm and its topography was assessed as being flat or only mildly concave in 28, markedly concave in 19 and showed a pit or aneurysm in 10. A clear Eustachian valve was seen on six angiograms.
Ablation data
Assessment of acute results showed bidirectional isthmus block, using the criteria mentioned, in 47 of 62 patients (76%). This was in 25 of 30 patients (83%) of the RF group and in 22 of 32 (69%) of the cryo group (NS). In 1 patient in the RF group the procedure was terminated because of recurrent AF with early recurrence and inability to assess isthmus block. This patient was taken as a failure which was confirmed at the time of a subsequent AF ablation. Procedural data for all patients are given in Table 2.
Table 2Procedure data and recurrent arrhythmias AllCryoRFp valueNumber623230Application number22 ± 1318 ± 1025 ± 160.05Ablation time (s)2,742 ± 1,9303,792 ± 1,9001,459 ± 950<0.001Acute success47 (76%)22 (69%)25 (83%)NSSingle line47 (76%)25 (78%)22 (73%)NS2 lines9 (15%)2 (6%)7 (24%)0.0733 lines4 (6%)3 (9%)1 (3%)NSReversal of block during application7 (11%)6 (19%)1 (3%)0,091Reversal of block during 30 min6 (10%)1 (3%)5 (7%)NSIsthmus length (mm)35 ± 1535 ± 1736 ± 11NSSheath usage7 (11%)0 (0%)7 (23%)<0.005Recurrent arrhythmiasFlutter (typical) After success1 (2%)0 (0%)1 (4%)NS After failure11 (73%)7 (70%)4 (50%)NSFlutter (atypical)1 (2%)0 (0%)1(3%)NSAtrial tachycardia5 (8%)2 (6%)3 (10%)NSAtrial fibrillation28 (45%)13 (41%)15 (50%)NSThe numbers are given with the standard deviation.Cryo Cryoablation; NS not significant; RF radiofrequency
In the successful patients the number of applications to ensure block in the whole group was 17 ± 11. It was 23 ± 13 in the RF group and 12 ± 6 in the cryo group (p < 0.005). Total ablation time was 1,283 ± 777 s and 2,905 ± 1,245 s (p = 0.0001). In those in whom bidirectional block could not be achieved, the total number of applications was 33 ± 14. For RF and cryo the values were 39 ± 21 applications with a total time of 2,724 ± 1,102 s vs 29 ± 8 applications with a time of 5,873 ± 1,337 s (p < 0.03 and < 0.0011, respectively).
In 22 of 30 in the RF group and 25 of 32 in the cryo group, a single line at approximately 6 o’clock was drawn. In the RF group two lines were drawn in seven, and three in one. In the cryo group two lines were made in four, and three lines in three. The need to draw more than one line was associated with failure of the procedure in four of eight in the RF group and five of seven in the cryo group (NS).
Short lived block, either recurring during the application or immediately thereafter, occurred in one patient in the RF group and in six in the cryo group, with a trend to statistical significance (p = 0.091), with it being a predictor of failure if it occurred more than three times. In five patients in the RF group and one patient in the cryo group, conduction recurred later during a waiting period (with a median of 15 min) requiring further applications (NS). In only one patient in the RF group was late recurrence associated with failure to induce bidirectional block. In no patient in whom bidirectional block was present at the end of the 30 min waiting period did isoprenaline change this. The average power applied in the RF group was 52 ± 6 W.
In the RF group the signal was significantly diminished at the end of each application whether isthmus block was present or not, while there tended to be preservation of signals on the cryoablation catheter after ablation across the isthmus until block occurred. Only then were low voltage signals seen.
Failures were not significantly associated with length of the isthmus (39.2 ± 23.5 vs 34 ± 9.0 mm in success, although there was a trend to this (p = 0.12) There was no significant difference in anatomy between the two groups.
Procedure data and complications
The overall procedure time was 160 ± 49 min, with no difference between 170 ± 48 min in the RF group and 151 ± 49 min in the cryo group. Overall fluoroscopy times were 28 ± 14 min, with a difference between 33 ± 15 min in the RF group and 23 ± 11 min in the cryo group (p < 0.02).
Change of catheter curve occurred in one patient in the RF group from large to standard curve. An SL1 sheath (Daig, Minnetonka, MN, USA) was used in seven patients in the RF group for stability, while no sheath usage occurred in the cryo group (p = 0.005). During the procedure, six patients required cardioversion for induced AF, four in the RF group and two in the cryo group (NS). There were two small pericardial effusions seen on echocardiography without further significance (one in each group).
For the initial 40 patients there was a significantly higher peak CK and CK-MB after cryo (Table 3). This remained so after 24 h. For the last 22 patients we observed the same for CK-MB mass, but not for Troponin T (Table 3).
Table 3Biomarkers after catheter ablation AllCryoRFp valueProcedure valuesCK (U/l)141 ± 96184 ± 10296 ± 600.02CK-MB (U/l)27 ± 1636 ± 1718 ± 80.011CK-MB mass (μg/l)18 ± 2133 ± 244.4 ± 1.20.004Troponin T (μg/l)0.49 ± 0.320.54 ± 0.380.39 ± 0.27NSValues after 24 hrsCK (U/l)264 ± 245289 ± 173136 ± 740.022CK-MB (U/l)37 ± 2851 ± 3018 ± 60.011CK-MB mass (μg/l)8.4 ± 7.312 ± 105.16 ± 3.30.07Troponin T (μg/l)0.44 ± 0.300.54 ± 0.360.38 ± 0.29NSThe numbers are given with the standard deviation.CK Creatine kinase; Cryo cryoablation; NS not significant; RF radiofrequency
In assessing the level of comfort during the procedure and the pain experienced by patients, similar numbers from both groups assessed sedation as adequate (67% for RF and 63% for cryo). The pain scores given at the end of the procedure were not significantly different (42.9 ± 24.0 for RF and 43.7 ± 15.8 for cryo). Diazepam was given as standard at the beginning of the procedure and as necessary thereafter for discomfort related to having to lie still for protracted periods. The usage of diazepam was statistically similar in both groups (7.4 ± 3.4 mg in the RF group and 8.0 ± 3.1 mg in the cryo group). However, a significantly higher usage of fentanyl in the RF group was observed (70.0 ± 44.9 μg vs 10.0 ± 22.1 μg; p < 0.01).
Follow-up results
Patients were followed up for between 90 and 411 days (138 ± 81 days, median 90 days), which was similar in both groups (p = ns).
A total of 6 patients were taking no antiarrhythmic medication prior to ablation and this increased to 13 post isthmus ablation.
Recurrent arrhythmias were frequent in both groups of patients. ECG documented recurrent flutter occurred only in one patient from the RF group. A further procedure confirmed recovery of isthmus conduction. This recurrence was seen at 14 months post ablation, whereas he had previously had monthly episodes of flutter. One other patient had symptoms with an apparently non-isthmus dependent flutter documented on ECG on day 1 post ablation and not since, and elected not to have a further procedure. We cannot exclude asymptomatic arrhythmias but at least during the first 6 weeks, patients sent in daily event monitor transmissions, which one would hope would have captured at least some asymptomatic recurrences, particularly of atrial flutter.
In 11 of 15 patients in whom the initial procedure failed, a redo procedure was performed after the elected period of 6 weeks because of documented recurrent flutter. The redo procedure required a small number of point touch ups in five patients (three in the RF group and two in the cryo group), while more extensive diffuse isthmus applications were required in six (one in the RF group and five in the cryo group). In those undergoing repeat ablation, success was achieved in all and during follow up no recurrent atrial flutter was noted.
Previously undocumented atrial tachycardias were also seen in both groups (three in the RF group, and two in the cryo group). ECG documented AF recurrence occurred in 28 patients, at a similar rate in both groups (15 in the RF group and 13 in the cryo group). The likelihood of asymptomatic recurrences is probably higher for atrial fibrillation but was not the primary focus of this study.
Discussion
Acute success
This study showed an acute success rate of 83% for an 8 mm tip RF catheter vs. 69% for an 8 mm tip cryocatheter. This non significant difference is clearly concerning. However, the cryoablation group required significantly less applications to achieve success, with a similar procedural duration, significantly lower fluoroscopy time, and with a much lower requirement for analgesia with fentanyl. Arrhythmia recurrences in the initially successful patients were similar with a very low flutter recurrence rate.
The acute success rate with 8 mm tip RF compares favorably with that found in the meta-analysis of Da Costa et al. [22]. However, the acute results for cryotherapy are lower than in previously published studies. The shorter fluoroscopy time using cryotherapy is related to cryoadherence during applications [26]. With the catheter firmly attached, no fluoroscopy to check position is required. This is also reflected in the fact that no long sheaths were necessary. A lower requirement for analgesia has previously been described and this study confirms this finding [17, 26].
Difficulties during ablation
Recurrence during or shortly after ablation occurring more commonly in the cryo group probably relates to a reversible cooling effect at the periphery of the ice ball which recovers during or shortly after termination of ablation, while the central lesion acutely formed is more permanent. The more common recurrence during the waiting period after RF ablation probably relates to a longer reversal time of the acute RF effects (either edema or temperature effect).
Cryoablation for flutter
Cryoablation may have some advantages over RF in addition to those mentioned above such as less thrombogenicity [27], and maintenance of tissue architecture with homogenous well delineated lesions [28]. A 10Fr, 6.5 mm tip catheter showed acute success of 94–100% and 6 month recurrence rates of 0–25% [17, 29, 30]. A system with 7Fr, 6 mm tip catheters showed success in 87% to 88% using a septal line, without symptomatic recurrences, but with resumption of isthmus conduction at repeat study at 6 months in 30–35% [31, 32]. Using a 9Fr, 8 mm tip increased success rates to100% with symptomatic recurrence of 0% to 10% and recurrent conduction at 1 to 3 months in 19% to 32% [32–34]. There is some discussion as to whether 3, 4 or 8 min of ablation are needed for adequate lesion formation [33, 35].
Higher peak CK levels, confirmed by the high CK-mass in the last patients, suggest that the damage caused by cryotherapy may be more extensive despite the lower number of applications. Several explanations are possible. The first is that the CK levels may be underestimated after RF, which denatures proteins in another way than cryotherapy. Troponin T levels are more accurate in estimation of myocardial damage, but they also tended to be higher, be it not significantly. This may suggest that lesions are equal with both approaches.
Future developments
Newer data suggest that ablation per point guided by the maximal voltage may be a useful technique, but this has not been confirmed for cryotherapy [36].
Study limitations
This study was a single centre, largely single operator study involving a relatively limited number of patients with a high population of AF patients, making follow-up rather complicated. In present practice and later in this series, patients with the combination are often treated in one session, which still leaves physicians the choice whether cryo should be used for flutter, following its use for the pulmonary veins [37]. We only studied the acute and short-term success, while flutter recurrence may happen at more than 4 months of follow-up. While we attempted to exclude asymptomatic recurrences especially in the first 6 weeks, we cannot exclude the possibility of these, especially of AF and also late after the ablation. Isthmus conduction recovery is also possible in asymptomatic cases. In the RF group we limited the power to 60 W to ensure safety and this may be too conservative. Further, in our practice routine use of long sheaths was discouraged. At the time of this study use of an 8 mm tip with RF was considered as effective as irrigated tip ablation, so we elected to use this approach for the RF group. More recent studies have shown higher success rates than we obtained and have also suggested alternatives in the approach. These include the use of irrigated tip catheters and long sheaths as well as other technical issues and newer techniques such as maximal voltage guided ablation. A higher success rate for RF would clearly have been even more prejudicial to cryotherapy in this study. We were also extremely critical in our assessment of isthmus conduction, using multiple criteria. Angiograms were also performed in fixed views and not based on other catheter positions as has been suggested by others [14].
Conclusion
Acute success with cryotherapy for atrial flutter ablation, while non-statistically less effective in this study, requires fewer applications and is associated with a significantly lower requirement for pain relief. While our acute results for both cryotherapy and RF may not be as high as those in some comparable studies the recurrence rate was only 2.5% in the RF group, with no clinical recurrence in the cryo group. While cryotherapy cannot perhaps be advocated as first line therapy, it may be useful in certain circumstances. Certainly, if cryoablation is performed for AF, and if the isthmus needs to be ablated, cryotherapy might be used as well [37].
|
[
"catheter ablation",
"atrial flutter",
"radiofrequency",
"cryothermy",
"arrhythmia",
"atrial fibrillation"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Histochem_Cell_Biol-4-1-2248607
|
Discovery and progress in our understanding of the regulated secretory pathway in neuroendocrine cells
|
In this review we start with a historical perspective beginning with the early morphological work done almost 50 years ago. The importance of these pioneering studies is underscored by our brief summary of the key questions addressed by subsequent research into the mechanism of secretion. We then highlight important advances in our understanding of the formation and maturation of neuroendocrine secretory granules, first using in vitro reconstitution systems, then most recently biochemical approaches, and finally genetic manipulations in vitro and in vivo.
Morphological era of discovery (1950s and 1960s)
Visualization of the cell interior by electron microscopy catalysed both morphologists and biochemists to initiate experiments and make observations to identify, define and understand the complex compartmentalization of specialized cells, most notably including the morphologists G.E. Palade, C. DeDuve, D.W. Fawcett, K.R. Porter, J. Rhodin and F.S. Sjostrand. The exocrine pancreatic acinar was a specialized cell type favoured by GE Palade, and it was with this cell type that the first observations on cells with specialized granule compartments were made.
The complex morphology of the cell compartments, and the desire to link a morphological observation with basic biochemical functions and pathways known at that time initiated the monumental effort to couple a biochemical approach with EM observations. Indeed in 1956, Palade observed intracisternal granules in the endoplasmic reticulum (ER) which resembled zymogen granules, and postulated that the material in the intracisternal granules was possibly the same as the zymogen granules, and these granules were related to the granules in the cells in the endocrine pancreas, β-cell granules (Palade 1956). Using the pancreatic acinar cell model, experiments by Siekevitz and Palade between the late 1950s and early 1960s defined the role of the ER and other subcellular compartments in the synthesis of zymogen granule proteins, and resulted in a series of papers published in the newly established Journal of Biophysical and Biochemical Cytology, soon to be renamed Journal of Cell Biology (Siekevitz and Palade 1958, 1962). The concept that proteins are transported from their site of synthesis, the ER, through the Golgi complex was also being recognized and visualized using EM autoradiography in the exocrine pancreas (Caro and Palade 1964), first at a low temporal level of resolution. Direct evidence for transport from ER through Golgi to zymogen granules was then obtained using more precise pulse-chase protocols on tissue slices by Jamieson and Palade (Jamieson and Palade 1967a, b). This technical advance was further exploited by Jamieson who applied subcellular fractionation techniques to pulse-labelled slices. Separation of the rough and smooth ER from condensing vacuoles and zymogen granules allowed a quantitative kinetic analysis and a direct demonstration that proteins are transferred from the ER to zymogen granules via condensing vacuoles (Jamieson and Palade 1967a, b).
Parallel studies in the anterior pituitary gland showed that discrete, small individual granules, called immature secretory granules (ISG) in mammotroph cells originated from the Golgi cisternae. These small ISGs coalesced into aggregates surrounded by a single membrane, finally becoming mature secretory granules (MSGs) (Smith and Farquhar 1966). Additional information about the secretory process in mammotrophs was obtained using high resolution EM autoradiography. ISGs (20–100 nm structures) were found to be maximally labelled after 30 min pulse with [3H]-leucine, and mature over the following 2 h to MSGs. This kinetic of maturation correlated well with the observations in the exocrine pancreas which demonstrated the condensing vacuoles were labelled after 37 min (Jamieson and Palade 1967a). The complexity of the secretory granule compartment was revealed in subsequent studies on dispersed pituitary cells where four types of secretory granules were detected (Salpeter and Farquhar 1981). After 15–55 min of chase the labelled protein was found in small (Type I) ISGs, and subsequently in type II and III polymorphic granules, and then larger Type IV MSGs after 55–185 min of chase (see Fig. 1, and Farquhar et al. 1978). A higher degree of resolution was obtained using newly developed fine-grain emulsion which allowed an accurate detection of small structures (20–100 nm), which combined with a more sophisticated analysis of silver grain distribution, led to the discovery that concentration of secretory protein in the ISG compartment was 200 times that of the adjacent Golgi cisternae (Salpeter and Farquhar 1981).
Fig. 1Mammotroph cell from the anterior pituitary gland of a lactating rat. This micrograph illustrates the morphological complexity of the regulated secretory pathway, and the different types of ISGs. ER endoplasmic reticulum; CM cell membrane; SG secretory granule; SV smooth vesicles; VE vesicle; LB lytic body. Reproduced from the J Cell Biol, 1966, 31:319–347. Copyright 1966 The Rockefeller University Press
During this period Tartakoff and colleagues outlined the concept that secretory cells may have different ways to secrete newly synthesized proteins; cells such as plasma cells, fibroblasts, and macrophages secrete in a “non-regulated” fashion in contrast to “regulated” cells such as exocrine pancreas, the hallmark of which is the storage of the secretory proteins (Tartakoff and Vassalli 1978). The nature of the secretory pathways in such “non-storage” secretory cells was also explored by the use of drugs that perturbed secretion, such as monensin, a Na+/K+ ionophore which causes a neutralization of acidic intracellular compartments, or drugs that alter the energy status of the cell (Tartakoff et al. 1977). The common requirement for energy, cyclic nucleotides, Ca+, and cytoplasmic Na+/K+, in both non-regulated and regulated cells, gave rise to the consensus that there is a common secretory pathway originating in the RER, through the Golgi complex (for review see Palade 1975).
Many of the observations made were obtained using careful morphological approaches and expanded using biochemical techniques, and subcellular fractionation. However, these early biochemical approaches were limited by the inability to manipulate the cell systems in use, typically tissues from mice or rats. The arrival of molecular tools and genetic manipulation provided the next wave of advances bringing the field to our current level of understanding.
Cell line model systems, the molecular age and the sorting problem
Regulated secretory cells, as well as other cell types such as liver cells, have specialized plasma membrane domains, which by definition have a unique composition. Therefore, in addition to classification of regulated versus non-regulated secretion (also called constitutive secretion (Kelly 1985), it was recognized that there must be multiple ways to reach distinct domains from the Golgi complex, in particular for membrane proteins. The first direct demonstration of multiple routes to the plasma membrane was from the work of Gumbiner and Kelly in 1982. Their observations using the AtT20 cell line, a mouse pituitary cell line, showed that the regulated secretory hormone ACTH was secreted with different kinetics from a viral model membrane protein, gp70, a glycoprotein of the endogenous murine leukaemia virus. Surprisingly, the kinetics of the secretion of a proportion of the ACTH precursor, POMC, which escaped the activity of pro-hormone processing enzymes present in the MSG was the same as gp70 (Gumbiner and Kelly 1982). These experiments solidified “the two pathways hypothesis” (Kelly 1985) which defined distinct post-Golgi pathways for proteins targeted for regulated and constitutive secretion. At this time it was also recognized that the mechanism for segregation of cargo proteins (either soluble or membrane associated), or sorting into a particular pathway was an important issue, as was the identification of the precise location for the initiation of the sorting process. A key technological advance in elucidating both sorting and location was the development of immunoelectron microscopy using thin, frozen sections (Tokuyasu 1980) which allowed the identification of proteins within subcellular compartments.
TGN and post-TGN sorting and processing in neuroendocrine cells
It was recognized in mid-1980s that the trans-Golgi network (TGN) might be the key exit point for proteins destined to the plasma membrane, constitutive secretory vesicles (CSV), endosomes, and ISGs (Griffiths and Simons 1986). It was proposed that sorting receptors might function in this compartment to segregate different cargos into different vesicles (Burgess and Kelly 1987). It also became apparent from studies on the mannose-6-phosphate receptor (M6PR), which with its bound lysosomal enzyme is sorted to endosomes, that sorting to endosomes from the TGN utilized clathrin and clathrin-coated vesicles (Geuze et al. 1985). At this time it was also proposed that CSVs were formed without the aid of clathrin coats (Griffiths et al. 1985). However, clathrin-coated regions were detected on the surface of β-cell secretory granules associated with the TGN after monensin treatment (Orci et al. 1984) or on ACTH-containing ISGs in AtT20 cells (Tooze and Tooze 1986) raising the possibility that 1) ISGs formation is through clathrin-coated regions of the TGN, or 2) that the biogenesis of ISGs involved a clathrin-coated vesicle dependent pathway.
Acidification of the TGN and post-TGN secretory compartments was increasingly recognized to be important for both transport (see Tartakoff et al. 1977) and sorting. A key contribution at this time was the direct demonstration, using DAMP labelling (Anderson et al. 1984), that the TGN and post-TGN compartments were acidic, and there was a gradient of acidification in the secretory pathway. It was proposed that acidification could play role in receptor-ligand uncoupling and recycling, thereby providing directionality to transport (Anderson and Pathak 1985). Although the concentration of DAMP, was thought to be proportional to the extent of acidification, later studies provided accurate measures of pH of the TGN and ISG (see review Moore et al. 2002).
Neutralization of acidic compartments caused mis-sorting of regulated secretory proteins to the constitutive pathway (Moore et al. 1983), a result which gave rise to the possibility that regulated secretory proteins may have sorting signals, and could be sorted by a receptor-ligand interaction in the TGN, or in a post-TGN compartment. The most controversial observation in support of this hypothesis was the proposed role for carboxypeptidase E as the sorting receptor (Cool et al. 1997), although this was immediately refuted (Irminger et al. 1997). Evidence in support for a sorting signal was the demonstration that regulated secretory proteins have transferable sorting signals, the first being in 1986 (Moore and Kelly 1986). An alternative hypothesis was that aggregation of regulated secretory proteins, best demonstrated for the Granins but also shown for a variety of other regulated secretory proteins, and favoured by low pH on the TGN, drives lumenal segregation of regulated secretory proteins. The segregation of the regulated secretory proteins maybe further enhanced by a homophilic interaction with membrane associated population of Granins (Gerdes et al. 1989). Furthermore, evidence for formation of multiple secretory granule populations with different hormone content supports the idea that aggregation drives sorting (Hashimoto et al. 1987). Some of these results form the basis for the “sorting for entry” hypothesis, which proposes that regulated secretory proteins are sorted into ISGs in the TGN, while constitutively secreted proteins enter CSVs (see review Tooze 1998).
Another model to explain sorting in the regulated pathway was developed from experiments in β-cells, where it was shown that insulin-containing secretory granule formation is driven by the hexamerization and condensation of insulin in the ISGs (Arvan et al. 1991). These studies revealed an essential difference between neuroendocrine cells and β-cell insulin granules, giving rise to the “sorting by retention” hypothesis (see review Arvan and Castle 1998) which proposes that rather than an active sorting in the TGN, the crucial sorting step occurs in the post-TGN ISGs. Both the “sorting for entry” and “sorting by retention” hypothesis agree on that non-regulated (possibly mis-sorted) secretory proteins are removed by ISG-specific clathrin-coated vesicles in a pathway called “constitutive-like secretion”, after secretory granules have formed from the TGN.
Most, if not all of the soluble regulated secretory proteins are processed by the endopeptidases, the pro-hormone convertases (PCs). Their discovery and final molecular characterization was driven by the original pioneering work of Steiner and colleagues working in the β-cell (Steiner et al. 1974), and reviewed by (Seidah et al. 1993). These enzymes are present in ISGs and MSGs, and are subjected to the same sorting machinery as are proteins such as the Granins. As their activity is pH-dependent, a large number of studies used the kinetics of their activation as monitors for studying where and when the sorting of regulated proteins occurred (see also below). A direct demonstration that the site of prohomone processing of proinsulin was the clathrin-coated ISG was obtained by immunocytochemistry (Orci et al. 1985b), which were subsequently shown to be acidic (Orci et al. 1986).
The progress from the 1950s until the early 1990s was the result of many researchers’ effort, most of which have not been covered in this brief review. The reader is directed to comprehensive reviews from the period (Burgess and Kelly 1987; Mains et al. 1987), and more recently (Borgonovo et al. 2006; Dannies 2001; Kim et al. 2006; Meldolesi et al. 2004; Solimena and Gerdes 2003). The main questions which we focus on in the next section are the molecular machinery of secretory granule formation and maturation, as addressed by the work started in the laboratory of W. Huttner and continued in the author’s laboratory which is largely based on in vitro reconstitution assays using isolated TGN, and post-TGN subcellular compartments, in particular ISGs as first developed using Golgi membranes by Rothman and colleagues (Fries and Rothman 1980).
In vitro reconstitution and biochemical analysis of secretory granule biogenesis
Cell-free reconstitution of ISG budding from TGN
While MSGs from a variety of tissues had been extensively characterized morphologically as well as biochemically, little was known about ISGs beyond their morphological appearance (Smith and Farquhar 1966). The cell-free reconstitution of secretory granule budding from the TGN brought new insights to understanding the formation of ISGs (Tooze and Huttner 1990). This assay was based on (1) the high fidelity sorting properties of PC12 cells which, unlike AtT20 cells, target greater than 90% of their regulated secretory protein to ISGs and MSGs, (2) the selective labelling of a regulated secretory marker, secretogranin II (SgII), and a constitutively secreted protein, heparan sulfate proteoglycan (hsPG) with radioative sulfate by sulfotransferases present only in the TGN, and (3) the ability to separate the TGN, ISGs and CSVs from one another by sequential velocity and equilibrium gradients centrifugations. This assay demonstrated that SgII and hsPG were present in two vesicle populations, SgII-containing regulated secretory vesicles (ISGs), and hsPG-containing CSVs. These results provided the first demonstration that regulated secretory proteins and constitutive secretory proteins were sorted into two distinct vesicle populations directly upon exit from the TGN, in support of the “sorting for entry” hypothesis. The formation of both ISGs and CSVs required GTP-binding proteins (Tooze et al. 1990), possibly heterotrimeric G-proteins (Leyte et al. 1992). The rate of formation of the ISGs was indistinguishable from the CSVs, occurring with a t1/2 of 15 min.
In addition, the subcellular fractionation protocol developed to distinguish ISGs and MSGs allowed the biochemical and morphological characterization of ISGs, in comparison to MSGs, isolated from PC12 cells (Tooze et al. 1991). ISGs were shown to have several components of the clathrin-coat machinery, including AP-1 (Dittié et al. 1996), and ARF1 (Austin et al. 2000), in addition to non-granule proteins such as the M6PR and furin (Dittié et al. 1997, 1999). In addition, it was possible to determine the size of ISGs and MSGs, results which provided direct support for the early EM autoradiography (Smith and Farquhar 1966) and became the basis for the experiments directed towards a molecular understanding of the change in size of the ISG through homotypic fusion and subsequent remodelling of the ISG through clathrin coats (see Fig. 2).
Fig. 2Key steps in secretory granule biogenesis. Model based on data discussed above, in addition to contributions from many researchers which have not been mentioned, but whose contributions are noted. Step 1 depicts the formation of ISGs from the TGN, Step 2 the homotypic fusion event between ISGs, and Step 3 the clathrin-mediated membrane remodelling event. See text for abbreviations
Determination of the pH of ISGs using an in vitro approach
While it was known from earlier studies that in endocrine cells MSGs are acidified it was not known what the precise pH of the ISG was, and if the pH of the ISG differed from the TGN. The establishment of a PC12 stable cell line expressing the PC2 allowed us to characterize the endopeptidase PC2 activity during granule maturation (Dittié and Tooze 1995). In this cell line, PC2 is correctly targeted to the ISGs and co-sedimented with SgII in fractions containing ISGs and MSGs. [35S] sulfate labelling demonstrated that SgII was proteolytically processed by PC2 in ISGs after approximately 30 min of chase into several lower-molecular-mass proteins, the major ones being a 18 and 28 kDa sulphated fragment.
As the efficiency of processing of SgII by PC2 was pH dependent, we used the ability to quantitate the extent of processing at different pHs to determine the pH in the ISGs (Urbé et al. 1997). Isolated ISGs and TGN, containing [35S] sulfate pulse-labelled SgII were incubated in presence of ATP at physiological pH and the extent of SgII processing in these compartments was compared to a standard curve prepared using ISGs equilibrated at a defined set of pHs. This allowed us to determine that the ISGs intra-granular pH was 6.3, similar to the TGN pH and clearly higher than the pH of MSGs (pH 5.5–5.0). Interestingly, no processing of SgII could be observed in the membrane fraction highly enriched in TGN under conditions for which processing was readily obtained in isolated ISGs. This data represent further evidence that the ISG is indeed a functionally distinct organelle from the TGN. Furthermore, the rate of SgII processing was strongly dependent on the intragranular pH, demonstrating that processing of SgII can be used as a pH indicator for granule interior.
ISG-ISG homotypic fusion
As shown originally by Farquhar (Smith and Farquhar 1966) in the anterior pituitary, and more recently in PC12 cells (Tooze et al. 1991), ISGs increase in size during maturation, and the increase in size was proposed to reflect homotypic fusion of ISG. To demonstrate that ISG–ISG fusion occurred, we developed an assay that provided the first biochemical evidence for such a fusion event and allowed us to dissect the molecular requirements of this process.
The cell-free assay to reconstitute homotypic fusion was performed by mixing two populations of ISGs, one containing the pro-hormone convertase PC2 and the other containing [35S]sulfate-labelled secretogranin II (SgII) (Urbé et al. 1998). The fusion was then measured by quantification of the 18 kDa PC2 cleavage product of SgII. This in vitro reconstitution of ISG-ISG fusion revealed that homotypic fusion is dependent on NSF (Urbé et al. 1998), α-SNAP and on the SNARE Syntaxin 6 but not on Syntaxin 1 or SNAP-25 (Wendler et al. 2001). More recently, Synaptotagmin IV (Syt IV), a member of the Synaptotagmin family of proteins involved in membrane fusion, has also been shown to be required for this step of granule maturation (Ahras et al. 2006).
ISG membrane remodelling
In addition to ISG-ISG homotypic fusion, ISG content and membrane remodelling is another important step during granule maturation. This remodelling is performed via budding of clathrin-coated vesicle from the maturing granule membrane, and is the pathway for non-regulated constitutive-like secretion, or possibly sorting to endosomes. This clathrin-mediated remodelling step is a common feature in neuroendocrine and endocrine secretory granules, and is thought to provide a mechanism for proof-reading the content and membrane composition of the maturing ISG to ensure the production of MSGs which contain biologically active hormones and can undergo efficient exocytosis.
ISGs in endocrine and neuroendocrine cells were shown by morphological techniques to be partially clathrin coated (Orci et al. 1985a; Tooze and Tooze 1986) but the recruitment mechanism and the composition of this clathrin coat were unknown until a biochemical approach was used. Our laboratory developed a fourth cell-free assay whereby the recruitment of the clathrin adaptor complex AP-1 to ISGs was reconstituted by the addition of bovine brain cytosol (Dittié et al. 1996). These experiments showed that AP-1 recruitment to ISGs was ATP-independent but GTP- and, ARF1-dependent. This study demonstrated that it is the AP-1 complex that is involved in the clathrin binding to ISGs. Cross-linking experiments demonstrated the direct interaction of AP-1 and ARF1 on the ISG (Austin et al. 2000). The M6PR and furin are likely part of the cargo targeted for removal from the ISGs by clathrin-coats as they possess sequences in their cytoplasmic domains which interact with AP-1, and are present on ISGs, but absent from MSGs (Dittié et al. 1997, 1999).
Further investigation into cargo sorting from the ISG showed that VAMP4, a SNARE involved in endosome to TGN vesicle trafficking, which is present on ISGs but removed from the maturing granule membrane, is also able to bind AP-1. The recruitment of AP-1 by VAMP4 is dependent on phosphorylation of the cytoplasmic domain of VAMP4 by the kinase CKII. CKII phosphorylation of VAMP4 allows the recruitment of PACS1 to ISGs enhancing the AP-1 dependent sorting event (Hinners et al. 2003). These data suggest that sorting non-regulated secretory proteins from ISGs, including the SNARE proteins, can be mediated by the recruitment of clathrin coats and may be essential for maturation of ISGs.
More recently, our laboratory demonstrated that the GGAs proteins are also involved in this clathrin-mediated membrane-remodelling step of secretory granule maturation. GGAs (Golgi-associated, γ-ear-containing, ADP-ribosylation factor-binding protein) are involved in recruitment of AP-1 and clathrin coats to membranes (for review see Robinson and Bonifacino 2001). siRNA mediated knock-down of GGA3 and over-expression of dominant negative form of GGA proteins resulted in a retention of the SNAREs Syntaxin 6 and VAMP4 in the MSGs. In addition, an alteration of the PC2 activity was also detected, suggesting that inhibition or alteration of clathrin-coat remodelling inhibited or altered intra-lumenal acidification (Kakhlon et al. 2006). Similar effects on PC2 activity were seen after siRNA mediated knock-down of Syt IV, suggesting that inhibition of either homotypic fusion or remodelling have similar detrimental effects on acidification of ISGs.
It was proposed that secretory granules, like other membranes have lipid micro-domains, or rafts (Tooze et al. 2001), and that the lipid composition of the secretory granule membrane is important for protein sorting during clathrin-coat mediated membrane remodelling. In fact, cholesterol was shown to be an important component in secretory granule formation (Wang et al. 2000). Recent work by Katsumata et al. (2007) has shown that the secretory granule membrane is composed of two kinds of micro-domains and that the proteins expected to be removed during the membrane remodelling (VAMP4 and Syntaxin6) are clustered together in the same micro-domain and separated from the proteins (VAMP2) that stay on the MSG membrane. These experiments validate our observations (see above), as well as the work by H.P. Moore in AtT20 cells (Eaton et al. 2000) and provide new information that opens new avenues for research into the role of lipids in sorting from ISGs.
Recent developments in secretory granule formation and maturation
In this last section we summarize the most recent developments in our understanding of secretory granule biogenesis obtained from biochemical screens, in vivo experiments based on genetic manipulation of cell in culture as well as in vivo transgenic mouse models. Only a small number of recent results are highlighted and readers are directed to recent reviews (Borgonovo et al. 2006; Dannies 2001; Kim et al. 2006; Meldolesi et al. 2004; Solimena and Gerdes 2003). The experiments we highlight show the direction that the field has taken, and the way less tractable questions can be addressed by new biochemical screens and transgenetic mouse models.
Role of cholesterol in secretory granule biogenesis
Secretory granule formation is dependent on lipid raft microdomains that are enriched in cholesterols (Wang et al. 2000). A recent investigation of the role in vivo of cholesterol in secretory granule biogenesis was done using Smith–Lemli–Opitz Syndrome (SLOS) and lathosterolis mouse models (Gondre-Lewis et al. 2006). Both diseases are a result of inborn errors in cholesterol synthesis. SLOS is a disease caused by a defect in the function of the enzyme, 7-dehydrocholesterol reductase (DHCR7), required for the final step of cholesterol biosynthesis. Patients with lathosterolosis lack lathosterol-5-desaturase (SC5D), the enzyme that catalyse the next-to-last step in cholesterol synthesis. Data obtained from DHCR7 and SC5D knock-out mice demonstrated that impairment of the cholesterol biosynthesis results in a significant decrease in secretory granules in pancreas, pituitary and adrenal glands. In addition, abberant zymogen granules were found in the exocrine pancreas, in which there was also observed a decrease in regulated secretion which could be rescued with exogenous cholesterol. The authors conclude that the presence of an elevated quantity of other sterols, 7-DHC and lathosterol, in Dhcr7-/- and Sc5d-/- mice, respectively, could not replace cholesterol in the regulated secretory pathway. They speculate that the abnormal properties of the secretory granule in these mice models may be attributed to the reduced rigidity of membranes containing these sterols instead of cholesterol. Importantly, this study demonstrates that cholesterol is essential during secretory granule biogenesis in vivo.
Regulation of secretory granule maturation
Peptidylglycine α-amidating monooxygenase (PAM) is an essential enzyme in the processing of many bioactive peptides and hormones (Prigge et al. 2000). PAM is targeted to the regulated secretory pathway in neurons and neuroendocrine cells. It is a type I membrane protein containing two enzymatic domains within the lumen of the secretory granule, a transmembrane domain, and a cytosolic domain containing sorting signals. PAM catalyses one of the final steps in peptide biosynthesis, and is retrieved from ISGs and the plasma membrane for re-utilization in newly forming secretory granules (Ferraro et al. 2005). Among the proteins able to interact with PAM are two Rho guanine nucleotide exchange factors (GEFs), Kalirin and Trio, identified by yeast two-hybrid screens using the cytosolic domain of PAM as bait (Alam et al. 1996; Xin et al. 2004).
In a recent study, Kalirin and Trio have been shown to be involved in the maturation of secretory granules (Ferraro et al. 2007). Overexpression of their N-terminal GEF domain enhances secretion from ISGs, reducing regulated secretory protein storage through constitutive-like secretion, in the absence of secretagogue stimulation of regulated exocytosis. Conversely, when GEF activity is inhibited the constitutive-like release is inhibited, resulting in the accumulation of cargo in MSGs above even normal levels. These results indicate that these two Rho-GEFs regulate sorting into secretory granules, constitutive-like secretion from ISGs, and perhaps constitutive secretion, and provide a novel mechanism for the regulation of secretory granule maturation.
Role of Chromogranin A in secretory granule formation
Chromogranin A (CgA) and B (CgB), members of the Granin family, have long been proposed to control the secretory granule biogenesis because of their pH-, calcium- and catecholamine-dependent aggregation properties, and widespread expression pattern. Recent results supports this hypothesis, however other recent results suggest that the regulated phenotype (Day and Gorr 2003; Meldolesi et al. 2004) is not simply conferred by expression of CgA.
In support of the hypothesis, in 2001, Kim et al. demonstrated that specific depletion of CgA, but not of CgB, by siRNA impaired secretory granule production in PC12 cells. In contrast, over-expression of CgA in 6T3 cells lacking CgA and the regulated secretory pathway, or fibroblastic CV-1 cells, not only triggered recovery of secretory granules but also regulated secretion (Kim et al. 2001). However, Malosio et al.(2004) refuted the hypothesis by demonstrating with very similar approaches that the “newly formed” vesicles which contained the CgA were in fact lysosomes, and overexpression of CgA did not induce the appearance of secretory granules, or alter the number of secretory granules.
Three groups used mouse models to determine the role of Chromogranin A in vivo by targeted ablation of the Chromogranin A gene. Mahapatra et al.(2005) confirmed the putative role of CgA in secretory granule biogenesis, in particular a decrease of chromaffin granule size and number. In addition, defects were observed in neurotransmitter storage and release and regulation of blood pressure. Similar results were obtained using anti-sense vectors specific for CgA in transgenic animals which resulted in a reduction of the CgA levels (Kim et al. 2005). In the adrenal medulla of these mice there was a large decrease in the number of secretory granules; in addition it was noted that the secretory granules present appeared to be swollen. This defect may be related to a need to maintain a proportional catecholamine and CgA concentrations.
Lastly, Hendy et al. (2006) found that the CgA null mutant mice had elevated secretion of epinephrine, norepinephrine and dopamine and that mRNA and protein level of other secretory granule proteins were up-regulated. No obvious abnormalities in development or neuronal and endocrine functions have been noticed. The authors suggested then that the increased expression of the other Granin family members is likely to compensate for the CgA deficiency.
Role of prohormone convertases in secretory granule maturation
Both in vitro and ex vivo the activity of the PC enzymes (notably PC1/3 and PC2 which will be the focus of the discussion here) have been extensively studied. PC null mice models have confirmed the in vivo function of the PCs and extended the characterization of their specific substrates and action in different tissues.
The first PC transgenic mice model obtained was the PC2 null mouse (Furuta et al. 1997). Mice lacking PC2 activity develop normally and are fertile, but exhibit a variety of neuroendocrine processing abnormalities in the brain and in pancreatic islets. These mice show elevated levels of pro-insulin (Furuta et al. 1998) and pancreatic islet pro-glucagon processing is completely blocked. Other defects in PC2 null animals include lack of production of α-MSH associated with accumulation of ACTH in the pituitary intermediate lobe (Laurent et al. 2002).
7B2, also known as SgV or SGNE-1, is a small acidic protein exclusively localized to neuroendocrine tissues. This protein has been shown to be associated to PC2 and it functions as a chaperone for PC2 (Braks and Martens 1994), and for review see Mbikay et al. (2001). A 7B2 null transgenic mice model has shown that 7B2 is required for PC2 activation in vivo, having in addition important functions in regulating pituitary hormone secretion (Westphal et al. 1999). 7B2 null mice have no demonstrable PC2 activity, and in general agreement with the PC2 null mouse (Furuta et al. 1997), are deficient in processing islet hormones and display hypoglycemia, hyperinsulinemia, and hypoglucanemia, with generalized islet cell expansion and altered islet cell morphology. However, the most important impairment of PC2-mediated peptide processing in 7B2 nulls involved the synthesis of the corticotrophin ACTH. ACTH is produced by cleavage of POMC through the action of PC1/PC3, further internal cleavage of ACTH occurs by the action of PC2 specifically in the neurointermediate lobe. In 7B2 nulls, ACTH remains intact in the intermediate lobe, resulting in extremely high levels of ACTH in this lobe (Westphal et al. 1999). The overall phenotype suggests that 7B2 may be involved in PC2 activity, as well as secretory granule biogenesis.
Finally, the role of the PC1/3 endopeptidase has also been investigated in a transgenic mouse model. Disruption of gene-encoding mouse PC1/PC3 results in a syndrome of severe post-natal growth impairment and multiple defects in processing many hormone precursors, including hypothalamic growth hormone-releasing hormone (GHRH), pituitary proopiomelanocortin to adrenocorticotropic hormone, islet proinsulin to insulin and intestinal proglucagon to glucagon-like peptide-1 and -2 (Zhu et al. 2002). In summary, the analysis of PC2, 7B2 and PC1/3 null mice has confirmed that PCs play a crucial role in the processing of many hormone precursors and the highlight the importance of PC2 and PC1/3 in neuroendocrine tissues, and regulated secretion.
Role of Rab3D in secretory granule maturation
Rab proteins are small GTPases that belong to the Ras protein superfamily and that function in membrane traffic. Rab3A is one of the best-characterized Rab protein. In mammals, three additional isoforms are expressed which are referred to as Rab3B, Rab3C and Rab3D. Rab3A, Rab3B and Rab3C are predominantly expressed in the nervous system, where they are localized to synaptic vesicles (Fischer von Mollard et al. 1994; Martelli et al. 2000). Rab3A-deficient mice studies suggested that Rab3A has a function intimately associated with fusion more than a function in vesicle tethering at the synapse (Geppert et al. 1997).
In contrast to Rab3A, Rab3B and Rab3C, Rab3D is predominantly expressed outside the nervous system, in peripheral tissues where the other isoforms either are expressed at low levels or are lacking. Originally identified in fat cells, Rab3D is present in several additional cell types including secretory cells such as pancreatic and parotid acinar cells, mast cells and peptide-secreting cell lines. In secretory cells Rab3D appears to be predominantly localized to secretory granules, thus mirroring the distribution of Rab3A in neurons and neuroendocrine cells. It was hypothesised that Rab3D could have the same function as Rab3A but in secretory cells.
Riedel et al. (2002) investigated the role of this Rab3 isoform by knocking out the Rab3D gene in mice. Rab3D-deficient mice are viable and fertile and showed no obvious defect. In particular, both kinetics and dose response of secretagogue-induced enzyme secretion of the pancreas were normal. However, these mice show a substantially increased size of secretory granules in both exocrine pancreas and the parotid gland with the volume being doubled, while the intragranular protein concentration appears to be unchanged. These data suggest that Rab3D functions in granule maturation but not in exocytic membrane fusion.
Conclusions and future perspectives
The early morphological and biochemical characterization of the regulated secretory pathway provided key insights to the secretory process. More recent experiments have provided a more detailed understanding of the functional properties, and the regulation of the maturation of neuroendocrine secretory granules. Many important issues need to be resolved, for example how the cytoplasmic events, such as membrane remodelling are coupled with the intralumenal biochemical changes such as PC activation, prohormone cleavage, and acidification. In addition, the biological relevance of the maturation process, as it affects hormone processing and secretion, remains to be determined in light of the differences between endocrine and neuroendocrine granules.
|
[
"immature secretory granule",
"mature secretory granule",
"secretogranin",
"chromogranin",
"prohormone convertase"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Invert_Neurosci-4-1-2257991
|
The cys-loop ligand-gated ion channel gene superfamily of the nematode, Caenorhabditis elegans
|
The nematode, Caenorhabditis elegans, possesses the most extensive known superfamily of cys-loop ligand-gated ion channels (cys-loop LGICs) consisting of 102 subunit-encoding genes. Less than half of these genes have been functionally characterised which include cation-permeable channels gated by acetylcholine (ACh) and γ-aminobutyric acid (GABA) as well as anion-selective channels gated by ACh, GABA, glutamate and serotonin. Following the guidelines set for genetic nomenclature for C. elegans, we have designated unnamed subunits as lgc genes (ligand-gated ion channels of the cys-loop superfamily). Phylogenetic analysis shows that several of these lgc subunits form distinct groups which may represent novel cys-loop LGIC subtypes.
Introduction
Members of the cysteine-loop ligand-gated ion channel (cys-loop LGIC) superfamily are neurotransmitter receptors that mediate synaptic transmission in vertebrates and invertebrates. Cys-loop LGICs are made up of five homologous subunits arranged around a central ion channel (Sine and Engel 2006). The characteristic cys-loop motif is situated in the N-terminal extracellular ligand-binding region of each subunit and consists of two disulphide-bond forming cysteines separated by 13 amino acid residues, several of which are highly conserved to form a signature sequence (Gene Ontology ID GO:0005230). The cys-loop plays roles in receptor assembly (Green and Wanamaker 1997) and gating of the ion channel (Grutter et al. 2005).
The subunit composition determines the functional and pharmacological properties of the cys-loop LGIC, thus receptor diversity is generated by multiple subunit-encoding genes. For instance, the human cys-loop LGIC superfamily consists of 45 genes encoding cation-permeable nicotinic acetylcholine receptors (nAChRs) (Kalamida et al. 2007), cation-permeable serotonin or 5-hydroxytryptamine type 3 (5-HT3) receptors (Reeves and Lummis 2002), anion-permeable γ-aminobutyric acid (GABA) type A and C (GABAA and GABAC) receptors (Bormann 2000; Darlison et al. 2005) and anion-permeable glycine receptors (Kirsch 2006). As demonstrated in the fruit fly (Drosophila melanogaster), honey bee (Apis mellifera) and red flour beetle (Tribolium castaneum), insect cys-loop LGIC superfamilies are more compact, consisting of just over 20 subunits (Jones and Sattelle 2006; Jones and Sattelle 2007; Littleton and Ganetzky 2000). Approximately 15 of these genes are known to encode cation-permeable nAChRs (Sattelle et al. 2005), anion and possibly cation channels gated by GABA (Buckingham et al. 2005), glutamate-gated anion channels (Vassilatis et al. 1997) and histamine-gated anion channels (Gisselmann et al. 2002; Zheng et al. 2002). The largest known cys-loop LGIC superfamily belongs to the simple nematode, Caenorhabditis elegans, where an early analysis of its completed genome revealed 90 ligand-gated ion channel genes (Bargmann 1998). In common with vertebrates, members of this large cys-loop LGIC superfamily include cation-permeable nAChRs (Jones and Sattelle 2004) and anion-permeable GABA receptors (Schuske et al. 2004). As is the case for insects, C. elegans has glutamate-gated anion channels (Wolstenholme and Rogers 2005) but so far no nematode histamine-gated ion channels have been found. However, C. elegans possesses cys-loop LGIC receptors not identified in vertebrates and insects which include anion channels gated by acetylcholine (Putrenko et al. 2005) and serotonin (Ranganathan et al. 2000). The considerable diversity of receptor subtypes in C. elegans may be broadened further as many cys-loop LGIC subunits have yet to be characterised. Here, we present an update on the genes present in the C. elegans cys-loop LGIC superfamily.
Methods
All C. elegans protein sequences bearing motifs particular to cys-loop LGICs were assembled from WormBase (http://www.wormbase.org release WS185 Dec 23 2007). These motifs and their source databases are neurotransmitter-gated ion-channel ligand-binding INTERPRO:IPR006202, neurotransmitter-gated ion-channel INTERPRO:IPR006201, neurotransmitter-gated ion-channel ligand binding domain PFAM:PF02931, neurotransmitter-gated ion-channel transmembrane region PFAM:PF02932 and ligand-gated ion channel KOG3644. Cys-loop and TM2 sequences were identified from an alignment of protein sequences of whole subunits constructed with ClustalX (Thompson et al. 1997) using the slow-accurate mode with a gap opening penalty of 10 and a gap extension penalty of 0.1 as well as applying the Gonnet 250 protein weight matrix (Benner et al. 1994). The protein alignment was viewed using GeneDoc (http://www.nrbsc.org/gfx/genedoc/index.html). The neighbour-joining method (Saitou and Nei 1987), available with the ClustalX program, was used to construct a phylogenetic tree, which was then displayed using the TreeView application (Page 1996).
Results and discussion
Our analysis of the C. elegans genome revealed the largest cys-loop LGIC superfamily so far characterised consisting of 102 genes. As we have previously reported the C. elegans nAChR and orphan subunits (Brown et al. 2006; Jones et al. 2007; Jones and Sattelle 2004; Mongan et al. 1998), we concentrate here on the remaining cys-loop LGIC genes. For unnamed subunit genes, we continued to use the nomenclature first used for orphan nAChR subunits (Jones et al. 2007). Thus, 24 genes were designated lgc for ligand-gated ion channel of the cys-loop LGIC superfamily (Table 1). As shown in Fig. 1, the C. elegans cys-loop LGIC subunits were grouped according to sequence homology. As with the C. elegans nAChR subunits (Mongan et al. 1998), each group was named, if possible, after a well characterised subunit. For a detailed comparison of C. elegans cys-loop LGIC subunits with those of other organisms, see (Dent 2006).
Table 1C. elegans cys-loop LGIC subunits that have been given names in this report. Amino acid residues preceding TM2 that are important for determining the charge selectivity of the ion channel (Jensen et al. 2005) are given. The cys-loop sequence is shown with the cysteine residues highlighted in black shading while gray shading indicates conserved residues. The presence of a putative second cys-loop (Dent 2006) is also notedFig. 1Tree showing the cys-loop LGIC gene superfamily of C. elegans. This figure focuses on cys-loop LGIC subunits other than nAChRs although representative subunits of the nAChR groups are shown. For the complete C. elegans nAChR gene family, see (Jones et al. 2007). Based on protein sequence homology, the subunits are divided into eight groups with three highly divergent subunits, LGC-32, LGC-33 and LGC-34, not belonging to any group
Ungrouped subunits
Three subunits, LGC-32, LGC-33 and LGC-34, were not placed into any groups as they are highly divergent (Fig. 1), only showing up to 10, 10 and 15% identity respectively with other C. elegans, human and D. melanogaster cys-loop LGICs. These subunits have not been functionally characterised so it is unknown what ligands act on them. The sequence of LGC-34 and a close homologue in the parasitic nematode, Dirofilaria immitis, has been previously described where it has been noted that even though the subunit possesses features common to cys-loop LGICs, such as conserved residues in the N-terminal extracellular region and four transmembrane domains, the cys-loop is absent (Yates and Wolstenholme 2004). Interestingly, a bacterial proton-gated ion channel (Glvi) resembling cys-loop LGICs but also lacking the cys-loop has been recently reported (Bocquet et al. 2007; Tasneem et al. 2005). LGC-34 may thus represent more ancestral members of the cys-loop LGIC superfamily. Unlike Glvi, LGC-34 possesses two cysteine residues which are conserved in glutamate-, histamine- and vertebrate glycine-gated anion channels (Dent 2006; Yates and Wolstenholme 2004). Known as the second cys-loop, these two cysteines flank loop C which is a region that contributes to ligand binding (Corringer et al. 2000).
The EXP-1 Group
This group consists of only two subunits, one of which is EXP-1, a GABA-gated cation channel that mediates enteric muscle contraction (Beg and Jorgensen 2003). While EXP-1 overall resembles more closely anion permeable cys-loop LGICs such as UNC-49 as opposed to cation-permeable nAChRs, it lacks the PAR motif preceding the second transmembrane domain (TM2) which is important for anion selectivity (Jensen et al. 2005). Instead it possesses the residues ETE where the presence of glutamic acid residues has been implicated in determining cation selectivity (Jensen et al. 2005; Wotring and Weiss 2008). The other subunit in this group, LGC-35, also has a glutamate residue preceding TM2 (Table 1) and thus may be a cation channel.
The AVR-14 Group
This group consists of six subunits (Fig. 1) that make up glutamate-gated chloride channels that are targeted by the avermectin and milbemycin anthelmintics (Wolstenholme and Rogers 2005). Examples of GluCl function include AVR-15 and GLC-2 in the inhibition of pharyngeal pumping (Dent et al. 1997; Laughton et al. 1997; Pemberton et al. 2001) and AVR-14, AVR-15, GLC-1 and GLC-3 in regulating locomotion (Cook et al. 2006). Recently, it has been shown GLC-3 plays a role in olfactory behaviour (Chalasani et al. 2007).
The UNC-49 Group
Products of the unc-49 gene form a GABA-gated anion channel that mediates body muscle inhibition during locomotion (Bamber et al. 1999; Richmond and Jorgensen 1999). There are four other subunits included in this group (Fig. 1). With regard to other species, members of the UNC-49 group are clearly most closely related to mammalian and insect GABA-gated anion channels (Dent 2006). GAB-1 forms functional GABA receptors when coexpressed in Xenopus laevis oocytes with either HG1A or HG1E, which are putative GABA receptor subunits from the parasitic nematode Haemonchus contortus (Feng et al. 2002). The remaining subunits have yet to be characterised but it is worth noting that LGC-36 possesses the amino acids AER instead of PAR before TM2 (Table 1). The presence of a glutamic acid residue at this position and the absence of a proline may result in a cation-selective channel (Jensen et al. 2005).
The GGR-1 Group
This group includes six subunits (Fig. 1), the functional characterisation of which has so far not been published. The neurotransmitters/ligands to which GGR-1 group subunits respond to cannot be easily inferred through phylogenetic analysis. For example, when comparing with human cys-loop LGICs they bear closest resemblance to glycine α subunits with approximately 25% identity while LGC-42 is slightly more similar to histamine-gated anion channels with 26% identity when comparing with insect cys-loop LGICs. Members of the GGR-1 group have been previously denoted as “ce_Group I” subunits which most closely resemble insect histamine-gated anion channels (Dent 2006). It would be of interest to determine if members of the GGR-1 group respond to glycine, histamine or other ligands not so far known to act on C. elegans cys-loop LGICs. The subunits of the GGR-1 group are likely to be anion-selective as they all possess the PAR motif before TM2 (Table 1) which is important for anion selectivity (Jensen et al. 2005). The only exception is GGR-1 which instead has the amino acid residues PGR. With the proline still present and an absence of an acidic residue (particularly glutamic acid) in this region, it is likely that GGR-1 will also be anion selective.
The LGC-45 Group
This group consists of three subunits (Fig. 1), the functional characterisation of which has so far not been published. As is the case for members of the GGR-1 group, LGC-43, LGC-44 and LGC-45 most closely resemble glycine α subunits and histamine-gated anion channels when compared with human and insect cys-loop LGICs respectively. However, unlike the GGR-1 group, the LGC-45 group subunits are likely to be cation selective as they lack the PAR motif before TM2 (Table 1) which is important for anion selectivity (Jensen et al. 2005). In particular, LGC-44 and LGC-45 possess glutamic acid instead of the proline residue which is likely to result in a cation-selective channel (Wotring and Weiss 2008). While LGC-44 shows notable sequence identity with other cys-loop LGIC subunits (up to 37% with LGC-45), it lacks the cys-loop (Table 1), thus it may represent an ancestral cys-loop LGIC subunit as has been noted for LGC-34.
The ACC-1 Group
This group includes ACC-1, ACC-2, ACC-3 and ACC-4 which are acetylcholine-gated anion channels (Putrenko et al. 2005). Acetylcholine-gated anion channels have also been characterised from the snail, Lymnae stagnalis, which appear to have evolved from cation channels through amino acid substitutions in the ion channel pore (van Nierop et al. 2005). In contrast, acetylcholine-gated anion channels of the ACC-1 group may have arisen from substitutions in the ligand-binding domain of anion channels (Putrenko et al. 2005; van Nierop et al. 2005).
The MOD-1 Group
MOD-1 modulates locomotory behaviour and is the only known invertebrate cys-loop LGIC to be gated by serotonin (Ranganathan et al. 2000). Unlike mammals which possess cation-permeable serotonin receptors (Reeves and Lummis 2002), MOD-1 is anion selective. One other subunit, LGC-50, is closely related to MOD-1 (Fig. 1).
The GGR-3 Group
This group includes six subunits (Fig. 1), the functional characterisation of which has so far not been published. All six subunits are likely to form anion-selective channels as they have the PAR amino acid motif before TM2 (Table 1) (Jensen et al. 2005). When comparing with human cys-loop LGICs they bear closest resemblance to GABA α, γ and ε subunits with approximately 22% identity while they most closely resemble histamine-gated chloride channels (21% identity) when comparing with insect cys-loop LGICs. As with the GGR-1 group, it would be of interest to determine if the GGR-3 group consists of ion channels responding to ligands not so far known to act on C. elegans cys-loop LGICs.
Cys-loop LGIC superfamilies of other nematode species
Genome sequencing projects are allowing the comparison of cys-loop LGIC superfamilies from different nematode species. For instance, Caenorhabditis briggsae, which is seemingly identical to C. elegans (Gupta et al. 2007; Stein et al. 2003), also has an extensive cys-loop LGIC gene superfamily although there are a few exceptions where there appears to be gene expansions in C. elegans, mainly within the nAChR orphan group. For example, lgc-23, lgc-24 and lgc-28 in C. elegans (Jones et al. 2007) may be paralogues resulting from gene duplication in the elegans species lineage since only one homologue (Cbr-lgc-28) is observed in C. briggsae. In line with this, lgc-23 and lgc-24 are situated close together within 6 kb in the C. elegans genome which may reflect a recent gene duplication event. Likewise, lgc-11 and lgc-12, as well as lgc-25 and lgc-26 have single homologues in C. briggsae, Cbr-lgc-11 and Cbr-lgc-26 respectively. Six C. elegans cys-loop LGIC subunits, lgc-13 to lgc-18, are tightly clustered within 20 kb of the T01H10 cosmid whereas in C. briggsae there are seven similar genes (CBG01219, CBG01221, CBG02706, CBG16212, CBG16226, CBG16227 and CBG16229), all of which are not clustered together. Outside of the nAChR orphan group, avr-15 and glc-1 may be paralogues arising in the elegans lineage as there is only a single homologue (CBG06688) in C. briggsae.
The cys-loop LGIC gene families of the parasitic nematodes, Brugia malyai and Trichinella spiralis, are much smaller than those of C. elegans and C. briggsae consisting of 30 and 19 subunits respectively (Williamson et al. 2007). The cys-loop LGIC subunits of both parasites have clear orthologous relationships with those of C. elegans with one exception, a nAChR subunit (ACR-26) which so far appears particular to B. malayi. As discussed by Williamson and colleagues, the striking difference in the cys-loop LGIC gene family sizes may reflect a free-living versus a parasitic lifestyle. Thus, the large complement of subunits in C. elegans and C. briggsae may be required for a small nervous system to be able to respond to many environmental cues which may be more restricted in the specific environmental niche occupied by parasitic nematodes. This shows that the extensive cys-loop LGIC gene superfamily observed for C. elegans is not a feature of all nematodes.
|
[
"ion channel",
"caenorhabditis elegans",
"acetylcholine",
"gaba",
"glutamate",
"serotonin"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Evid_Based_Complement_Alternat_Med-4-1-1810371
|
Does the Consumption of Green Tea Reduce the Risk of Lung Cancer Among Smokers?
|
Experimental and epidemiological studies were reviewed to assess whether the consumption of green tea could reduce the risk of lung cancer in smokers. Articles published since 1990 were located by searching electronic databases PubMed, Ovid and Science Direct, using keywords ‘lung cancer’, ‘tea’ and ‘smoking’ without any restriction on language. After relevant articles had been located, further papers were obtained from their reference lists. Evidence from experimental studies (in vitro animal and human trials) suggested that regular intake of green tea may be protective against tobacco carcinogens. However, the mechanism behind the protective effect is only partly understood. In most of the epidemiological studies reviewed, the green tea exposure was within 5 years of the interview or follow-up, which would coincide with the induction period and latent period of lung cancer. Longer term studies are thus needed to further quantify the cancer risk. There is some evidence suggesting regular intake of green tea at high level (>3 cups per day) may reduce the risk of smokers developing lung cancer. Improvement in measuring green tea intake is required in order to confirm the evidence from epidemiological studies.
Introduction
Lung cancer has become one of the leading causes of death in both developed and developing countries (1,2). Most of the deaths from lung cancer are due to smoking tobacco products. Worldwide in 2000, tobacco smoking contributed 77% of the lung cancer deaths in males aged 30–69 years, and 82% in males over 70 years (1,2). In 2003, for example, the prevalence of smoking in China was estimated to be 58.9% among adult males, while Chinese smokers are four times more likely to die from lung cancer than non-smokers (3).
Although elimination of tobacco smoke exposure is the obvious strategy to control lung cancer, no country has been completely successful in implementing an effective intervention. The efficacy of smoking cessation programs may be enhanced if they are conducted in conjunction with other complementary therapies. Epidemiological and experimental studies have suggested that green tea may be protective against cancer in different sites (4–21). In particular, recent evidence from in vitro, animal and human trial studies indicated the possibility that the consumption of green tea may reduce the risk of lung cancer among smokers (2,22–35), yet findings from epidemiological studies remained inconsistent.
The aim of this review is to assess evidence from experimental and epidemiological studies whether green tea is a risk modifier for those exposed to tobacco smoke. The focus is on the measurement of green tea exposure undertaken in these studies.
Methods
Articles published since 1990 were located by searching electronic databases PubMed, Ovid and Science Direct, using keywords ‘lung cancer’, ‘tea’ and ‘smoking’ without any restriction on language. After relevant articles had been located, further papers were obtained from their reference lists. A total of 78 relevant articles were eventually found.
Results and Discussion
In vitro and Animal Studies
Both in vitro and animal studies have demonstrated that green tea or tea polyphenols have a chemopreventive effect against carcinogens from tobacco and is dependent on both the period and dose of the exposure.
Period of exposure
The period of exposure to green tea or tea polyphenols and the period of exposure to carcinogens must overlap. In one in vitro study, green tea polyphenols were introduced once into the living cell environment 2 h before exposing the cells to a smoke solution for 30 min. Result showed that green tea polyphenols can reduce DNA strand breakage induced by cigarette smoke in cultured human bronchial cells A549 (32). In another in vitro study, human lymphoid cells were exposed to one single dose of green tea extracting solution 2 h before adding benzo(a)pyrene-diol-epoxide (BPDE)—a metabolic production of benzo(a)pyrene (B(a)P), which is a major component of tobacco. The results demonstrated that green tea can significantly reduce DNA damage caused by BPDE (22). Similarly in four other animal studies (23,29,33,34), mice were given green tea prior to or within 48 h of exposure to tobacco carcinogens, and the green tea was continuously administered for 16 weeks or 9 months. Other experimental conditions such as diet, humidity, temperature and dark/light cycle were controlled. Three out of the four studies found a reduction in the incidence and the size of lung cancer tumors due to drinking green tea (23,33,34).
Dose of exposure
The preventive effect of green tea is likely to be related to the concentration of green tea as well as the dose of carcinogens. In experimental studies both cells and animals were often treated with relatively high doses of carcinogens and green tea. For example, in a study which showed a protective effect of green tea, the concentration of tobacco carcinogens used was equivalent to the amount contained in the sidestream smoke from one cigarette dissolved in 24 ml of water, while the concentration of green tea polyphenol was set at 250 μg ml−1, corresponding to 116.5 μg ml−1 of epigallocatechin-3-gallate (EGCG) (32). Moreover, the bioavailability of different green tea polyphenols has been a concern in animal and human studies. Green tea polyphenols are not usually stable, specifically, in water, EGCG can oxidize more than 50% within 12 h, yet green tea used in experimental studies was typically prepared every 48 h (33,34) to every 7 days (29).
In some studies where mice were exposed to both NNK [4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone] and green tea, the NNK dosage administered was usually equal to that experienced by a smoker in a lifetime (36). In one experiment mice were randomized into different groups and given green tea solution at concentrations of 0.2%, 0.4% and 0.6% as drinking fluid 24 h after treating with NNK. Comparing with the control group that drank only water, the protective effect of green tea was only evident at the 0.6% concentration level (33). In another study (29), mice were exposed to high levels of environmental tobacco smoke for 5 months, while being given 1.25% green tea (2.5 g tea leave mixed with 200 ml boiling water). The results indicated green tea had little effect on reducing tumor incidence or multiplicity of the tumors. However, none of the animals in the green tea treatment group experienced early death while >10% of the mice in the control group experienced early deaths, probably due to an overdose of carcinogens exposure (29). Findings from these experiments suggest that green tea may not be effective to inhibit the formation or growth of lung tumor under high level of carcinogens exposure and over a long duration.
Human Trials
Short-term effect
Because a single dose of green tea intake only produces a short-term antioxidant effect, regular consumption seems necessary. In a human study (37), 21 non-smokers consumed 300 ml of green tea solution or about three cups. Ferric reducing ability of plasma (FRAP-assay) was used to determine the antioxidant activity in plasma. The results of this study showed that green tea consumption produced a significant increase in antioxidant activity within 30–120 min of exposure, but after 120 min the antioxidant activity appeared to decrease. Another human trial with 10 healthy subjects similarly found a short-term increase of plasma antioxidant induced by drinking two to three cups of green tea at one time (38). Such a short-term increase seemed to correspond with the relatively short half-life of tea polyphenols (30). In one pharmacokinetic study (39), 30 healthy participants ingested 200 mg of EGCG equivalent to two cups of green tea. Blood samples were collected at 0.5, 1, 2, 4, 6, 8 and 24 h after the EGCG administration. This experiment found that the mean half-life of EGCG in plasma was only 118 min.
Protective effect from regular consumption
Regular green tea consumption could be protective over an extended period. In one pilot study (35), three non-smokers and three smokers (consumed more than 10 cigarettes per day) drank two cups of green tea in the morning, two cups in the afternoon and one cup at night daily for 4 weeks. Each cup of green tea contained 400–500 mg of green tea extract. Oral keratinocytes were harvested from each subject for analysis. Outcomes included a reduction in smoking-induced DNA damage, increased apoptosis and a further decrease in the number of cells carrying damaged DNA (35). In a randomized control trial (27), 133 smokers who smoked at least 10 cigarettes per day during the past year were randomized into three groups consuming at least four cups daily of either (i) decaffeinated green tea, (ii) decaffeinated black tea, or (iii) water. After 4 months of intervention, smokers who drank decaffeinated green tea had a significant decrease in their urinary 8-hydroxydeoxyguanosine (8-OHdG) level, but no significant changes were observed in both water and black tea groups. Similar reduction in urinary 8-OHdG level was observed among HbsAg positive volunteers who took daily capsules containing green tea polyphenols equivalent to two to four cups of tea for 3 months (40). In an intensive observational study (25), the frequencies of sister-chromatid exchange (SCE) in peripheral lymphocytes were measured as markers of smoking-induced DNA damage. The frequencies of SCE were significantly higher among smokers who were non-tea drinkers, than those of non-smokers and smokers who consumed green tea at least two to three cups per day during the past 6 months. The differences between the latter two groups, however, were not significant. It is possible that consuming high levels of green tea over a long period may reduce the DNA damage caused by tobacco smoking.
Lung tumor treatment
Only one clinical trial used green tea for treatment of lung cancer (41). In this Phase I study, 17 advanced lung cancer patients were given green tea extract at a single dose of 3 g m−2 per day for 4 weeks. The dosage equates to 20 cups of green tea, which has been found to be the maximum tolerated dose. Unfortunately, no effect of the treatment was observed. Therefore, green tea is unlikely to be an effective cytotoxic agent against existing tumor.
Limitation of human trials
Except two studies (27,40), most human intervention studies involved small sample sizes generally less than 50 subjects. In addition to ethical concerns, there were neither direct measurement on the concentration of green tea polyphenols in lung tissue nor on the change of lung tissue after green tea exposure. Instead, the effects of green tea were measured by the change in DNA damage and apoptosis of oral cells (35), changes in biomarkers of DNA damage such as 8-OHdG and SCE (25,27,40), and changes in plasma antioxidant capacity (30,37,38). Trials with sufficient sample size and longer period of follow-up are needed to confirm the protective effect of green tea.
Mechanisms Based on Experimental Studies
As shown in Fig. 1, carcinogens contained in cigarette smoke are activated by Phase I enzymes in the body, leading to the formation of DNA adducts (36). If DNA adducts escape cellular repair mechanisms, it may result in miscoding and eventually a stable mutation. Blocking any step in the pathway would reduce the incidence of lung cancer among smokers (36). Table 1 provides the possible mechanisms of chemoprevention effect of green tea. The evidence suggested that catechins in green tea, especially EGCG, may prevent the formation of a mutated cell.
Figure 1.Mechanism of tobacco carcinogens inducing lung cancer.
Table 1.Possible mechanism of actionPossible mechanism of chemoprevention effect of green teaReferencesActing as antioxidant to prevent DNA damage and reduce the formation of DNA adducts(11,22,25,27,35,40)Increasing apoptosis and preventing the persistence of miscoding DNA (possibly by inducing gene regulation)(11,26,31,33,35,40,43,44)Activating Phase II detoxifying enzymes, and increasing the excretion of carcinogens (possibly by inducing gene regulation)(2,8,28,31)
In both animal and in vitro studies, green tea was found to increase the activities of Phase II enzymes (2,8,28,31). Phase II enzymes transform carcinogens into non-toxic molecules for subsequent excretion (42). The amount of carcinogens that may be further activated, as well as DNA adducts, are thus reduced. The antioxidant properties of green tea have also been demonstrated in trials involving smokers. It reduced the urine concentration of 8-OHdG, a strong biomarker of DNA damage (27). Moreover, green tea extract could inhibit the formation of human DNA adducts (11,22,35).
Once carcinogen DNA adducts have formed, persistent mutation is likely unless cells with DNA adducts are removed by apoptosis (36). Human (35), animal (11,33) and in vitro studies (26,43,44) have shown that green tea can induce apoptosis among injured cells exposed to carcinogens from tobacco smoke. However, smoking may be regarded as a chronic exposure, since long duration of tobacco smoking has greater impact on lung caner risk than the amount smoked per day (45). Therefore, a long exposure to green tea is needed in order to reduce the damage caused by tobacco carcinogens.
Epidemiological Studies
Induction time and latent period of lung carcinoma
Latent period refers to the time between disease occurrence and detection, while tumor induction time commences when the effect of the exposure occurs until all of the component causes involved in one sufficient cause are satisfied—the time of disease occurrence (46). Exposure within the induction and latent periods is irrelevant to the formation of a tumor (47). Moreover, the incidence of lung cancer does not appear to be lower among ex-smokers who quit smoking within 5 years than current smokers (48,49). This indicates that the sum of induction period and latent period of lung cancer caused by tobacco smoking may be longer than 5 years. Logically, no reduction in incidence of lung carcinoma is expected among current smokers who drink green tea for 5 years prior to the calculation of incidence, regardless of their consumption level.
Measurement of green tea exposure and other limitations
Table 2 summarizes epidemiological studies on tea and lung cancer. Three case–control studies (50–52) found a significantly lower risk of lung cancer among non-smokers with a high consumption of green or black tea, but no protective effect was evident among smokers (50,51). However, the period of tea consumption referred to 5 years or less before the diagnosis or interview. Although the subjects might have been drinking tea for a long time, the duration and quantity of tea consumption was not assessed, so the effect of green tea could not be fully justified. Cohort studies, on the other hand, were based on information of tea consumption at the beginning of the study (53–55) and suggested that tea consumption increased with age (54,55). In particular, the Japanese study (53) with a long follow-up period of 14 years showed that people consuming green tea two to four times per day had a relative risk of 0.78 (95% CI = 0.6–1.0), when compared with those drinking once or less per day and after adjustment for potential confounders. But for subjects drinking green tea more than five times per day, the relative risk was only 0.79 (95% CI = 0.59–1.1), the non-significance result probably due to insufficient sample size at this high level of consumption (53).
Table 2.Epidemiology studies on tea and lung cancerCountryStudy designSample sizeTea type and highest consumption levelReference period of intakeSmoking status of subjectsResultsChina (50)Population-based case–control studyCases: 649 Controls: 675Green tea: >1500 g tea leaves per year5 years before interviewSmokers and non-smokersOR = 0.65; 95% CI = 0.45–0.93 for non-smokers OR = 0.94; 95% CI = 0.40–2.22 for smokersCanada (52)Population-based case–control studyCases: 582 Controls: 582Type unspecified: >7 cups per day2 years before studyNon-smokersOR = 0.4; 95% CI = 02–0.7USA (51)Population based case–control studyCases: 161 Controls 483Green tea: >1 cup per day1 year before diagnosis or interviewSmokers and non-smokersOR = 0.9; 95% CI = 0.5–1.6Japan (53)Cohort study (1980–94)Subjects: 38 540 Person-years at risk: unspecified Cases: 436Green tea: >5 times per dayUnspecified; survey taken at start of studySmokers and non-smokersRR = 0.78; 95% CI = 0.6–1.0 for people drinking tea 2–4 times per dayJapan (54)Cohort study (1986–95)Subjects: 8552 Person-years at risk: 71 248.5 Cases: 384Green tea: >10 cups per dayUnspecified; survey taken at start of studySmokers and non-smokersCancer in all sites RR = 0.68 for males; 95% CI = 0.39–1.21 RR = 0.57 for females; 95% CI = 0.33–0.98OR, odds ratio; RR, relative risk; CI, confidence interval.
In summary, there is only limited evidence from epidemiological studies that green tea is protective against lung cancer among smokers. Moreover, green tea exposure was generally taken to be within 5 years of interview or follow-up, which would have little impact on the incidence of lung cancer.
Conclusion
Our review suggests that regular intake of green tea at high levels (>3 cups per day) may offer protection against tobacco carcinogens for smokers, provided that the duration of green tea consumption is sufficiently long to cover the smoking period. For epidemiology studies, improvement in measuring green tea intake is required in order to confirm the chemopreventive effect of green tea observed in in vitro animal and human trials.
|
[
"smoking",
"tobacco",
"polyphenols",
"carcinoma"
] |
[
"P",
"P",
"P",
"P"
] |
Eur_Spine_J-2-2-1602186
|
Adjacent level discitis after anterior cervical discectomy and fusion (ACDF): a case report
|
This report describes a case of spondylodiscitis occurring adjacent to levels at which anterior cervical discectomy and fusion was performed. The objective is to describe a rare cause of spondylodiscitis and discuss its successful management. Post-operative discitis involving the same level is a known occurrence. We report an interesting case of spondylodiscitis occurring at the adjacent level of fusion, and to our knowledge this is the first such case reported in literature. A two-level decompression and fusion was performed at C5–6 and C6–7 levels with PEEK cages and anterior cervical plating in a middle-aged gentleman for persistent axial neck pain and left-sided radiculopathy involving C6 and C7 distribution. After 6 weeks, the patient presented to us with complaints of mild paresthesia in the abdomen and extremities. Radiological investigations including plain radiographs and MRI revealed a surprising finding of discitis at C4–5 level with an associated epidural abscess. In view of the patient’s myelopathic symptoms, surgical debridement and decompression of the spinal cord was performed. The plate and screws were removed, the cages were left intact, and the C4–5 disc level was reconstructed with tricortical iliac crest autograft. No further instrumentation was performed. The biopsy specimen from the disc at C4–5 level grew Serratia marcescens. It was contemplated that C4–5 discitis was initiated by inoculation of bacteria at the superior endplate of C5 by contaminated vertebral pins/drill-bit or screws. Adjacent level discitis is a rare but potentially serious complication of anterior cervical fusion. A high index of suspicion of infection is necessary if the patient complains of new symptoms after anterior cervical fusion. Thorough assessment and aggressive treatment is necessary for successful management.
Introduction
Post-operative discitis occurring at the operated level is a known complication of spinal surgery. The most common organisms implicated in the pathogenesis are Staphylococcus aureus and Staphylococcus epidermidis. Degeneration occurring at the adjacent level is also a well-established phenomenon after spinal fusion. We present a unique case of discitis occurring at a level adjacent to the levels that underwent anterior cervical fusion. The case becomes interesting because of numerous factors e.g. the site of involvement, the likely etiology, the organism concerned, the subtle nature of the symptoms, problems with diagnosis and its management.
Case report
A healthy 56-year-old male patient presented with neck pain of 6 months duration and associated radicular pain and numbness in the left upper-limb. The history was negative for fever, night pain or constitutional symptoms. MRI revealed left-sided foraminal stenosis at C5–6 and C6–7 levels secondary to disc herniation. A decision to perform surgical decompression and stabilization was taken after a failure of conservative trial of 12 weeks. A standard anterior cervical discectomy and fusion (ACDF) was performed at both levels using a right-sided approach. Based on anatomical landmarks and radiological guidance, adequate care was taken to expose only the presumed C5–6 and C6–7 disc levels and avoid inadvertent injury to other discs. The Casper distractor system was subsequently utilized throughout the surgery by obtaining anchorage of the Casper vertebral pins in the C5 and C7 vertebral bodies. Iliac crest autograft-packed PEEK (Solis, Stryker, Allendale, NJ, USA) cages were placed at both levels and cervical plating (Orion, Medtronic-Sofamor-Danek, Memphis, TN, USA) was performed. Prophylactic intravenous antiobiotics (cefazolin) was administered 1 h before surgery and continued for 48 h post-surgery. The immediate post-operative period was uneventful and the patient had good recovery from his neurological symptoms.
The patient presented 6 weeks later with mild neck and bilateral shoulder pain associated with paresthesia in his upper limbs, left lower limb and abdomen. On examination, the patient was afebrile and there was no warmth or tenderness in the region of the neck. The neck movements were terminally restricted, but there was no motor weakness. Hyperreflexia could be elicited in both lower limbs, Babinski’s sign was negative and there were no objective sensory deficits. Hematological investigations revealed a mild increase in ESR (30 mm/h), normal CRP (1.1 μg/ml) and a mild increase in total WBC count (8.18×109/l). Radiographs revealed persistently enlarged pre-vertebral soft-tissue shadow, decreased disc height at C4–5 level with erosion of end-plates and segmental C4–5 kyphosis, apart from the presence of halo around the screws as shown in Fig. 1. MRI revealed features of discitis at C4–5 level with associated pre-vertebral and epidural abscesses from C4 to C6 levels as well as artifacts secondary to instrumentation and cages as shown in Fig. 2a and b.
Fig. 1Lateral X-ray of cervical spine at 6 weeks after index surgery. Persistence of pre-vertebral soft-tissue shadow, erosion and irregularity of C4 and C5 end plates apart from segmental kyphosis and loosening of screws can be seenFig. 2a Sagittal T-2 weighted MRI image showing hyperintensity signal at C4–5 disc space. Anteriorly located epidural abscess at the same level can be appreciated. b Axial T-2 weighted MRI image displaying anterior compression of the spinal cord by the epidural abscess
The patient underwent an aggressive surgical debridement and stabilization through a left-sided approach. The loose screws and the plate were removed. A thorough decompression of the epidural abscess along with surgical debridement of C4–5 disc space was performed. No pus or evidence of infection was seen at the C5–6 and C6–7 levels and the cages were left behind as they appeared stable and secure. Tricortical iliac-crest autograft was impacted at the prepared C4–5 disc space and thorough lavage with antibiotic solution (cloxacillin and gentamycin) was carried out. The wound was then closed over a drain and an empirical course of intravenous antibiotic (cefazolin) was initiated. The drain was taken out after 3 days and the culture specimen grew Serratia marcescens, sensitive to ceftriaxone but resistant to the more commonly used antibiotics including cefazolin. Intravenous ceftriaxone was initiated for the first 3 weeks, and then the patient was discharged on oral antibiotics for another 3 weeks. The patient’s condition improved dramatically with complete neurological recovery within the first week. At the 8-month mark, he is asymptomatic and the X-rays performed after 5 months reveal solid fusion at all levels and obliteration of the pre-vertebral soft-tissue shadow as shown in Fig. 3.Fig. 3Lateral X-ray at 5 months after the revision surgery. Solid fusion at all levels and obliteration of the pre-vertebral soft-tissue shadow can be seen
Discussion
Post-operative discitis in cervical spine is not a frequent occurrence due to the rich vascularity and the routine use of antibiotic prophylaxis [5]. The prevalence ranges from 0 to 1.1%. The use of metal implants may increase the possibility of infection [4]. Discitis is a known complication of cervical discography [2, 11]. Occasional cases of cervical discitis secondary to a fishbone lodged in the throat have been reported [4, 7].
A case of post-operative discitis adjacent to the operated level has not been previously reported. The potential causes of adjacent level discitis at C4–5 were scrutinized. One of the possibilities was an accidental inoculation of bacteria into the C4–5 disc space intra-operatively by a contaminated spinal needle that was used as a radiological marker. When we reviewed the intra-operative localizing X-ray, it was evident that the needle was placed into the C5–6 intervertebral disc and hence this possibility was ruled out. The possibility of discitis secondary to an esophageal perforation was ruled out since the patient neither had dysphagia nor the severe systemic symptoms that usually accompany esophageal injury in the post-operative period. Additionally, no esophageal tear was noticed during the second surgery. Discitis secondary to annular injury of C4–5 intrevertebral disc was ruled out since great care was taken to avoid inadvertent exposure of uninvolved levels during surgery. Hematogenous spread is another possibility but the only positive cultures came from the C4–5 intervertebral disc during the operation. Other samples, e.g. blood, sputum, and urine cultures demonstrated no growth and chest X-ray appeared normal. The only other possibility could be inoculation of bacteria into the superior endplate of C5 and initiation of an infective process in the C4–5 peridiscal region. This could be through the contaminated Caspar pins or the drill-bit utilized to create holes for the screws used for anterior instrumentation. This is a possibility since the organism that was cultured (S. marcescens) is known to contaminate and breed in normal saline solution [3, 6], and the operative instruments in question are generally placed in seemingly sterile large bowls filled with normal saline during surgery.
Post-operative spinal infection due to S. marcescens is very rare. In a review of 2,391 consecutive index procedures on the spine, 46 cases of wound infection were identified [8]. Amongst them, only one case of Serratia infection was recognized and that was a part of mixed infection. Serratia spondylodiscitis has been reported recently involving the lumbar spine [3]. One of the patients had discitis at the same level that underwent a micro-discectomy. The second case was interesting in that her discitis occurred at the same level that had earlier undergone a posterior decompression (laminotomies and foraminotomies) but the disc was untouched. These two patients presented acutely within a week of index surgery with a fiery presentation, described as life threatening by the authors. The ESR, WBC counts and CRP levels were elevated and the blood culture grew S. marcescens. However, our patient had a subacute course and was afebrile throughout. The only abnormal features were subtle axial neck pain with subjective paresthesia in the extremities and abdomen. The ESR and WBC count were modestly raised but the CRP was normal. A high index of suspicion along with reliance on radiological investigations was essential to diagnose the condition and institute appropriate treatment. Persistence of pre-vertebral soft-tissue shadow apart from reduction in the C4–5 disc height, segmental kyphosis and irregularity of the endplates at C4–5 pointed towards a diagnosis of C4–5 discitis and prompted us to perform MRI in order to confirm the diagnosis and decide on the further management.
The literature is not very clear with respect to the duration of antibiotics to be given for spinal fusions. The protocols recommended range from no antibiotics for anterior spinal surgery [9] to 3 days of antibiotics for all spinal surgeries [10]. A meta-analysis performed to identify the efficacy of prophylactic antibiotics in spinal surgery found that there was no difference in the efficacies of various antibiotic regimens, provided at least one dose of pre-operative antibiotic with gram-positive coverage was administered [1]. Our protocol for antibiotic prophylaxis for spinal surgeries consists of a pre-operative dose followed by 2-day course of intravenous cefazolin. The protocol followed in our division is based on the fact that the drainage tube as well as the urinary catheter is removed by 48 h and hence the antibiotics are usually continued till then. S. marcescens is highly resistant to cefazolin and sensitive to third-generation cephalosporins [3], and this was evident in the culture sensitivity reports of our patient.
An aggressive approach to debride the C4–5 disc space as well as to decompress the epidural space was necessary in the wake of myelopathic symptoms and in order to isolate the organism. The spine was approached from the opposite side (left side) for the second surgery since inadvertent esophageal or neurovascular injury as a result of post-operative scarring was a possibility on the right side. The issue of leaving the PEEK cages at C5–6 and C6–7 levels behind during the revision surgery is debatable. We did not replace them because of a number of reasons. The infection was principally arising from the C4–5 disc level, the cages were absolutely stable and we wanted to minimize graft site morbidity. Most importantly, we felt that infection would not thrive in the presence of stability. This approach has been advocated by other authors too [8]. In the second surgery, a plate was not utilized for instrumentation for the following reasons. Firstly, the bones were soft as a result of inflammatory edema resulting from infection and hence were not ideal for screw implantation. Secondly, since we were also not sure what the organism was and as most of the times it is Staphylococcus aureus, the risk of persistent infection as a result of glycocalyx membrane formation is high. Additionally, the construct appeared stable after the C4–5 level was impacted with the autologous iliac crest graft and the lower two levels were already stable as aforementioned.
Conclusions
The possibility of infection should be considered in patients with an alteration in the nature of their symptoms after index surgery. Unusual presentations can be a challenge to diagnose and treat and a high index of suspicion and aggressive approach are prudent for a successful outcome.
|
[
"adjacent level discitis",
"epidural abscess",
"serratia marcescens",
"cervical spine",
"cervical spondylodiscitis"
] |
[
"P",
"P",
"P",
"P",
"R"
] |
Graefes_Arch_Clin_Exp_Ophthalmol-4-1-2206251
|
Incidence and visual outcome of endophthalmitis associated with intraocular foreign bodies
|
Purpose To determine the risk factors and visual outcome of endophthalmitis associated with traumatic intraocular foreign body (IOFB) removal and its allied management.
Introduction
Intraocular foreign body (IOFB) caused ocular trauma is a significant and unique type of trauma that requires skillful investigation and an early intervention. Endophthalmitis is an uncommon but potentially catastrophic complication of penetrating ocular injury with retained intraocular foreign bodies (IOFB) [1]. Studies have reported incidence of endophthalmitis ranging from none to as high as 13.5% [1–5]. Management of endophthalmitis associated with retained IOFB is a challenging subject. With advancement in vitreous surgery, a large number of eyes with ocular trauma and IOFB are being saved. Visual prognosis, however, is affected by the complexity of the confluent factors surrounding the IOFB, which include the size, site, material, trajectory, reactivity of foreign body, inflammatory response, degree and type of tissue damage, length of time since injury and any associated endophthalmitis [2]. Limited information is available regarding ocular trauma and IOFB-associated endophthalmitis outside of the United States and other developed countries [6–10]. In particular, very limited information on this entity is available from the Middle East. The current study was undertaken to analyze and report our experience at a large tertiary eye care referral center in the Middle East regarding the management of eyes with IOFBs and associated endophthalmitis. We identify the risk factors for the development of clinical endophthalmitis and visual outcome of such eyes at a major referral center in the Middle East.
Methods
A detailed retrospective review was conducted of all patients who presented to King Khaled Eye Specialist Hospital, a JCIA (Joint Commission International Accreditation, USA) accredited tertiary eye care referral center in Riyadh, Saudi Arabia, from January 1983 to August 2004 with penetrating ocular trauma and retained intraocular foreign bodies (IOFB). Methods of IOFB extraction has varied during these years. For simplicity, the study period was divided into the first decade from 1983 to 1993 and the second decade from 1994 to 2004. For the past decade pars plana vitrectomy has been used more often for the posteriorly located IOFBs. Prior to removal of the foreign body (FB) all adhesions around the FB were released and it was freed from encapsulation where indicated. All IOFBs were removed by using IOFB forceps. Where necessary the sclerotomy was enlarged to facilitate easy removal of the FB. Endophotocoagulation was applied to the retina adjacent to the site of the IOFB. After removal of the FB, vitrectomy was utilized to remove any remains of FB capsule or fibrous tissue with the vitrectomy cutter. Only those eyes which showed any clinical evidence of endophthalmitis after trauma and retained IOFB were included in the study. Patients demographic studied included, age at presentation, sex, place of trauma, occupation, mode of injury and time between injury and repair. Other parameters included initial and final best corrected Snellen visual acuity, entry and location of IOFB and associated cataract, vitreous hemorrhage, retinal detachment, development of endophthalmitis, diagnostic studies performed, treatment rendered, type and size of IOFB. Complications such as cataract, retinal detachment and secondary procedures performed were also noted. Causes of visual loss such as corneal scarring, cataract, retinal detachment, retinal scars and loss of an eye were also investigated. In particular, all eyes with a retained IOFB and associated endophthalmitis were investigated and analyzed in detail.
Statistical analysis was carried out using SPSS version 10 (SPSS Inc., Chicago, IL, USA). We studied the association between the variables using the chi-square test because the data was categorical. Further multiple logistic regression analysis was conducted to predict the factors that were associated with favorable or poor visual outcome. A P < 0.05 was taken as a level of statistical significance.
Results
Among the 589 eyes of 565 patients (90.3% males; 9.7% females; 24 bilateral) with retained intraocular foreign bodies (IOFB) after trauma, 44 eyes (7.5%) had evidence of clinical endophthalmitis at the time of initial evaluation or subsequent to removal of the IOFB (Table 1). The setting of the injury resulting in IOFB varied, with the majority occurring at the place of employment in younger patients. The size of the IOFBs recovered ranged from 0.5 mm to 18 mm (average 4.2 mm). At presentation visual acuity (VA) of 20/200 or better was found in 8 eyes (18.1%), 20/400 to counting fingers in 6 (13.6%) and hand motion (HM) to light perception (LP) in 30 eyes (68.2%) (Table 2). Delay in treatment was mostly due to late patient referral after the trauma or lack of understanding on the part of the patient about the urgency in presentation to the primary care center. Endophthalmitis was recognized preoperatively in 32 eyes (72.7%) and post-operatively in 12 eyes (27.3%). In general, the size of the IOFB posed no significant risk of causing endophthalmitis except in cases where larger IOFB was associated with a significant trauma to the eye. The most important predictive factor of developing endophthalmitis was delayed repair of the globe and removal of IOFB. A delay in intervention of more than 24-hours was associated with a risk of clinical endophthalmitis (Fig. 1). In fact, only 11 eyes (25%) which developed endophthalmitis were repaired and IOFB removed within 24 hours after their trauma compared with 33 eyes (75%) that were repaired and IOFB removed more than 24 hours after trauma (range 2–42 days). In other words, a delay of more than 2 days in the repair of the traumatic globe and removal of IOFB was associated with a significant risk of endophthalmitis development (P < 0.05). The composition of IOFB had no significant effect on the development of clinical endophthalmitis. In particular, eyes with wood IOFB did not appear to be associated with increased risk of endophthalmitis compared to eyes with metallic IOFBs. The age of the patient had no bearing on the development of clinical endophthalmitis. However, when age and delayed repair of the trauma and removal of IOFB was correlated, an association with increased evidence of endophthalmitis was observed. The risk of endophthalmitis development and poor visual outcome was less common in eyes where the location of IOFB was in the anterior chamber or lens compared to the eyes having IOFB in the vitreous or retina. Among the 10 eyes where the final vision of no light perception (NLP) was recorded after the treatment of IOFB trauma and associated endophthalmitis, only 2 eyes had anterior location of foreign bodies compared with 8 eyes where IOFB were found in the posterior chamber, the difference being significant (P < 0.05). There were 24 cases of endophthalmitis during the first decade (1983–1993) and 20 cases of endophthalmitis during the second decade (1994–2004) of the study; the difference however was not statistically different.
Table 1Reported incidence of endophthalmitis with IOFBStudiesEyes with endophthalmitis (total eyes)PercentBrinton et al. [11]11 (103)10.7Khan et al. [12]10 (198)5.1Williams et al. [3]14 (105)13.3Behrens-Baumann and Praetorius [5]14 (297)4.7Thompson et al. [1]34 (492)6.9El-Asrar et al. [4]13 (96)13.5Present study (Chaudhry et al.)44 (589)7.5Table 2Presenting and final visual acuity of patients with IOFB and endophthalmitisVisual acuity (VA)PresentingFinal20/20–20/6031120/80–20/2005920/400–CF64HM–LP3010NLP010VA visual acuity, CF count finger, HM hand motion, LP light perception, NLP no light perceptionFig. 1Patient with endophthalmitis associated with retained intraocular foreign body
Positive cultures were obtained from 17 eyes (38.6%) with clinical signs of endophthalmitis (Fig. 1). In addition, 10 eyes showed evidence of microorganisms on the Gram stain but no growth was observed in cultures. Staphylococcus and Streptococcus species were most often recovered (9 eyes); the other species included Hemophilus, Bacillus, Pseudomonas, Eikenella, Corynebacterium, Propionebacterium acnea and Escherichea coli. None of the eyes with Bacillus, Pseudomonas or Corynebacterium attained any useful vision. All eyes suspected of endophthalmitis were administered intravitreal antibiotics and in some cases the treatment was repeated to obtain resolution of the signs. Most common intraocular antibiotics administered included vancomycin, ceftazidime and amikacin. Systemic preoperative antibiotics were administered in eyes with signs of endophthalmitis at approximately the same frequency as eyes without any signs of clinical endophthalmitis. However, systemic antibiotics were continued postoperatively for longer duration in eyes with signs of infectious endophthalmitis.
Thirty-one eyes (70%) underwent primary pars plana vitrectomy and 20 eyes (45%) had pars plana lensectomy at the time of IOFB removal and primary repair. Secondary procedures included scleral buckle around 8 globes, endolaser or external cryotherapy in 10 eyes and gas tamponade using SF6/C3F8 in 10 eyes for retinal breaks or retinal detachments. In addition, 11 eyes required cataract surgery, 12 eyes developed retinal detachment requiring additional treatment after the removal of IOFB and treatment of endophthalmitis. Vitreous hemorrhage was noted in 24 eyes that were associated with posteriorly located IOFBs but none among the eyes with IOFBs located in the anterior part of the eye. The final VA was 20/200 or better (defined as a good visual outcome) in 20 eyes (45%). Twenty-four eyes (55%) had final VA of less then 20/200 (defined as poor visual outcome). Twenty eyes (45%) had a final VA of hand motion or worse (Fig. 2). The causes of final vision of HM or worse in 20 eyes were phthisis (Fig. 3) or enucleation in 10 eyes, retinal detachment with proliferative vitroretinopathy in 5 eyes and corneal or macular scar in 5 eyes. Overall, 22 eyes (50%) that developed endophthalmitis due to trauma and retained IOFB had improvement in their VA. Ten eyes (22.7%) with endophthalmitis became NLP compared to 58 eyes (10.6%) from the larger group which became NLP but had no evidence of endophthalmitis, the difference being significant (P < 0.05). Among the 10 eyes with endophthalmitis where the final vision of NLP was recorded after the treatment of IOFBs and associated endophthalmitis, only 2 eyes had anterior location of FBs compared to 8 eyes where FBs were found in the posterior chamber. Among the patients with endophthalmitis, 4 eyes (12.9%) from the vitrectomy group (31 eyes) had final vision of NLP, while from the eyes without vitrectomy (11 eyes), 5 eyes (45.4%) had final vision of NLP, the difference between these two groups being significant (P < 0.05).
Fig. 2Patient with light perception vision after treatment of endophthalmitis and removal of intraocular foreign bodyFig. 3Patient with phthisical left eye and no light perception after treatment of endophthalmitis and removal of intraocular foreign body
Discussion
Ocular trauma associated with retained intraocular foreign bodies (IOFB) constitutes a significant proportion (18–40%) of all ocular injuries requiring surgical management [1, 2]. Despite advances in vitreoretinal and microsurgical techniques, the management of these injuries remains a challenge. The risk of endophthalmitis developing after penetrating ocular injury with retained IOFB is relatively low with the current techniques of surgical repair [1]. The incidence of infectious endophthalmitis after retaining an IOFB has been reported to be as low as 0% to as high as 13.5% [3–5, 13], the mean incidence being 6.8% (Table 1). The results from the present study are in agreement with the above studies of endophthalmitis associated with retained IOFB. A delay in management of more than 24 hours was found to increase the risk of infectious endophthalmitis in the present study. From a large study of the National Eye Trauma System Registry, 91% of the eyes in which infectious endophthalmitis developed had such evidence at the time of initial evaluation [1]. Delay in primary repair at our tertiary care eye hospital in the Middle East was either due to the patient presenting late or being referred to our facility later in the course of trauma and retained IOFB. Despite methods of improved transportation, early diagnosis and referral to properly trained eye care professionals, morbidity due to IOFB associated trauma has not decreased over the last decade.
The risk of an infectious endophthalmitis developing with IOFB has been reported to increase with age [1]; although, no such trend was observed in the present study. However, when age and delayed repair of the trauma and removal of IOFB was correlated, an association with increased evidence of endophthalmitis in our group of patients was observed. These results are in agreement with previous observations that older patients with delayed primary repair do appear to have an increased susceptibility to developing infectious endophthalmitis compared with younger patients [1].
Visual acuity (VA) appears to be significantly affected in the eyes with IOFB that subsequently developed endophthalmitis. From a previous large study of retained IOFB, Roper-Hall [14] reported 555 cases over a 19 year period, where there were no bilateral cases: in 60 eyes IOFB was not removed, size of IOFB was recorded in only 89 eyes and the enucleation rate was 20%. From a more recent report of 96 cases of posterior segment IOFBs studied over a 14 year period with 8.6 months of average follow-up, endophthalmitis was recorded in 13.5% of cases [4]. The overall rate of 7.5% cases with endophthalmitis in the current study are within the reported range of larger series [1, 3–5, 14, 12].
Several preoperative and operative factors have been found to have prognostic value in the final visual outcome of the traumatized eye with retained IOFB and endophthalmitis. Initial VA is an indicator of final VA and is a well recognized predictor of visual outcome in most of the previously published studies on IOFBs [4, 15]. In our study, poor initial VA of HM or light perception was significant in predicting a poor visual outcome (Table 2). The risk of endophthalmitis development and poor visual outcome appeared to be significantly increased when the IOFBs were recovered from the posterior chamber (vitreous or retina), compared to the IOFBs recovered from the anterior chamber. Among the 10 eyes with endophthalmitis where the final vision of NLP was recorded after the treatment of IOFBs and associated endophthalmitis, only 2 eyes had anterior location of foreign bodies compared to 8 eyes where foreign bodies were found in the posterior chamber.
In our study, positive cultures were obtained in slightly more then one-third of the eyes with clinical signs of endophthalmitis. In general, the culture is often positive in clinically diagnosed endophthalmitis secondary to trauma than due to intraocular surgery. However, considering that our hospital is a tertiary eye care referral center, most of these patients were on topical as well as systemic antibiotics prior to their referral and obtaining cultures at the time of IOFB removal. This view is corroborated by the observations that 10 additional patients who had evidence of microorganisms on Gram stain showed no growth in culture. The microorganisms isolated from the eyes in the present study were similar to those reported in previous studies on the development of infectious endophthalmitis after IOFB trauma [1, 4]. Although, from many eyes suspected of infectious endophthalmitis, no definite organism was recovered. The fact that these eyes showed clinical improvement after intravitreal injection of antibiotics supports the notion that an infectious process was involved. In a recent study, reasonable visual outcomes with immediate intravitreal injection of antibiotics and delayed vitrectomy in patients with clinical features of endophthalmitis and retained IOFB have been achieved [16].
The role of vitrectomy in the treatment of penetrating ocular trauma has been considered a major advance in the last two decades [17]. Early vitrectomy has helped in attaining useful functional vision in 25–51% of the eyes with posterior segment trauma [17, 13], and in some series it has helped in significant survival of eyes without improvement of final visual outcome [17]. In our hospital, during the last decade, pars plana vitrectomy had been used more frequently for the posteriorly located IOFBs. Although there were fewer cases of IOFB-associated endophthalmitis during the second decade of the study period, statistically we did not find a difference between the two decades. In our study of endophthalmitis, only 12.9% of the eyes in which vitrectomy had been performed initially did the vision later become NLP compared to 38.5% in which the vision became NLP in the eyes which had not undergone vitrectomy as an initial procedure at the time of IOFB removal and repair of ocular trauma. Vitrectomy removes the vitreous and blood clot scaffold that provides a framework to the visual localization of IOFB; it helps to find retinal breaks and detachments which may not be visible in the setting of vitreous hemorrhage. Our results corroborate previous reports of beneficial effects of early vitrectomy in the setting of ocular trauma associated with IOFBs [13]. Removal of lens was not necessary in all cases with corneal entry wound because the lens injury was localized without interfering with view for vitrectomy. Removal of posterior segment IOFB by pars plana vitrectomy has been advocated because it provides direct viewing and controlled removal of the IOFB [3, 18]. With the advent of advanced instrumentation and viewing systems used in vitrectomy, most vitrectomies in our hospital were performed in the second decade compared to the first.
The size of IOFB has been found to be a significant predictive factor of poor visual outcome in the previous studies of IOFB removal [18]. A large IOFB is more likely to inflict severe damage at the time of entry because of its higher kinetic energy, leading to a poor visual prognosis [18]. However, when considering similar sized IOFBs, no particular association between the visual outcome and the size of IOFB in eyes which developed endophthalmitis was found in our study.
A preoperative retinal detachment may be present in 5–21% of the eyes with IOFB and has been reported to be an important risk factor for poor visual outcome [3, 19]. The timing of surgery in these eyes for the removal of IOFBs has been found to be an important prognostic factor for better visual outcome. Without delay, removal of IOFB has been found to reduce the chances of endophthalmitis [15, 19, 20]. In the absence of endophthalmitis, many studies have found no significant difference on the visual outcome when IOFB removal was delayed for several weeks [3, 18, 21]. Ahmadieh et al. reported a delay in IOFB removal of more then 4 weeks in 85% of their patients because a majority of their patients had war related IOFBs [22]. Lack of suspicion of IOFBs by primary care physicians has led to the delay in the treatment of up to 44% of patients resulting in a delay of up to 3 years in the diagnosis, treatment, or both, of IOFB [18, 21].
In conclusion, the data from this retrospective study reiterate the importance of prompt recognition of retained IOFBs and early repair of related injuries to prevent endophthalmitis and associated complications of visual loss. Final vision appears to be considerably determined by the presenting VA and severity of injury. Eyes with IOFBs in the anterior segment appear to have better prognosis compared to eyes having IOFBs in the posterior segment. Early vitrectomy at the time of IOFB removal may be beneficial for overall visual outcome of the operated eyes. Owing to the retrospective nature of our study and most of the reported studies and the various types of injuries in different settings, results from these studies are difficult to compare. It is hoped that with the advent of modern instruments such as wide-angle viewing systems and high-speed cutters, the chances of complications such as retinal detachment may be greatly reduced. A multi-center prospective study may be required to address some of the confounding factors in the management of IOFB and associated endophthalmitis.
|
[
"endophthalmitis",
"foreign bodies",
"risk factors",
"trauma",
"visual acuity"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Neurosci_Lett-1-5-1885995
|
The Notch-1 intracellular domain is found in sub-nuclear bodies in SH-SY5Y neuroblastomas and in primary cortical neurons
|
Notch signalling affects most aspects of development, not least the determination of neural stem cell fate. Here, we describe the presence of the Notch-1 intracellular domain (N1ICD) in sub-nuclear bodies in SH-SY5Y neuroblastomas and in primary rat cortical neurons as well as several other mammalian cell lines. We also demonstrate that these N1ICD-positive sub-nuclear bodies are distinct from premyelocytic leukaemia (PML) and SC35 bodies. Furthermore, using Notch deletion constructs we determined that a region C-terminal of amino acid 2094 is involved in targeting the N1ICD into sub-nuclear bodies. These findings have ramifications for nuclear architecture and gene transcription.
Notch signalling regulates many aspects of invertebrate and vertebrate development ranging from somite formation [7] to epithelial–mesenchymal interactions [20] and the determination of neural stem cell fate [23]. Notch signalling is also active in adult tissues including the brain where it is involved in synaptic plasticity and memory [3,19]. There are four mammalian Notch receptors (Notch 1–4), which undergo a series of sequential proteolytic cleavages. The first cleavage (S1) occurs in the Golgi by a furin-like convertase generating two fragments that remain non-covalently attached to form the mature Notch receptor at the cell surface. Ligands of the Jagged and Delta families, expressed on neighbouring cells, bind to the Notch receptor leading to the second cleavage (S2) by tumor necrosis factor alpha converting enzyme (TACE), which results in the shedding of the extracellular domain. The second cleavage is followed by a constitutive cleavage (S3) within the transmembrane domain of Notch by γ-secretase. This cleavage releases the Notch intracellular domain (NICD), which migrates to the nucleus [20,14,15] where it interacts with the C-promoter binding factor-1 (CBF-1), the mammalian homologue of Suppressor of hairless (Drosophila) and Lag-1 (C. elegans), also known as RBP-Jκ [16]. In the absence of NICD, CBF-1 forms a complex with repressor proteins, resulting in the suppression of gene transcription. Interaction with NICD displaces the repressor proteins and allows the entry of transcriptional activators such as p300/CBP and Mastermind, which convert the CBF-1 repressor complex into an activator of gene transcription [8]. Notch/CSL target genes include the hairy and enhancer of split (HES) family and the HES related proteins (HERP/Hey) that encode basic helix-loop-helix transcription factors, which suppress differentiation by antagonising the expression of down-stream lineage-specifying genes. In this report, we demonstrate the presence of the N1ICD in sub-nuclear bodies in SH-SY5Y neuroblastomas and in primary rat cortical neurons.
To undertake this study the following cell lines, antibodies and plasmid constructs were used: SH-SY5Y neuroblastoma cells and chinese hamster ovary cells (CHO) were obtained from ECACC (UK) and human embryonic kidney 293a cells (HEK293a; an adherent clone of HEK293 cells) were purchased from Quantum Biotechnologies (Canada). Mouse monoclonal anti-splicing factor SC-35 antibody and rabbit polyclonal anti-FLAG tag antibody were from Sigma (UK). Mouse monoclonal anti-myc antibody was from Cell Signalling Technology (UK). Mouse monoclonal anti-promyelocytic leukaemia protein antibody, goat anti-Notch-1, goat anti-mouse IgG and donkey anti-goat IgG were from Santa Cruz Biotechnology (USA). Alexa Fluor 488 and 594 goat anti-mouse and donkey anti-goat IgG conjugates were from Molecular Probes (UK). C-terminally tagged myc-His human N1ICD, ΔTAD-N1ICD (lacking the transactivation domain TAD; the OPA domain and the proline (P), glutamate (E), serine (S), threonine (T) rich sequence—PEST) and ΔRA-N1ICD (lacking the RAM domain and the Ankyrin repeats) were gifts from Dr. Tom Kadesch (University of Pennsylvania, USA) [13]. Enhanced green fluorescent protein (EGFP) C-terminally tagged human N1ICD (N1ICD-EGFP) was from Dr. Allan Levey (Emory University School of Medicine, USA) [11]. The Notch-CBF-1 reporter, 4xwt-CBF-Luc, which contains four tandem repeats of the consensus CBF-1 DNA binding sequence, GTGGGAA and N-terminally tagged FLAG CBF-1 were generous gifts from Dr. Diane Hayward (Johns Hopkins University School of Medicine, USA) [5].
Cells were maintained in 5% CO2 in air in a humidified incubator at 37 °C. All culture media and supplements were from Invitrogen, UK, unless otherwise stated. SH-SY5Y neuroblastomas were grown in a 50/50 mix of F12/Eagle's minimal essential medium (EMEM) supplemented with 15% (v/v) fetal bovine serum (FBS; Autogen Bioclear, UK), 2 mM l-glutamine, non-essential amino acids (Sigma), 100 IU penicillin and 100 mg/ml streptomycin. HEK293a were cultured in low glucose Dulbecco's modified Eagle's medium (DMEM) containing 10% (v/v) FBS, 2 mM l-glutamine, 100 IU penicillin and 100 mg/ml streptomycin. CHO cells were grown in F12 media supplemented as described above for HEK293a cells. Primary cortical neurons were prepared and cultured in Neurobasal media plus B27 supplement as previously described from E19 rat embryos [17].
Cells on coverslips were transfected with 500 ng of N1ICD-EGFP or 500 ng of myc-His N1ICD using FuGene 6 according to the manufacturer's instructions. Cells were fixed in ice-cold methanol 24 h after transfection then stained according to standard protocols. In brief, cells were incubated with goat anti-Notch-1 antibody (1:100), mouse monoclonal anti-promyelocytic leukaemia protein antibody (1:100) or mouse monoclonal anti-splicing factor SC-35 antibody (1:250) before being incubated with the appropriate fluorescent secondary antibody (1:200). Nuclei were stained with Hoescht 33342. Immunofluorescence was visualized and captured using a Zeiss LSM510 meta confocal microscope. Images were processed using LSM5 image examiner software (Zeiss). All experiments were performed in triplicate. Figures shown are representative images from a single experiment.
For luciferase reporter gene assays SH-SY5Y neuroblastomas were transfected with 400 ng of CBF-Luc (firefly luciferase based reporter DNA) and 500 ng of N1ICD-EGFP or 500 ng of myc-His N1ICD using FuGene 6 according to the manufacturer's instructions (Roche, UK). To control for transfection efficiency 50 ng of phTK-Renilla luciferase (Promega, UK) was also included in transfections. Empty vector DNA was included where necessary to maintain constant DNA concentrations. Twenty-four hours post-transfection the cells were lysed and firefly and Renilla luciferase activities were sequentially measured using Dual-Glo reagents (Promega) in a Wallac Trilux 1450 Luminometer (Perkin-Elmer, UK). Values were normalized by dividing firefly values by the Renilla value from the same well. Data for each set of four replica transfections was averaged, the control in each set normalized to 1 and data presented as fold increases over control. All experiments were performed in triplicate.
Endogenous S3 cleaved, nuclear, NICDs are expressed at very low levels and are difficult to detect in cell culture experiments [9], therefore we expressed exogenous N1ICD-EGFP in SH-SY5Y neuroblastomas to model N1ICD in vitro. In these cells N1ICD-EGFP exhibited a diffuse nuclear staining pattern with intense sub-nuclear bodies (Fig. 1A and B). In contrast, expression of exogenous p53 or β-catenin resulted in diffuse nuclear staining (data not shown), providing evidence that the N1ICD-positive sub-nuclear bodies are specific. The sub-nuclear bodies were not attributable to the EGFP protein since SH-SY5Y cells transfected with the EGFP expression vector containing no insert did not display a speckled nuclear staining pattern (data not shown). Furthermore, a myc-His tagged form of N1ICD was found in sub-nuclear bodies in SH-SY5Y neuroblastoma cells (Fig. 1C) as demonstrated using an anti-Notch-1 antibody, providing additional evidence that the observed speckled staining pattern is specific and not due to the presence of EGFP. Control cultures that were not exposed to primary antibody did not exhibit any positively stained cells (data not shown). To ensure the presence of nuclear bodies was not an artefact of the fixation method used (ice-cold methanol), SH-SY5Y cells were transfected with both EGFP and myc-His tagged Notch constructs and fixed with either 3.7% formaldehyde or 4% paraformaldehyde. With both methods very similar nuclear bodies were seen with both constructs (data not shown). Of note, both N1ICD-EGFP and myc-His tagged N1ICD were functionally active as demonstrated using CBF-Luc (Fig. 1D) and we have previously shown that exogenous N1ICD expression promotes the inhibition of retinoic acid-induced differentiation in SH-SY5Y neuroblastomas [4].
N1ICD-positive sub-nuclear bodies were not cell-type specific. We observed identical sub-nuclear structures in primary rat cortical neurones (Fig. 2A), HEK293a cells (Fig. 2B) and in CHO cells (Fig. 2C) with N1ICD-EGFP and myc-His tagged N1ICD (data not shown). As the punctate structures were detected with both constructs in a variety of cell lines and in primary neurons they are unlikely to be an artefact of tagging or a result of cellular transformation. Furthermore, we found that exogenous CBF-1 exhibited a diffuse nuclear staining pattern in SH-SY5Y neuroblastomas (Fig. 2D), but in the presence of N1ICD CBF-1 re-located into sub-nuclear bodies (Fig. 2E), indicative of functional gene transcription within these nuclear domains.
Several sub-nuclear structures, similar to those shown here, have been previously described including premyelocytic leukaemia (PML) bodies and SC35 bodies. These foci are believed to be areas of high transcriptional activity [2] and pre-mRNA splicing, respectively [10]. We therefore transfected SH-SY5Y cells with N1ICD-EGFP and used antibodies to detect endogenous PML and SC35 bodies. Both PML (Fig. 3A, mid left panel) and SC35 reactive bodies (Fig. 3B, mid left panel) were detected in SH-SY5Y cells. However, in both cases these formations were distinct from the N1ICD-EGFP positive foci (Fig. 3A and B, far right panels). The PML and SC35 bodies were more consistent in size and typically smaller and more numerous than the formations in which N1ICD was detected.
Next we examined which region of N1ICD was involved in targeting the protein to sub-nuclear bodies by using two myc-tagged N1ICD deletion constructs, ΔTAD-N1ICD and ΔRA-N1ICD. ΔTAD-N1ICD runs from a.a. 1760 to 2093 and lacks the transactivation domain, the OPA domain and the proline (P), glutamate (E), serine (S), threonine (T) rich sequence (PEST). ΔRA-N1ICD extends from a.a. 2094 to 2556 and lacks the RAM domain and the Ankyrin repeats (Fig. 4A). ΔTAD-N1ICD was detected diffusely throughout the nucleus in SH-SY5Y neuroblastomas (Fig. 4B, left panel), while ΔRA-N1ICD was located in sub-nuclear bodies very similar in appearance to those seen with N1ICD (Fig. 4B, right panel). This indicates that a region C-terminal of a.a. 2094, lying within the TAD, the OPA domain or the PEST domain or an unidentified motif within the intervening regions, is involved in targeting N1ICD into sub-nuclear bodies.
In summary, we demonstrate that exogenous N1ICD is expressed in sub-nuclear bodies in SH-SY5Y neuroblastomas, primary rat cortical neurons as well as other mammalian cell lines and these sub-nuclear structures were distinct from SC35 or PML bodies. N1ICD foci were detectable as early as 5 h after transfection (cells were routinely fixed 24 h post-transfection in all other experiments described in this report), which suggests that the nuclear inclusions are not merely a consequence of protein aggregation with time (data not shown). The presence of the N1ICD in sub-nuclear bodies has previously been demonstrated when associated with the Epstein-Barr virus derived protein, RPMS [24] and the Mastermind family of Notch/CBF-1 co-activators [18,21,22]. Here, we report that the N1ICD appeared in foci independently of other exogenously expressed proteins. A number of endogenous Notch associated proteins, including CBF-1, have recently been observed in similar nuclear foci in primary hippocampal neurons [12], although the presence of NICDs was not examined. Interestingly, we found that exogenously expressed CBF-1 exhibited a diffuse nuclear staining pattern, however in the presence of exogenous N1ICD CBF-1 was found in sub-nuclear bodies. A number of previous reports [1,6] concerning Notch cytochemistry contain figures in which N1ICD appears to be in punctate formations, which with greater resolution could have been interpreted as distinct sub-nuclear bodies. It is feasible then that the N1ICD-positive bodies form part of a physiologically relevant, multimeric, protein complex involved in gene transcription. Moreover, the presence of such sub-nuclear bodies provides evidence that the nucleus contains defined domains and is not merely a homogenous pool of nucleic acids and proteins.
|
[
"notch",
"sub-nuclear bodies",
"transcription",
"cbf-1",
"luciferase"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Purinergic_Signal-1-2-2096532
|
Guanosine stimulates neurite outgrowth in PC12 cells via activation of heme oxygenase and cyclic GMP
|
Undifferentiated rat pheochromocytoma (PC12) cells extend neurites when cultured in the presence of nerve growth factor (NGF). Extracellular guanosine synergistically enhances NGF-dependent neurite outgrowth. We investigated the mechanism by which guanosine enhances NGF-dependent neurite outgrowth. Guanosine administration to PC12 cells significantly increased guanosine 3-5-cyclic monophosphate (cGMP) within the first 24 h whereas addition of soluble guanylate cyclase (sGC) inhibitors abolished guanosine-induced enhancement of NGF-dependent neurite outgrowth. sGC may be activated either by nitric oxide (NO) or by carbon monoxide (CO). -Nitro-l-arginine methyl ester (l-NAME), a non-isozyme selective inhibitor of nitric oxide synthase (NOS), had no effect on neurite outgrowth induced by guanosine. Neither nNOS (the constitutive isoform), nor iNOS (the inducible isoform) were expressed in undifferentiated PC12 cells, or under these treatment conditions. These data imply that NO does not mediate the neuritogenic effect of guanosine. Zinc protoporphyrin-IX, an inhibitor of heme oxygenase (HO), reduced guanosine-dependent neurite outgrowth but did not attenuate the effect of NGF. The addition of guanosine plus NGF significantly increased the expression of HO-1, the inducible isozyme of HO, after 12 h. These data demonstrate that guanosine enhances NGF-dependent neurite outgrowth by first activating the constitutive isozyme HO-2, and then by inducing the expression of HO-1, the enzymes responsible for CO synthesis, thus stimulating sGC and increasing intracellular cGMP.
Introduction
The roles of extracellular purine nucleosides and nucleotides as neurotransmitters and modulators are well documented [1]. Extracellular purines also exert trophic effects on cells; influencing growth, division [2, 3] differentiation, and even apoptosis [4–6]. Purines are released from cells under physiological conditions, acting as neurotransmitters and neuromodulators [7–9]. Trauma and other insults to the central nervous system are also potent stimuli that cause the release of purines [3], most of which are converted by ectoenzymes to adenosine and guanosine [3, 5]. After insults to cells, more guanine-based than adenine-based purines are released [10]. Furthermore, the extracellular concentration of guanosine remains elevated for prolonged periods in vitro [10] and for up to a week after central nervous system (CNS) injury [11], implying that extracellular guanosine may exert trophic effects in vivo.
Sprouting of neurites, which later become axons and dendrites, is an important change associated with neural development and differentiation [12, 13]. Neurite sprouting through regeneration or collateral sprouting also plays an important role in the functional recovery following injury to the central or peripheral nervous systems [14]. Amongst its trophic effects, guanosine stimulates neurite outgrowth and enhances NGF-dependent neurite outgrowth [3, 15].
PC12 cells serve as a useful model for studying cell signaling [16]. They respond to many growth factors, neurotrophins and hormones, which initiate multiple signaling pathways [17–20]. The specific cellular targets of these signaling pathways mediate the distinct responses of differentiation, proliferation and survival, all of which can be assessed [16]. The addition of NGF to PC12 cells causes a sustained activation of ERK, a mitogen-activated protein kinase, through activation of the TrkA receptor [16, 19]. This leads to the development of many phenotypic characteristics in PC12 cells, which are also associated with mature sympathetic neurons [13, 21].
Extracellular guanosine not only stimulates neurite outgrowth from primary cultures of rat hippocampal neurons [3], and PC12 cells [15] but also enhances the neuritogenic effects of NGF on PC12 cells [15, 22, 23]. Both in astrocytes [24] and in PC12 cells [22], guanosine increases intracellular adenosine 3′,5′-cyclic monophosphate (cAMP). Some of the neuritogenic effects of guanosine in PC12 cells appeared to be mediated through this mechanism [22]. However, guanosine also has cAMP-independent effects, synergistic with NGF, because its effects are not abolished by an adenylate cyclase inhibitor [22] and are synergistic with substances that increase intracellular cAMP [15].
The cAMP-independent component of the signal transduction mechanism through which guanosine synergistically enhances NGF-dependent neurite outgrowth is unknown. However, several lines of evidence indicate a role for guanosine 3′,5′-cyclic monophosphate (cGMP). First, extracellular guanosine increases intracellular cGMP in rat mesenteric artery [25]. Second, nitric oxide (NO) donors enhance NGF-dependent PC12 cell neurite outgrowth via a cGMP-dependent mechanism; and thirdly, the neuritogenic effects of the NO donors are abolished by guanylyl cyclase inhibitors and mimicked by cGMP analogs [26]. The mechanism by which cGMP enhances neurite outgrowth is currently unclear.
Nitric oxide synthases are a family of enzymes, which synthesize NO through the catalytic conversion of l-arginine to l-citrulline. The constitutively expressed forms are the endothelial (eNOS) or neuronal (nNOS), which are regulated by the cytosolic concentration of Ca2+ [27]. The inducible isoform (iNOS) is widely distributed, and becomes active only hours after an inducing event [28]. NO is involved in cell communication and signal transduction in many systems including the CNS [29–31]. NGF induced different isoforms of NOS in PC12 cells after 4 days in vitro [32, 33]. Peunova and Enikolopov [32] reported that the NO produced preceding the development of the differentiated phenotype is due predominantly to iNOS.
The diffusible gas carbon monoxide (CO) is a putative neurotransmitter [34]. Heme oxygenase (HO) synthesizes CO from the biologically active substrate biliverdin, which is rapidly reduced to bilirubin, and iron from intracellular heme [34]. CO can modulate activities attributed to cGMP in the nervous system [35–38]. It has been proposed that CO production is responsible for baseline cGMP levels in the hippocampus [36]. In cerebellar granule cell cultures, CO produced by HO affects intracellular cGMP concentrations by modulating the NO-soluble guanylate cyclase signaling system [39]. Conversely, NO, synthesized by NOS may induce the expression of HO-1 [40–42], indicating a close and reciprocal interaction between the NOS-NO and the HO-CO signaling systems [35, 36]. Based on these findings it has been proposed that a possible role for HO-1 is to counteract NO toxicity [43].
Together, these data led us to question whether guanosine enhanced NGF-dependent neurite outgrowth is through a mechanism involving cGMP, and if so, whether it was attributable to the stimulation of either NO or CO synthesis.
Materials and methods
Cell culture and treatments
Tissue culture supplies were from Life Technologies. All other supplies were obtained from Sigma RBI unless otherwise stated. 2.5S NGF was a generous gift from Dr. M. Coughlin, Department of Medicine, McMaster University, 6-(phenylamino)-5,8-quinolinedione (LY83583) was obtained from (Calbiochem), copper protoporphyrin from (Porphyrin Products), and zinc protoporphyrin IX from (Research Biochemical).
PC12 cells were maintained in either RPMI 1640 medium supplemented with 5% heat-inactivated (HI) fetal calf serum (FCS), 5% HI-horse serum (HS) and 1% antibiotic-antimycotic (Anti-Anti; 10,000 units of penicillin, 10,000 µg of streptomycin, 25 µg amphotericin B/ml in 0.85% saline) [15] or F-12K (Kaighn's Modification) medium supplemented with 15% HI-HS, 2.5% HI-FCS, and 1% Anti-Anti at 37 °C in a 5% CO2 environment.
Neurite outgrowth assay
To evaluate the effect of test compounds on neurite outgrowth in PC12 cells, they were added to cultures for 48 h as previously described [15]. Briefly, PC12 cells were plated onto poly-d,l-ornithine (PORN)-coated 24-well plates at a density of 2.5 × 104 cells/well. Cells were cultured in RPMI 1640 supplemented with 1.5% HI-HS, 1.5% HI-FCS and 1% Anti-Anti. Guanosine was dissolved in 10% sodium hydroxide (1N NaOH) and when added to the culture medium the final concentration was 0.01% sodium hydroxide. In experiments in which guanosine (300 µM) was added to each well first, followed within 15 min by the addition of 2.5S NGF (40 ng/ml). Methylene blue, LY83583, hemoglobin, Nω-nitro-l-arginine methyl ester (l-NAME), copper protoporphyrin-IX (CuPP) and zinc protoporphyrin-IX (ZnPP), were added 15 min prior to the addition of guanosine. All test substances remained in the medium throughout the 48-h test period. Cells were maintained at 37 °C in a humidified atmosphere of 95% air, 5% CO2. After 48 h, cells were fixed with 10% formalin in phosphate buffered saline (PBS), pH 7.4 for 10 min and stored at 4 °C in PBS (pH 7.4) containing sodium azide (0.1%, w/v) until counted (usually within 2 weeks). Total cell number was determined by counting two randomly selected areas in each well. The observer was blind to the treatments the cells had received. At least 50 cells per well were counted, and the number of cells bearing one or more neurites determined using a NIKON Diaphot microscope equipped with phase contrast optics. Neurites were defined as processes extending at least one cell body diameter from the cell with growth cones at their tips [44].
Western immunoblot analysis
The expression of HO-1, HO-2, iNOS, nNOS, and β-actin in PC12 cells was determined by Western immunoblot analysis. PC12 cells were seeded onto PORN-coated plates at a density of 1 × 106 cells/plate and maintained at 37° for 48 h prior to treatments. Cells were placed in a medium containing 3% HI-FCS and 3% HI-HS 12 h before the start of treatments. Cells were then treated with various agents for 0, 6, 12, 24, or 48 h. At the end of the treatment period cells were washed once with PBS and then harvested at 4 °C using a lysis buffer (25 µM Tris/HCl pH 7.4, 10 µM NaCl, 10 µM EDTA, 100 µl/10 ml Tween 20, 10 µM sodium pyrophosphate decahydrate, 10 mM sodium orthovanadate, 5 µg/ml leupeptin, 10 mM glycerophosphate). Cells were disrupted by sonication and aliquots (25 µl) were removed for the determination of protein concentrations using the bicinchoninic acid (BCA) protein assay (Pierce, Illinois, USA). Cells lysates were separated on 12% SDS-polyacrylamide gels and electrophoretically transferred to nitrocellulose membranes (PALL, Michigan, USA). Membranes were incubated with a specific primary antibody overnight at 4 °C then were exposed to a secondary antibody for 1 h at room temperature. Positive control peptides for nNOS and iNOS (obtained from Cayman Chemical, Ann Arbor. MI) were used as electrophoresis standard. The following antibodies were used: monoclonal antibody to brain NOS (Sigma, dilution: 1:1,000), monoclonal antibody to the inducible NOS (Sigma, dilution: 1:1,000), anti-iNOS/bNOS, goat anti-mouse IgG-HRP (StressGen Biotechnologies, dilution 1:100,000), monoclonal antibody to β-actin (abCAM, dilution: 1:20,000), monoclonal anti-β-actin, goat anti-mouse IgG-HRP (Novus Biological, dilution 1:100,000), monoclonal antibody to HO-1 (StressGen Biotechnologies, dilution 1:2,000), anti-HO-1, goat anti-mouse IgG-HRP (StressGen Biotechnologies, dilution 1:200,000), rabbit polyclonal antibody to HO-2/HO-1 (StressGen Biotechnologies, dilution 1:1,000), anti HO-2/HO-1, goat anti-rabbit IgG-HRP (StressGen Biotechnologies, dilution 1:200,000). Immunocomplexes were then visualized using a chemiluminescence substrate (Sigma-Aldrich). Immunoblots were quantified by densitometric analyses, using the Northern Eclipse program (EPIX). All bands on the immunoblots were normalized to their corresponding β-actin bands prior to statistical analysis.
Determination of cyclic GMP
PC12 cells were seeded onto PORN-coated 12-well plates at a density of 5 × 105 cells/well and maintained at 37 °C for 48 h prior to treatments. Cells were placed in a medium containing 3% HI-FCS and 3% HI-HS 12 h before the start of treatments. Cells were then treated with various agents for different times (0, 6, 12, 24, or 48 h). At the end of the treatment period cGMP was extracted and analyzed using a cGMP enzyme immunoassay (EIA) kit (Amersham Biosciences). The assay is based on a competition between unlabeled cGMP in the sample or (standard) and a fixed quantity of peroxidase-labeled cGMP for a limited number of binding sites on a cGMP specific antibody. With fixed amounts of antibody and cGMP-peroxidase conjugate, the amount of bound cGMP-peroxidase conjugate is inversely proportional to the concentration of unlabeled cGMP.
Statistical analysis
Statistical analysis was carried out using a two-way ANCOVA, when applicable, or a two-way ANOVA followed by Fischer's LSD test for multiple comparisons.
Results
Guanosine enhances NGF-dependent neurite outgrowth via activation of soluble guanylate cyclase
As we have reported previously, we have evaluated the effect of guanosine on NGF-dependent neurite outgrowth during the first 48 h, as guanosine most likely exerted its effects within the first 1–2 h after treatment [23]. Therefore, all experiments reported in this manuscript were performed during the first 48 h of PC12 cell treatment with various agents.
PC12 cells contain both soluble, and particulate isoforms of guanylate cyclase (GC) [26, 45]. To determine whether activation of GC was necessary for the neuritogenic effect of guanosine, inhibitors of GC were added for 48 h to PC12 cells treated with NGF plus guanosine. Methylene blue at concentrations ranging from 0.1 to 1 mM was added to PC12 cell cultures that contained NGF alone (40 ng/ml), or guanosine (300 mM) plus NGF (40 ng/ml). Cultures were then evaluated for neurite outgrowth. At the concentrations used in these experiments, methylene blue inhibits sGC but not particulate GC [46, 47]. Methylene blue had no effect on the outgrowth of neurites induced by NGF alone (Figure 1). However, in cultures treated with both guanosine and NGF, the addition of methylene blue reduced significantly (P < 0.01) the proportion of neurite-bearing cells in a concentration-dependent manner (Figure 1). LY83583 inhibits both, particulate GC and sGC [48]. When LY83583 (10 nM) was added to PC12 cell cultures, it inhibited neurite outgrowth elicited by guanosine plus NGF but had no effect on neurite outgrowth in cultures treated with NGF alone (data not shown). These data support the hypothesis that cGMP plays a role in enhancing the effect of guanosine on NGF-mediated neurite outgrowth. Moreover, since methylene blue at the concentrations used inhibits soluble but not particulate GC, these data imply that guanosine activates sGC.
Figure 1
Methylene blue attenuates guanosine-enhanced nerve growth factor-dependent neurite outgrowth in PC12 cells. PC12 cells were cultured in RPMI 1640 medium supplemented with 5% heat-inactivated fetal calf serum, 5% heat-inactivated horse serum and 1% antibiotic-antimycotic for 48 h. Cultures were then treated with NGF (40 ng/ml) or NGF (40 ng/ml) plus guanosine (300 µM) and with increasing concentrations of methylene blue (0–1 µM). After 48 h, the total cell number and number of cells bearing one or more neurites were determined by counting two random areas in each well. The mean proportion of neurite-bearing cells in cultures treated with NGF was approximately 25%–35%. Because this value varied slightly between experiments, all experimental values are expressed relative to the NGF treated cultures, which was defined as 100%. Open bars: NGF treatment, closed bars: NGF plus guanosine treatment. Cultures treated with guanosine plus NGF had a significantly (P < 0.01, two-way independent ANOVA) greater proportion of neurite-bearing cells than those treated with NGF alone. Methylene blue (0.1 to 1.0 µM) had no effect on the proportion of neurite-bearing cells in cultures treated with NGF alone, but at concentrations from 0.1 to 1.0 µM, it significantly (**P < 0.01, two-way independent ANOVA) reduced the proportion of neurite-bearing cells in cultures treated with guanosine plus NGF. Data represent the mean ± SEM of 12 determinations from two replicate experiments.
Guanosine may activate sGC either directly, or indirectly by stimulating the formation of NO, CO or hydroxyl radicals, which are all physiological activators of sGC [49]. To distinguish between these possibilities we added hemoglobin (100 nM) to cultures of PC12 cells. Extracellular hemoglobin directly scavenges NO and CO because these substances avidly bind to the heme moiety, thereby inhibiting their GC activating activity [46]. Hemoglobin reduced the synergistic effects of guanosine on NGF-mediated neurite outgrowth but had no effect on neurite outgrowth elicited by NGF alone (data not shown). Although these data imply that guanosine stimulates the synthesis of a sGC-activating factor, they do not show whether this factor is NO, CO or hydroxyl radical.
Inhibition of nitric oxide synthase (NOS) has no effect on guanosine-enhanced NGF-dependent neurite outgrowth
In the rat mesenteric artery, guanosine stimulates NO formation [25]. Therefore, we investigated if guanosine could also stimulate NO synthesis in PC12 cells. To examine whether NO was involved in the signal transduction pathway that mediated the enhancement of neurite outgrowth by guanosine through sGC, we inhibited the enzyme NOS, which catalyzes the conversion of l-arginine to l-citrulline plus NO. The l-arginine analog, Nω-nitro-l-arginine methyl ester hydrochloride (l-NAME) [50] inhibits NOS competitively [25, 51]. In this experiment, we reduced the concentration of l-arginine in the culture medium from 1 mM to 80 µM, thus permitting more effective competition of the NOS inhibitor with l-arginine. Under these conditions, L-NAME inhibits NOS at 10 µM [25, 51]. When added to the culture medium of PC12 cells at 10 µM l-NAME did not inhibit NGF-dependent neurite outgrowth, and had no effect on the guanosine-mediated enhancement of NGF-dependent neurite outgrowth (Figure 2a). To confirm that NO was not involved in the signal transduction pathway that mediates the enhancement of neurite outgrowth by guanosine, we determined the expression of two subtypes of NOS isoenzymes: iNOS, the inducible isoform (Figure 2b), and nNOS, the constitutive isoform (Figure 2c). PC12 cultures were exposed to guanosine alone (300 µM), or NGF alone (40 ng/ml), or to the combination of guanosine plus NGF for 48 h, and NOS expression was determined by Western immunoblot analysis. We analyzed various protein concentrations (5–50 µg/ml) and determined β-actin expression on the same Western immunoblots. Using protein concentrations as high as 50 µg/ml neither nNOS nor iNOS expression could be detected in PC12 cells treated with guanosine, or NGF, or the combination of the two agents. These data indicate that since NOS is not expressed during the first 48 h of treatment, thus NO cannot mediate the effect of guanosine on neurite outgrowth.
Figure 2
(a) Inhibition of nitric oxide synthase has no effect on the proportion of neurite-bearing PC12 cells cultured for 48 h in the presence of NGF, or NGF plus guanosine. PC12 cells were cultured with NGF (40 ng/ml) or NGF (40 ng/ml) plus guanosine (300 µM) as described in Figure 1. Cultures were treated with the general nitric oxide synthase inhibitor, l-NAME (0.1–10 mM) for 48 h and the number of cells bearing one or more neurites were determined as described in Figure 1. Open bars: NGF treatment, closed bars: NGF plus guanosine treatment. Neurite outgrowth in cultures treated with guanosine plus NGF was significantly (P < 0.01, two-way independent ANOVA) greater than in cultures treated with NGF alone. l-NAME (0.1–10 µM) had no effect on the proportion of neurite bearing cells in cultures treated with NGF alone, or in combination with guanosine. Data represent the mean ± SEM of 12 determinations from two replicate experiments. (b) Inducible nitric oxide synthase is not expressed in PC12 cells cultured for 48 h in the presence of guanosine, or NGF, or guanosine plus NGF. PC12 cells were cultured on plates coated with poly-d,l-ornithine for 72 h. Cells were then grown in serum-reduced medium (3% heat-inactivated fetal calf serum and 3% heat-inactivated horse serum) for 12 h, and were treated with guanosine (G, 300 µM) or NGF (N, 40 ng/ml) or guanosine (300 µM) plus NGF (40 ng/ml) (G + N), or with no added treatments (C). The expression of inducible nitric oxide synthase was determined at various time points (6, 12, 24 and 48 h) by Western immunoblot analysis. Recombinant inducible nitric oxide synthase protein (50 ng) was used as a positive control (P). Immunoblots were quantified by densitometric analysis and were normalized to the corresponding β-actin bands as described in the Materials and methods. Statistical analysis was performed using a two-way ANCOVA followed by Fischer's LSD post-hoc comparison test. Data presented are representative of at least three independent experiment. (c) Neuronal nitric oxide synthase is not expressed in PC12 cells cultured for 48 h in the presence of guanosine, or NGF, or guanosine plus NGF. PC12 cells were cultured as described in panel (b), and the expression of neuronal nitric oxide synthase was determined at various time points (6, 12, 24 and 48 h) by Western immunoblot analysis. Recombinant neuronal nitric oxide synthase protein (50 ng) was used as a positive control (P). Immunoblots were quantified by densitometric analysis and were normalized to the corresponding β-actin bands as described in the Materials and methods. Statistical analysis was performed using a two-way ANCOVA followed by Fischer's LSD post-hoc comparison test. Data presented are representative of at least three independent experiment.
Inhibition of heme oxygenase (HO) attenuates guanosine-enhanced NGF-dependent neurite outgrowth
Since NO apparently did not mediate the effects of guanosine, we examined the possibility that guanosine stimulated CO synthesis. Heme oxygenase (HO), the enzyme which synthesizes CO, has been detected in rat adrenal glands [52]. Zinc protoporphyrin-IX (ZnPP) inhibits both the constitutive HO (HO-2) and the inducible HO (HO-1) isoforms of this enzyme [52, 53]. We, therefore, tested the effect of increasing concentrations of ZnPP on the ability of NGF or, NGF plus guanosine to induce neurite outgrowth. Addition of ZnPP (0.01, 0.1, 1 µM) to PC12 cell cultures did not significantly reduce outgrowth of neurites induced by NGF (Figure 3), when compared to control cultures. In contrast, ZnPP attenuated the ability of guanosine to enhance NGF-stimulated neurite outgrowth (Figure 3). As a control for non-specific effects of metalloporphyrins we used copper protoporphyrin-IX (CuPP), which does not inhibit heme oxygenase [54]. The addition of CuPP to PC12 cells did not inhibit either NGF-dependent neurite outgrowth or the ability of guanosine to enhance NGF-dependent neurite outgrowth (data not shown).
Figure 3
Inhibition of heme oxygenase attenuates guanosine-enhanced NGF-dependent neurite outgrowth in PC12 cells. PC12 cells were cultured with NGF (40 ng/ml) or NGF (40 ng/ml) plus guanosine (300 µM) as described in Figure 1. Some cultures were treated with the selective inhibitor of heme oxygenase zinc protoporphyrin-IX (0.01Y1 µM) for 48 h and the number of cells bearing one or more neurites was determined as described in Figure 1. Open bars: NGF treatment, closed bars: NGF plus guanosine treatment. Cultures treated with guanosine plus NGF had a significantly (P < 0.01 two-way independent ANOVA) greater proportion of neurite-bearing cells than cultures treated with NGF alone. Zinc protoporphyrin-IX significantly decreased (**P < 0.01) the neurite growth from cultures treated with guanosine plus NGF, but had no significant effect on neurite outgrowth in cultures treated with NGF alone. Data represent the mean ± SEM of 12 determinations from two replicate experiments.
Guanosine induces heme oxygenase-1 (HO-1) expression
PC12 cells have been previously shown to express both HO-1 and HO-2 proteins under basal conditions [55]. Since the addition of ZnPP to PC12 cultures decreased NGF-dependent neurite outgrowth, we attempted to establish whether guanosine treatment had any effect on the expression of the HO-1 protein during the first 48 h of this process. We treated PC12 cultures with guanosine alone (300 µM), or with NGF alone (40 ng/ml), or with a combination of guanosine plus NGF for 48 h, and determined HO-1 expression by Western immunoblot analysis. We used the monoclonal antibody directed against rat HO-1, which recognizes a single protein band, at 32 kDa, the reported molecular weight for this enzyme [36] (Figure 4b). Western immunoblots were quantified and normalized against their corresponding β-actin band. Statistical analysis revealed no differences in β-actin expression for any of the treatments at the different time points (data not shown); therefore, this was used as a control for normalizing HO-1 protein expression. There was no significant change in HO-1 expression in untreated PC12 cells during the 48-h period. In cells exposed to guanosine alone HO-1 expression, unexpectedly, was significantly reduced at 6 h compared to untreated cells at 0 time (P < 0.05), and at 6 h (P < 0.01) (Figure 4). A possible explanation for this result may stem from the observation that HO-1 gene expression is suppressed by elevated intracellular calcium concentrations [56] and guanosine has been shown to increase intracellular calcium concentrations in astrocytes [57]. Since increases in calcium are rapid and transient, this may account for the reduced HO-1 expression at 6 h. The addition of NGF, however, is sufficient to override this effect and so at later time points HO-1 expression is up-regulated by the combination of NGF and guanosine. The addition of guanosine plus NGF significantly increased HO-1 expression at 12 h (P < 0.05), compared to that of control cells at 0 time (Figure 4). After 24 h all three treatments (guanosine alone, or NGF alone or guanosine plus NGF) led to a significant increase in HO-1 expression compared to control cells at 0 time (P < 0.01) (Figure 4). In cells exposed to guanosine alone the enhanced HO-1 expression at 24 h was also significantly different from that detected in untreated cells at 24 h (P < 0.05). Whereas in cells treated with NGF alone, or with the combination of NGF plus guanosine this difference was significant at P < 0.01 (Figure 4). After 48 h, HO-1 expression was elevated significantly only in cells exposed to NGF alone when compared to untreated cells at 0 time, and at 48 h (P < 0.05) for both comparisons. In cells exposed to guanosine alone or to NGF plus guanosine HO-1 expression by 48 h declined and was similar to untreated control values at 0 time and at 48 h.
Figure 4
Guanosine induces the expression of heme oxygenase-1 protein in PC12 cells. PC12 cells were cultured on plates coated with poly-d,l-ornithine for 72 h. Cells were then grown in serum-reduced medium (3% heat-inactivated fetal calf serum and 3% heat-inactivated horse serum) for 12 h, and were treated with guanosine (G, 300 µM) or NGF (N, 40 ng/ml) or guanosine (300 µM) plus NGF (40 ng/ml) (G + N), or with no added treatments (C) as described in Figure 2b. The expression of heme oxygenase-1 was determined at various time points (6, 12, 24 and 48 h) by Western immunoblot analysis. Immunoblots were quantified by densitometric analysis and were normalized to the corresponding β-actin bands as described in the Materials and methods. Open bars: untreated controls (C), closed bars: guanosine plus NGF (G + N) treatment, stippled bars: NGF (N) treatment, hatched bars: guanosine (G) treatment. Statistical analysis was performed using a two-way ANCOVA followed by Fischer's LSD post-hoc comparison test (○ P < 0.05 compared with control time point 0); (○○ P < 0.01 compared with time point 0); (* P < 0.05 compared with control at same time point); (** P < 0.01 compared with control at same time point). (a) Data represent the mean optical density ± SEM obtained in six independent experiments. (b) Results are representative Western immunoblots obtained in these experiments.
In parallel experiments, we determined whether guanosine had an effect on the expression of the constitutive isoform, HO-2 in these cells. We treated PC12 cultures with guanosine alone (300 µM), or NGF alone (40 ng/ml), or in combination guanosine plus NGF for 48 h and determined HO-2 expression by Western immunoblot analysis. Using the monoclonal antibody directed against rat HO-2 we detected a single protein band, of molecular weight 36 kDa, as described for HO-1 [36] (Figure 5b). Western immunoblots were quantified and normalized against their corresponding β-actin band. Again, statistical analysis revealed no differences in β-actin expression for any of the treatments at the different time points (data not shown), therefore, this was used as a control for normalizing HO-2 protein bands. HO-2 expression was detectable at 0 time in untreated PC12 cells, but none of the treatments had any significant effect on its expression at any time point determined. In the guanosine-treated cells HO-2 expression was elevated slightly at 12 h, but this was not statistically significant.
Figure 5
Guanosine has no effect on heme-oxygenase-2 protein expression in PC12 cells. PC12 cells were cultured on plates coated with poly-d,l-ornithine for 72 h. Cells were then grown in serum-reduced medium (3% heat-inactivated fetal calf serum and 3% heat-inactivated horse serum) for 12 h, and were treated with guanosine (G, 300 µM) or NGF (N, 40 ng/ml) or guanosine (300 µM) plus NGF (40 ng/ml) (G + N), or with no added treatments (C) as described in Figure 2b. The expression of heme oxygenase-2 was determined at various time points (6, 12, 24 and 48 h) by Western immunoblot analysis. Immunoblots were quantified by densitometric analysis and were normalized to the corresponding β-actin bands as described in the Materials and methods. Open bars: untreated controls (C), closed bars: guanosine plus NGF (G + N) treatment, stippled bars: NGF (N) treatment, hatched bars: guanosine (G) treatment. Statistical analysis was performed using a two-way ANOVA. (a) Data represent the mean optical density T SEM obtained in three independent experiments. (b) Results are representative Western immunoblots obtained in these experiments.
Guanosine increases intracellular cGMP concentrations in PC12 cells during guanosine-enhanced NGF-dependent neurite outgrowth
CO is a known activator of sGC [36]. Since we have shown that HO-1 expression is elevated in PC12 cells exposed to guanosine, we next investigated whether this increased HO-1 expression is accompanied by elevation in intracellular cGMP concentrations. PC12 cell cultures were treated with guanosine alone (300 µM), or NGF alone (40 ng/ml), or in combination guanosine plus NGF for 48 h. Intracellular cGMP concentrations were determined after 6, 12, 24 and 48 h of treatment. In cells exposed to guanosine alone, cGMP concentrations increased significantly after 6 h compared to untreated cells at this time point (P < 0.05). In PC12 cells treated with guanosine plus NGF, cGMP concentrations increased significantly after 6 h compared to untreated cells at 0 time (P < 0.05), remained significantly elevated at 12 h (P < 0.05) and at 24 h (P < 0.05), and declined to values comparable to control by 48 h.
Inhibition of heme oxygenase (HO) attenuates intracellular cGMP concentrations in PC12 cells during guanosine enhanced NGF-dependent neurite outgrowth
Previous experiments have demonstrated that the HO inhibitor, ZnPP, blocked guanosine enhanced NGF-dependent neurite outgrowth of PC12 cells (Figure 3). We have also shown that HO-1 expression and cGMP concentrations were elevated during the first 12–24 h of this neurite outgrowth process (Figures 4 and 6, respectively). We therefore examined whether pre-treatment of cells with the HO inhibitor ZnPP had any effect on cGMP concentrations. Although metalloporphrins, such as ZnPP-IX have been used widely to demonstrate the role for this enzyme in numerous physiological situations, their selectivity for HO has been questioned; as they may also inhibit haemoproteins such sGC and NOS [58]. However, metalloporphrins used at a concentration below 10 µM are more selective for HO [59]. Since we found that ZnPP, at 100 nM concentration significantly attenuated guanosine-enhanced neurite outgrowth (Figure 3), we used this concentration of ZnPP to test its effect on intracellular cGMP concentration. PC12 cell cultures were exposed to guanosine alone (300 µM), or NGF alone (40 ng/ml), or in combination guanosine plus NGF for 12 h. ZnPP (100 nM) was added to some cultures prior to the addition of guanosine, or NGF, or NGF plus guanosine, and intracellular cGMP concentrations were determined. In cells exposed to guanosine alone, or to guanosine plus NGF, cGMP concentrations were significantly elevated compared to untreated cells (P < 0.01) and to cells treated with NGF alone (P < 0.01) (Figure 7). In cultures pretreated with ZnPP, cGMP concentrations were significantly reduced in cells exposed to guanosine alone (P < 0.05), or guanosine plus NGF (P < 0.05) compared to the corresponding treatments without the inhibitor. Treatment of PC12 cells with NGF did not increase intracellular cGMP concentrations, and the addition of ZnPP had no effect on cGMP concentrations.
Figure 6
Guanosine increases cGMP concentrations during guanosine-enhanced NGF-dependent neurite outgrowth in PC12 cells. PC12 cells were cultured on plates coated with poly-d,l-ornithine for 72 h. Cells were then grown in serum-reduced medium (3% heat-inactivated fetal calf serum and 3% heat-inactivated horse serum) for 12 h, and were treated with guanosine (G, 300 µM) or NGF (N, 40 ng/ml) or guanosine (300 µM) plus NGF (40 ng/ml) (G + N), or with no added treatments (C) as described in Figure 2b. Cells were lysed at time points 0, 6, 12, 24, and 48 h and cGMP concentrations were determined by a competitive enzyme immunoassay. Open bars: untreated controls (C), closed bars: guanosine plus NGF (G + N) treatment, stippled bars: NGF (N) treatment, hatched bars: guanosine (G) treatment. Statistical analysis was performed using a two-way ANCOVA followed by Fischer's LSD post-hoc comparison test (○ P < 0.05, compared to time point 0); (* P < 0.05, relative to control); (** P < 0.01, relative to control); (Δ P < 0.05, relative to NGF). Data presented represent the mean relative optical density T SEM obtained in six independent experiments.
Figure 7
Inhibition of heme oxygenase attenuates cGMP concentrations during guanosine enhanced NGF-dependent neurite outgrowth in PC12 cells. PC12 cells were cultured on plates coated with poly-d,l-ornithine for 72 h. Cells were then grown in serum-reduced medium (3% heat-inactivated) fetal calf serum and 3% heat-inactivated horse serum) for 12 h, and were treated with guanosine (G, 300 µM) or NGF (N, 40 ng/ml) or guanosine (300 µM) plus NGF (40 ng/ml) (G + N), or with no added treatments (C) as described in Figure 2b. The selective inhibitor of heme oxygenase, zinc protoporphyrin-IX (100 nM), was added to some cultures prior to the addition of guanosine, or NGF, or guanosine plus NGF. Cells were lysed after 12 h and cGMP concentrations were determined by a competitive enzyme immunoassay. Open bars: no zinc protoporphyrin-IX added, closed bars: zinc protoporphyrin-IX added. Statistical analysis was performed using a one-way ANOVA followed by Fischer's LSD post-hoc comparison test (** P < 0.01, relative to control); (ΔΔ P < 0.01, relative to NGF); (○ P < 0.05, relative to treatment without zinc protoporphyrin-IX). Data presented represent the mean relative optical density T SEM obtained in six independent experiments.
Discussion
Addition of guanosine to cultures of undifferentiated, NGF-naïve PC12 cells modestly increases the proportion of cells with neurites after 48 h and, in cultures treated with maximally effective concentrations of NGF, addition of guanosine produces a significant further increase in the proportion of neurite-bearing cells [15]. This implied that the effects of NGF and guanosine are mediated through distinct signaling systems. We reported previously that, whereas some of the neuritogenic effects of guanosine were mediated by increases in intracellular cAMP [22, 23], there was also a cAMP-independent component [22]. Several points led us to consider that a signaling pathway involving cGMP may be responsible for the cAMP-independent component. Vuorinen et al. [25] had shown that extracellular guanosine increases intracellular cGMP in rat mesenteric artery through a NO-dependent mechanism; nitric oxide donors enhance NGF-dependent PC12 cell neurite outgrowth via a cGMP-dependent mechanism [26]; and the neuritogenic effects of the NO donors are abolished by guanylyl cyclase inhibitors and mimicked by cGMP analogs [26]. Our data indicated that indeed cGMP is involved in the neuritogenic effects of guanosine in PC12 cells, but, surprisingly, that NO is not.
Increases in intracellular cGMP are often due to activation of GC by NO [35, 45]. Schulick et al. [60] demonstrated that activation of GC by NO donors or atrial natriuretic peptide enhanced NGF-dependent neurite outgrowth in PC12 cells. Further, in PC12 cells NGF upregulates NOS expression, and NO plays an important role in their cellular differentiation [32]. However, NGF-induced differentiation of undifferentiated NGF-naïve PC12 cells could not be abolished by inhibitors of sGC, and NO alone is not sufficient to induce PC12 cell differentiation [26, 33].
In the experiments reported here, pre-treating undifferentiated, NGF-naïve PC12 cells with the NOS inhibitor l-NAME, had no effect on the guanosine-enhanced NGF-induced neurite outgrowth. Moreover, like others [61, 62], we found that neither nNOS nor iNOS isoforms were expressed during the first 48 h of cellular differentiation of PC12 cells. Therefore, NO cannot mediate the neuritogenic effects of guanosine on PC 12 cells during the first 48 h.
Rather than NO, our data indicated that CO might activate sGC in response to guanosine in undifferentiated PC12 cells. The heme oxygenase inhibitor, ZnPP attenuated both the ability of guanosine to enhance NGF-dependent neurite outgrowth, and its ability to increase intracellular cGMP, alone or in the presence of NGF. This led us to conclude that guanosine either activated constitutive HO-2 or induced the expression of either HO-1 and/or the HO-2 isoforms of this enzyme.
Our data indicated that, as others have reported [55], both HO-1 and HO-2 were expressed constitutively in PC12 cells under basal cell culture conditions. Guanosine induced a significant increase in the expression of HO-1 protein at 24 h in undifferentiated PC12 cells, but had no significant effect on HO-2 expression. Guanosine, in the presence of NGF, produced an even more rapid and pronounced enhancement of HO-1 expression, which was detectable after 12 h. Treatment of PC12 cells with NGF alone also increased the expression of HO-1. Several authors reported recently that NGF promotes HO-1 gene transcription in PC12 cells [63, 64] via activation of the MEK pathway. HO-1 mediates the anti-oxidative and antiapoptotic effects of NGF in serum-deprived PC12 cells [63]. In PC12 cells exposed to oxidative stress induced by 6-hydroxy dopamine, NGF promotes HO-1 upregulation, in this case via the PI3K/Akt/PKB pathway [64].
Others [63, 64] have reported increased HO-1 protein by 6–9 h after exposure to NGF, whereas in our study we observed no change in HO-1 protein concentration until 24 h after NGF addition. This discrepancy may be due to a difference in the PC12 cell line, or to the experimental conditions to which the cells were exposed.
We did not expect to find inhibition of HO-1 protein expression by guanosine at 6 h. It has been reported however, that the HO-1 gene expression is suppressed by the elevation of intracellular calcium concentration [56]. Since we have found that guanosine increases intracellular calcium concentrations in astrocytes [57], similar mechanisms may downregulate HO-1 expression in PC12 cells at the 6-h time point. As the guanosine-stimulated calcium increases are rapid and transient, at later time points (12–24 h) other intracellular pathways, such as the MAPK and PI-3 kinase pathways are activated, and these are known to promote HO-1 gene expression. Most likely, by 48 h these pathways are no longer activated.
In contrast to HO-1, neither guanosine nor NGF, alone or in combination with guanosine, affected HO-2 expression. This confirms the results of others who have found that HO-2 is not regulated by NGF [63, 64]. The HO-1 isozyme is induced by diverse intracellular signals (reviews: Maines [36], Immenshchuh and Ramadori [65]), that include some of the pathways which activate protein kinases, such as PKA [66], or PKC [67, 68], or the second messenger cGMP [42] most likely via PKG [69]. These pathways likely converge on nuclear effectors such as the nuclear factor-κB (NF-κB), activator protein-1 (AP-1) and heat shock factor, to enhance HO-1 gene transcription by interacting with the regulatory elements in the promoter region the gene [36, 65].
We reported previously that inhibition of transport of guanosine into PC12 cells did not affect its neuritogenic activity [23], raising the possibility that its effects are mediated by activation of a cell surface, membrane-bound receptor. Indeed, recently we have identified and pharmacologically characterized a specific binding site for guanosine in rat brain membranes [70], in cultured astrocytes [71] and in PC12 cells (F. Caciagli et al., unpublished observations), with characteristics of a Gi-protein coupled receptor [5, 70]. Occupation of this putative receptor by guanosine activates the MAPK pathway [5] and the PI3K/Akt/PKB pathway [72, 73]. These effects are most likely mediated by the β,γ-subunit of the Gi-protein, coupled to the guanosine receptor [74–76]. Guanosine has also been shown to increase intracellular cAMP concentrations in several cell types [22–24] and phosphorylate the cAMP binding protein, CREB (F. Caciagli et al., unpublished observations).
In PC12 cells NGF-promoted induction of HO-1 expression requires the activation of the MEK/ERK pathway [63] and the PI3K/Akt/PKB pathway [64]. Since guanosine, via its putative receptor may also stimulate these intracellular pathways, HO-1 expression in PC12 cells may be regulated by similar mechanisms.
Guanosine plus NGF in combination induce HO-1 expression to a similar extent than NGF alone at 6, 12 and 24 h. At 48 h, however the effect of guanosine plus NGF on HO-1 expression is comparable to that of control and lower than NGF alone. Since the mechanisms that induce HO-1 expression are complex, the roles played by guanosine and NGF, and the time course of these effects are unclear. We have shown, however that guanosine alone is insufficient to cause major neurite outgrowth. It does however enhance the NGF-induced neuritogenesis. Neuritogenesis is not solely due to HO-1 expression. Either HO-1 or HO-2 could be responsible for these early events since ZnPP, a non-selective HO inhibitor reduces the concentration of cGMP elicited by guanosine or guanosine plus NGF by 12 h.
NGF activates a number of signalling pathways in PC12 cells, including the MAP kinase and PI3-kinase pathway [18] and the c-JUN N-terminal kinases (JNKs.) [20]. Induction of the HO-1 gene by these pathways seems to override the inhibitory effect of guanosine on HO-1 expression. HO-1 expression at 12 and 24 h is upregulated to a similar extent by guanosine alone, or by NGF alone or by guanosine plus NGF, so there is no additive effect of the combination treatment. As the mechanisms involved in HO-1 gene induction are complex and not clearly defined, our results suggest that the activation of these processes by the three treatments is comparable.
At 48 h HO-1 expression in cells treated with guanosine alone, or guanosine plus NGF is similar to control. In the NGF-treated cells, however HO-1 expression is still significantly upregulated compared to control. We attribute this to a difference in the time course of HO-1 gene activation by guanosine, and NGF. Since NGF-activated intracellular signalling pathways are complex and manifold [18, 20] these effects may persist for a longer time. These results are different from those we obtained at 6 h. At this time point, the only significant effect on HO-1 expression is its inhibition by guanosine alone, and this may be due to downregulation of HO-1 expression by guanoisne-stimulated elevations in intracellular Ca2+ concentrations as described above.
In several types of cells, HO-1 expression is regulated by the NO-cGMP pathway (reviewed by Maines [36]). In the case of guanosine this mechanism is not likely to play a significant role, as neither iNOS, nor nNOS was detected in NGF-treated PC12 cells during the first 48 h of guanosine-promoted neurite outgrowth.
In contrast, upregulation of HO-1 expression by activation of the constitutive HO-2 via the transcription factor NF-κB, is a very plausible alternative mechanism [36]. HO-2 may be activated by PKC, which in turn is stimulated by increased intracellular calcium concentrations [77]. PKC may also activate sGC directly, by phosphorylation, thus providing a receptor-mediated elevation of cGMP concentrations in PC12 cells [78]. We have shown that guanosine increases intracellular calcium concentrations under certain conditions [57], and this in turn may activate PKC. Thus, guanosine may activate HO-2 in a calcium-dependent manner. The time course of intracellular cGMP elevation elicited by guanosine, or by guanosine plus NGF, is consistent with this hypothesis. In PC12 cells exposed to guanosine alone, or to guanosine and NGF, cGMP concentrations were significantly elevated by 6 h and until 12 h in cells exposed to guanosine plus NGF. Although the addition of NGF to PC12 cells induces HO-1 expression at 24 and 48 h, this treatment does not increase cGMP concentrations. Initially we were also puzzled by these data, as we expected a parallel increase in the cGMP concentrations with this treatment. It has been reported by Liu et al. [79] that NGF treatment of PC12 cells rapidly decreases sGC mRNA and protein in a transcription- and translation-dependent manner. So our interpretation of these results is that although HO-1 protein is upregulated by NGF, there is a concurrent suppression of sGC expression, so no increase in cGMP is detected under these conditions. In addition, NGF may also inhibit the activities of some phosphodiesterase isoforms, and thus decrease intracellular cGMP concentrations [80].
Since HO-1 expression in the guanosine treated cells was not observed until 24 h, and since neither iNOS nor nNOS was detectable at these times, our data is most parsimoniously explained by activation of HO-2.
Although we have not determined HO-2 activity, we have shown, using Western immunoblot analysis that this isozyme is expressed in PC12 cells constitutively, throughout the time course of the experiment (from 0 to 48 h) during all treatment conditions.
CO synthesized by HO-1 and HO-2 will bind and activate sGC, increasing intracellular cGMP [35–38] and this in turn may activate protein kinase G (PKG) [81, 82] and regulate phosphodiesterases (PDE) [81, 82]. Our data are compatible with the neuritogenic effects of guanosine being mediated by cGMP, possibly from early activation of HO-2, and later the induction of HO-1, leading to CO generation, and activation of sGC. But how these molecular events enhance neuritogenesis is unknown. CO itself may directly activate several intracellular signaling pathways, including the MAPK pathway [83] and p38 MAPK [84], the transcription factor NF-κB [85] and induce the transcription of immediate early genes [37, 38].
Recently, Xiang et al. [86] proposed that navigation of growing axons in response to guidance cues, is mediated by Gi protein coupled receptors, via elevation of cytosolic cGMP [86]. This is another mechanism by which guanosine may promote differentiation of PC12 cells.
In conclusion, we have demonstrated the importance of HO, CO and cGMP in mediating the neuritogenic effect of guanosine. The relationship of this signaling mechanism to the putative Gi-protein-linked receptor for guanosine is currently the subject of active investigation in our laboratories.
|
[
"guanosine",
"neurite outgrowth",
"pc12 cells",
"heme oxygenase",
"cyclic gmp",
"ngf",
"carbon monoxide",
"nitric oxide synthase"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
J_Gastrointest_Surg-3-1-1852372
|
Duodeno-Gastric-Esophageal Reflux—What is Pathologic? Comparison of Patients with Barrett’s Esophagus and Age-Matched Volunteers
|
Introduction The aim of the study was to analyse pH- and bile-monitoring data in patients with Barrett’s esophagus and in age- and gender-matched controls.
Introduction
The incidence rates for adenocarcinoma (AC) of the esophagus and gastric cardia have risen rapidly in Western industrialized countries.1 Besides nicotine and alcohol abuse, nutritional factors, high body mass index, acidic gastric reflux, and Barrett’s esophagus are believed to be critical factors of carcinogenesis.2–4 Recent studies have shown that the presence of biliary reflux in combination with acidic gastric reflux damages the esophageal mucosa and causes complications of gastro-esophageal reflux disease (GERD), e.g., development of Barrett’s mucosa (BM).5,6 Duodeno-gastric reflux into the esophagus (DGER), in particular, appears to be important to the pathogenesis of Barrett’s esophagus.7 Prolonged esophageal aspiration studies have documented increased bile acids in patients with severe esophagitis and Barrett’s esophagus.8 Eighty percent of patients with Barrett’s esophagus on proton-pump inhibitors show a normal esophageal pH profile, but 60% show abnormal esophageal exposure to bile as measured by Bilitec 2000.9
In the past, direct and prolonged quantification of duodenoesophageal reflux has been difficult to achieve. Now, bilirubin concentration can be directly measured by spectrophotometry, based on the specific absorption at a wavelength of 453 nm. Biliary reflux can be measured with a transnasally passed, ambulatory fiberoptic probe (Bilitec 2000), which records bile absorption. A number of papers have already been published on the exposure of the esophagus10–13 and stomach14–16 using this technique. However, in these studies, the control patients were between 25 and 35 years old. In clinical practice, patients with Barrett’s mucosa tend to be older. In addition, the authors of each study used varying reference values to measure biliary reflux in the esophagus, making comparison of the measured values difficult.
The aim of the present study was to analyze data of pH and bile monitoring in a collective of healthy age- and gender-matched controls and patients with Barrett’s esophagus.
Subjects and Methods
Subjects
Selection of controls was carried out according to a strict protocol. Healthy volunteers treated from 1999 to 2000 between the ages of 40 and 60 years were included in the study. None of the controls were on acid suppressing or gut motility medications, had a history of upper gastrointestinal disease, had undergone upper or major abdominal surgery, or had had therapeutic endoscopic procedures of the upper gastrointestinal tract. Diagnostic endoscopy and barium swallows were not performed, but gallstone disease was excluded by ultrasound scan.
From 1999 to 2002, 24 patients with histologically confirmed Barrett’s mucosa were included in the study. For additional comparison, we include a group of patients with esophagitis (stage I to III according Savary and Miller) without Barrett’s esophagus, which had the same diagnostic procedures before planned laparoscopic fundoplication. During the aforementioned time span, 21 patients age older than 40 years were available for this study. Exclusion criteria were history of esophageal, gastric, or biliary surgery, history of abdominal or thoracic radiotherapy, or presence of peptic ulcer disease, active gastrointestinal bleeding, esophageal or fundic varices, esophageal or upper small intestine chronic disease, or neoplastic disease. All drugs potentially affecting gastrointestinal motility and secretion were discontinued at least 1 week before the study.
Upper Gastrointestinal Endoscopy
All patients underwent classical upper gastrointestinal endoscopy. If sedation was necessary, intravenous administration of propofol (up to 200 mg) was normally used, or occasionally, midazolam (up to 5 mg) was used. During endoscopy, the presence and extent of esophagitis, Barrett’s esophagus, and hiatal hernia was noted. Biopsies were taken from the Barrett’s mucosa.
Ambulatory Esophageal/Gastric pH and Bile Monitoring
All groups underwent esophageal manometry and 24-h pH and simultaneous bile monitoring using a standardized protocol. Ambulatory pH monitoring was performed using a transnasally inserted antimony pH electrode with a separate skin reference electrode (Synetics Medical, Stockholm, Sweden). The data were stored on a portable digital recorder (Digitrapper MkIII, Synectics Medical Stockholm, Sweden). Before each study, the pH probe was calibrated in buffer solutions of pH 7 and 1. An episode of acid reflux was defined as a decrease in esophageal pH to less than 4 for more than 10 s.
To quantify duodeno-esophageal reflux, a transnasally passed, ambulatory fiberoptic spectrophotometer (Bilitec 2000, Synectics, Sweden) was used. The system consists of a miniaturized probe of 1.5-mm diameter that carries light signals into the esophagus and back via a plastic fiberoptic bundle. Before each study, the probe was calibrated in water. Corresponding to the current literature, a bilirubin absorption >0.25 was used as a reference for noxious biliary reflux.14
The bile and pH probes were taped together and passed transnasally into the esophagus and stomach, as described in detail elsewhere.17 The upper tips of the probes were positioned 5 cm above the upper border of the lower esophageal sphincter as defined by esophageal manometry. The distal pH electrode and fiberoptic sensor were placed in the fundus of the stomach, 10 cm distal to the lower esophageal sphincter (Fig. 1). Controls and patients were asked to follow a strict protocol of three meals per day, with no liquids between meals. Recumbent phases of recording were permitted only at night. Patients were asked to keep a diary recording of the exact nature of meals, the supine and erect phases of measurement, and the sensations of heartburn and regurgitation.
Figure 1Position of the pH- and bilirubin-probes in the stomach and in the esophagus. (UES = upper esophageal sphincter; LES = lower esophageal sphincter).
The simultaneous biliary and pH monitoring was done with administration of a colorless “white diet” (WD) including liquid and solid foods with a maximum in vitro bile absorbance of 0.25 [absorbance scale ranging from 0 (plain water) to 1 (total screen)]. The meals included water, milk, toast, potatoes, chicken, dry biscuits, and fish.
Data Acquisition and Interpretation
After completion of the measurements, probes were withdrawn from the patients, and data were stored via interface on an IBM-compatible computer equipped with Polygram® software (Medtronic). The data of each second of the 24-h measurements were used for analysis. To assess the presence of gastric or esophageal biliary reflux, the percentage of time when absorbance was greater than 0.25 was calculated for the following periods: total supine, upright, and postprandial. The postprandial period was defined as 2 h after the end of meals. The percentage of time with esophageal pH lower than 4 and median gastric pH and the percentage of time with gastric pH measuring 1, 2, 3–7, and >7 was also calculated for the above periods. The mean duration of the ambulatory pH and Bilitec monitoring study was 22 h, 40 min in patients and 23 h, 44 min in the controls.
Statistical Analysis
The SPSS (version 11.0, Chicago, Illinois) program was used to analyze the results. For graphical presentation, we used the program MedCalc for Windows, (Version 9.0, MedCalc Software, Belgium). Median, interquartile range (IQR or 25th to 75th percentile) values were established. The nonparametric tests (Mann–Whitney and Kruskal–Wallis analysis) were used to assess the relationship between variables. Box and Whisker plots were used to present some of the data. In these plots, the box represents the IQR, and the Whiskers represent the highest and lowest values. Outliers are also plotted, defined as more than 1.5 times the IQR from the 75th centile. Extreme values were defined as more than three times the IQR from the 75th centile.
A receiver operator characteristic (ROC) curve was used to find a cutoff value for optimal sensitivity and specificity according Zweig and Campbell.18 The area under the curve (AUC) as a measurement of diagnostic performance of the test was used. The results are given as point with the 95% confidence interval (95% CI) and graphically for presentation of all data. As the positive group, we used the patients with Barrett’s mucosa, and the negative test group was defined by the healthy volunteers. A nonparametric distribution of the area under the curve was assumpted.
The assumptions for calculation of the required sample size were alpha = 0.05, beta = 0.80, and that a test is only valid for daily use if less than 20% of the healthy controls and at least 80% of the patients have positive test results. The calculated sample size for each group was 20.
Ethics
The study protocol was approved by the ethics committee of the University of Cologne. Each subject gave written informed consent.
Results
Twenty-four patients with Barrett’s esophagus (mean age: 58 years), 21 patients with esophagitis (mean age: 57 years), and 19 healthy controls with a mean age of 51 years were included in the study. Patients with BM showed esophagitis grade 0 (4 cases), grade I (12 cases), and grade II (8 cases). The control group of patients with esophagitis showed nine cases with grade I, eight cases with grade II, and three cases with grade III. Demographics of patients and volunteers are displayed in Table 1 (the data of one volunteer was not usable due to technical problems).
Table 1Demographic Data of Patients with Barrett-Mucosa or Esophagitis and Healthy VolunteersParametersPatients with Barrett’s Esophagus (n = 24)Patients with Esophagitis (n = 21)Controls (n = 19)Significance Pat. with Barrett vs ControlsAge (median) 57 years58 years51 years–Min–max29–75 years42–77 years39–62 yearsGender m:f16:811:1011:8n.s.BMI (median) min–max27.0 kg/m2 (18.6–33.1)26.9 kg/m2 (17.9–31.5)24.1 kg/m2 (19.62–27.34)p = 0.003Smokers (%)n = 5 (20.8)n = 4 (19.0)n = 6 (31.6)n.s.No alcohol %n = 3 (12.5)n = 5 (23.8)n = 5 (26.3)n.s.BMI Body mass index
Acidic Gastric Reflux (AGR)
Patients with Barrett’s esophagus, 19 of 24 (79%) and 20 of 21 control patients with esophagitis (95%) had pathologic AGR [pH < 4 in >5% of total measuring time (TMT)], but also 6 of the 19 healthy controls (32%) showed pathologic AGR without any symptoms (p = 0.002). During the TMT, the median AGR was 10.6% for Barrett patients and 3.2% for controls (p < 0.01). In particular, measurements of long acid reflux (LAR), defined as reflux pH < 4 lasting longer than 5 min, showed significant differences between patients and controls. Pathologic AGR was found in patients during both the supine and upright fasting measuring periods. In contrast, pathologic AGR in healthy controls occurred only in the upright position (Table 2).
Table 2Median of Acidic Gastric Reflux into the Esophagus in Patients with BM or with Esophagitis and in Healthy ControlsParametersPatients with Barrett’s Esophagus (n = 24) [median (LQ-UQ)]Patients with Esophagitis (n = 21) [median (LQ-UQ)]Controls (n = 19) [median (LQ-UQ)]Significance Pat. with Barrett vs ControlsPercentage of total measuring time pH < 4 (%)10.6 (6.2–38.3)19.9 (1.6–71.7)3.2 (0.9–5.5)p = 0.01Percentage of upright measuring time pH < 4 (%)11.7 (6.03–6.4)18.9 (8.7–60.8)2.4 (0.9–6.1)p < 0.05Percentage of supine measuring time pH < 4 (%)10.9 (0.4–27.1)6.3 (0.0–13.3)0.3 (0.0–4.2)p = 0.004LQ Lower quartile, UQ upper quartile
24-h Intragastric pH and Bile Monitoring
Gastric pH monitoring showed no significant differences between patients and controls for all measuring periods (Table 3). Gastric bilirubin exposure, indicating biliary reflux, was significantly more frequent in patients than in controls during all measuring periods (Table 3). Biliary exposure in the supine position typically occurred during the early hours of the morning during sleep, represented by increased absorbance over 2–3 h, with a rapid return to baseline values around the time the subject resumed the upright position. Over the same time period, gastric pH monitoring showed increased pH levels to greater than 2 (Fig. 2).
Table 3Results of 24-H Intragastric pH and Bile Monitoring in Patients with Barrett Esophagus and Healthy ControlsParametersPatients (n = 24) [Median (LQ–UQ)]Controls (n = 19) [Median (LQ–UQ)]SignificanceMedian of intragastric pH during TMT1.3 (1.0–1.4)1.4 (1.1–1.7)n.s.Bilirubin exposure percentage (%) of TMT 7.8 (1.6–17.8)0.0 (0.0–1.0)p = 0.001Bilirubin exposure percentage (%) of upright time6.9 (0.1–12.9)0.0 (0.0–1.3)p < 0.01Bilirubin exposure percentage (%) of supine time2.0 (0.0–28.6)0.0 (0.0–0.0)p = 0.001LQ Lower quartile, UQ upper quartile, TMT total measuring timeFigure 224h intragastric pH- and bile monitoring in a patient with Barrett’s esophagus demonstrating the duodenogastric reflux in the early morning. a. Bilitec®-monitoring, b. pH-monitoring.
Bilirubin Exposure of the Esophagus
Over the TMT, the median of esophageal biliary reflux was 7.8% for patients with Barrett’s esophagus (LQ–UQ = 1.6–17.8%) and 3.5 (LQ–UQ = 0.1–13.5) for control patients, in contrast to 0% for the controls (LQ–UQ = 0–1.0%), p = 0.001). Figure 3 shows that esophageal bile monitoring in patients with Barrett’s esophagus and healthy controls varied during the total measuring and supine periods.
Figure 3Results of the esophageal bile-monitoring in 24 patients with Barrett’s esophagus, 21 patients with esophagitis and 19 healthy controls a) total measuring period (Kruskal-Wallis Test = 0.01) b) supine period (Kruskal-Wallis Test p = 0.01).
The receiver operating curve, plotting the true positive rate (patients with Barrett’s esophagus identified by bilirubin exposure) in function of the false positive rate (healthy controls with high bilirubin exposure) is shown in Fig. 4. With an area under the curve of 0.78 (95% CI = 0.56–0.89), the ROC analysis of biliary monitoring showed the optimal value for patients at 1% of the TMT [75% sensitivity (95% CI = 53–90%), 84% specificity (95% CI = 60–96%)]. Therefore, the cutoff value to distinguish normal vs pathologic biliary reflux using 24-h biliary monitoring in the esophagus (absorbance threshold >0.25) should be fixed at 1% of TMT.
Figure 4ROC-curve with 95% confidence intervals for pathologic bile-monitoring in patients with Barrett esophagus compared to age and sex matched healthy controls.
Barrett patients, 18 of 24 (75%), 15 of 21 control patients with esophagitis (71%), and 3 of 19 controls (16%) showed biliry reflux into the esophagus more than 1.1% of the TMT (p < 0.001). Using this cutoff value, none of the controls, 10 of the control patients (48%), and 11 of 24 Barrett’s patients (46%) had pathologic bilirubin exposure during sleep.
Discussion
The results of our study confirm that patients with Barrett’s esophagus have significantly more frequent duodenogastric reflux into the esophagus than age- and sex-matched healthy controls. In addition, this reflux, measured by acid and bilirubin exposure, remains longer in the esophagus, especially during sleep.
The role of acid and nonacid reflux into the esophagus as a causative factor of symptoms and mucosal lesions has been addressed in a number of studies. Not only the duration, but possibly the composition of the reflux, is instrumental in the development of such lesions.13 Twenty-four-hour intragastric bile monitoring has provided the clinician with unequivocal evidence of excessive duodenogastric reflux (DGR) in 41% of patients with endoscopic esophagitis, gastroesophageal reflux (GER) symptoms, and gastric symptoms suggestive of DGR.19 Reflux of duodenal contents into the stomach, especially postprandially, is a physiological event;20 however, biliary reflux is a large contributor to mucosal lesions in the whole stomach.21
In our study, the control group of patients with different grades of esophagitis showed no significantly different measurements of acidic or bile reflux into the stomach or the esophagus compared to Barrett’s patients. This may be caused by selection of patients with esophagitis, which were candidates for fundoplication, but both groups of patients differed significantly compared to healthy controls. Therefore, our results are of great clinical relevance especially for preoperative diagnostic.
Marshall et al. compared healthy controls to patients with different grades of reflux-esophagitis and Barrett’s esophagus with regard to bile measurements in the stomach.22 In this study, the average age of the control patients was 25 years, and that of the patients in Groups I, II, and III was 42, 50, and 60 years, respectively. The bilitec-probe was positioned 10 cm below the lower esophageal sphincter (LES). The threshold of bilirubin absorbance was 0.14, and although no difference was found between groups over the TMT, gastric bilirubin exposure was higher in the supine than in the upright position. In the current study, the control group was older than that of the Marshall study. More duodeno-gastric reflux was recorded in both study and control patients during all periods of measurement. These findings may be due to improved study conditions.
We used an esophageal threshold of 0.25 for bilirubin absorbance. Fein et al.,14 in an in vitro study of absorption of different white meals, showed that the least food interference during bile monitoring was measured with an absorbance > 0.25.
Tack et al. reported the influence of meal consistency on Bilitec measurement results in healthy subjects.23 They compared two groups of young controls. The subjects took either liquid meals only, not absorbing light of the same wavelength as bilirubin, or solid food, avoiding diets that interfere with bilirubin absorbance. The authors found significant differences between the two groups using a bilirubin absorbance threshold >0.14 with a median percentage (interquartile range) over the TMT of 10.9 % (6.7–19.3) for solid meals and 0.3% (0.0–2.8) for liquid meals. Major meal artifacts were present in two-solid-meal (10%) and no-liquid-meal subjects. In our study, we found such a meal artifact in one patient and one control, but the values of bilirubin absorbance were lower than 0.25, and therefore, not relevant to our results.
It is not unusual for gastro-esophageal reflux to contain bile, duodenal, and pancreatic secretions. Utilization of the Bilitec spectrophotometric probe has demonstrated a higher prevalence of abnormal esophageal bilirubin exposure in patients with Barrett’s metaplasia when compared to those with erosive injury or without signs of esophagitis.6,8,10,13 In those studies, patients were consistently older than volunteers included in the control group. However, other studies have shown an increased prevalence of gastroesophageal reflux with age.24 For these reasons, we studied age and sex matched healthy controls and patients with Barrett’s esophagus or with esophagitis. We found pathologic acid and biliary reflux of the esophagus in one-third of the controls. Perhaps, this may be caused by artefacts or by violation of the protocol by the volunteers. But in a previous published study, we could show that younger healthy controls had no such pathologic reflux.17 Possibly, these phenomena are caused by relaxations which occur more often in older people. In contrast, nearly all patients with Barrett’s esophagus (87%) and all patients with esophagitis (100%) showed pathologic acidic reflux and/or bile reflux measured with combined pH and bile monitoring. Bile reflux into the esophagus during sleep, in particular, was only found in patients with BM or with esophagitis.
In our study, we measured the intragastric pH and the bile reflux from the duodenum into the stomach (DGR). The median of the intragastric pH was similar in both groups. But patients with Barrett’s esophagus had significantly longer duodenogastric reflux during the 24-h measuring period than controls. More DGR was demonstrated at night than during the day in both groups of study patients and in healthy controls. This could be associated with an alkaline shift in the pH, according to previously published studies.25–27 The precise mechanism by which nocturnal DGR occurs and the roles posture plays remain unclear.
Bowrey et al. were unable to establish either gastro-esophageal or duodenogastric reflux as the predominant cause of inflammation in gastric cardiac mucosa with use of the Bilitec 2000 device.16 This is understandable, as the amount of reflux into the stomach (DGR) does not necessarily correlate with DGER into the esophagus. In this study, the authors demonstrated more DGR in females during the supine period, while males presented more DGER. At the same time, there was no correlation between bile levels in the stomach and esophagus. The controls were, however, much younger than the patients. We found significant differences in bile measurements of the stomach and esophagus between BM patients and controls. In contrast to Bowrey et al., we saw more DGER in females during the supine period and more DGR in male patients.
In contrast, Banki reported similar esophageal exposure to refluxed acid and bilirubin in females and in males with Barrett’s mucosa.28 Pfaffenbach et al.29 studied esophageal bile and acid reflux in patients with long segment Barrett’s esophagus (LSBE), short segment Barrett’s esophagus (SSBE) and patients with gastro-esophageal reflux disease (GERD). Subjects underwent esophageal manometry and simultaneous 24-h pH and bile monitoring (Bilitec 2000) with an absorbance value >0.2 for 10.9% of the total period. GER did not differ between the groups. However, DGER differed between patients with LSBE (14.7%), SSBE (2.1%), and GERD (2.1%).
In summary, the analysis of reference values of esophageal acid and bile-reflux measurements in a collective of healthy, age- and gender-matched controls compared to patients with BM led to the following conclusions:
Although about 30% of the healthy controls showed acid reflux in pH monitoring, patients with BM had significantly more acid reflux during all measured periods.Healthy controls did not have relevant duodeno-gastric-esophageal reflux measured by bilirubin absorbance. Especially during the supine period, there was no bile reflux.The optimal threshold for pathological bile reflux is 1.1 % (bile monitoring with an absorbance of 0.25).
|
[
"esophagus",
"barrett’s mucosa",
"spectrophotometry",
"acid reflux",
"bile reflux",
"reference value bilitec"
] |
[
"P",
"P",
"P",
"P",
"P",
"R"
] |
J_Interv_Card_Electrophysiol-4-1-2292477
|
Long term outcome of cavotricuspid isthmus cryoablation for the treatment of common atrial flutter in 180 patients: A single center experience
|
Objective Recent literature has shown that common type atrial flutter (AFL) can recur late after cavotricuspid isthmus (CTI) catheter ablation using radiofrequency energy (RF). We report the long term outcome of a large group of patients undergoing CTI ablation using cryothermy for AFL in a single center.
Introduction
Ablation of the cavotricuspid isthmus (CTI) for the treatment of atrial flutter (AFL) has become standard practice. Most of the procedures are performed using radiofrequency energy (RF) [1–5]. High chronic success rates are described but the majority of the data comes from a relatively short term follow-up [1–13]. A recent study by Chinitz et al. [14] reported some interesting data regarding the long term follow-up of 80 patients with common type AFL who underwent CTI ablation using RF. They found a 12.5% (ten patients) recurrence rate at an average of 21 months after the procedure with most patients having a recurrence after the first year post ablation. Our prior experience using cryothermy in a limited number of patients [15] also showed that AFL may recur 1 year after CTI cryoablation.
The purpose of this study was to evaluate the long term outcome of CTI cryoablation in a large patient population with common type AFL in a single center.
Methods
One hundred and eighty patients with sustained symptomatic common type AFL referred for CTI ablation were enrolled prospectively from July 2001 to July 2006 in our institution. Signed written consent, approved by the local ethics committee, was obtained from all participants.
Before CTI cryoablation, anticoagulation with warfarin aiming a therapeutic international normalized ratio of 2 to 3 was kept for at least 3 weeks. Antiarrhythmic drugs were not discontinued before ablation.
Electrophysiologic study and ablation
We focused our study on the clinical aspects and long term follow-up of patients with AFL who were submitted to CTI cryoablation.
Our CTI cryoablation protocol has been already reported [15]. Briefly, our methodology was as follows: under local anesthesia and via the femoral route, a decapolar catheter is positioned in the distal coronary sinus (for evaluation of left atrial activation), a duodecapolar catheter (2-mm interelectrode spacing, Halo catheter, Biosense Webster, Baldwin Park, CA) for mapping the right atrial lateral wall and a quadripolar catheter in the His bundle area. A deflectable, 10.5 F, 6.5 mm tip cryoablation catheter (CryoCor Inc., San Diego, CA) is inserted into the right atrium through a 12F, 65-cm-long sheath (DAIG, St. Jude Medical Inc., St. Paul, MN) [15–19].
Entrainment to confirm the isthmus dependence of the AFL circuit was performed in every patient in whom AFL was present or could be induced at the start of the procedure, according to previously published techniques [2, 10, 15, 16, 20, 21]. If AF, requiring cardioversion, was present or developed during stimulation or if we were unable, even under isoproterenol infusion, to induce AFL, CTI ablation was performed during sinus rhythm. Linear lesions were created by use of a point-by-point technique with gradual pullback of the cryocatheter in a ventricular atrial fashion. The first application was delivered at the ventricular insertion of the isthmus and applications were continued for an average of 3 min. In patients in whom ablation of the posterior isthmus proved insufficient, an attempt was made to ablate the septal isthmus (four patients). After documentation of bidirectional isthmus conduction block, the atrial pacing (from the proximal coronary sinus) protocol (up to three atrial extrastimuli at three pacing cycle lengths and incremental atrial pacing) was performed without and with isoproterenol infusion (1 to 3 μg/min). Acute success was defined as bidirectional isthmus conduction block, 30 min after the last application without and with isoproterenol infusion [22].
All patients were studied in the fasting state without sedation. Those presenting in AF while on the catheterization table were converted to sinus rhythm by internal or external cardioversion. During the procedure intravenous heparin was given as a 100-IU/kg bolus dose after the venous sheaths were inserted. The 12-lead ECG and intracardiac electrograms were recorded and stored by the BARD Labsystem PRO.
Post ablation management
All patients were monitored in hospital for 24 h and oral anticoagulation was started the day of ablation. Antiarrhythmic drugs (AAD) were stopped after the procedure in patients without a history of AF; in those with AF/AFL the same AAD were continued. All patients had anticoagulation for at least 1 month. Subsequently, the need for chronic anticoagulation was assessed by the amount of recurrences of AFL/AF and the presence of risk factors for thromboembolic events.
All patients had a 12-lead ECG and a 24 h Holter recording at discharge and during each clinic visit (1 month, 3, 6, 9, 12 months and yearly thereafter) or earlier if they had symptoms.
Due to the logistics of the Maastricht area—and also the presence of a dedicated research nurse (S. P.) who was available to address patients’ concerns and questions at any time—we were able to follow every patient on an individual basis.
Statistical analysis
Continuous variables are presented as mean ± SD, where appropriate. In cases of a non-Gaussian distribution, medians and quartiles are given. Categorical variables are expressed as numbers and percentages of patients.
Statistical analysis was performed using the Student t test for unpaired data. All values were considered significant at P < 0.05.
The authors had full access to the data and take responsibility for its integrity.
Results
Of the 180 enrolled patients, 39 patients (22%) were women with a mean age of 58 (from 18 to 80) years. More than half of the patients (92 patients, 52%) had structural heart disease: arterial hypertension: 57 patients, coronary artery disease: 22 patients, valvular heart disease: 13 patients, congenital heart disease: 11 patients, idiopathic cardiomyopathy: 18 patients. Counterclockwise AFL was documented in 91% (164) of the patients and clockwise AFL in 9% (16 patients).
The mean left atrial diameter and the mean left ventricular ejection fraction were 44 ± 7 mm and 57 ± 7%, respectively. A prior history of AF was present in 123 (69%) of the patients. The clinical characteristics of the patients, related to the presence or absence of AF before ablation, are described in Table 1. Note that AF during follow-up is significantly higher in the group with a prior history of this arrhythmia.
Table 1Characteristics of the 180 patients with atrial flutter referred for CTI cryoablation related to the presence or absence of atrial fibrillation AF before ablation AF/AFL patients (123 patients), 69%AFL only (57 patients), 31%p valueAge (year)57 ± 1358 ± 13nsWomen19% (23 patients)28% (16 patients)nsNo SHD55% (68 patients)32% (18 patients)< 0.05LAd (cm)4.44.5nsLVEF (%)5855nsAcute failuresa5% (6 patients)5% (3 patients)nsAF in long term follow up69% (85 patients)35% (20 patients)< 0.05aPatients in whom CTI cryoablation did not result in bidirectional block (failed procedure).AF Atrial fibrillation, AFL atrial flutter, CTI cavotricuspid isthmus, Lad left atrium diameter, LVEF left ventricular ejection fraction, SHD structural heart disease
Total fluoroscopic—mean of 30 ± 18 min (range, 12 to 152 min)—and procedure times—mean of 2.6 ± 1.1 h (range, 1 to 6.5 h)—decreased throughout our study with a long duration of a procedure and fluoroscopy being attributed mainly to the learning curve of a new technology. An average of 7 (3 to 20) applications per patient were delivered with a mean temperature and duration of −88°C and 3 min, respectively.
The acute success rate for cryoablation of the CTI was 95% (171 patients). There were no complications. Of the nine patients in whom bidirectional CTI block was not achieved, three underwent a successful reablation. The other six patients had much improvement of their symptoms (despite an incomplete line) and preferred not to have another procedure.
After a mean follow-up of 27 ± 17 (range from 12 to 60) months, recurrences of symptomatic AFL occurred in 15 patients (9%) resulting in a 91% chronic success rate. Those recurrences occurred in six patients within the first 3 months post ablation, in four patients from 3 to 6 months post ablation, in four patients from 6 to 9 months post ablation and in one patient at 14 months post ablation (Fig. 1). Ten of those 15 patients underwent a second successful cryoablation of the CTI. The other five had not only reablation of the isthmus but also pulmonary vein cryoisolation (PVI) for AF during the second procedure.
Fig. 1Percentage of patients (171 successfully ablated) free of common type atrial flutter over time
Despite the success as far as AFL was concerned, AF was still present in 85 patients (69%) with a prior history of this arrhythmia. Those patients were treated by AAD (69 patients, 81%), PVI (14 patients, 16%) or AV nodal ablation with pacemaker implantation (two patients, 3%). New episodes of AF developed in 20 (35%) of those 57 patients without documented AF prior to CTI cryoablation and were all controlled by AAD.
Discussion
Main findings
Our current study showed a 91% chronic success rate of CTI cryoablation in a large population (180 patients) with AFL followed for a long period of time (1 to 5 years, mean of 27 months).
According to the most recent ACC/AHA/ESC guidelines for the management of supraventricular arrhythmias CTI ablation is the only therapy with a class I indication for the long term management of AFL [2]. The majority of those ablations are done using RF energy. Concerning the follow-up length of those procedures, most of the literature available reports on a relatively short period, the majority of them being less than 2 years. However Chinitz et al. [14] published a study of 80 patients with AFL submitted to CTI ablation using RF that had up to 6 years follow-up. Interestingly, they found a 12.5% recurrence of AFL that occurred on an average of 21 months post ablation, ratifying the need of data with a longer follow-up. It is important to keep in mind though, that arrhythmias after ablation do not always correlate with symptoms [23, 24] and one could question if those patients with such a late recurrence had those episodes much earlier than what is reported. That might be one reason why our results, where most of our recurrences happened within 1 year, are discrepant with those from their study. The intrinsic characteristics of our hospital, our clinical follow-up and the population of Maastricht could be responsible for our ability to find those recurrences earlier.
Regarding cryothermy as energy source, a 9 months follow-up study in 39 patients undergoing CTI cryoablation with a different system was reported [25]. They achieved a chronic success rate of 100% in the cryoablated group, despite documenting reconduction through the isthmus in 31% of patients during a 3 month follow-up electrophysiologic study. The CRAAFT trial presented the results in 32 patients with AFL submitted to either cryo- or RF ablation of the CTI [26]. They report an 84% success rate after a follow-up of 14 months. Those two small studies, with a relatively short follow-up, reported similar success rates with RF and cryothermy for AFL ablation. The results from our study—which included a significant larger patient population (180) with a longer follow-up (1 to 5 years, mean of 27 months)—reinforces the effectiveness of cryothermy for the treatment of AFL, being the outcomes comparable to those reported in the literature using RF (where most outcome data also comes from non invasive follow-up) [1, 4–7, 11, 12, 27, 28]. Therefore, cryothermy can be considered as an efficient energy source for the treatment of common type AFL.
The relation of AF with AFL
The close relation between AF and AFL is well described [1, 5–14, 21, 29–39]. Our data showed a high prevalence of AF (123 out of 180 patients, 69%) in patients with predominant AFL referred for CTI ablation. A recent study by Ellis and colleagues [40] strengthened even more this association. They followed 363 patients with lone AFL who underwent CTI RF ablation—during a mean follow-up of 39 months—and reported an 82% incidence of drug refractory AF in their patient population.
The new development of AF in our patient population without a prior history of AF preceding AFL ablation may be a sign of an already present electrical and morphological change in the right and left atria. If in addition, a functional or anatomical line of block between the two venae cavae (or elsewhere) occurs, atypical AFL(s) may develop either in the right (because the CTI is already ablated) or in the left atrium [11, 14, 21, 31, 32, 37–39].
Study limitations
Our recurrence data rely mostly on the subjective assessment by the patient, like the great majority of RF data [1, 4, 6–8, 12]. Only an objective measurement (such as a repeat electrophysiological study with documented bidirectional block) will determine the long lasting effect of CTI ablation for the treatment of common type AFL.
Conclusions
In this prospectively studied large population of patients with common type AFL, cryoablation of the CTI has a 91% chronic success rate during long term follow-up (range 1 to 5 years, mean of 27 months). These results are similar to those obtained with RF, validating cryothermy as an efficient alternative energy source. We also were able to ratify the frequent association of AF with AFL.
|
[
"cavotricuspid isthmus",
"cryoablation",
"atrial flutter",
"long term follow-up",
"atrial fibrillation"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Pediatr_Nephrol-4-1-2259256
|
Neurogenic bladder: etiology and assessment
|
A review of the various causes of neurologic impairment to the lower urinary tract in children was the aim of this presentation. The emphasis was on diagnosis, pathophysiology, and treatment that strive to maintain as normal a function as possible in order to achieve eventual urinary continence and health of the upper urinary tract. The latest principles based on the most up to date evidence are promulgated but with an eye towards historical prospective. The reader should gain an adequate understanding of various disorders that comprise this condition and feel comfortable with proposing options for management when faced with the responsibility of caring for an affected child.
Introduction
Neurogenic bladder dysfunction in children is an ever-evolving condition. The expansion of its understanding and treatment over the past 50 years has been just remarkable. In the mid 1950s there were few insights and minimal alternatives to the child’s being in diapers or wearing an appliance over an abdominal wall stoma. Starting with the development of adequate X-ray assessment and reliable urodynamic investigation, the advent of clean intermittent catheterization, artificial sphincter implantation, continent urinary conduits and a plethora of drug therapies that modulate lower urinary tract function, we have learned a great deal about the pathophysiology, pathogenesis and treatment of this disorder and the evidence specific ways to manage it. With the promise of tissue engineering and stem cell therapy, new vistas for treatment seem to be on the horizon.
The most common cause of neurogenic bladder dysfunction in children is neurospinal dysraphism, primarily an open back lesion, but an occult or closed dysraphic state is being diagnosed with more frequency as neonatal spinal ultrasound and magnetic resonance imaging (MRI) are used with increasing regularity to visualize any lower midline spinal cutaneous or gluteal cleft malformation. It was thought that folic acid deficiency was a cause of this disease and that its replacement in women of childbearing age would practically eliminate the condition, but this has not happened. There has been a definite decrease in its incidence; however, in parts of the USA its prevalence has not diminished at all [1–4].
Other causes of neurogenic dysfunction involving the spine include sacral agenesis, tethered spinal cord associated with imperforate anus, cloacal malformations, and spinal cord injuries from sporting injuries and motor vehicle accidents. Central nervous system abnormalities include spastic diplegia (cerebral palsy) and learning disabilities, i.e. attention deficit hyperactivity disorder (ADHD) or attention deficit disorder (ADD). Terminology used throughout this manuscript will conform to the standardization document recently published in the Journal of Urology [5].
Urodynamic studies
A urodynamic study consists of the following components. The child is catheterized with a triple-lumen urodynamic catheter after a small amount of liquid lidocaine (1%) has been injected into the urethra and held in place for a moment or two. First, intravesical pressure is recorded; then, the bladder is drained and the residual urine carefully measured, yielding a pressure at residual volume (this helps determine detrusor compliance at natural filling and is more accurate than cystometric compliance measured during even slow filling of the bladder). A small balloon catheter is passed into the rectum to measure intra-abdominal pressure during the cystometrogram to identify artifacts of motion and monitor increases in abdominal pressure during the filling and emptying phases of the study. The side-hole port of the urethral pressure channel is positioned at the highest point of resistance in the urethra and kept in place, measuring this resistance throughout bladder filling and emptying to determine the leak point pressure. External urethral sphincter electromyography (EMG) is performed using a 24-gauge concentric needle electrode inserted perineally in boys or para-urethrally in girls and advanced into the skeletal muscle component of the sphincter until individual motor unit action potentials are seen or heard on a standard EMG recorder. The characteristics of the individual motor unit potentials at rest, in response to various sacral reflexes (i.e. bulbocavernosus, anocutaneous, Valsalva and Credé maneuvers) and bladder filling and emptying are recorded to detect degrees of denervation. Next, the bladder is filled through the second port while intravesical pressure is monitored via the third port of the tri-lumen urodynamic catheter. The rate of filling is set at 10% of expected capacity for age [age (in years) +30 × 30 = expected capacity in milliliters] [6]. Detrusor pressure measurements are continuously recorded throughout filling to calculate compliance, and during voiding or leaking to denote emptying pressure. Detrusor overactivity is defined as any short-lived pressure rise of >15 cm H2O from baseline before capacity is reached [6]. Sometimes, the urodynamics study is combined with fluoroscopic video-imaging using a dilute radio-opaque contrast agent to visualize the appearance of the bladder wall and bladder neck or to detect the presence of vesicoureteral reflux during the test. Alternatively, a radionuclide agent is instilled, with the patient lying above a nuclear camera, to determine at what pressure reflux occurs, when it is known to be present beforehand. The study is not considered complete until the child actually urinates or leaks and the ‘voiding’ pressure is measured. The small size of the urodynamic catheter does not seem to affect the voiding pressure adversely, even in very young children. The normal end filling pressure should be <10 cm H2O, while the normal voiding pressure varies from 55 cm to 80 cm H2O in boys and from 30 cm to 65 cm H2O in girls [5]. Detrusor overactivity is considered an abnormal finding at any time [6].
The examination findings are considered normal when there is an appropriate capacity, good compliant bladder, with no overactivity, and normal innervation of the sphincter with normal sacral reflexes, and an increase in sphincter activity during filling and complete silencing during emptying. An upper motor neuron lesion is present when there is detrusor overactivity and/or hyperactive EMG responses to sacral reflexes and/or a failure of the sphincter muscle, on EMG, to relax (either partially or completely) with a bladder contraction or leaking at capacity. A lower motor neuron lesion is noted when there are no contractions of the detrusor muscle and/or there is a degree of denervation, either partial or complete, in the sphincter muscle, with characteristic EMG changes in the motor units or no motor unit activity at all, respectively, and little or no response in the sphincter to sacral reflexes and/or bladder filling or emptying [7].
Neurospinal dysraphism
Myelomeningocele
With regard to embryology, the developing spinal canal begins on the 18th day of gestation and is completed by day 35, closing in a caudad direction from the cephalic end of the body. Failure of mesodermal in-growth over the developing spinal cord results in an open lesion, most commonly seen in the lumbo-sacral area and, with decreasing regularity, in the thoracic and cervical areas (Table 1). The exposed spinal cord and its nerve roots, some of which may protrude into the meningocele sac, and tension on the spinal cord as the cord ‘rises up’ the canal with elongation of the fetus (from L2, L3 in mid- to late fetal life, to L1 at birth), contribute to a variable picture of neural injury to the lower urinary tract and lower extremities [8]. Coupled with obstruction of the aqueduct to the fourth ventricle (Chiari malformation), with possible herniation of the brainstem and the center for micturition coordination (the pontine mesencephalic center), additional layers of dysfunction are added to those nerve pathways already affected.
Table 1Spinal bony level of myelomeningocele (uppermost vertebral abnormality)LocationIncidence (%)Cervical-high thoracic2Low thoracic5Lumbar26Lumbosacral47Sacral20
The over-riding issue at birth is whether or not the child has detrusor external urethral sphincter dyssynergy and whether the infant can empty the bladder completely at low pressure. All newborns with radiologic abnormalities at birth (5–10% with hydronephrosis, reflux) have active obstruction of the bladder outlet that had caused these changes to occur in utero [9]. The presence of elevated detrusor filling pressure, bladder sphincter dyssynergy or high voiding or leaking pressures (above 40 cm H2O) at capacity, which can result in upper urinary tract deterioration in as many as 63% of children [9], warrants early intervention with clean intermittent catheterization (CIC) and anticholinergic drugs [10–13]. Thus, the investigation of these children when newborn includes a renal and bladder ultrasound, a catheterized measurement of urine residual after voiding or leaking, a determination of serum creatinine concentration after 7 days of life and a urodynamic study that incorporates both detrusor pressure measurements and urethral sphincter electromyography. Voiding cystography is undertaken when hydronephrosis is present and/or urodynamic studies indicate bladder outlet obstruction with either increased pressure at capacity or bladder sphincter dyssynergy. The incidence of reflux when there is functional obstruction of the bladder outlet can range as high as 50% [14, 15].
Although some clinicians still preach watchful waiting if the renal ultrasound findings are normal, and do not recommend urodynamic studies in the newborn period, opting instead for instituting CIC and drug therapy only with the first sign of ureteral or renal pelvic dilation [16, 17], most centers within the USA now advocate full investigation of the lower urinary tract and initiate prophylactic treatment if there are signs of outlet obstruction and/or elevated bladder filling or voiding pressure [18–20]. The incidence of urinary tract deterioration can be greater than 50% [11, 12, 20] when children with the potential for deterioration are followed with expectant and not preemptive therapy. Even though the ‘watchful waiters’ demonstrate that they can reduce the presence of hydronephrosis and, possibly reflux, the changes in detrusor dynamics are not as easily reversed, and the need for subsequent aggressive management of the bladder to control incontinence and a poorly compliant bladder is commonplace (Table 2). Therefore, CIC is begun when detrusor sphincter dyssynergy, elevated leak point pressures greater than 40 cm H2O, and/or reflux grade 3 or higher (on a scale of 1 to 5) are present. Instituting CIC and anticholinergic therapy in infancy has revealed many advantages over time [19–21]: the parents and the child adapt to the routine of CIC much easier than they would have if it were to be begun when the child is older; the bladder often remains very compliant, expanding as the child grows and maintaining appropriate wall thickness as noted on bladder echography; hydronephrosis and vesicoureteral reflux develop in fewer than 10%; continence is readily achieved in greater than 50% with no additional maneuvers; and the need for augmentation cystoplasty to maintain a reasonable organ for storage is markedly reduced from almost 60% to 16% when compared with that in children followed expectantly [22].
Table 2Surveillance in infants with myelodysplasia (until age 5 years) (IVP intravenous pyelogram, ECHO sonogram, UDS urodynamic study, VCUG voiding cystourethrogram, RNC radionuclide cystogram)Sphincter activityRecommended testsFrequencyIntact-synergicPost-void residual volumeEvery 4 monthsIVP or renal ECHOEvery 12 monthsUDSEvery 12 monthsIntact-dyssynergicaIVP or renal ECHOEvery 12 monthsUDSEvery 12 monthsVCUG or RNCbEvery 12 monthsPartial denervationPost-void residual volumeEvery 4 monthsIVP or renal ECHOEvery 12 monthsUDScEvery 12 monthsVCUG or RNCbEvery 12 monthsComplete denervationPost-void residual volumeEvery 6 monthsRenal ECHOEvery 12 monthsaPatients receiving intermittent catheterization and anticholinergic agentsbIf detrusor hypertonicity or reflux is already presentcDepending on degree of denervation
When vesicoureteral reflux is present, CIC effectively lowers the intravesical emptying pressure when the bladder is drained. In addition, anticholinergic medication can be added to lower detrusor filling pressure, increasing compliance without fear of causing urinary retention when combined with CIC. The lowered filling and emptying pressures has proven to be very beneficial; in 30–50% of children reflux is resolved within 2–3 years of its discovery and initiation of therapy.
The disadvantages are few and include a higher rate of bacteriuria (60–70% versus 30%) but a lower rate of symptomatic urinary tract infection (20% versus 40%) during childhood than in children followed expectantly [23, 24]. Because the subsequent risk of reflux is lower, the effect on renal function and the development of scarring is reduced when these prophylactically treated children are compared to those monitored with watchful observation [16, 25].
Credé voiding is not an efficacious form of bladder emptying in children with myelodysplasia, especially if the urethral sphincter is partially or fully innervated. Because most children have intact motor function above L1, any increase in abdominal pressure from a Credé maneuver can lead to a reflexive increase in urethral sphincter activity, thus producing an increase in bladder outlet resistance resulting in “high voiding pressure”. This can be particularly noxious in children with moderate or severe grades of reflux. In addition, as the child grows, the bladder resides more in the pelvis and not intra-abdominally, further reducing the effectiveness of the Credé maneuver.
The key to a stable bladder and, consequently, renal function is the maintenance of a good capacity, highly compliant, detrusor muscle, combined with periodic and complete emptying of the bladder at low pressure. Anticholinergic medication (primarily oxybutynin, tolterodine, glycopyrrolate, hyoscyamine or trospium) and CIC achieve that in a majority of children and provide an added benefit of continence if the child has reasonable bladder outlet resistance [9]. Several alpha sympathomimetic agents, such as phenylpropanolamine, ephedrine or pseudoephedrine, are used to increase bladder outlet resistance when it is not sufficient to maintain continence between CICs. A variety of surgeries has been devised when these conditions cannot be met. A detailed description of the surgical treatment is beyond the scope of this discussion, but, suffice it to say, bowel augmentation onto the bladder has been a common practice to lower pressure and increase capacity in order to make the bladder a useful organ for storage. Disadvantages abound, including problematic mucus production, recurrent urinary infection, electrolyte imbalance, stone formation and the recently documented risk for the late occurrence of cancer in the augmented segment [26].
A plethora of surgical procedures has been designed to increase bladder outlet resistance in those children with a level that is insufficient to maintain continence between catheterizations, that include implantation of an artificial urinary sphincter, bladder neck tightening, using adjacent tissue, a fascial sling, and various bulking agents. All can provide more resistance, but no one procedure is ideally suited for every patient.
Creating a catheterizable urinary stoma has become fashionable in those children with intractable urethral incontinence (with obliteration of the bladder neck) or inability to catheterize their urethra easily due to obesity, poor eye–hand coordination or caretaker issues surrounding genital organ privacy. Long-term success has been achieved that provides the individual with a degree of independence, but problems with stomal stenosis can occur.
Occult spinal dysraphism
Occult spinal dysraphism has been diagnosed with increasing frequency since the advent of spinal ultrasound and MR imaging [27]. An intraspinal lipoma or lipomeningocele, a diastematomyelia, a fatty filum with tethering, or a dermal sinus tract make up the extent of these disorders. Most pediatricians now will image the back of any newborn with a cutaneous lower midline back lesion, which can be detected in 90% of affected individuals [28]. These lesions include a subcutaneous mass, dermal vascular malformation, hypertrichosis, a midline dimple or sinus tract, a skin tag or an asymmetric gluteal cleft. These lesions often signify an underlying bony and/or spinal cord malformation. Ultrasound within the first 3 months of the infant’s life can easily visualize the intraspinal space. After the infant has reached that age MR imaging is needed to diagnose and/or confirm the presence of a dysraphic state.
Most infants have no other manifestation of this disease (other than the cutaneous lesion) because lower extremity neurologic function is normal. In the past, before intraspinal imaging was feasible, these lesions often went undetected until urinary and/or fecal incontinence became problematic or lower extremity difficulties became evident [28–31]. This frequently occurred around the time of a pubertal growth spurt, when increased traction on the spinal cord took place [32]. As the imaging techniques evolved and urodynamic studies in children were performed at an early age, it became clear that most babies had minimal neurologic impairment at first but that the neurologic lesion progressed with advancing age [33].
The pathophysiology involves apparent tension on the lower end of the spinal cord as the child grows. Normally, the conus medullaris ends at L1, L2 at birth but ‘rises’ cephalad to T12, L1 at puberty. The differential growth rate between the spinal cord and the vertebral bodies stretches the lower cord and cauda equina due to fixation of the filum terminale to the bottom of the vertebral canal, or the nerves roots emanating from the cord become compressed by an expanding intraspinal lipoma. With time, this stretching and/or compression affects the oxidative process of the neural tissue that then leads to impaired function of the lower extremities and/or lower urinary tract [34, 35].
Initial investigation of the affected child includes a urodynamics study and a renal and bladder ultrasound. The findings of urodynamics studies in children less than 1 year old are invariably normal, but when abnormal they are not coupled with abnormalities in the lower extremities [33, 36]. When findings are abnormal, partial denervation in the urethral sphincter muscle or failure of the sphincter to relax during a detrusor contraction are the most common findings in infancy, while extensive denervation of the sphincter and/or an acontractile detrusor combined with changes in lower extremity function are the most common abnormalities in an older child [37]. As noted previously for myelomeningocele in children, a voiding cystourethrogram is warranted only when the urodynamics parameters suggest risk to the upper urinary tract from increased bladder outlet resistance or poor detrusor compliance. Vesicoureteral reflux, hydronephrosis and urinary incontinence are all managed in the same fashion as one would treat children with similar neurologic impairment due to an open spinal abnormality.
The abnormal condition tends to improve following spinal cord de-tethering in the infant, but this normalization is unlikely when an older child is operated on [33, 37–39]. Therefore, specific urologic therapy is not instituted following delineation of an abnormality in infants until urodynamics studies have been repeated 3 months after the condition has been repaired. If the findings have not changed, or the child is older (when the chance that the neurologic abnormality will not improve), treatment based on principles outlined in the section on open spinal lesions is instituted.
Thirty percent of children will have secondary spinal cord tethering over time, some as late as puberty or just beyond it, when the last growth spurt occurs [28, 37, 39]. Therefore, careful surveillance and repeated assessment at the first sign of incontinence or changing lower extremity function is warranted. No child is considered risk free until he or she has reached full adult height.
Sacral agenesis
Partial or complete absence of the lowermost vertebral bodies is labeled sacral agenesis. The condition can range from the absence of just the last two or three sacral bodies to the absence of sacral and several lumbar bones as well (sirenomyelia). This can be seen in offspring of insulin-dependent diabetic mothers (1%) [40], but it may be part of a genetic disorder due to a deletion of part of chromosome 7 (7q36), leading to absence of an important transcription factor that plays a role in the development of the caudal end of the spinal cord and vertebral column [41]. In familial cases of sacral agenesis associated with the Currarino triad syndrome (presacral mass, sacral agenesis and anorectal malformation), deletions in chromosome 7 (7q) resulting in HLXB9 genetic mutations have been found [42]. A mutation in HLXB9, a homeodomain gene of a 403 amino acid protein, which appears to be responsible for neural plate infolding, has been identified in 20 of 21 patients with familial Currarino triad syndrome and in two of seven sporadic cases of this syndrome [43, 44]. Heterozygote carriers within these families have also been identified [45]. Thus, sacral agenesis may represent one point on a spectrum of abnormalities that encompass a sacral meningocele and ano-rectal malformations [46].
In the newborn period (and even afterwards) these infants appear normal, with no lower extremity abnormality. Unless it is thought of during the examination of a newborn of a diabetic mother, such babies often go undiagnosed. With time, as they have difficulty in toilet training or have urinary infection, it becomes evident that there is a problem [40, 47]. The pathognomic sign is absence of the upper end of the gluteal cleft, with flattened buttocks. When the diagnosis is considered a lateral spine film (or a spinal ultrasound in infants) will confirm the abnormality. A spinal MR reveals a sharp cut off to the cord at about T-12, with nerve roots streaming from it. Approximately 90% of children develop neurogenic bladder dysfunction [48].
An equal number of children have an overactive detrusor with sphincter dyssynergy or an acontractile detrusor with complete denervation in the urethral sphincter [47, 49]. The former is often associated with recurrent urinary infection and vesicoureteral reflux, whereas the latter produces continuous incontinence. The type of neurologic impairment affecting the lower urinary tract cannot be predicted from the level of absent or abnormal vertebral bones [47]. Obviously, management depends on the type of dysfunction present. CIC, anticholinergic medication and antibiotics are instituted in those with an upper motor neuron type lesion, while surgical measures with CIC are needed in those with an incompetent sphincter mechanism.
Associated conditions
Imperforate anus
Failure of the lowermost portion of the developing colon in the fetus to canalize fully, resulting in a closed rectum that does not open onto the anal skin verge, is a rare anomaly, but one that needs immediate attention after delivery. In boys there is a fistula from the end of the rectal canal to the posterior urethra, whereas a stenotic, anteriorly placed anal canal that ends at the posterior aspect of the vestibule, may be noted in girls (Table 3). Usually, a colostomy is performed, or, more recently, definitive repair is undertaken in the first 24 h of life in boys, while anal dilatation is begun in girls.
Table 3Wingspread classification of anorectal malformationsFemaleMaleHighHigh Anorectal agenesis Anorectal agenesis With rectovaginal fistula With rectourethral (prostatic) fistula Without fistula Without fistula Rectal atresia Rectal atresiaIntermediateIntermediate Rectovestibular fistula Rectovestibular urethral fistula Rectovaginal fistula Anal agenesis without fistula Anal agenesis without fistulaLowLow Anovestibular fistula Anocutaneous fistula Anocutaneous fistula Anal stenosis Anal stenosis Rare malformationCloacal malformationRare malformation
From a urologic standpoint, this condition is often part of a constellation of abnormalities known as the VATER or VACTERL association (the mnemonic denotes all affected organs: V vertebral; A anal; C cardiac; TE tracheoesophageal fistula; R renal; L limb). Unilateral renal agenesis, vesicoureteral reflux and spinal cord tethering are the most common abnormalities affecting the urinary tract [50]. Urinary, and fecal, incontinence issues that become prominent as the child matures, due to progressive denervation of the lowermost nerves modulating bladder and urethral and anal muscle function, are likely clinical scenarios. This is most commonly seen in children where the rectum has ended above the levator ani muscle (50%), but it is also seen when the rectum ends below that pelvic floor muscle (18%) [51]. A spinal vertebral bony image is not a reliable sign that spinal cord abnormality is present [52]. A spinal ultrasound in newborns, and MRI of the spine in older children, are a mandatory part of the investigation in these children. An upper motor neuron lesion with detrusor overactivity and/or detrusor sphincter dyssynergy are most likely to develop, but an acontractile detrusor and sphincter denervation are also seen as a result of spinal cord tethering [49, 53]. It has been shown that the earlier neurosurgical intervention is undertaken, the better the individual’s chances of having normal sacral spinal cord and lower urinary tract function [54]. Therefore, early detection of this condition is necessary to improve the child’s chance of maintaining healthy kidneys and becoming continent.
Central nervous system disorders
Cerebral palsy
This condition results from a non-progressive injury to the brain, occurring in the perinatal period, that produces a neuromuscular disability or a specific symptom complex of cerebral dysfunction [55]. It is caused by perinatal infection or a period of anoxia (or hypoxia) affecting the tissues of the central nervous system [56] (Table 4). It appears in babies who are less than 2 kg at birth, have had intraventricular hemorrhage, experienced a neonatal seizure or received mechanical ventilation for a prolonged period of time in the postnatal period [57]. The incidence of cerebral palsy is increasing, as more severely premature infants are surviving and going home.
Table 4Perinatal risk factors in cerebral palsy. Adapted from [61] used with permission. UMN upper motor neuron lesion, LMN lower motor neuron lesionFactorUMN (no. of patients)LMN (no. of patients)Prematurity101Respiratory distress/arrest/apnea92Neonatal seizures5–Infection51Traumatic birth5–Congenital hydrocephalus3–Placenta previa/abruption22Hypoglycemia seizures2–Intracranial hemorrhage2–Cyanosis at birth13No specific factor noted15–
Affected children have delayed gross motor development, abnormal fine motor performance, altered muscle tone, abnormal stress gait, tight heel cords and exaggerated deep tendon reflexes [58]. These findings can vary substantially, from being quite obvious to very subtle, unless a careful neurologic examination is performed. These abnormalities may not be manifest in the early postnatal period, but become evident over time, because myelination of axons and maturation of neurons in the basal ganglia are required before spasticity, dystonia, and athetosis become apparent. Some less affected children have even milder forms of the disease, with only learning disabilities, attention deficit, or attention deficit hyperactivity disorders being seen.
Most children with cerebral palsy develop total urinary control, albeit at an age later than most unaffected children. Incontinence is a feature in some (24% of surveyed families in one study) [59], but the exact incidence has never been truly determined. The presence of incontinence is often related to the extent of the physical impairment, primarily because the handicap prevents the child from reaching the bathroom on time, causing an episode of wetting [60]. Urinary infection and vesicoureteral reflux are not features of this disease, and the kidneys are invariably normal on ultrasonic imaging.
An upper motor neuron type of bladder dysfunction with detrusor overactivity (80%), but not necessarily with detrusor sphincter dyssynergy (5%), is invariably present, but lower motor neuron injury with sphincter denervation due to spinal cord involvement can be seen as well (11%) [61] (Tables 5 and 6). Volitional control over sphincter function is present, and affected children have some ability to prevent leaking from an overactive detrusor by tightening the muscle for a variable period of time. Children with milder forms of dysfunction, with just learning disabilities without spasticity, have an overactive detrusor leading to either urgency (with or without incontinence) and nocturia, or day and night wetting. Therefore, modulating the overactivity with anticholinergic medication is the treatment of choice. However, this must be done judiciously, with careful monitoring of residual urine to prevent the development of retention.
Table 5Lower urinary tract function in cerebral palsyTypeNumberUpper motor neuron lesion49Mixed upper + lower motor neuron lesion5Incomplete lower motor neuron lesion1No urodynamic lesion2Table 6Urodynamics findings in cerebral palsy (some patients had more than one finding)Type of LesionNo. of PatientsUpper motor neuron (detrusor or sphincter)Detrusor overactivity35Detrusor sphincter dyssynergy7Overactive sacral reflexes6No voluntary control3Smaller than expected bladder capacity2Poorly compliant2Lower motor neuron (abnormal motor unit potentials)Excessive polyphagia of sphincter5↑ Amplitude + ↑ duration potentials4
Trauma
Traumatic injuries to the spine
Fortunately, spinal cord injuries in children are rare (2.6 per million) [62]. The incidence tends to increase geometrically with age [63]. When an injury does occur, it is more likely to happen in a boy than in a girl, and it is usually the result of a motor vehicle or bicycle accident (24–52%), a fall from a high place, a gunshot wound, or a diving or sports incident. Injuries may also occur iatrogenically after surgery to correct scoliosis, kyphosis or other intraspinal processes, congenital aortic anomalies, or patent ductus arteriosus operations [62, 64–67]. Newborns are particularly prone to a hyperextension injury during high forceps delivery [68].
Spinal cord injuries in children are intrinsically different from those in adults, owing to a variety of factors, including the mechanism of injury and the difference in configuration of the cord in children compared with that in adults. In addition, the horizontal versus vertical orientation of the facet joints in vertebral bodies that predisposes to anteroposterior subluxation in children, the delayed supportive effect of the paraspinous musculature and ligaments, and the relative heaviness of the head, which causes a fulcrum of maximal flexion of the upper cervical region in infants and young children, all contribute to a high degree of hypermobility that places the child’s spinal cord at risk for ischemic necrosis [65].
The lower urinary tract dysfunction that ensues is not likely to be an isolated event but is usually associated with loss of sensation and paralysis of the lower limbs. Radiologic investigation of the spine may not reveal any bony abnormality, although momentary subluxation of osseous structures resulting from the elasticity of the vertebral ligaments can result in a neurologic injury. This condition has been seen only in children (usually younger than 8 years old) and has been labeled SCIWORA (spinal cord injury without radiologic abnormality) [69]. Overall, SCIWORA can account for up to 38% of spinal cord injuries in children [70]. Often, what appears to be a permanent lesion initially turns out to be a transient phenomenon with time. Although sensation and motor function of the lower extremities may be restored relatively quickly, the dysfunction involving the bladder and rectum may persist considerably longer.
During the acute phase of the injury, the bladder is often acontractile and the urethral sphincter nonreactive, although normal-appearing bioelectric potentials can be recorded on sphincter EMG (spinal shock). Over a variable but unpredictable period of time, detrusor contractility and sphincter reactivity return as spinal cord edema subsides. With this return of function, an overactive detrusor and bladder-sphincter dyssynergy may develop if the lateral reticulospinal cord pathways to and from the brainstem have been disrupted. When the lesion affects the cauda equina, there is probably little to no return of bladder or sphincter function. Sacral sensation and peripheral reflexes are not good indicators of ultimate lower urinary tract function [71]. Over time, the predominant urodynamic pattern in patients with a thoracic-level lesion is an overactive detrusor with sphincter dyssynergy, high voiding pressures, eventual hydronephrosis and vesicoureteral reflux. Patients with an upper thoracic or cervical lesion are likely to exhibit autonomic dysreflexia with a spontaneous discharge of α1 stimulants during bladder filling and with contractions of the detrusor that require careful monitoring of their blood pressure during any investigational studies of the lower urinary tract [72, 73].
Foley catheter draining in the immediate post-injury phase is needed, but CIC should be begun as soon as feasible after that time [74–76]. When a child starts volitional voiding, CIC can be tapered and stopped as residual urines are measured and found to be insignificant (less that 5 ml). Urodynamics studies should be undertaken no earlier than 6 weeks after the injury, to allow for the manifestation of the extent of the neurologic injury [77]. Periodic reassessment of bladder and sphincter function is appropriate up to 2 years after the injury, due to the potential for change during that time. Renal ultrasonography should be part of the assessment, but voiding cystography is only needed when there are signs of potential risk (i.e. sphincter dyssynergy or poor detrusor compliance). Low detrusor filling and voiding pressures with complete emptying are the goal [78]. When this is not present during testing, starting or continuing CIC and adding anticholinergic medication is paramount to insuring the long-term health of the lower and upper urinary tract [79].
Conclusions
Neurogenic bladder dysfunction in children takes in a very wide spectrum of conditions that include congenitally acquired conditions that may even be preventable today, conditions that are associated with specific anatomic abnormalities, and acquired conditions that may occur perinatally or from accidents or sports or motor vehicle related injuries. Despite the etiology, the guiding principles for management are similar; insuring and maintaining an adequate sized, normally compliant, reservoir that evacuates urine completely, at a relatively low pressure, is the key to maintaining a healthy environment for the kidneys. A plethora of methods has come into existence, especially since the advent of clean intermittent catheterization and the advancement of pharmacologic understanding and manipulation. Future prospects look bright for affected children, with the overall health of the individual the most paramount goal to be achieved.
|
[
"myelodysplasia",
"pediatric neurogenic bladder",
"diagnosis and evaluation"
] |
[
"P",
"M",
"M"
] |
Plant_Mol_Biol-3-1-1805041
|
A large-scale collection of phenotypic data describing an insertional mutant population to facilitate functional analysis of rice genes
|
In order to facilitate the functional analysis of rice genes, we produced about 50,000 insertion lines with the endogenous retrotransposon Tos17. Phenotypes of these lines in the M2 generation were observed in the field and characterized based on 53 phenotype descriptors. Nearly half of the lines showed more than one mutant phenotype. The most frequently observed phenotype was low fertility, followed by dwarfism. Phenotype data with photographs of each line are stored in the Tos17 mutant panel web-based database with a dataset of sequences flanking Tos17 insertion points in the rice genome (http://tos.nias.affrc.go.jp/). This combination of phenotypic and flanking sequence data will stimulate the functional analysis of rice genes.
Introduction
Rice is the most important staple crop for half the world’s population. Improvements in rice yield and quality beyond the benefits of the green revolution of 30 years ago are required to meet the demands of an increasing global population. At the beginning of the 21st century, with the hope of finding creative solutions to the problems of food production, nutrition and transportation, nearly the entire nucleotide sequence of the rice genome, Oryza sativa L. cv. Nipponbare, was determined through a world-wide collaborative effort (International Rice Genome Sequencing Project 2005). The sequence data provide an important resource for promoting the discovery of important genes for crop improvement. Currently, a Rice Annotation Project (RAP) using nucleotide sequences of full-length cDNAs (The Rice Full-Length cDNA Consortium 2003) is in progress to position functional genes on the rice genome map (Ohyanagi et al. 2006); however, many genes remain “unknown” due to lack of experimental evidence or sufficient similarity with characterized genes from other organisms. The next great challenge after completion of genome sequencing is the functional characterization of these genes and discovery of genes that affect vital developmental, agronomic or biochemical plant functions (Hirochika et al. 2004).
Gene annotation is mostly based on the sequence similarity to known genes from other species. The limitations of this method are that every organism may have unique genes that do not have homologues even in closely related species, and assignment of a protein’s function based on similarity may only give a partial description. For example, a gene may contain a domain that is conserved among protein kinases, but the actual substrate of the enzyme would be difficult or impossible to determine without experimental evidence. Categorization using cladistic associations (i.e. a phylogenetic tree) is more sensitive and is able to detect BLAST mis-hits or false positives (Sjolander 2004). However, characteristics supported by experimental data, e.g., correlation between the phenotype and the function of the protein kinase, are indispensable to the exact annotation, because all of the associative algorithms ultimately depend on experimental data.
Mutational analysis through gene disruption is one of the most efficient methods for identifying gene function. The related approaches of interrupting gene expression with RNAi (Miki and Shimamoto 2004), and overexpression are also effective because the functions of genes can be determined by the correlation of disrupted genes and their associated phenotypes. Currently, more than 130,000 T-DNA insertion lines of Arabidopsis thaliana have been created and are publicly available (Alonso et al. 2003). Phenotyping of Ds insertion lines in Arabidopsis is also in progress (Kuromori et al. 2006). For characterizing monocot genomes, we have produced more than 50,000 disruption lines of Oryza sativa cv. Nipponbare (Japonica), using the endogenous retrotransposon Tos17, which has a “copy and paste” type of transposition activity (Miyao et al. 2003). There are two native copies of Tos17 in the Nipponbare genome that are activated specifically in cultured cells (Hirochika et al. 1996). On average, ten new copies of Tos17 are transposed in each cell during 5 months in culture. When plants have been regenerated from the cultured cells, Tos17 retrotransposition is immediately inactivated, and Tos17 copies become fixed and segregate in a Mendelian fashion in the next generation.
There are many advantages to the Tos17 disruption system for mutational analyses. Because the flanking regions of Tos17 insertion points are easily amplified by TAIL-PCR or a suppression PCR method using a 3′-end primer of Tos17, insertion sites in the rice genome can be easily determined. Furthermore, mutants can be screened by PCR using the Tos17 end primer and a primer from any desired genomic sequence. The distribution of transposed Tos17s is not random, and a large number of transpositional “hot spots” are detected throughout the rice genome. The insertion frequency of Tos17 into genic regions is three-fold higher than that into non-coding regions. Owing to this polarization, Tos17 insertion lines have great advantages for the functional analysis of rice genes (Miyao et al. 2003). Because Tos17 is an endogenous retrotransposon, regenerated lines can be grown in the field, and seeds can be exported without regulations associated with genetically modified organisms such as rice lines that have undergone “transgenic-” insertion of T-DNA (Jeon et al. 2000; Sallaud et al. 2003; Wu et al. 2003), Ds (Greco et al. 2003; Kim et al. 2004), or En/Spm (Kumar et al. 2005).
Mutant lines generated by Tos17 retrotransposition have already been used for the functional analysis of rice genes. For example, Tos17 insertion mutants with a dwarf or viviparous phenotype were used to identify and analyze genes for gibberellin and abscisic acid metabolism (Sakamoto et al. 2004; Agrawal et al. 2001). Phenotypic characteristics of most of the Tos17 insertion lines, however, remain to be described. A large number of plant scientists working with rice or other monocotyledonous species could benefit from a systematic phenotypic analysis of many Tos17 insertion lines and the creation of a public database. To promote the functional analysis of rice genes, the phenotypes of all of our mutant lines have been observed in rice fields through the collaboration of seven laboratories. Collected phenotypic data are useful for predicting the function of a disrupted gene. In this paper, we report the phenotypic statistics of a Tos17 insertion mutant population for the discovery of agronomically important genes.
Materials and methods
Plant materials
Nipponbare calli derived from embryos were grown in N6 liquid medium (Otsuki 1990) containing 1 mg/l 2,4-D for 5 months. In total, about 50,000 plants were regenerated. Seeds of the M2 generation were harvested from each M1 plant. To check the activity of Tos17 under various hormone conditions, N6 liquid media containing 1, 2, 5, 10 or 20 mg/l 2,4-D with or without 0.1 mg/l BA were used.
Phenotyping
Ten to twenty-five seeds were planted per line. Germination rates and seedling phenotypes were observed in the nursery. After 1 month from seeding, the seedlings were transferred to the paddy field. Phenotypes in the field were observed at the vegetative stage, near the heading stage, at the seed maturation stage, and at harvest. Lines segregating abnormal plants at about 25% frequency were digitally photographed and assigned a phenotype ID.
Database
All phenotype data were stored into a relational database on the PostgreSQL relational database managing system (http://www.postgresql.org/) with a FreeBSD 5.5 operating system (http://www.freebsd.org/). FreeBSD 6.1 was used for the statistical analysis. Tabular structure for phenotypic description consists of the line name, plant serial number, observation date, observing person, a link to the photograph file, phenotype ID, and detailed descriptions. Additional data for each line are inserted as a new row. Data can be modified only by the observing person. With this table structure, observation logs for each line by many persons over the course of many years can be stored without conflict. For counting lines of each mutant phenotype, Perl script, which can connect with the PostgreSQL server with a Pg.pm module, reads the phenotype table, stores line names and phenotype IDs into an associative array (“hash”), to convert from a redundant number of phenotype IDs to a unique number for each line. The number of lines showing each phenotype ID in the hash was counted. Flanking sequence data of Tos17 insertion sites are also stored in the same database. Loci of Tos17 insertions are determined by BLASTN searches against rice genome sequences of the International Rice Genome Sequencing Project (IRGSP) Build3. Loci of annotated genes are also stored in the relational database.
SOM analysis
Software for SOM analysis was downloaded from the Neural Networks Research Center of Helsinki University of Technology (http://www.cis.hut.fi/research/som_pak/). For the initialization program, randinit, parameters: 24 for x dimension, 16 for y dimension, hexa for topology, bubble for neighborhood function, and 123 for seeds were used. For the map training program, vsom, parameters for first learning: 1000 for learning length in training, 0.05 for initial learning rate, 10 for initial radius of the training area in som-algorithm, parameters for second learning: 10000 for learning length, 0.02 for initial learning rate, 3 for initial radius. The analyzed data were visualized using the umat program.
Results
Flow of phenotypic analysis
Nipponbare calli from 92 seeds were independently cultured for 5 months, and about 50,000 plants were regenerated. Seeds (M2 generation) from the regenerated plants were independently harvested and labeled with a line name. Ten to twenty-five M2 plants of each line were observed over a full developmental cycle in the nursery and paddy.
Each line was designated by “N” for “Nipponbare” followed by a letter A-G that indicates the yearly lot, and four figures, e.g. NA1234. Lines named with “NA” and “NB” were used in a small-scale pilot study and “NC” to “NG” were used in large-scale studies.
To evaluate the effect of hormone (auxin and cytokinin) concentration in the medium on Tos17 retrotransposition activity, the NC line was subcultured in media containing 1, 2, 5, 10, or 20 mg/l of 2,4-dichlorophenoxyacetic acid (2,4-D) with or without 0.1 mg/l benzyl adenine (BA). There was no significant difference in observed transposition events among lines derived from the subcultures. In lots, ND, NE, NF, and NG, calli were induced in a medium containing 2 mg/l 2,4-D, and maintained in a medium supplemented with 1 mg/l 2,4-D for 5 months. The cultural conditions for each line are available with the phenotypic description list from the mutant panel database.
Co-segregation of mutant phenotypes with Tos17 insertion can be detected by DNA blot hybridization, which will help the further analysis of gene function. If co-segregation is detected, the flanking region of co-segregated Tos17 can be isolated by TAIL-PCR or suppression PCR (Miyao et al. 1998).
Classification of phenotypes
To provide a classification system and a database useful for most rice scientists, phenotype scoring was limited to 53 phenotype descriptors belonging to 12 classes. Because environmental conditions such as day length, temperature and soil conditions differ significantly among the seven fields in which this project was conducted, some variability in traits such as heading date or leaf color may have occurred between different fields. Thus, some of the present data may vary with environmental conditions, although most of mutant phenotypes are stably expressed. Phenotype classifications and a summary of observations are shown in Table 1. Each subclass has a phenotype ID code to enable data entry as a barcode with a portable recording device and to enable data compilation from all seven laboratories.Germination. This trait was evaluated by measuring germination rate under defined conditions. Since wild type (cv. Nipponbare) showed a germination rate higher than 95%, lines that showed germination rates less than 75% were recorded in the present project. Since several laboratories measured germination rates of all lines, all primary data are also stored in the database. In total, 3489 lines showed a low germination rate, and 525 lines showed germination rates less than 50%. Some of the lines with low germination rates may be embryo mutants.Growth. Growth was observed at the seedling stage in the rice nursery or at an early stage after transfer to the field. “Weak” refers to mutants that formed slim seedlings with retarded growth, probably caused by a deficiency in some housekeeping gene product (Fig. 1, NG0352). Some of the “Weak” lines were reclassified eventually to “Lethal”. An example of “Abnormal shoot” is shown in Fig. 1 (NG0356). Another example is represented by NE3024. This seedling is smaller than wild type and its leaves are short and wide, resembling mutants defective in genes associated with gibberellin biosynthesis or signaling pathways (Sakamoto et al. 2004; Uegchi-Tanaka et al. 2005).Leaf color. Frequently appearing pigmentation phenotypes are “Albino” (Fig. 1, NG1048) and “Virescent” (Fig. 1, NE1517). Completely white or yellow (Fig. 1, NG1469) seedlings died within 3 weeks after seeding. If green and white segments coexisted on leaves, a condition called virescent, the seedlings survived in the field. Zebra mutants that show repetition of a white or a pale green band and a green band in the longitudinal direction were also observed (Fig. 1, NF6044). However, in most lines, the “Zebra” phenotype was limited to young stages. The “Stripe” phenotype (Fig. 1, NE4001) often showed extremely biased segregation, e.g., only one plant in 25, and often was not stably inherited.Leaf shape. Many different kinds of abnormally shaped leaves were observed. “Short Leaf” and some “Wide leaf” phenotypes were eventually reclassified as “Dwarf” or “Severely dwarf” phenotypes (Fig. 1, NE8114). The “Short leaf” phenotype is similar to mutants defective in gibberellin biosynthesis or signal transduction. The line NE5022 could not develop a normal, flat leaf blade and eventually died (Fig. 1). Several lines showed pleiotropic phenotypes of the shoot. The line NE8329 did not develop tillers, showed dwarfism and formed rolled or twisted leaves (Fig. 1). Since these phenotypes are difficult to characterize exactly by only phenotype codes, additional remarks were presented in the comment column of the database. Frequency distribution of the leaf width mutants was biased toward the narrow type (Fig. 1, NG0754), rather than the wide type.Culm shape. Dwarfism is the most abundant mutant phenotype, along with sterility. The “Semi-dwarf” condition is characterized by plant heights that are 70–80% of wild type values. “Dwarf” describes plants with heights smaller than “Semi-dwarf” but larger than “Severely dwarf”. Plants are classified “Severely dwarf” when the plant height is smaller than 30 cm at maturity. Some dwarf phenotypes often co-segregated with other abnormal phenotypes such as “Fine leaf”, “Wide leaf”, “Spiral leaf”, or “Abnormal panicle shape”. The dwarf mutants accompanying other shoot/panicle abnormalities are expected to be involved in hormonal signaling or synthesis pathways. Other culm phenotypes such as “Thick culm” (Fig. 1, NG9874) appeared infrequently.Spotted leaf/lesion mimic. Various types of “Spotted leaf/lesion mimic” phenotypes, e.g., small and scattered spots, large and dispersed spots, were observed. Heavy lesions caused the early death of leaves (Fig. 1, NG0752). Mutants with large white lesions were also observed. Brown spots were more frequently observed than white ones.Tillering. In the tillering class, “Low tillering” mutants were abundant. The “High tillering” phenotype tends to co-segregate with the “Fine leaf” phenotype. The “Lazy” (Fig. 1, NG0667) mutants have open and recumbent tillers. A quite unusual phenotype among “High tillering” lines is NG2552 that produced tillers reiteratively from upper internodes and failed to produce panicles.Heading date. In this category, lines whose heading date deviated more than 7 days from normal were classified as heading date mutants. This category includes “Early heading”, “Late heading”, and “Non-heading?” mutants. A small number of lines segregated as “Non-heading?” mutants that did not form panicles even at harvest (five to 6 months after sowing). The number of “Early heading” lines (1797) is a little larger than that of “Late heading” lines (1244).Spikelet. Each spikelet of rice is composed of two rudimentary glumes, two empty glumes and one floret comprised of one lemma, one palea and three kinds of floral organs (two lodicules, six stamens and one pistil). “Abnormal hull” refers to any glume abnormality and “Abnormal floral organ” refers to any abnormality in floral organs (e.g. an abnormal pistil). In the “Abnormal hull” class, there is a mutant that failed to close hulls after flower opening. On the other hand, mutants that could not open flowers due to underdeveloped stamens were often re-classified as “Completely sterile” mutants. Another mutant produced an extra glume. Abnormal flower phenotypes showing shoots growing from floral organs as in Fig. 1 (NC7672), were also detected at relatively low frequency.Panicle. The most frequently observed abnormal panicle phenotype was the precocious germination of seeds while still attached to the maturing panicle, or “Viviparous”. As for panicle shape, “Dense panicle” and “Short panicle” mutants (Fig. 1, NE6778) were also frequently observed, and “Long panicle” and “Lax panicle” were somewhat rare. Incomplete emergence of the panicle from the flag leaf sheath was categorized as “Neck leaf”.Sterility. “Sterile” is the third most frequent mutant in this Tos17 insertion mutant population after low seed fertility and dwarf mutants. The “Sterile” and “Low fertility” conditions correspond to lines with less than 2% and ca. 50% seed fertility, respectively.Seed. Only “Large grain”, “Small grain”, and “Slender grain” were distinguished in this category. The frequency of “Small grain” was two and four times higher than the frequency of “Large grain” and “Slender grain”, respectively. Among the “Others”, white or dull kernel phenotypes were often observed.Table 1Summary of phenotype dataClassPhenotypeID codeNCNDNENFNGTotal1GerminationLow germination rate13275311326100530034892GrowthLethal21353006892712341629Abnormal shoot31891722198233841787Weak411420434155639016053Leaf colorAlbino112643842302752541407Yellow129613918631771809Dark green 1329528535458691061Pale green 142704293722653951731Virescent1535278206131184834Stripe16464064132102384Zebra171215133929108Others18915232216854Leaf shapeWide leaf21382851421142Narrow leaf221653244322042511376Long leaf2333144529Short leaf2424258140Drooping leaf258140351231226Rolled leaf2631521532383342Spiral leaf27102355318109Brittle leaf/culm28514405310122Abnormal lamina joint angle29142713142896Withering3011227025188147868Others31375446113933435Culm shapeSemi-dwarf416458226616339003661Dwarf428031550135188211235709Severely dwarf432484113551432171374Long culm44258217506883676Fine culm45116118Thick culm46312421462Others47114286Spotted leaf/lesion mimicSpotted leaf/lesion mimic5111521119726930210947TilleringHigh tillering611127252127111Low tillering624186786966094212822Lazy63701155380563748Heading dateEarly heading6535297699297651789Late heading661115612472151101244Non-heading?6722423218969SpikeletAbnormal hull71412683652111508Abnormal floral organ721212843111721110PanicleLong panicle8113193329Short panicle82517942692104752Lax panicle831816152332104Dense panicle844118976543264Viviparous85854732671421381105Shattering861326Neck leaf873851481024171Abnormal panicle shape882143265792543311SterilitySterile9144110268186588823825Low fertility9228773127935299126121254212SeedLarge grain101557833425195Small grain10247959614544427Slender grain1032028182114101Others10497911178831433473469Lots NC through NG were harvested in 1997 through 2001, respectively. Numbers of mutant lines with the corresponding phenotypes are listedFig. 1Representative phenotypes of Tos17 insertional mutants. Phenotype and line designation are indicated on each photograph
Correlation between phenotypes
To understand how the 53 phenotypic descriptors are related, lines showing two or more abnormal phenotypes were selected, and a matrix with 53 rows and 53 columns of phenotypes with a number of lines showing each pair of phenotypes was devised. Values for the number of respective phenotypes were changed to “x” in the matrix, (“x” is ignored by the SOM program), and data were subjected to self-organizing map (SOM) analysis (Kohonen 1995). The SOM algorithm is used for visualization of multidimensional complex data using an unsupervised learning method based on a grid of artificial neurons. The 53-dimensional correlation data of phenotypes was reflected in a two-dimensional map (Fig. 2).
Fig. 2Self-organizing map for correlation of phenotypes. Lines that showed two or more abnormal phenotypes were analyzed by SOM. Distances between phenotypes indicate the magnitude of correlation of phenotypes. Col_Oth., Color Others; Lg_lf, Long leaf; L_joint, Abnormal lamina joint angle; Lg_pa, Long panicle; Ab_floral_organ, Abnormal floral organ; Sl_grain, Slender grain; Thick_cm, Thick culm. Grayscale levels of each node represents the distance between references
Topologies (not distances) of phenotypes on the SOM coordinates coincide with the correlation between phenotypes. Distance between two phenotypes is indicated by grayscale color on the SOM. For example, assume that phenotype A, B, and C appeared in 5000, 50, and 10 lines, respectively. Seven lines showed phenotype A and C. Three-lines showed phenotype B and C. The color between phenotype A and C should be darker than the color between phenotype B and C, because the line number of phenotype A lines is much greater than that of phenotype B lines, although the phenotype correlation of C with A is stronger than that with B. It is difficult to explain multidimensional data completely on a two dimensional map, but a SOM map indicates generally that neighboring phenotypes have relatively tight correlations. The SOM analysis in Fig. 2 shows that there are many pairs of phenotypes that are apparently correlated. That is, among 53 phenotypes, several phenotypes have a high probability to emerge simultaneously with other specific phenotypes. This phenotypic correlation could be caused by pleitropic expression of a single gene, or could reflect developmental causality of the two abnormalities. For example, “Long leaf” and “Long panicle” (lg_lf, lg_pa) located in the upper left corner on the SOM map in Fig. 2 have a very strong correlation. This result suggests that the gene involved in leaf length also affects panicle length. On the contrary, long leaf and dwarfism, at the opposite corner of the map, do not have any correlation. Both dwarf and semi-dwarf phenotypes have a relatively strong correlation with sterile and low fertile phenotypes. Since this correlation is not so strong, a part of “Dwarf” and “Semi-dwarf” mutants do not affect seed fertility, but other “Dwarf” and “Semi-dwarf” mutants may be regulated by genes associated with some housekeeping functions.
Correlations between mutant phenotype and Tos17 insertion
If allelic insertion lines show the same phenotype, it is highly probable that the mutant phenotype is caused by disruption of the target gene by a Tos17 insertion. We searched for genes that were disrupted by Tos17 in at least two lines. A total of 391 genes were detected based on the Rice Annotation Project (Ohyanagi et al. 2006). Table 2 is a partial list of such loci and their mutant phenotypes. For example, five lines have insertions of Tos17 in the Os06g027500 gene (Fig. 3A), of which two lines have the Tos17 insertions in exons and show an early heading phenotype. On the other hand, the other three lines have Tos17 insertions in introns and are not early heading. (see http://tos.nias.affrc.go.jp/). The Os06g027500 gene is Hd1, a key gene for determining heading date (Yano et al. 2000). Another example is the magnesium chelatase subunit gene, Os07g0656500 in Fig. 3B (Jung et al. 2003). Tos17 insertion into this gene was correlated with an albino phenotype, because mutants defective in this gene lack active chlorophyll and eventually become albino. All correlations between phenotypes and disrupted genes by Tos17 insertion are displayed on the mutant panel database website (http://tos.nias.affrc.go.jp/).
Table 2List of loci that have Tos17 insertions in exons in at least two linesLocus nameDescriptionPhenotypesOs01g0113200LRK14Pale green leaf, low fertilityOs01g0113300Receptor-like kinase ARK1ASDwarf, spotted leaf/lesion mimicOs01g0147800Protein of unknown function DUF547 domain containing proteinPale green leaf, semi-dwarf, long culm, short panicle, sterile, low fertileOs01g068590065 kD Microtubule associated proteinNarrow leaf, semi-dwarfOs02g05526008-Oxoguanine DNA glycosylaseLethalOs04g0464200Betaine-aldehyde dehydrogenase (EC 1.2.1.8) (BADH)Low fertilityOs04g0680400Allantoinase (EC 3.5.2.5)Dwarf, low fertilityOs05g0318600Protein kinase domain containing proteinNarrow leafOs05g0548900Phosphoethanolamine methyltransferaseEarly headingOs05g0552400Zn-finger, RING domain containing proteinEarly headingOs06g01768002OG-Fe(II) oxygenase domain containing proteinDark green leaf, dwarf, severely dwarfOs06g0275000Hd1Dwarf, early headingOs06g0680500Glutamate receptor 3.1 precursor (Ligand-gated ion channel 3.1) (AtGLR2)Low fertilityOs07g0197100HexokinaseDwarf, sterileOs07g0646500SWIM Zn-finger domain containing proteinLate headingOs07g0656500Protoporphyrin IX Mg-chelatase subunit precursorLethal, albino, dwarfOs09g0278300Phosphatidylinositol-4-phosphate 5-kinase family proteinSemi-dwarf, dwarf, low tillering, sterile, low fertilityOs10g0567100Chlorophyll b synthase (Fragment)Dark green leaf, pale green leaf, withering, semi-dwarf, severely dwarf, early heading, late heading, low fertilityOs12g0127600WRKY transcription factor 57Sterile, low fertilityOs12g0566000HCO3-Transporter domain containing proteinSeverely dwarfOs12g0572500Protein of unknown function XH domain containing proteinEarly headingLocus name and description are from RAP dataFig. 3Examples of genes that have insertions of Tos17. (A) Os06g0275000, Hd1 gene. (B) Os07g0656500, Magnesium chelatase subunit gene. The structure and insertion points of these genes are drawn on demand from our web site (http://tos.nias.affrc.go.jp/). Coding regions are indicated with broad lines. Positions of Tos17 insertions are indicated by flags. The direction of the flag is the direction of the Tos17 insertion. Detailed descriptions of insertion points, sequence names, accession numbers, and line names will appear on the web page. When a sequence name is clicked, the figure will be re-drawn with a flag labeling the sequence name to distinguish closely inserted Tos17s
A large-scale phenotypic characterization of an M2 population generated by Tos17 retrotransposition has revealed that this population contains a large number of mutants covering many easily scored phenotypes. In addition, linkage between the mutant phenotype and a specific Tos17 insertion facilitates greatly the isolation of the causal genes and the elucidation of the gene functions.
Discussion
Three-years of collaboration among seven laboratories has led to a large-scale phenotypic characterization of about 50,000 M2 plants generated by Tos17 retrotransposition. We have examined 53 kinds of abnormal phenotypes in rice from the seedling to harvest stages that are easily evaluated in the field. This project has revealed that this population contains a large number of mutants covering a wide range of phenotypes. Although several mutant phenotypes may be environment-sensitive and their expression may be unstable, most of the mutant phenotypes were stable. We also deposited additional information about certain phenotypes in a comment field within the phenotype description table, if this information was provided. A text search box on the phenotype list web page is envisioned to facilitate use of the resource.
Because M2 plants were observed, inserted Tos17s segregate among individuals. An average of 10 new copies of Tos17 are inserted into each regenerated plant. If the phenotype co-segregates with an insertion of Tos17, the phenotype is likely to be caused by disruption of the target gene. Of course, there are lines whose mutant phenotypes are not correlated with Tos17 insertion. These mutants might be caused by insertions of other native transposons, by chromosomal aberrations, or by other mutations during tissue culture. In this case, their causal genes could be isolated by usual positional cloning methods.
There were some difficulties in clustering phenotypes using standard relational algorithms, because each of the 53 phenotypes has as many as 52 kinds of frequencies against other phenotypes, respectively. SOM analysis is a kind of clustering method suitable for such multi-dimensional non-linear data (Kohonen 1995). SOM analysis confirmed some associations due to hormonal or developmental constraint, but some additional intriguing correlations were also detected, e.g., between leaf color/shape and heading date, between dwarfism and sterility. In Fig. 2, the “Long leaf” phenotype is strongly correlated with the “Long panicle” phenotype. The “Weak” phenotype at the seedling stage has a relatively strong correlation with “Lethal”, “Abnormal shoot”, “Pale green” Leaf and “Spotted leaf/lesion mimic” phenotypes, indicating that the “Weak” plants at the seedling stage likely show “Lethal”, “Abnormal shoot”, “Pale green” and/or “Spotted leaf/lesion mimic” phenotypes at later stages. “Dwarf”, “Semi-dwarf”, “Sterile”, “Low fertile” and “Low tillering” phenotypes showed relatively high correlations. The SOM map is quite useful for the overview of phenotype correlations. This method of analysis will be useful not only for considering the major and side-effects of gene disruption but also for re-categorization of phenotypes. The advantage of SOM analysis is the ability to analyze many different kinds of data simultaneously. Currently, metabolome data have been obtained from Arabidopsis (Hirai et al. 2004), tomato (Schauer et al. 2006) and other plants. It is reasonable to predict that the combinatorial analysis of phenotype and metabolome data of rice will reveal correlations between metabolic pathways and phenotypes without additional genetic information.
Phenotype databases of rice have been developed in many countries, e.g., Oryzabase (Kurata and Yamazaki 2006), RMD (Zhang et al. 2006), and IRRS (Wu et al. 2005). Our system includes both phenotypic and insertion data on the rice genome in a relational database. The number of photographs on file is more than 58,000, and the number of phenotype description records is more than 158,000. Our database structure enables direct access of Perl script to the database and the extraction of many kinds of correlations. Flanking sequences of insertion points and phenotypes of the insertion lines provide a provisional assignment of the function of a disrupted gene. We have already sequenced more than 25,000 flanking sequences from 20% of the insertion lines, and the number of flanking sequences is still increasing. Of the 27,448 total annotated loci based on RAP1/IRGSP Build3, 391 loci have more than two insertions in exons. When two or more lines have Tos17 insertions in a common gene, they usually exhibited similar phenotypes. These data are a strong indication of the function of a disrupted gene. However, several phenotypes such as “Dwarf” and “Low fertile” were observed in many lines. It might be difficult to assign a gene function based solely on correlation for these frequently observed phenotypes, and a complementation test would be required to confirm the correlation. If flanking sequence data from all insertion lines can be obtained, this correlation will be more useful for the annotation of genes. We are continuing to sequence the flanking regions of all our mutant lines.
In this study, we collected a large amount of phenotypic data in seven fields under natural conditions. If phenotypes are observed under other conditions such as drought or temperature stress conditions or under pathogen pressure, phenotypic description of this population would be much more enriched. In addition, the present study evaluated only a limited number of traits easily scored on above-ground organs. Thus, other traits such as roots and seed storage composition remain to be investigated in the future. Expansion of this study to include new traits would enable investigators to find new correlations with disrupted genes. Furthermore, integration of phenotype data with those of microarray experiments, metabolic profiling, and other approaches will be a powerful tool for revealing new aspects of plant physiology.
All phenotype and flanking sequence data can be obtained via http://tos.nias.affrc.go.jp/. At this site, all annotated rice genes and locations of Tos17 insertions are shown on a clickable chromosome map. Details containing illustrated gene structures with insertion points, phenotypes of corresponding lines, nucleotide sequences flanking the disrupted region, candidate primer sequences for segregation analysis, and annotation from RAP data corresponding to the selected position are displayed. Mutant seeds are available for scientific use from the Genome Resource Center at NIAS (http://www.rgrc.dna.affrc.go.jp/).
|
[
"phenotyping",
"retrotransposon",
"database",
"oryza sativa",
"insertion mutagenesis"
] |
[
"P",
"P",
"P",
"P",
"M"
] |
Pituitary-3-1-2045692
|
Acromegaly caused by growth hormone-releasing hormone-producing tumors: long-term observational studies in three patients
|
We report on three newly diagnosed patients with extracranial ectopic GHRH-associated acromegaly with long-term follow-up after surgery of the primary tumor. One patient with a pancreatic tumor and two parathyroid adenomas was the index case of a large kindred of MEN-I syndrome. The other two patients had a large bronchial carcinoid. The first patient is still in remission now almost 22 years after surgery. In the two other patients GHRH did not normalize completely after surgery and they are now treated with slow-release octreotide. IGF-I normalized in all patients. During medical treatment basal GH secretion remained (slightly) elevated and secretory regularity was decreased in 24 h blood sampling studies. We did not observe development of tachyphylaxis towards the drug or radiological evidence of (growing) metastases. We propose life-long suppressive therapy with somatostatin analogs in cases with persisting elevated serum GHRH concentrations after removal of the primary tumor. Independent parameters of residual disease are elevated basal (nonpulsatile) GH secretion and decreased GH secretory regularity.
Introduction
Acromegaly caused by ectopic extracranial growth hormone-releasing hormone (GHRH) secretion is a very rare disorder occurring probably in less than 1 % of the acromegalic patients [1]. The majority of the GHRH-secreting tumors are bronchial carcinoids. Other GHRH-secreting tumors in decreasing order of occurrence are pancreatic adenomas, gastro-intestinal tumors, thymic tumors, and tumors associated with the MEN-I syndrome [2].
Here we report our clinical experience in three patients with the ectopic GHRH-syndrome derived from a population of over 200 acromegalic patients, diagnosed and treated at the Leiden University Medical Center from 1976 till 2002. Long-term follow-up studies in these patients and results of medical therapy with somatostatin analogs are scarce. The purpose of this report is to expand our knowledge of this rare clinical entity. In addition, we report detailed results of diurnal GH secretion, before and after removal of the GHRH-producing source in order to investigate whether pulsatile and basal GH secretion due to GHRH overproduction differs from that of a primary pituitary somatotropinoma.
Methods
Basal concentrations of hormones, including prolactin, free thyroxin, triiodothyronine, cortisol, testosterone, estradiol, progesterone, IGF-I, and IGFBP3 were measured. In addition the following tests were performed: Oral glucose loading test (75 g), TRH test (200 μg i.v. ), and a GHRH test (50 μg i.v. ) and the following hormones were measured: glucose tolerance test: GH, insulin and glucose at 0, 30, 60, 90 and 120 min, TRH test: TSH, prolactin and GH at −15, 0, 15, 20, 30 45, 60, 90 and 120 min; GHRH test: GH and prolactin at 0, 20, 30, 45, 60, and 90 min. For the 24-h GH secretion profile the patients were hospitalized, and an indwelling i.v. cannula was inserted in a forearm vein, and blood samples were withdrawn at 10-min intervals. The patients were free to move around, but not to sleep during daytime. Meals were served at 0800, 1230 and 1730 h. Lights were turned off between 2200 and 2400 h.
Assays
Plasma GH was measured with a sensitive time-resolved fluoro-immunoassay (Wallac Oy, Turku, Finland). The assay is specific for the 22 kDa GH. The standard was biosynthetic recombinant human GH (Genotropin, Pharmacia & Upjohn, Uppsala, Sweden), and was calibrated against the WHO First International Reference Preparation 80/505 (to convert μg/l to mU/l multiply by 2.6). The limit of detection of this assay (defined as the value 2 SD above the mean value of the zero standard) was 0.01 mU/l (0.0038 ng/ml). The intraassay coefficient of variation varied between 1.6% and 8.4% in the range from 0.01 μg/l to 18 μg/l and interassay coefficient of variation was 2.0–9.0% in the same range.
Total IGF-I was determined by RIA (Incstar, Stillwater, MN) after extraction and purification on ODS-silica columns. The intraassay coefficient of variation was less than 11%. The detection limit was 1.5 nmol/l. Age-related normal data were determined in the same laboratory. The measurement of IGFBP3 was performed by RIA (Nichols Institute Diagnostics, San Juan Capistrano, CA). The limit of detection of this assay was 0.08 mg/l, and the interassay coefficient variation was below 6.8%.
Deconvolution analysis
A multiparameter deconvolution technique was used to estimate relevant measures of GH secretion from the 24-h serum GH concentration profiles, as described previously [3]. Initial estimates of basal GH secretion rate were calculated to approximate the lowest 5% of all plasma GH concentrations in the time series. Peak detection entailed application of 95% statistical confidence intervals to two thirds of all GH secretory peaks considered jointly and individual 95% statistical confidence intervals to the remaining one third smaller pulses, as validated in simulations [4]. The following four secretory and clearance measures of interest were estimated: (1) the number and locations of secretory events; (2) the amplitudes of secretory bursts; (3) the durations of randomly dispersed GH secretory bursts; and (4) the endogenous single component subject specific plasma half-life of GH. It was assumed the GH distribution volume and half-life were time and concentration invariant. The following parameters were calculated: Half-duration of secretory bursts (duration of the secretory burst at half-maximal amplitude), hormone half-life, burst frequency, amplitude of the secretory burst (maximal secretory rate attained within a burst), mass secreted per burst, basal secretion rate, pulsatile secretion rate (product of burst frequency and mean burst mass) and total secretion (sum of basal and pulsatile).
Approximate entropy
The univariate approximate entropy (ApEn) statistic was developed to quantify the degree of irregularity, or disorderliness, of a time series [5]. Technically, ApEn quantifies the summed logarithmic likelihood that templates (of length m) of patterns in the data that are similar (within r), remain similar (within the same tolerance r) on next incremental comparison and has been formally defined elsewhere [6]. The ApEn calculation provides a single non-negative number, which is an ensemble estimate of relative process randomness, wherein larger ApEn values denote greater irregularity, as observed for ACTH in Cushing’s disease, GH in acromegaly, and PRL in prolactinomas [7–9]. In the present analysis, we calculated ApEn with r = 20% of the SD of the individual time-series and m = 1. This choice of parameters affords sensitive, valid and statistically well-replicated ApEn metrics for assessing hormone time-series of this length. ApEn results are reported as absolute values or as the ratio of the absolute value to that of the mean of 1,000 randomly shuffled data series. Ratio values that approach 1.0 thus denote mean empirical randomness.
Copulsatility
Copulsatility between the GHRH and GH time-series was quantified by the hypergeometric (joint binomial) distribution [10]. This program calculates the probability that hormone pulses in time-series occur randomly. We used a time-window of 20 min, with GHRH as leading hormone series. Because details of the secretion characteristics of GHRH are not well established, we estimated significant pulses in both time-series with Cluster, which is largely model-free [11].
Clinical findings at diagnosis and initial treatment
Case 1. A 50-year-old male was referred in 1982 because of recurrent kidney stones, hypercalciuria (24 h urinary calcium excretion between 14.5 mmol and 16.3 mmol, normal upper value 6 mmol/24 h) and hypercalcaemia (serum Ca between 2.86 mmol/l and 2.95 mmol/l, normal values between 2.25 mmol/l and 2.60 mmol/l). The referring internist suspected the patient of having mild acromegaly, because of the coarse facial features. The diagnosis primary hyperparathyroidism was confirmed and the patient underwent parathyroid surgery, and two large adenomas were removed, after which the patient became normocalcaemic until now. The patient had noted increase in size of his feet and hands since several years, but otherwise he had no complaints. Glucose loading decreased GH from 12 mU/l to below 0.5 mU/l (normal value during glucose suppression is below 2.5 mU/l with RIA). Intravenous administration of 200 μg TRH, however, increased GH from 3.5 mU/l to 58.0 mU/l and PRL increased from 7.0 μg/l to 13 μg/l. CT scanning of the pituitary gland with the first generation CT-scanner (in 1983) did not show abnormalities, and CT scanning of the thorax and abdomen also failed to show the presence of a tumor. Because of progressive complaints of fatigue after cure for hyperparathyroidism and strong clinical suspicion of acromegaly, the patient underwent transsphenoidal pituitary exploration. At surgery a small suspect lesion was removed with part of the surrounding pituitary gland tissue. Histology of the lesion was compatible with somatotrope hyperplasia. After the patient had recovered from surgery, CT, MRI and arteriographic studies were repeated, and serum samples were sent to St. Bartholomew’s Hospital, London, UK (Dr L.H. Rees) for GHRH measurement. The abdominal CT-scan showed a 4 cm mass in the middle section of the pancreas (Fig. 1) and the mass was also visible with selective arteriography of the superior mesenterial artery (Fig. 1). The fasting GHRH concentration amounted to 3,810 pg/ml (normal range 10–60 pg/ml). GHRH concentration in the arterial supply to the tumor was 12,470 pg/ml, and in the venous tumor outflow 31,120 pg/ml, while in the systemic venous system the concentration was 8,900 pg/ml. After removal of the pancreatic tumor, peripheral GHRH concentration decreased to normal values of 16–33 pg/ml. In retrospect, the first abdominal CT-scan, 2 years before, already showed the pancreatic tumor with a similar size.
Fig. 1Upper panel shows the abdominal CT at the level of the pancreas and the lower panel the selective arteriography of the superior mesenterial artery
The patient was the index case of a large kindred affected by MEN-I syndrome. Later, a gene mutation was revealed in exon 2 of chromosome 11q13. During long-term follow-up GH and IGF-I concentrations remained normal.
Six years later, in 1988, the patient developed diabetes mellitus, initially treated with oral hypoglycemic drugs and with insulin from 1992 onwards. The patient also had mild bilateral nodular adrenal hyperplasia since 1988, with no evidence of growth during the last recent 10 years. No excess of adrenal (cortex and medulla) hormones or precursors was demonstrable during follow-up. During the last 3 years the plasma level of pancreatic polypeptide (PP) increased to 350 nmol/l (normal < 100 nmol/l). Repeat CT and MRI scanning of the upper abdomen failed to reveal the presence of a pancreatic tumor thus far.
Random serum GH during follow-up ranged from 0.16 mU/l to 1.81 mU/l and IGF-I ranged from 12 nmol/l to 17 nmol/l from 1994 till now (see Fig. 2). All these values are perfectly normal for his age and gender. The last serum GHRH measurement in 2006 was normal with 32 pg/ml.
Fig. 2Follow-up of GH and IGF-I concentration in patient 1 after removal of the pancreatic source of GHRH. Fasting serum GH concentrations are depicted as circles, and IGF-I as triangles
Case 2. A 27-year-old acromegalic female patient was referred in 1993 to our center for octreotide treatment. She had a 5-year history of hyperhydrosis, fatigue, paraesthesias, and acral enlargement. After delivery of a healthy daughter in 1992 she had persisting amenorrhea and galactorrhea. Physical examination revealed mild, although characteristic features of acromegaly. Serum GH concentration was elevated at 99 mU/l and decreased insufficiently to 55 mU/l following oral glucose administration. IGF-I concentration amounted to 86 nmol/l (normal upper level for her age 32 nmol/l) Thyroid and adrenal functions were normal, but prolactin concentration was elevated to 20 μg/l (normal upper limit for females 12 μg/l). Other investigations performed (TRH test, GHRH test, and the i.v. octreotide test) are summarized in Fig. 3. TRH bolus injection (200 μg iv) caused a ∼8-fold increase of GH and a 3.5-fold increase of PRL. Intravenous injection of 50 μg GHRH (1-40) caused a decrease of GH while PRL concentrations remained unchanged. Octreotide (50 μg i.v.) caused a 98% inhibition of serum GH but no effect on PRL. MRI scanning of the pituitary gland showed global enlargement without signs of an adenoma (Fig. 4). Further investigations revealed a large tumor in the right lower lobe of the lung (Fig. 5). The tumor and the pituitary gland were positive on 111In-labeled-octreotide scintigraphy (Fig. 6). The plasma GHRH concentration was increased 50-fold by 2,519 pg/ml (normal values <50 pg/ml).
Fig. 3Dynamic GH (circles) and PRL (triangles) tests in two patients with a GHRH-secreting lung carcinoid before (closed symbols) and after surgery (open symbols). Note the GH increase after TRH (200 μg) administration in both patients, the decrease in serum GH concentration in patient 2 after GHRH (50 μg), and the moderate GH decrease after i.v. octreotide (50 μg)Fig. 4MRI scans of the pituitary gland of patient 2 (left panel) and of patient 3 (right panel) before treatmentFig. 5Chest X-rays of patient 2 (left panel) and patient 3 (right panel), showing the large bronchial carcinoidFig. 6Octreoscans of the patients 2 and 3. In the female patient (left panel) the tumor is seen in the right lower lobe of the lung and also the positive staining of the pituitary gland. The male patient (right panel) had a large tumor in the left lung, without pituitary staining
Under the diagnosis of ectopic GHRH-producing lung tumor (carcinoid) the patient underwent thoracotomy and the tumor was resected completely. The plasma GHRH disappearance profile after removal of the tumor is shown in Fig. 7, but the GHRH concentration did not normalize. GHRH concentrations in simultaneously withdrawn blood samples from the arterial supply to the tumor, venous tumor outflow and the systemic venous system were 1,890, 2,180, and 1,680 pg/ml, respectively.
Fig. 7Decrease in serum GHRH concentration in two patients before and immediately following removal of the lung tumor. Patient 2 is shown by triangles and patient 3 by circles. Note that the GHRH concentration is shown on a logarithmic scale. The GHRH concentration did not normalize in the patients after 2 and 3 days, respectively. Upper normal GHRH is 50 pg/ml
Two weeks after surgery, endocrine investigations revealed a persisting paradoxical increase (9-fold) of GH to TRH, insufficient GH suppression after oral glucose loading (52–16 mU/l) and a GH increase after GHRH injection. Octreotide caused a 98% decrease of GH from 41 mU/l to 0.9 mU/l (see Fig. 3). Treatment with octreotide was started in 1994, about 8 months after surgery, because of persisting elevated GHRH concentration (between 360 pg/ml and 488 pg/ml), increased GH( 31–42 mU/l), insufficient suppression by glucose loading (minimum GH concentration 9.13 mU/l), an increased IGF-I concentration ( 40–43 nmol/l), and lesions in the liver on repeat CT scans, suspect for metastases (one lesion in segments 2 and 8, and two lesions in segment 7).
Case 3. A 27-year-old male was referred in 1997 to our center. He had a 9-year history of hyperhydrosis, acral enlargement and headaches. Examination revealed moderately advanced features of acromegaly. The circulating GH concentration was elevated at 110 mU/l and decreased insufficiently to 52 mU/l following glucose administration, and IGF-I was elevated to 63 nmol/l (normal upper level at this age 32 nmol/l). Thyroid and adrenal functions were normal, but prolactin concentration was elevated to 21 μg/l (normal upper limit for males 6 μg/l). Other investigations performed (TRH test and the i.v. octreotide test) are summarized in Fig. 3. TRH bolus injection (200 μg i.v.) caused a ∼8-fold increase of GH and a 2-fold increase of PRL. Octreotide (50 μg i.v) caused a 94% inhibition of GH (Fig. 3). MRI scanning of the pituitary gland showed a macroadenoma II AE (Hardy classification, modified by Wilson [12, 13] see Fig. 4). He was treated for 6 months with sc octreotide (100 μg tid). The size of the adenoma decreased slightly, but because of the moderate clinical response to medical treatment and failure of normalization of GH and IGF-I, pituitary surgery was advised. One day before surgery a chest X-ray was taken, showing a large parahilar lesion (Fig. 5). CT scanning of this lesion was classified by the consultant pulmonologist and radiologist as a bronchial cyst and apparently unrelated to acromegaly. During pituitary surgery the adenoma was removed completely, but the neurosurgeon remarked that the consistency of the adenoma was firmer than normal. Since we suspected that the lung tumor was potentially a GHRH-producing source, postoperative investigations were focused on this possibility. After surgery, MRI scanning of the pituitary region did not show residual tumor, GH concentrations decreased considerably and PRL concentrations became normal. GH dynamic tests did not completely normalize: GH increased after TRH injection from 4.63 mU/l to 22.1 mU/l and after glucose loading GH decreased from 4.33 mU/l to 2.56 mU/l (normal value below 1 mU/l). During an i.v. octreotide test GH decreased from 6.21 mU/l to 0.85 mU/l, but treatment with this drug was not reinstituted. On 111In-labelled-octreotide scanning the tumor was positive (see Fig. 6).
Under the diagnosis of ectopic GHRH-producing lung tumor (carcinoid) the patient underwent thoracotomy and the tumor was removed completely. GHRH concentrations in the arterial supply to the tumor were 48,290 pg/ml, in the venous outflow 94,000 pg/ml and in the systemic venous system 49,000 pg/ml. After removal of the tumor GHRH concentrations remained slightly elevated, as shown in Fig. 7.
Histopathological studies
Pituitary gland
The removed part of the anterior pituitary gland of patient 1 consisted of hyperplastic cells, immunostaining positively for GH. The removed tissue of the third patient consisted of a mixture of hyperplasia and adenoma formation. The cells stained positively for both GH and PRL.
GHRH-producing tumors
Patient 1. The pancreatic tumor had a diameter of 5 cm. Amorphous material was present between the cells, staining as amyloid. On electronmicroscopy, the cells contained neurosecretory granules with a diameter between 100 nm and 200 nm. The tumor stained positively, but sparsely for somatostatin, insulin and glucagon and negatively for cytokeratine, vimentine, neurofilaments, desmine and GH. In the removed part of the pancreas three additional small adenomas with identical staining characteristics were present.
Patient 2. The diameter of the removed lung tumor was 5 cm, and contained centrally calcified material. The cells were layered in nests, slightly polymorphic, but without mitotic figures. The tumor cells stained positively for keratine, vimentin, synaptophysin, SCCL (N-CAM), leu 7, and chromogranin and negatively for calcitonin, GH, pancreatic polypeptide, insulin, prolactin, somatostatin, gastrin, ACTH, CEA, and neurofilaments.
Patient 3. The dimensions of the tumor were 8 × 7×7 cm3. The tumor showed clear proliferation of neuroendocrine cells with three mitotic figures per high power field, staining positively for NSE, CD56, and synaptophysin and negatively for keratine, chromogranin, serotonin, somatostatin, prolactin, insulin, glucagons, gastrin, ACTH, GH, and insulin.
Somatostatin analog therapy
After removal of the lung carcinoid, patients 2 and 3 received long-term treatment with octreotide, because GHRH was not normalized. Patient 2 received the medication via chronic sc infusion (300 μg/24 h) till the end of 1998. Her complaints quickly disappeared and the menstrual cycle was restored. GH and IGF-I concentrations normalized and are detailed in the left panel of Fig. 9. The size of the pituitary gland decreased markedly (Fig. 8). PRL concentration also normalized from 15.2 ± 0.2 μg/l to 6.8 ± 0.2 μg/l. From 1999 onwards the medication was changed to octreotide long-acting repeatable (Sandostatin LAR), 20 mg in 4-weekly i.m. injections. Growth hormone and IGF-I concentrations remained unchanged (see Fig. 9). During chronic treatment with the short-acting octreotide formulation treatment was withheld several times for GHRH measurement. Invariably, the concentration was (slightly) elevated (range 116–363 ng/ml) so that in combination with the radiological suspicion for liver metastases, treatment is continued until now. During treatment with octreotide long-acting repeatable (Sandostatin LAR) GHRH concentration ranged from 63 ng/ml to 108 ng/ml. Repeat CT scanning of the liver during the successive years demonstrated the stabilization of size and number of lesions until now.
Fig. 8MRI of the pituitary gland of patient 2 during therapy with octreotide. These pictures were taken after 12 months treatment with 300 μg octreotide given as a continuous subcutaneous infusionFig. 9GH (circles) and IGF-I (triangles) concentrations during long-term treatment with octreotide. Patient 3 received only the long-acting repeatable form, but patient 2 was treated initially with chronic sc octreotide infusion. The time of change into the slow-release formulation is indicated by the arrow. Normal values for IGF-I for this age: <32 nmol/l. Normal value for random GH < 5 mU/l
After removal of the carcinoid tumor in patient 3, GHRH, GH, and IGF-I concentrations remained elevated, and therefore he was also treated with the long-acting repeatable octreotide (20 mg/4 weeks). GH and IGF-I concentrations are shown in Fig. 9, right panel, showing clinically normal values. Repeat investigations with CT, 111In-labeled octreotide and 131I-MIBG however, did not reveal suspect (liver) metastases. At the end of 2001, GHRH was elevated to 1,725 ng/ml, so that thereafter the dose of Sandostatin was increased to 30 mg at 4-weekly intervals, which normalized IGF-I concentrations. Last year the patient stopped medication. Subsequently, IGF-I levels increased, as shown in Fig. 9. GHRH concentration became larger than 2,000 pg/ml. Detailed localizing studies with octreotide and CT revealed a small metastasis in the superior anterior mediastinum for which surgery is scheduled.
Fig. 10Serum GH concentrations obtained by 10 min blood sampling for 24 h. Patient 2 was studied before therapy and after surgical removal of the lung tumor. Note that GH concentration decreased more than 10-fold and that the secretion pattern became more regular, but basal GH concentration remained slightly elevated. The left lower panel represent the profile of patient 3 after pituitary surgery, but before removal of the carcinoid tumor. Nadir values were clearly increased. After thoracic surgery and under octreotide treatment GH secretion pattern visually normalized
GH secretory profiles
Detailed GH secretory profiles were obtained before removal of the GHRH-producing bronchial carcinoids (Fig. 10). The secretory patterns were irregular, showing increased burst frequency and increased basal concentrations. The GH secretory parameters as estimated by multiparameter deconvolution are listed in Table 1 with normal values obtained in healthy adults of comparable age. The distinct and persisting abnormality in both patients after removal of the carcinoid and while on octreotide treatment was the increased basal (nonpulsatile) GH secretion.
Table 1Deconvolution of the 24 hour serum GH profiles in patients with ectopic GHRH syndrome and controlsPatient 2 before surg.Patient 2 after surg.Female controlsPatient 3 before surg. Patient 3 after surg. Male controlsJaffe’s patientVance’s patientPulse frequency (no/24 h)301617 (14–21)302412 (7–14)3021Half-life (min)15.515.912.9 (12.0–15.5)15.615.717.5 (15.2–19.7)24.616.2Pulse half-duration (min)20.932.927.7 (22.7–29.8)25.323.225.3 (19.5–34.4)30.541.3Pulse height (mU/l /min)4.920.4420.387 (0.198–0.881)0.2270.1600.141 (0.087–0.422)6.01.87Pulse mass (mU/l)10114.410.2 (7.3–18.5)5.663.684.44 (2.42–9.55)18176.5Basal secretion(mU/l /24 h)2,26130.29.1 (5.5–17.6)18.014.93.4(1.4–5.3)3,9004,60Pulsatile secretion (mU/l /24 h)3,047230173 (122–312)1708843.5 (17.8–103)5,4001,600Total secretion(mU/l /24 h)5,309260182 (132–325)18810347 (19.7–107)9,3002,060Blood samples were taken at 10-min intervals for 24-h and analyzed by multiparameter deconvolution. The female patient (no. 2) was studied before surgical removal of the GHRH-secreting bronchus carcinoid and repeat sampling study was done after thoracic surgery under octreotide LAR. The male patient (no. 3) was studied first after adenomectomy of the pituitary tumor, but before thoracic surgery. The second sampling study was performed after removal of the bronchial carcinoid during octreotide-LAR treatment. The serum profiles of the patients reported in literature were digitized and deconvoluted with the assay precision according to the authors. The GH data were subsequently transformed from μg-mass units into mU using the conversion factor 2.0. Reference values were obtained in nine males and 10 females healthy controls. Values shown are medians and 95% confidence intervals between brackets
In addition, the secretory regularity was quantified with the approximate entropy statistic, ApEn. In patient 2, ApEn was 1.256 before removal of the carcinoid, and after surgery and under somatostatin analog treatment ApEn was still increased to 0.686 (median normal for women 0.400, 95% confidence interval 0.300–0.440). In patient 3 ApEn also remained abnormal: preoperative 1.256 and after surgery 0.687, median normal for males 0.240, 95% confidence interval 0.160–0.350 (see Table 2). In addition, the serum GH profiles of two patients reported in literature were digitized and analyzed in a similar way [14, 15]. The results of these analyses are also displayed in Table 1. In these male patients basal GH secretion was much higher than in our healthy controls and pulsatile secretion was augmented via increased pulse frequency and pulse amplitude. ApEn of GH secretion was 1.533 in Jaffe’s patient and 1.248 in the patient reported by Vance (increased SD scores by 8- and 6-fold, respectively). ApEn for the serum GHRH-time series were 1.759 and 1.223, respectively. Copulsatility of the GH and GHRH hormone series was highly significant in both patients (P < 0.0001).Table 2Approximate entropy of GH secretion in ectopic GHRH syndromePatientBefore removal of the ectopic GHRH sourceAfter surgery and during octreotide treatmentReference values, Median and 95% CINo. 2 (female)1.2560.6860.400 (0.300–0.440)No. 3 (male)0.8420.5610.240 (0.166–0.350)Jaffe’s patient1.5330.240 (0.166–0.350)Vance’s patient1.2480.240 (0.166–0.350)Calculations were performed on GH data series consisting of 145 samples obtained at 10 min intervals during 24 h. Normal values were derived from nine males and 10 females healthy controls
Discussion
In this study we described in detail the clinical and biochemical characteristics of three patients with ectopic extracranial GHRH secretion diagnosed in a series of about 200 acromegalics investigated and treated in our center during the last 25 years. The incidence in the present series agrees with that mentioned in literature [1, 16]. Faglia summarized the clinical findings of 39 reported acromegalic patients with proven ectopic GHRH secretion [17]. Subsequently, van den Bruel and colleagues [18] mentioned 52 reported patients in 1999, including their own patient and since then 14 other patients have been reported, bringing the total number reported to 66 patients [19–32]. From a conservative estimation of the total number of newly diagnosed acromegalic patients in Europe, Japan and the USA, it is evident that most patients with ectopic GHRH syndrome are either not reported or remain undiagnosed.
The criteria for demonstration of ectopic extracranial GHRH-induced acromegaly are summarized by Losa and von Werder [2], and include the presence of high circulating concentration GHRH by specific radioimmunoassays, the presence of GHRH in the tumor, the presence of mRNA for GHRH by in situ hybridization and/or a significant arterio-venous gradient across the ectopic source. The second requirement to be fulfilled is the reversibility of acromegaly after complete removal of the ectopic-hormone producing tumor. All our patients met at least two criteria, although the tumors were not investigated for the presence of GHRH.
The clinical symptomatology in ectopic GHRH-induced acromegaly is not different from that of the primary pituitary adenomatous form. However, symptoms due to the underlying neoplasm or cosecretion of other substances by the tumor might suggest the ectopic origin of acromegaly [33]. Specific dynamical tests for GH excess do not allow classification with certainty in either category, although most patients with ectopic GHRH syndrome exhibit a paradoxical increase of GH after TRH and glucose (i.e. >50%) and a blunted GH rise (<100%) after exogenous GHRH injection [33]. In addition, most patients also exhibit a moderate increase in serum prolactin concentration, which regresses after removal of the ectopic GHRH source, supporting the notion that GHRH is directly responsible for the hyperprolactinemia, a finding also present in hGHRH-transgenic mice [34]. The patients reported here all exhibited GH increase after TRH, and two had mild hyperprolactinemia. After surgery, hyperprolactinemia normalized, but GH still increased after TRH administration, suggesting that GHRH was still being produced by tumor remnants or metastases in two patients (nos. 2 and 3).
The most frequent source of ectopic GHRH is the bronchial carcinoid, followed by pancreatic islet tumors [17], as we also found in the present evaluation of our patient series. The pancreatic tumor in the first patient was found by CT-scanning and MRI. The large bronchial carcinoids were easily seen by chest X-ray examination, and showed uptake of radio-labeled octreotide, thus demonstrating in vivo the presence of somatostatin receptors. Although positive somatostatin receptor-scintigraphy is usually found in carcinoids in general, its demonstration in ectopic GHRH tumors is limited. Nevertheless, this examination could be particularly useful in the diagnostic phase of patients with suspected, but not yet proven, ectopic GHRH secretion.
Histological investigation of surgically removed pituitary tissue revealed pituitary hyperplasia in most cases (i.e. with an intact reticulin fiber network), but in other patients adenomatous transformation can be found [35]. Indeed, the histological findings in hGHRH transgenic mice resemble those in the human: in young animals diffuse hyperplasia was found, while those of older mice showed adenomas [34, 36]. The pathological changes in the pituitary gland are the result of GHRH per se and not of GH, since hyperplasia and tumor formation still occur in the absence of GH signaling [37]. A hypothalamic GHRH-producing tumor (gangliocytoma) is generally associated with a pituitary adenoma rather than with pituitary hyperplasia and is attributed to a much higher (assumed) local (pituitary) concentration of GHRH or the presence of other growth stimulating factors [38]. In addition, the combination of an intrasellar GHRH-containing gangliocytoma and a somatotropinoma has been described in rare patients [39, 40]. In two of our patients pituitary histology was available: one patient (no. 1) showed hyperplasia in the absence of an enlarged gland, while histology in the other patient (no. 3) showed a mixture of hyperplasia and adenomatous transformation.
Neuroimaging studies of the pituitary gland are variable: in about half of the patients no tumor or a slight enlargement of the sella is detected, while in other patients intrasellar adenomas or macroadenomas with suprasellar extension are found [18]. The CT-scan of patient 1 was normal and MRI scanning in patient 2 showed diffuse enlargement of the pituitary gland without a clear adenoma and with diffuse uptake of gadolinium DTPA. In patient 3 asymmetric enlargement was present, suggestive of a macroadenoma, but uptake of gadolinium DTPA throughout the whole pituitary gland was diffuse, illustrating the variability of imaging results. Although the pituitary MRI studies in the ectopic GHRH syndrome are non-specific, in the absence of a clear adenoma, ectopic GHRH secretion should be considered and appropriate investigations installed to prevent unnecessary pituitary surgery. On the other hand, primary medical treatment for acromegaly in these patients might obscure the real cause.
Carcinoids, especially of the lung and gastrointestinal tract constitute about two-thirds of all cases reported so far, followed by pancreatic islet cell tumors [17, 33]. Metastases of carcinoid tumors were present in about 30%, a frequency similar to lung carcinoids in general [35]. Two of our patients had a large primary lung carcinoid, but had no histological proven metastases. Because serum GHRH concentration did not normalize after removal of the GHRH source we assumed (liver) metastases.
Immunochemistry of the GHRH-secreting tumors has regularly demonstrated cosecretion of other hormones, including gastrin, gastrin-releasing peptide, calcitonin, pancreatic polypeptide, VIP, glucagon, insulin, and somatostatin, pointing to the multihormonal nature of these tumors. Particularly, cosecretion of somatostatin may modify the clinical picture and severity [31–44].
Only few patients with a GHRH-producing pancreatic islet cell tumor and with MEN I syndrome have been described [41, 43, 44]. In agreement with our patient 1, they had multiple tumors; however, the pancreatic tumors in our patient showed an identical immunohistochemical profile in contrast with the other reported patients. The tumor stained positively for somatostatin, which might explain the only moderately increased GH concentration and possibly also the normal GH response to glucose loading, as alluded to above. In the follow-up of more than 20 years, no recurrence of acromegaly occurred, but the (modest) increase of pancreatic polypeptide could point to the development of a new pancreatic adenoma.
GH secretory characteristics of two patients were investigated before removal of the GHRH source. We established that pulsatile and non-pulsatile (basal) secretion, pulse frequency and secretory regularity (ApEn) were comparable to untreated acromegalic patients with primary pituitary adenomas [8, 45]. Because of the highly significant copulsatility of GHRH and GH in two other patients reported in literature, it is reasonable to postulate that pulsatile GH secretion in these patients was generated via episodic GHRH secretion of the primary tumor, while in patients with a pituitary GH-secreting adenoma pulsatile secretion is probably a tumorous feature [8].
Approximate entropy calculations of serum GHRH profiles disclosed high values, suggesting highly irregular secretion, a feature common with many other hormone-secreting tumors [7–9]. It is reasonable to assume that the irregular secretion of GH in the four patients who were investigated till now was mainly the result of the irregular GHRH-input signal, and analogous to the irregular cortisol secretion driven by ACTH in pituitary-dependent Cushing’s syndrome [7, 46]. Nevertheless, part of the irregularity of GH secretion might also be due to increased GHRH per se, since GHRH infusion in healthy subjects augments irregular GH secretion [47]. After surgery and on effective octreotide therapy aimed at normalizing IGF-I, GH secretion became almost normal. GH secretion remained however irregular (increased ApEn), notwithstanding the increased feedback signal of octreotide, and the diminished forward input signal of GHRH by octreotide [48, 49]. Similarly, octreotide treatment in (classical) acromegaly represses secretory-burst mass and non-pulsatile secretion but does not restore event frequency or orderly GH secretion [50]. Collectively, these results suggest that even a modestly increased GHRH concentration can maintain irregular GH secretion in ectopic acromegaly.
Therapy
Surgical removal of the GHRH source is obviously the treatment of choice and results in the resolution of acromegaly when no metastases are present, as we observed in the first patient with more than 20 year follow-up. In the presence of metastases octreotide treatment is often, but not invariable, successful in suppressing GH and GHRH, because of its dual action on the pituitary gland and the GHRH-secreting tumor [51]. Most published reports concerned subcutaneously administered octreotide mostly by three daily injections and in a minority by chronic subcutaneous infusion [48, 52] and summarized by van den Bruel et al. [18]. In the majority of patients GH decreased substantially, but IGF-I remained elevated in eight of 13 patients in whom IGF-I was measured. The serum concentration of GHRH generally remained elevated [2], as we found in patient 3. Publications of treatment results with the slow release formulation of octreotide and lanreotide are scarce and follow-up short-term. We could find only four cases in literature: three patients were treated with lanreotide and one with octreotide-long acting repeatable (octreotide LAR) [19, 21, 24, 25]. In the present study we reported the long-term effect of octreotide-LAR treatment in two patients, which resulted in suppressed GH secretion and normalized IGF-I concentration in both patients, but with still slightly elevated GHRH concentrations, as long as the patients continued suppressive therapy. The latter finding is indeed often mentioned in other reports as well, as discussed above. Other treatment modalities applied in single cases are the use of a GHRH-receptor antagonist and chemotherapy in a patient with severely metastasized disease [15, 53].
During octreotide treatment no visible growth of (eventual) metastases was noted for many years with high-sensitivity liver CT scanning and CT scanning of the thorax in our patients, which suggests that in selected cases octreotide can inhibit growth of metastases from a carcinoid tumor. In addition, pituitary hyperplasia, the likely cause of the preoperative enlargement in patient 2, slowly disappeared completely, as previously described in other patients under medical therapy or after complete surgical removal of the GHRH source [27, 54]. Nevertheless, in patient 3 GHRH levels increased slowly, suggesting growing metastasis under octreotide restraint. Recently, we could localize a suspect lesion in the anterior mediastinum.
GHRH and its receptor are normally present in several organs, including the ovary and testis, playing a role in the regulation of steroidogenesis, and the pancreas [55–57]. It has been found that some GHRH-producing carcinoids also express GHRH-receptors, which may stimulate tumor growth via an autocrine mechanism [22]. Particularly interesting was the description of a single acromegalic patient whose large pituitary adenoma co-expressed GHRH-receptor and GHRH together with a 100-fold increased plasma GHRH concentration. After pituitary adenomectomy plasma GHRH normalized. It is likely that the high local concentration of GHRH contributed to the growth of the adenoma [58].
GHRH and its receptor are also expressed in cancers of the breast, ovary, and endometrium and small-cell lung cancer, and may function as a paracrine/autocrine growth factor in regulating local IGF-I and/or IGF-II secretion. Other studies have shown that GHRH receptor-antagonists inhibit the growth of colorectal, prostatic, mammary, lung, and pancreatic cancers, partly by direct suppression of IGF-I and IGF-II production, or acting directly without influencing these growth factors [59–63]. When more potent GHRH-receptor antagonists will become available for clinical application patients with advanced disease may benefit from this new approach, because they respond at best to somatostatin analogue therapy by suppressing GH secretion and partly that of GHRH, but the tumor and metastases will usually grow with fatal outcome.
|
[
"acromegaly",
"ectopic ghrh",
"octreotide",
"gh secretion",
"approximate entropy"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Semin_Immunopathol-4-1-2315691
|
The role of complement in ocular pathology
|
Functionally active complement system and complement regulatory proteins are present in the normal human and rodent eye. Complement activation and its regulation by ocular complement regulatory proteins contribute to the pathology of various ocular diseases including keratitis, uveitis and age-related macular degeneration. Furthermore, a strong relationship between age-related macular degeneration and polymorphism in the genes of certain complement components/complement regulatory proteins is now well established. Recombinant forms of the naturally occurring complement regulatory proteins have been exploited in the animal models for treatment of these ocular diseases. It is hoped that in the future recombinant complement regulatory proteins will be used as novel therapeutic agents in the clinic for the treatment of keratitis, uveitis, and age-related macular degeneration.
Complement system
Complement has long been recognized as a critical component of the innate immune system [1–6]. It comprises of soluble and surface proteins that play a central role in host defense against infection and in the modulation of antigen-specific immune and inflammatory responses [4, 5]. Thus, complement also serves as a bridge between innate and adaptive immunity. The complement system can be activated by three proteolytic cascades namely, the classical, the alternative, and the lectin pathways (Fig. 1). Although the initial signal for the activation of each cascade differs, complement activation via these pathways triggers a sequence of biologic reactions in which each component is activated by the upstream component [3, 5]. This sequential activation of complement components by all three pathways leads to the formation of membrane attack complex (MAC). Under normal conditions, the activation of the complement system is kept under tight control by the coordinated action of soluble and membrane-associated complement regulatory proteins (CRegs). However, interference with the function and/or expression of CRegs leads to unwanted host tissue damage and tissue pathology due to unregulated complement activation [7, 8].
Fig. 1Complement system and complement regulatory proteins (CRegs). The complement system can be activated through three pathways and complement activation triggers a sequence of biological reactions which result in the generation of effector molecules (e.g., C3a, C3b, C5a, MAC) with potent immunological activities. The classical pathway can be activated by antigen-antibody complexes or by substances such a C-reactive protein. C1, C2, C4, and C3 are the components of the classical pathway. Activation of the alternative pathway does not require the presence of an antibody; C3, factor B, factor D, and properdin are the components of the alternative pathway. Like the alternative pathway, the activation of the lectin pathway is also antibody independent and can be achieved by interaction of serum lectins (e.g., mannose-binding protein, MBL), with sugar residues present on the bacterial cell wall. All three pathways converge on a common reaction that results in the formation of an enzyme—C3 convertase. This enzyme cleaves C3 into C3a and C3b. C3b forms C5 convertase that ultimately leads to the formation of membrane attack complex (MAC). The complement cascade is regulated at different check points (shown with black blocks) by CRegs. The soluble CRegs are shown in white boxes and the membrane-bound CRegs are shown in gray boxes
CRegs can be broadly categorized into two classes—membrane-bound CRegs and soluble CRegs (Fig. 1). Membrane-bound CRegs (Fig. 1) include decay-accelerating factor (DAF, CD55), membrane co-factor protein (MCP, CD46), complement receptor 1 (CR1, CD35), and membrane inhibitor of reactive lysis (MIRL, CD59). DAF, a glycosylphosphatidylinositol (GPI)-anchored protein prevents the activation of the complement system by inhibiting the formation of C3 and C5 convertases of both the classical and alternative pathways and accelerating the decay of these convertases [9, 10]. However, DAF does not irreversibly inactivate C3b and C4b and, after interaction with DAF, these molecules still possess cytolytic activity. MCP a transmembrane glycoprotein regulates the complement cascade by serving as a co-factor for serine protease factor I, which irreversibly inactivates the hemolytic potential of C3b and C4b [11, 12]. CR1 has been reported to possess both decay-accelerating and co-factor activities [13]. Like DAF, CD59 is also attached to the cell surface by GPI anchors. CD59 acts at the terminal step of complement activation and prevents the formation of membrane attack complex (MAC, C5b-9 complex), the final activation product of all three complement cascades by blocking the incorporation of C9 [14, 15]. Crry (5I2 antigen) is a C3 convertase inhibitor that possesses both decay-accelerating and membrane co-factor activities and present only in rodents [16, 17]. However, DAF, MCP, and CD59 have also been identified and characterized in rodents [18–23].
C1 inhibitor (C1INH), C4 binding protein (C4bp), complement factor H (CFH), complement factor I (CFI), S-protein and clustrin, are the examples of important soluble CRegs (Fig. 1). C1INH regulates activation of the classical complement pathway by inactivating the protease function of activated C1 complex. It also prevents spontaneous activation of C1 and binding to the zymogen forms of C1r and C1s [24]. C4bp acts as a co-factor for CFI, catalyzing the cleavage and inactivation of C4b [25]. CFH inhibits C3 activation by binding to C3b and acting as a co-factor for factor I-mediated cleavage of C3b and also has decay-accelerating activity for the alternative pathway C3 convertase, C3bBb. It also competes with factor B for surface-bound C3b [26]. Both clustrin and S-protein prevent the formation of MAC [6]. Soluble forms of MCP, DAF, CR1, CD59, and Crry have been reported to be present in various biological fluids in both humans and rodents [27–30]. Thus, a wide variety of CRegs (both membrane bound and soluble) have evolved to create efficient checkpoints at critical steps of complement cascades so that the host could be protected from the inadvertent activation of complement on its own tissue (Fig. 1).
Since complement is a powerful cytotoxic defense system, regulation of complement activation is crucial for host tissue damage control during inflammation especially in a sensitive organ like the eye. The eye is an immunologically privileged site and ocular immune privilege does not refer to the absence of immune system but, rather its fine regulation [31]. The studies discussed in this article highlight the role of complement activation and regulation in the protection of the normal eye. In addition, the studies discussed below demonstrate the importance of interplay between complement activation and complement regulation in the development of vision-threatening complications such as keratitis, uveitis, and macular degeneration.
Role of complement in the protection of the normal eye
It is important that we discuss the role of complement and CRegs in ocular protection under normal conditions before discussing their importance in ocular pathology. As mentioned above, the eye is an immune-privileged organ [31] and is highly vulnerable to various immunological insults. Therefore, it is not surprising that the complement system, which represents the first line of defense, plays an important part not only in protecting the eye from immunological insults but also in maintaining the immune-privileged state of the eye. Our results and several other independent studies have demonstrated the presence of various components of the complement system in the normal eye [2, 32–37]. We have further shown that ocular complement components are not just silent bystanders but are chronically active at a low level in normal eye [32]. We reported that both iC3b (activation product of C3) and MAC (end product of complement activation) are present in normal rat eye indicating that the complement system is continuously activated at a low level in the normal eye [32]. In 2003, we documented that the interaction between complement activation product iC3b and its receptor is vital for the creation of immunosuppressive environment that leads to the induction and the maintenance of ocular immune privilege and protection of intraocular structures that are critical for vision [38]. The presence and expression pattern of CRegs in the normal human eye was first described by us [33]. We reported that membrane-bound CRegs—MCP, DAF, and CD59—are differentially expressed in normal human and eyes [33]. Our studies along with other independent studies that were performed later demonstrated that Crry and CD59 (membrane bound and soluble) are also present in the normal rodent eye [32, 34, 35]. We have further shown that ocular CRegs are functionally active and tightly regulate the activation of intraocular complement [27, 32]. Functional inhibition of ocular CRegs using specific antibodies resulted in unregulated complement activation leading to severe intraocular inflammation in normal animals [32]. Furthermore, complement inhibitory activity in normal human intraocular fluid was blocked by inhibiting the function of MCP, DAF, and CD59 [27].
Thus, our results clearly established that within the normal eye the complement system seems to perform two important functions. First, the chronic low level of complement activation serves as a primary defense mechanism of the eye against pathogenic infection and is finely regulated by the soluble and membrane-bound intraocular complement regulatory proteins. This enables destruction of the putative pathogen without inadvertent damage of ocular tissue, which is vital for the maintenance of vision. Second, complement activation products (such as iC3b) generated as a result of the low level of complement activation results in the selective suppression of harmful T cell responses. In this role complement protects the eye from innocent bystander destruction associated with the T cell response to the invading pathogens. Complement components and CRegs present in normal tears have also been reported to be functionally active and have been proposed to serve as first line of defense in protecting the eye [36]. However, additional studies are required to determine the role of the complement system in normal tears.
Role of complement in corneal disease
The cornea is constantly exposed to various physical as well as immunological insults. Indeed the normal cornea is well equipped to protect itself from these insults and has been shown to possess functional complement activity [39–46]. In addition, low levels of C3, C3 split products, and MAC were detected in the normal cornea [32, 39, 46]. This observation further proves that low level of complement activation is taking place in the normal cornea since complement activation is necessary for the generation of these products. Studies reported in the literature suggest that complement cascade can be aggressively activated in the cornea during an immune and inflammatory reaction [45, 46]. Two reports by Mondino and colleagues described interesting sets of experiments [45, 46]. In these experiments, the normal cornea from human donors was exposed to lipopolysaccharide (LPS), ribitol teichoic acid immune complex, acid (HCl), or alkali (NaOH) separately and the generation of anaphylatoxins—C3a, C4a, and C5a—as well as MAC in the corneas was monitored. It was reported that anaphylatoxins—C3a, C4a, and C5a—could be generated when the cornea was injured with LPS, immune complexes, acid, or alkali. Interestingly MAC could only be generated when the cornea was exposed to LPS or immune complexes. Cornea failed to generate MAC when insulted with acid or alkali. The immune response mounted to LPS or immune complex is similar to that generated against infectious agents like gram-negative bacteria. Indeed the complement system has been shown to play a critical role in protection against Pseudomonas aeruginosa infection that causes keratitis [47, 48]. Additionally, complement activation is believed to play an important role in ulceration of human cornea induced by gram-negative bacteria [49].
The aforementioned observations have important clinical implications. If in future anti-complement agents were to be considered for the treatment of corneal pathology, anaphylatoxins would be an ideal target for the patient presenting with chemical injury such as exposure to acid or alkali. On the other hand, in the case of bacterial infection both anti-anaphylatoxin and anti-MAC therapy should be considered. Although as discussed below the cornea possesses the ability to express different CRegs to protect itself from complement-mediated damage, it may not be able to upregulate various CRegs to an appropriate level during an acute episode of complement activation as observed during chemical insults or acute bacterial infection. In such cases, topical application of recombinant CRegs could be beneficial and serve as a better alternative for intervention.
Enhanced complement activation can cause damage to the autologous corneal tissue during an infection or inflammation. Membrane-bound CRegs such as MCP, DAF, Crry (Fig. 2a), and CD59 (Fig. 2b) are expressed in the different layers of the cornea to protect this tissue from complement-mediated damage [27, 33, 50]. Interestingly, compared to corneal stroma and corneal endothelium, various CRegs are very strongly expressed on the corneal epithelium [33] (Fig. 2a and b). This high expression of CRegs on the corneal epithelium is crucial for the protection of the cornea because the cornea is continually exposed to various pathogens including bacteria. These bacteria produce phospholipase and other enzymes which can remove GPI-anchored DAF and CD59 from the corneal surface [51]. Thus, in the absence of a very high expression of CRegs on the corneal epithelium, bacterially induced loss of DAF and CD59 could be deleterious to the cornea and lead to vision loss after the putative pathogen has been destroyed by the complement system. To investigate if a deficiency or abnormality in the expression of CRegs may play a role in ocular surface disease, systemic examination of corneal tissue from patients with different diseases such as pseudophakic bullous keratopathy, HSV-1 keratitis, and herpes zoster scleritis is needed.
Fig. 2Complement regulatory proteins in normal cornea. The figure shows immunofluorescent staining for Crry (a) and CD59 (b) in the cornea of naive Lewis rat. Objective magnification ×20. Epi Epithelium, S stroma, Endo endothelium, AC anterior chamber
Role of complement in autoimmune uveitis
Uveitis, the inflammation of the uveal tract, can be classified anatomically as anterior, intermediate, posterior, or diffuse (panuveitis) depending on the segment of the eye that is affected. Each year, approximately 17% of active uveitis patients experience some degree of vision loss [52]. Uveitis may be idiopathic, associated with systemic diseases such as Behçet’s disease, and Vogt–Koyanagi–Harada disease or resulting from a variety of infectious agents. Inflammation resulting from uveitis may lead to conditions like cataract, glaucoma, and cystoid macular edema that may cause irreversible vision loss. Anterior uveitis is the most common form of uveitis, and the most common form of anterior uveitis is of unknown (i.e., idiopathic) etiology [52].
Although the exact cause of the disease is still unknown, a wide array of studies using different animal models have established that the inflammation in uveitis is due to an autoimmune response to various ocular antigens [53–56]. Unfortunately, almost all of the studies in the past three decades have focused on the role of T cells in the pathogenesis of uveitis [54, 57, 58]. Very few laboratories have investigated the role of the complement system in the development of autoimmune uveitis [32, 58–66].
Through our recent studies, the critical role of complement and complement regulatory proteins (CRegs) in the pathogenesis of idiopathic anterior uveitis was addressed and recognized. Our laboratory has used experimental autoimmune anterior uveitis (EAAU)—an animal model of human autoimmune anterior uveitis—to investigate the role of the complement system, a significant and previously neglected facet of uveitis [53, 62, 63]. EAAU is induced by injecting bovine melanin-associated antigen emulsified in Freund’s complete adjuvant, in the foot pad of Lewis rats [53]. The onset of the disease (days 12–14 post immunization) is marked by inflammation of the iris/ciliary body followed by massive infiltration of the inflammatory cells in the anterior segment of the eye during the peak (days 16–19 post immunization) of EAAU (Fig. 3). The inflammation resolves after approximately 3 weeks, and the eye returns to the normal state with no apparent sign of any tissue damage (Fig. 3). Interestingly, studies from our laboratory have demonstrated that the activation of the complement system is critical for the development of EAAU [62]. During EAAU, the complement system is activated (detected by Western blot analysis for iC3b) in the eye, and the kinetics of complement activation follows a pattern that is similar to the clinical course of the disease. Furthermore, depletion of the complement system of the host resulted in complete inhibition of EAAU. In complement-depleted rats, the levels of IFN-γ, IP-10, ICAM-1, and LECAM-1 were extremely low compared to the complement-sufficient rats during EAAU. Our observations suggest that the activation of the local complement system plays a critical role in the development of inflammation during EAAU and suppression of intraocular complement system, thus may provide a successful strategy for uveitis therapy. At present, the exact mechanism and pathways involved in local complement activation during uveitis are not known. It is also not known if the increase in complement components in the anterior segment of the eye is due to the upregulation of their synthesis by resident ocular cells or is a result of spillover from the systemic complement system. Previous studies from our laboratory established that a functionally active complement system is present in the anterior segment of the eye because when a well-known activator of complement such as zymosan was injected in the anterior chamber, severe anterior uveitis was induced. This zymosan-induced uveitis was due to the activation of the complement system because anterior uveitis was completely inhibited when the complement system of the host was depleted [32]. Furthermore, we and others have reported the presence of various complement components in aqueous humor as well as other parts of the normal eye [32, 39, 41, 59, 61]. These observations taken together suggest that ocular complement is activated during the course of EAAU and the inhibition of the intraocular complement system may lead to more effective management and/or treatment of autoimmune uveitis.
Fig. 3Clinical and histologic pictures of rat eye during different stages of experimental autoimmune anterior uveitis (EAAU). No inflammation could be detected in the iris (I) and the ciliary body (CB) after clinical (a) and histologic (d) examination at day 8 post immunization (before the onset of EAAU). At the peak of disease (day 19 post immunization) the anterior chamber (AC) appeared cloudy and opaque due to the presence of proteinaceous material and inflammatory cells (b). Histopathologic picture of rat eye at the peak of EAAU (e) revealed severe inflammation of the iris (I) and ciliary body (CB). Heavy infiltration in the anterior chamber (AC) with spillover into the anterior vitreous was also observed at this time point. The eye appeared normal both clinically (c) and histologically (f) after the resolution of EAAU (day 30 post immunization). Objective magnification ×10
When considering anti-complement therapy, we cannot ignore the role of CRegs in the resolution of autoimmune uveitis. We have demonstrated that the functionally active CRegs downregulate intraocular complement activation and are critical to protect the ocular tissues from complement-mediated damage during EAAU [63]. We observed that various CRegs were upregulated during the course of EAAU on the resident ocular tissue and this upregulation was not due to the infiltrating inflammatory cells because the levels of CRegs remained upregulated after the resolution of the disease when all the infiltrating cells were cleared from the anterior segment of the eye [63]. Furthermore, when the function or expression of CRegs was inhibited in vivo, more severe EAAU with early onset and delayed resolution was observed. These observations demonstrate that ocular tissue can protect itself by upregulating various CRegs during uveitis and provide strong evidence for the use of recombinant CRegs as therapeutic agents in the treatment of autoimmune uveitis.
Activation of the complement system has also been reported in other animal models of anterior uveitis such as endotoxin-induced uveitis (EIU) [64, 65]. In EIU, it was demonstrated that the complement system is critical for the development of inflammation, and there is a local activation of the complement system as evident by accumulation of complement components (C1q, C3, and C4) in various parts of the eye during inflammation [66]. Activation of the complement system was also implicated in experimental allergic uveitis [67]. Using experimental autoimmune uveoretinitis (EAU) animal model, the complement system has been reported to play an important role in the pathogenesis posterior uveitis [68]. Recently, Read and co-workers demonstrated that the transgenic mice that express soluble Crry in the retina had decreased incidence and severity of EAU [68]. In these transgenic mice, EAU was not completely suppressed. This may be due to relatively low levels of Crry expressed compared with the levels that are required for the complete inhibition of the complement system. It is possible that a higher dose of exogenous Crry or similar CReg may result in complete inhibition of EAU.
Thus, these studies provide strong evidence that the complement system and CRegs play an important role in the pathogenesis of autoimmune uveitis and provide alternative approaches for the development of effective therapy. More research in the future will present us enough information to develop anti-complement therapy for uveitis. Hopefully, in the future, complement inhibitors might be used as novel anti-uveitic agents in the clinic for the treatment of this important form of human ocular disease.
Role of complement in age-related macular degeneration
Age-related macular degeneration (AMD) is the most common cause of legal blindness worldwide among the elderly over the age of 50. Approximately 1.8 million people in the United States alone have AMD and it is projected that by 2020, roughly 2.9 million people will develop this disease [69–71]. In AMD, there is a progressive destruction of the macula leading to the loss of central vision. AMD is classified into non-exudative (dry type) and exudative (wet type) [69–71]. In the dry type of AMD, drusen—small yellow deposits—are formed between the retinal-pigmented epithelium (RPE) and Bruch’s membrane [72]. In wet-type AMD, abnormal vessel growth takes place from choroid, termed choroidal neovascularization (CNV), under the retinal pigment epithelium [71, 72]. Although several risk factors are associated with AMD, the exact pathogenesis still remains unknown. The last few years have witnessed an amazing series of research establishing the central role of the complement system in the pathogenesis of AMD both in humans and in experimental animals. Various studies in the literature have indicated a potential role for complement in drusen formation in the non-exudative form of AMD in humans [73–75]. Complement components, complement activation products (C3a, C5a, MAC), and complement regulatory proteins (CD46, Vitronectin) have been localized in drusen in patients with AMD [73–75].
The discovery of allelic variants of complement factor H (CFH) as a major risk factor for AMD has become a landmark study [76–79]. The polymorphism that gives rise to this allelic variant of CFH results from a tyrosine to histidine replacement at position 402. The “high risk” variant, i.e., homozygous for histidine residue at position 402 is associated with five- to sevenfold increased risk of AMD [76–79]. CFH consists of 20 short consensus repeats (SCRs) and the polymorphic site 402 resides in SCR7. SCR7 contains the binding site for C3b, glycosaminoglycans (GAG) and C-reactive protein (CRP). The polymorphism at position 402 in SCR7 reduces the binding affinity to CRP [80] and GAG [81]. Binding of CFH to CRP and GAG plays an important role in the regulation of the alternative pathway. Thus the polymorphic form of CFH with histidine at position 402 may have reduced ability to regulate the activation of the alternative pathway. This may lead to uncontrolled activation of alternative pathway which may result in the development of AMD. The loss of regulatory activity of CFH due to polymorphism can have a direct effect on the predisposition of the eye to AMD as it has been shown recently that CFH is expressed locally in the optic nerve, retina, and retinal pigment epithelium [82]. The importance of CFH in AMD was further demonstrated in a recent study. It was reported that the aged mice genetically deficient in CFH (cfh−/−) develop spontaneous changes in retina (e.g., photoreceptor dysfunction, increased deposition of C3 in retina, thinning of Bruch’s membrane) which is consistent with those seen in AMD [83].
Polymorphic variants of other complement components have also been shown to be associated with AMD. In a recent study, Gold et al. reported that complement components—factor B (CFB) and complement component 2 (C2)—are also associated with AMD [84]. In this report, the investigators identified high risk allele and two protective alleles for CFB and C2 that were associated with AMD. An independent study further established a similar association between CFB, C2, and AMD [85]. More recently, investigators compared 847 patients with AMD with 701 unaffected people and observed that a variant in the complement C3 gene affected the risk of developing AMD [86].
The aforementioned studies establish that the polymorphism in the genes of certain complement components and regulators genetically predisposes some individuals to AMD. However, we should also pay attention to the fact that other risk factors for AMD such as alcohol consumption and smoking nicotine can directly affect the level of complement activation and CRegs. For instance, chronic ethanol consumption upregulates CFB but downregulates CFH in mice [87]. Thus, association of alcohol consumption with AMD may be due to its direct effect on the complement system. Indeed we have reported that chronic alcohol consumption increases the size of CNV complex in rat model of laser-induced CNV [88]. Similarly, cigarette smoking has been shown to decrease the affinity of CFH for C3 and activate the complement system [89].
In our laboratory, we use a mouse model of laser-induced CNV to study the role of complement in the pathogenesis of wet-type AMD. A reliable way to produce CNV in animals is to rupture Bruch’s membrane with laser photocoagulation [90–92]. We and others have found that choroidal neovascularization induced in rodents by laser photocoagulation is useful to gain insights into the pathogenesis of new vessel growth from the choroid [90–96]. In 2005, we demonstrated for the first time using this animal model that complement activation and generation of MAC is critical for the development of laser-induced CNV. In this study, C3−/− mice and anti-C6-treated mice failed to develop CNV [93]. As mentioned above, the importance of C3 in the pathogenesis of AMD was recently confirmed by Yates et al., who established an association of C3 polymorphism with AMD [86]. We further observed that in the absence of MAC deposition, release of angiogenic factors such as vascular endothelial growth factor (VEGF), TGF-β2, and β-FGF was inhibited. This is a crucial observation as it suggests that the complement system is a potential therapeutic target for the treatment of CNV. Subsequently, we reported that the activation of complement by alternative pathway is responsible for MAC formation in laser-induced CNV. This alternative pathway activation was due to the decrease in CFH levels and increase in CFB levels in the RPE/choroid after laser treatment [94]. In a separate study, we used recombinant-soluble CD59-Fc chimeric protein to inhibit the formation of MAC after laser photocoagulation [95]. Treatment of mice with recombinant-soluble CD59-Fc inhibited the development of choroidal angiogenesis in the laser-induced mouse model by blocking MAC formation which in turn blocked the induction/release of angiogenic growth factors such as VEGF, β-FGF, and TGF-β2 (Fig. 4). Thus, our studies suggest that the recombinant forms of CD59 could be used as a potential drug in the future to treat CNV. Another potential target could be anaphylatoxins—C3a and C5a. It was reported that C3a and C5a are present not only in human drusen but are also localized in the choroid of mice with laser-induced CNV [97].
Fig. 4Role of CD59 in laser-induced choroidal neovascularization (CNV). Laser photocoagulation results in complement activation and MAC formation in the posterior segment of the eye. MAC is a complex assembly of C5b to C9 components with multimers of C9 forming a pore in the cell membrane. This causes increased release of growth factors—VEGF, TGFβ2, and β-FGF—which in turn leads to the development of CNV. The membrane-bound as well as recombinant-soluble (rs) CD59, bind to C9 and C8 and inhibit the polymerization of C9 which is prerequisite for MAC formation. In the absence of MAC, release of angiogenic factors is downregulated. This inhibits the growth of new vessel from the choroid (CNV)
Using the mouse model of laser-induced CNV, we observed that the CH50 levels in the serum of lasered mice varied when compared to non-lasered mice [93, 94]. Another study demonstrated that the plasma complement C3a des Arg levels increased in AMD patients compared to controls [98]. This increase in plasma C3a des Arg concentration reflects systemic complement activation. Thus systemic activation of the complement system may also contribute to the pathogenesis of AMD.
Currently, several different approaches are being used to treat wet AMD in humans [99]. Unfortunately, these treatment options have limitations due to their short-term and serious side effects. Although, photodynamic therapy (PDT) reduces the rate of vision loss in most patients, it does not lead to significant improvement in vision. Furthermore, repeated PDT can cause severe damage to the posterior segment of the eye and is not cost effective. Two drugs Lucentis and Macugen are approved by Federal Drug Administration for the treatment of AMD and both drugs inhibit vascular endothelial growth factor. One of the major drawbacks of Lucentis or Macugen is that repeated injections of these drugs are needed to treat AMD. Repeated injections can cause some serious side effects including hemorrhage of the eye membrane, eye pain, and vitreous floaters. Therefore, alternative therapeutic strategies with minimum side effects are required to better treat AMD patients. We believe that the results derived from the studies reviewed above would open up the window of opportunities for the development of anti-complement therapy in the treatment of AMD.
In view of important developments relative to the role of the complement system in AMD during the past several years, we believe that the next decade would be critical for such studies. We hope that the interest will continue to grow in the future so that an alternative therapy could be developed by inhibiting complement activation. Since untreated CNV leads to the irreversible loss of central vision, it is important for the pharmaceutical companies to utilize the knowledge available to them from different research laboratories and should make every effort to transfer this knowledge to develop new drugs and therapies for AMD patients.
Conclusions and clinical implications
During the last decade, a lot of progress has been made to understand the pathophysiology of corneal diseases, uveitis, and macular degeneration. Specifically, the studies defining the role of the complement system in ocular diseases have drawn a lot of attention. It is an exciting area for future investigations and the basic questions such as what triggers the complement activation and which specific pathway is involved in corneal diseases, uveitis and macular degeneration need to be addressed. This understanding will help us to develop better therapies for ocular diseases in the future so that blindness could be prevented.
The evidence derived from the studies reviewed here supports the concept that recombinant complement regulatory proteins may be therapeutically useful in the prevention of ocular tissue damage and in the treatment of various ocular diseases. Agents that specifically inhibit the complement system or complement activation products have been proven beneficial in the treatment of both human and experimental diseases over the past few years. Thus, the therapy based on complement inhibition has great potential in the future for the treatment of various ocular diseases such as keratitis, uveitis, and macular degeneration.
|
[
"complement system",
"complement regulatory proteins",
"eye",
"uveitis",
"age-related macular degeneration",
"corneal inflammation"
] |
[
"P",
"P",
"P",
"P",
"P",
"R"
] |
Int_J_Cardiovasc_Imaging-4-1-2233707
|
Evaluation of left atrial systolic function in noncompaction cardiomyopathy by real-time three-dimensional echocardiography
|
Background Noncompaction cardiomyopathy (NCCM) is a rare disorder with persistance of the embryonic pattern of myoarchitecture. NCCM is characterized by loosened, spongy myocardium associated with a high incidence of systolic and diastolic left ventricular (LV) dysfunction and heart failure (HF). It is known that LV dysfunction contributes to elevated left atrial (LA) and pulmonary vascular pressures, however atrial function has not been examined in NCCM. The objective of the present study was to assess LA systolic function characterized by LA ejection force (LAEF) in NCCM patients using real-time three-dimensional echocardiography (RT3DE) and to compare to control subjects. Methods The study comprised 17 patients with an established diagnosis of NCCM and their results were compared to 17 healthy age-matched controls with no evidence of cardiovascular disease. Forty-one percent of NCCM patients were in NYHA functional class II / III HF. Previously proposed echocardiographic diagnostic criteria for NCCM were used. All patients underwent conventional two-dimensional echocardiography and RT3DE. LAEF was measured based on MA annulus diameter (LAEF3D-MAD) and area (LAEF3D-MAA) using RT3DE. Results The presence and severity of mitral regurgitation were more frequent in NCCM patients than in control subjects. LV diameters and mitral annulus were significantly increased in NCCM patients. Compared with control subjects, both LAEF3D-MAD (3.8 ± 2.2 vs 2.3 ± 1.0 kdyne, P < 0.05) and LAEF3D-MAA (12.7 ± 7.6 vs 4.9 ± 2.1 kdyne, P < 0.01) were significantly increased in NCCM patients. Conclusions LAEF as a characteristic of LA systolic function is increased in NCCM patients compared to normal individuals. These results can suggest compensating left atrial work against the dysfunctional LV in NCCM patients.
Introduction
Noncompaction cardiomyopathy (NCCM) is a rare disorder with persistance of the embryonic pattern of myoarchitecture, and has been assumed to occur as a consequence to intrauterine arrest of compaction [1]. NCCM is characterized by loosened, spongy myocardium associated with a high incidence of systolic and diastolic left ventricular (LV) dysfunction and heart failure. It is known that LV dysfunction contributes to elevated left atrial (LA) and pulmonary vascular pressures, however atrial function has not been examined in NCCM. It is difficult to assess non-invasively LA contraction, but Manning et al. reported a non-invasive method for the evaluation of LA ejection force (LAEF) as a characteristic of LA systolic function [2]. Anwar et al. demonstrated the usefulness of real-time three-dimensional echocardiography (RT3DE) for the evaluation of LAEF [3]. The objective of the present study was to assess LAEF in NCCM patients using RT3DE and to compare to control subjects.
Methods
Patient populations
The study comprised 17 patients with an established diagnosis of NCCM and their results were compared to 17 healthy age-matched controls with no evidence of cardiovascular disease. All patients and control subjects were in sinus rhytm and all of them were examined by two-dimensional echocardiography (2DE) and RT3DE.
Diagnostic criteria for NCCM
Previously proposed echocardiographic diagnostic criteria for NCCM by Jenni et al. were used to include patients into the present study from the Outpatient Clinic of Erasmus Medical Center, Rotterdam during 3 years [4]: (1) absence of coexisting cardiac anomalies, (2) segmental, excessive thickening of the LV wall with a two-layered structure: a thin, compacted epicardial layer and a much thicker, noncompacted layer with the characteristic appearance of numerous, prominent trabeculations (meshwork) and deep intertrabecular recesses, (3) colour Doppler evidence of deeply perfused intertrabecular recesses, and (4) predominant localization of thickening in the LV apical, midlateral, and midinferior walls. Hypertensive heart disease was excluded by clinical and echocardiographic examinations (septal thickness < 13 mm).
2DE examination
The 2DE was undertaken with the patient lying in the left lateral decubitus position using both apical and parasternal views. The 2DE studies were performed using a 3.5-MHz transducer and a commercially available ultrasound system (Sonos 7500, Philips, Best, The Netherlands). The following measures were obtained: (1) LV end-diastolic diameter (LV-EDD) and LV end-systolic diameter (LV-ESD) by M-mode echocardiography; (2) LV fractional shortening (LV-FS); (3) peak velocities of transmitral E and A waves with pulsed Doppler and (4) mitral regurgitation graded as mild (jet area < 4 cm2), moderate (jet area 4–8 cm2), and severe (jet area > 8 cm2) according to maximum jet area by color Doppler.
RT3DE examination
RT3DE was performed using the same ultrasound system, and performed with a X-4 matrix transducer capable of providing real-time B-mode and colour Doppler. The 3D images were collected within 5–7 s of breath holding in full-volume mode. The 3D data were transferred to an offline analysis system (TomTec Inc., Munich, Germany). Data were stored digitally and subsequently evaluated by two echocardiographers. Data analysis of RT3DE imaging was based on a two-dimensional approach relying on the echocardiographic images obtained from the apical views and on manual tracing of inner border of the mitral annulus (MA). Once this was completed the surface area was automatically delineated and could be visualized from different points of views. Manual modification was done to correct any image if necessary. The following measures were obtained: (1) MA diameter (MAD3D) and area (MAA3D) measured at end-diastole (just before mitral valve closure) and (2) LAEF3D-MAD and LAEF3D-MAA calculated by formula as 0.5 × 1.06 × (MAD3D or MAA3D) × V2, where V is the peak velocity of A wave, 1.06 gm/cm3 is blood viscosity and 0.5 is a coefficient factor [2] (Fig. 1).
Fig. 1Automatically cropped two-dimensional image from real-time three-dimensional echocardiographic dataset demonstrating left ventricular cavity in a patient with noncompaction cardiomyopathy (A). Measurement of mitral annulus area and mitral inflow A wave velocity can help calculation of left atrial ejection force as a characteristic of atrial systole. Real-time three-dimensional echocardiography allows the `en-face` evaluation of mitral annulus helping exact measurement of mitral annulus diameter and area (B). Abbreviations: MA: mitral annulus, MAD: mitral annulus diameter, E and A: mitral inflow E and A wave velocity measured by pulsed Doppler
Statistical analysis
All data obtained by 2DE or RT3DE were presented as mean ± SD. Data analysis was performed using statistical software (SPSS, Version 12.1, SPSS Inc, Chicago, USA). Independent sample t test was performed to compare between means of variables of groups to determine the statistical significance of difference. The level of significance was set to p < 0.05. In our department, an excellent correlation (r = 0.82, P < 0.0001) has been described between MAD3D and MRI measurements of MAD [5]. There was no significant difference between MAD3D and MADMRI. Similarly, there was no difference between MAA3D and MAAMRI.
Results
Clinical and 2D echocardiographic data
Clinical and echocardiographic data of NCCM patients and controls are shown in Tables 1 and 2. The presence and severity of mitral regurgitation were more frequent in NCCM patients than in control subjects. LV diameters were significantly increased, while LV-FS was significantly decreased in NCCM patients. The number of noncompacted segments was 5.4 ± 2.9 in NCCM patients, which did not show any correlation with any echocardiographic parameters.
Table 1Demographic and clinical dataParameterNCCM patients (n = 17)Controls (n = 17)Age (years)48 ± 2043 ± 10Male sex (%)8 (47)11 (65)Diabetes mellitus (%)9 (53)10 (59)Hypercholesterolaemia (%)10 (59)11 (65)NYHA-class II/III HF (%)3 (18)/4 (24)0 (0)Abbreviations: NCCM: noncompaction cardiomyopathy, NYHA-class II/III HF: New York Heart Association functional class II/III heart failureTable 2Echocardiographic dataParameterNCCM patients (n = 17)Controls (n = 17)2D echocardiography LV-EDD (mm)60.6 ± 11.0 **51.2 ± 6.6 LV-ESD (mm)48.1 ± 13.0 **33.0 ± 6.7 LV-FS (%)22.1 ± 17.1 **35.7 ± 6.4 Mitral inflow A wave velocity (cm/s)50.1 ± 16.2 *41.8 ± 9.2Mitral regurgitation No8 (47) *15 (88) Mild6 (35) *2 (12) Moderate-severe3 (18) *0 (0)RT3DE examination MAD3D (cm)3.5 ± 0.5 *3.2 ± 0.3 MAA3D (cm2)12.0 ± 2.8 *7.9 ± 1.4 LA-EF3D-MAD (kdyne)3.8 ± 2.2 *2.3 ± 1.0 LA-EF3D-MAA (kdyne)12.7 ± 7.6 **4.9 ± 2.1* P < 0.05 vs controls**P < 0.01 vs controlsAbbreviations: 2D: two-dimensional, 3D: three-dimensional, A: mitral inflow A wave velocity measured by pulsed Doppler, D: diastolic, EDD: enddiastolic diameter, ESD: endsystolic diameter, FS: fractional shortening, NYHA-class II/III HF: New York Heart Association functional class II/III heart failure, LA-EF: left atrial ejection force, LV: left ventricular, MAA: mitral annulus area, MAD: mitral annulus diameter, NCCM: noncompaction cardiomyopathy
RT3DE data
Acquisition and postprocessing of RT3DE data were performed successfully with clear delineation of the MA in all patients. Compared with control subjects, MAD3D and MAA3D were significantly larger, while LAEF3D-MAD and LAEF3D-MAA were significantly increased in NCCM patients (P < 0.05) (Table 2).
Discussion
Left atrial ejection force (LAEF), based on classic Newtonian mechanics, provides a physiological assessment of atrial systolic function and is a potentially useful index for assessing atrial contribution to diastolic performance. To the best of authors` knowledge this is the first time to demonstrate increased LAEF during RT3DE indicating high LA workload in NCCM patients. NCCM is known to be a relatively new clinicopathologic entity characterized by a pattern of prominent trabecular meshwork and deep intertrabecular recesses communicating with the LV cavity [1]. Results of recent studies confirmed the hypothesis that isolated ventricular noncompaction can be a part of a more widespread cardiomyopathy, involving both the morphologically normal and dysmorphic LV myocardial segments [6–9]. However, the pathophysiological mechanisms of LV dysfunction remained unclear and, moreover little is known about LA function in NCCM.
Manning et al. surmised that LAEF is equal to left ejection volume times acceleration of LA flow according to Newton’s law of motion (force equals mass times acceleration) [2]. Anwar et al. confirmed that RT3DE is a useful method for the evaluation of LAEF3D-MAA [3]. In recent studies MAA3D has been confirmed to be traced successfully with good interobserver agreement and comparable to magnetic resonance imaging [5]. RT3DE allowed exact ‘en-face’ measurement of MAA3D suggesting overstretching and dilatation in dilated [10], hypertrophic [3, 10] and noncompaction cardiomyopathy (present study). LAEF was higher in hypertrophic cardiomyopathy (HCM) patients compared with control subjects, with the highest value found in patients with obstructive HCM due to higher LV resistance [3]. LAEF3D-MAA calculated by the MAA3D-derived formula was higher in HCM patients than by the ‘mitral valve area’-derived formula either by 2D echocardiography or RT3DE [3].
In the present study, 41% of NCCM patients were in NYHA functional class II / III HF. Recently, Triposkiadis et al. demonstrated that in patients with HF, LAEF and LA kinetic energy are higher than in normal controls [11]. This is most likely due to LA dilation, which compensates for the depressed LA systolic shortening and increases LA active emptying volume and LA contribution to LV stroke volume, maintaining LV stroke volume and cardiac output [11, 12]. However, it should be noted that other authors found decreased LA kinetic energy in patients with advanced HF (NYHA functional class III HF) [13, 14]. It has been concluded that LA systolic failure contributes to the transition from moderate to advanced HF.
Limitations
During evaluation of LA systolic function, consideration of other factors affecting LAEF is important. According to recent findings, age considerably influences LAEF [15]. This fact could also affect our results, but no significant differences were found in age between NCCM patients and controls. LA systolic function is deteriorated in patients with sick sinus syndrome and in those with paroxysmal atrial fibrillation [16]. However, all of our patients were in sinus rhythm. Cioffi et al. demonstrated that LV geometry can also influence LA size and function [17]. The concentric LV pattern was associated with the greatest LA volumes and with the strongest LAEF, suggesting that chronic pressure overload more than volume overload stimulates the LA in increasing its workload. Left atrial dilation can be resulted from chronic mitral regurgitation, which can also affect LA systolic function. Another important limitation is that left atrial contraction produces both the forward blood flow toward the LV through the mitral valve and backward flow towards the pulmonary veins. Unfortunately, only the forward blood flow was examined during the present study.
Conclusions
LAEF is increased in NCCM patients compared to normal individuals. These results can suggest compensating left atrial work against the dysfunctional LV in NCCM patients. However, further studies are warranted evaluating LA function by RT3DE in patients with different stages of NCCM.
|
[
"noncompaction",
"three-dimensional echocardiography",
"atrial function"
] |
[
"P",
"P",
"P"
] |
Virchows_Arch-4-1-2329735
|
Mixed-type gastric carcinomas exhibit more aggressive features and indicate the histogenesis of carcinomas
|
To investigate the pathobiological behaviors of gastric mixed-type (MT) carcinomas and gastric carcinogenesis, the clinicopathological characteristics of MT carcinomas were analyzed and compared with intestinal-type (IT) and diffuse-type (DT) carcinomas. The expression of Ki-67, caspase-3, p53, fragile histine triad (FHIT), maspin, extracellular matrix metalloproteinase inducer (EMMPRIN), vascular growth factor (VEGF), MUC-2, 4, 5AC and 6, CD44, E-cadherin, β-catenin, and phosphorylated glycogen synthase kinase 3β-ser9 (P-GSK3β-ser9) was examined on tissue microarrays using immunohistochemistry. It was found that MT carcinomas exhibited large size, deep invasion, frequent local invasion, and lymph node metastasis in comparison with IT and DT carcinomas (p < 0.05). All the markers except MUC-5AC showed higher expression in IT than DT carcinomas (p < 0.05). The expression of maspin, EMMPRIN, VEGF, MUC-4, and membrane E-cadherin was stronger in MT intestinal than diffuse component (p < 0.05). Immunoreactivities to Ki-67, EMMPRIN, and VEGF were weaker in IT carcinoma than in the MT intestinal portion (p < 0.05), while the opposite was true for CD44, MUC-2, and MUC-6 (p < 0.05). The MT diffuse component displayed a higher expression of FHIT, VEGF, and P-GSK3β-ser9 than DT carcinoma (p < 0.05). The accumulative survival rate of the IT carcinoma patients was higher than the other types (p < 0.05). The invasive depth, venous invasion, lymph node, peritoneal or liver metastasis, and Lauren's classification were independent prognostic factors for gastric carcinomas (p < 0.05). These findings suggested that MT carcinomas were also indicated to be more aggressive than IT and DT carcinomas. Significant differences were observed in the proliferation, apoptosis, angiogenesis, mucin secretion, and cell adhesion between IT and DT carcinomas, whereas only a few of these characteristics showed differences between the MT intestinal and diffuse parts, thus suggesting that both the MT components might originate from the stem cells with similar genetic traits, but follow different histogenic pathways.
Introduction
Gastric carcinoma is the second leading cause of cancer-related death behind lung carcinoma despite a worldwide decline in both its incidence and mortality since the late half of the twentieth century [11]. It continues to be a major health problem because of the slow decrease in incidence in Asia and high mortality of diagnosed gastric carcinoma in the West [23]. Generally, the prognosis of patients with gastric carcinoma is somewhat dependent on its histological type, such as intestinal type (IT) or diffuse type (DT) as defined by Lauren [16].
Although gastric carcinoma is a malignant tumor originating from the same gastric epithelium, its morphological features vary substantially among individual patients. Histologically, IT carcinoma principally includes papillary, well-differentiated, and moderately-differentiated or mucinous adenocarcinoma without signet ring carcinoma (SRC) cells, whereas DT mainly consists of the SRC, poorly differentiated and undifferentiated adenocarcinoma of WHO classification [7, 23]. Although Kushima and Hattori [12] firstly proposed an entity of gastric-type carcinoma and terminologically seemed homologous to the IT one, the carcinoma cells of this type closely resemble normal foveolar cells on the basis of tissue morphology and mucin properties and are considered to have been derived from foveolar hyperplasia or pyloric gland adenoma [12, 13]. Recently, our group also focused on Lauren's classification and found that IT carcinoma frequently occurred in old men, while DT was more frequent in comparatively young women. The latter group was more inclined to invasion into the muscularis propria, lymphatic invasion, and lymph node metastasis, and belonged to a higher TNM staging, in comparison to its IT counterparts. Further analysis demonstrated that IT gastric carcinomas with a more favorable prognosis were prone to high levels of proliferation and apoptosis and also always accompanied by a strong expression of the fragile histine triad (FHIT), phosphatase and tensin homology deleted from human chromosome 10 (PTEN), p53, and extracellular matrix metalloproteinase inducer (EMMPRIN) [30]. Additionally, many genetic and epigenetic changes, such as mutation, deletion, loss of heterozygosity, methylation, and microsatellite instability were differentially observed in both histological types of gastric carcinoma [27, 36]. These data indicate that different carcinogenic pathways exist for gastric IT and DT carcinomas. Although carcinogenesis is a multistage process which consists of a multi-factorial etiology resulting from gene–environment interaction, the intestinal type is frequently related to environmental factors like diet and predominates in areas with a high incidence of this disease. In contrast, the diffuse type is thought to be of genetic origin and evenly distributed worldwide [30, 24]. It is commonly believed that Lauren's classification is valuable for both the epidemiological studies and gastric carcinogenesis.
However, a small number of gastric carcinomas remain characterized as unclassified or mixed-type (MT) ones, which are comprised histologically of non-homogenous mixtures of IT and DT carcinomas. Stelzner and Emmrich [26] found that 28 MT carcinomas showed a deep infiltration of the gastric wall, frequent regional lymph node metastasis, and high staging in comparison to the other two types. Machado et al. [21] reported that inactivating E-cadherin mutations were exclusively observed in the diffuse component of the tumors, thus suggesting that MT carcinomas displayed phenotypic divergence. A recent study showed that bone-marrow-derived cells (BMDCs) could progress through metaplasia and dysplasia to intraepithelial cancer under the induction of the carcinogen, Helicobacter pylori [9]. Therefore, the analysis of distinct MT carcinoma components can be of remarkable pathogenic significance if it would be true that carcinoma cells originate from BMDCs. In the present study, 814 cases of gastric carcinomas were collected to further clarify the pathological characteristics of MT carcinoma and the gastric carcinogenesis pathways using a combination of tissue microarray (TMA) and immunohistochemical techniques.
Materials and methods
Subjects
A total of 814 gastric carcinomas were collected from surgical resection in Kouseiren Takaoka Hospital between 1998 and 2006. The patients with carcinomas included 573 men and 241 women (29–91 years, mean = 65.7 years). Among them, 312 cases were demonstrated with lymph node metastasis and 24 with liver metastasis. None of these cases underwent either chemotherapy or radiotherapy before surgery. All patients gave their informed consent for the use of tumor tissue specimens for clinical research and the University Ethical Committee approved the research protocol. All patients were followed up by consulting their case documents and through telephone interviews.
Pathology
All tissue specimens were fixed in 4% neutralized formaldehyde, embedded in paraffin, and cut into 4-μm sections. These sections were stained by hematoxylin and eosin (HE) to confirm their histological diagnosis and other microscopic characteristics. The staging for each gastric carcinoma was evaluated according to the TNM system of the Internationale Contre le Cancer (UICC) indicating the extent of tumor spread [25]. The histomorphological architecture of the tumors was expressed according to Lauren's classification [7, 16, 30]. The growth patterns were divided into five groups based on a modification of Emmrich's method [26]. Briefly, group I showed a combination of two components with diffuse distribution. Group II displayed both components with the border clearly visible. Group III represented some signet cells in the intestinal and diffuse components. Group IV were mucinous carcinomas with signet cells. Group V exhibited separately intestinal and diffuse components in the gastric wall. In addition, the depth of invasion, lymphatic and venous invasion, and peritoneal dissemination were all determined.
Tissue microarray
IT and DT carcinomas, as well as both components of the mixed type, were identified in HE stained sections of the selected tumor cases and a 2-mm-diameter tissue core of each donor block was punched out and transferred to a recipient block with a maximum of 48 cores using a Tissue Microarrayer (AZUMAYA KIN-1, Tokyo, Japan). Four-micrometer-thick sections were cut from the recipient block and transferred to poly-lysine-coated glass slides. HE staining was performed on TMA for confirmation of the tumor tissue (Fig. 1).
Fig. 1HE staining of TMA of gastric carcinomas
Immunohistochemistry
Consecutive sections were deparaffinized with xylene, dehydrated with alcohol, and subjected to antigen retrieval by irradiating in target retrieval solution (TRS, DAKO, Carpinteria, UAS) for 15 min in a microwave oven (Oriental rotor Lmt. Co. Tokyo, Japan). Five percent bovine serum albumin (BSA) was then applied for 15 min to prevent non-specific binding. The sections were incubated with primary antibodies for 15 min, then treated with the anti-mouse or anti-rabbit Envison-PO (DAKO, USA) antibodies for 15 min. All incubations were performed in the microwave oven for intermittent irradiation as described previously [15]. After each treatment, the slides were washed with TBST (10 mM Tris–HCl, 150 mM NaCl, 0.1% Tween 20) three times for 1 min. The primary antibodies are summarized in Table 1. All immunostaining was visualized with 3, 3′-diaminobenzidine (DAB) with 5-min reaction and counterstained with Mayer's hematoxylin. Omission of the primary antibody was used as a negative control.
Table 1 Primary antibodies used in this studyNamesSourceCompanyDilutionKi-67RabbitDAKO, Carpinteria, USA1:25Caspase-3RabbitDAKO, USA1:150p53MouseDAKO, USA1:100FHITRabbitNeomarkers, Fremont, USA1:200MaspinMouseNovocastra, Newcastle upon Tyne, UKRead-to-useEMMPRINMouseNovocastra, UK1:100VEGFRabbitLabvision, Fremont, USA1:50P-GSK3β-ser9RabbitSAT, USA1:300MUC-2MouseNovocastra, UK1:100MUC-4MouseNovocastra, UK1:100MUC-5ACMouseNovocastra, UK1:100MUC-6MouseNovocastra, UK1:100CD44MouseDAKO, USA1:50E-cadherinMouseTakara, Otsu, Japan1:100β-CateninMouseCalbiochem, CA, USA1:200
The immunoreactivity to Ki-67 and p53 was localized in the nucleus; FHIT, caspase-3, vascular growth factor (VEGF), maspin, phosphorylated glycogen synthase kinase 3β-ser9 (P-GSK3β-ser9), MUC-2, MUC-5AC, and MUC-6 in the cytoplasm; EMMPRIN and MUC-4 in the cytoplasm and membrane; CD44 and E-cadherin in the membrane; and β-catenin in the nucleus, cytoplasm, and membrane (Fig. 2). All evaluations were performed blindly by two independent observers (Takano Y and Zheng HC).
Fig. 2Immunostaining of TMA of gastric carcinomas. The immunoreactivity to Ki-67 (a) and p53 (c) was localized in the nucleus; FHIT (d), caspase-3 (b), VEGF (g), maspin (e), P-GSK3β-ser9 (h), MUC-2 (i), MUC-5AC (k), MUC-6 (l) were seen in the cytoplasm, EMMPRIN (f) and MUC-4 (j) were seen in the cytoplasm and the plasma membrane, CD44 (m), and E-cadherin (n) in the plasma membrane; and β-catenin (o) in the nucleus, cytoplasm, and plasma membrane. These molecules were strongly expressed in the intestinal-type gastric carcinomas
Statistical analysis
The statistical evaluation was performed using Fisher's exact possibility to differentiate the rates and the Mann–Whitney U test to differentiate the means of the different groups. Kaplan–Meier survival plots were generated and comparisons between the survival curves were made with the log-rank statistic. The Cox proportional hazards model was employed for multivariate analysis. p < 0.05 was considered to represent a statistically significant difference. The SPSS 10.0 software program was employed to analyze all data.
Results
Clinicopathological characteristics of gastric IT, DT, and MT carcinomas
As summarized in Table 2, 415 IT cases were identified (51.0%), 221 DT (27.1%), and 178 MT (21.9%) among the 814 gastric carcinomas in this study. Patients with IT carcinoma were found to be significantly older than those with DT (p < 0.05). The male and female ratios in the carcinoma cases gradually decreased from IT to DT through the MT groups (p < 0.05). It was the same for lower TNM staging in the three groups (p < 0.05). The positive rate of peritoneal spread was higher in the DT and MT cases than IT (p < 0.05). MT carcinomas showed aggressive characteristics, such as large size, deep invasion, frequent local invasion, and lymph node metastasis, compared to the IT and DT (p < 0.05). There was no significance between the three types of gastric carcinomas in the incidence of the liver metastasis (p > 0.05).
Table 2 Comparison of the clinicopathological features in gastric IT, DT, and MT carcinomasClinicopathological featuresIntestinal-type carcinomaDiffuse-type carcinomaMixed-type carcinomaCase number415 (51.0%)221 (27.1%)178 (21.9%)Age (mean ± SD, years)67.09 ± 10.7362.03 ± 11.3*65.80 ± 13.28Sex (male:female)324:91125:96124:54**Tumor size (mean ± SD, cm)3.82 ± 3.215.50 ± 6.955.56 ± 3.08**Invasion into muscularis propria153 (36.9%)121 (54.8%)123 (69.1%)**Lymphatic invasion (+)117 (28.2%)83 (37.6%)99 (55.6%)**Venous invasion (+)49 (11.8%)32 (14.5%)39 (21.9%)**Lymph node metastasis (+)113 (27.2%)94 (42.5%)105 (59.0%)**Liver metastasis (+)10 (2.4%)8 (3.6%)6 (3.4%)Peritoneal spread (+)14 (3.4%)***22 (10.0%)21 (11.8%)TNM staging (O, I)302 (72.8%)69 (31.2%)112 (63.0%)***p < 0.05 compared with intestinal- and mixed-type carcinomas**p < 0.05 compared with intestinal- and diffuse-type carcinomas with both analyzed***p < 0.05 compared with diffuse- and mixed-type carcinomas
Pathological behaviors of gastric MT carcinomas with different growth patterns or histological distribution of components
According to the growth pattern, the intestinal and diffuse components diffusely distributed without absolute border in group I of the MT carcinomas, which occurred more frequently (60.0%, 106/178) than the other four groups. As indicated in Table 3, most groups III and V gastric carcinomas were generally of early stage with less local invasion, infrequent lymph node metastasis and low TNM staging. In this study, only four advanced gastric carcinomas belonged to group IV, among which three cases were women and all were accompanied with lymphangiogenic invasion, lymph node metastasis, and high TNM staging. Although the main histological component was different from that in lymph node metastasis, the statistical data revealed that there was histological consistency between primary and lymph node metastasis foci (p < 0.05). There were no remarkable differences in the pathological behaviors of MT carcinomas with different ratios of intestinal and diffuse portions (p > 0.05).
Table 3 Pathological behaviors of gastric MT carcinomasClinicopathological featuresnSex (male)Invasion into MPLymphatic invasionVenous invasionLymph node metastasisStagingnI>DI=DI<D(O, I)Growth pattern I10673 (68.9%)82 (77.4%)64 (60.4%)26 (24.5%)69 (65.1%)32142335 (33.0%) II2518 (78.3%)21 (84.0%)16 (64.0%)7 (28.0%)16 (64.0%)5567 (28.0%) III2922 (75.9%)11 (37.9%)9 (31.0%)3 (10.3%)10 (34.5%)43318 (62.1%) IV41 (25.0%)4 (100.0%)4 (100.0%)1 (25.0%)4 (100.0%)0220 (0.0%) V1410 (71.4%)5 (35.7%)6 (42.9%)2 (14.3%)6 (42.9%)2049 (64.3%)Histological appearance Intestinal>diffuse8059 (73.8%)54 (67.5%)43 (53.8%)20 (25.0%)48 (60.0%)3081033 (41.3%) Intestinal=diffuse2015 (75.0%)12 (60.0%)9 (45.0%)5 (25.0%)10 (50.0%)3528 (40.0%) Intestinal<diffuse7850 (64.1%)57 (73.1%)47 (60.3%)14 (18.0%)47 (60.3%)10112628 (35.9%)MP Muscularis propria; in lymph node metastatic foci: I>D intestinal>diffuse, I<D intestinal<diffuse, I=D intestinal=diffuse
Immunohistochemical analysis of intestinal or diffuse components from three types of gastric carcinomas
Table 4 demonstrates that all the markers except MUC-5AC were expressed at higher levels in IT carcinomas than in MT ones (p < 0.05). The expression of maspin, EMMPRIN, VEGF, MUC-4, and E-cadherin was stronger in the intestinal component of MT carcinomas than in their diffuse counterpart (p < 0.05). Immunoreactivities to Ki-67, EMMPRIN, and VEGF were weaker in the intestinal-type carcinomas than in the intestinal component of the MT (p < 0.05), while it was the opposite for CD44, MUC-2, and MUC-6 (p < 0.05). In contrast, the diffuse components of the MT carcinomas showed greater expression of Ki-67, FHIT, VEGF, and P-GSK3β-ser9 than DT carcinomas (p < 0.05).
Table 4 Immunohistochemical analysis in gastric IT, DT, and MT carcinomasBiological markersIntestinal-type carcinomaMixed-type carcinomaDiffuse-type carcinomaIntestinal partDiffuse partKi-6788/146 (60.3%)*,**81/110 (73.6%)78/110 (70.9%)*57/119 (47.9%)Caspase-384/152 (55.3%)*52/119 (43.7%)37/119 (31.1%)24/115 (20.9%)p5376/151 (50.3%)*37/118 (31.4%)30/118 (25.4%)25/120 (20.8%)FHIT90/150 (60.0%)*63/113 (55.8%)48/113 (42.5%)*27/116 (23.3%)Maspin71/150 (47.3%)*55/112 (49.1%)***34/112 (30.4%)38/120 (31.7%)EMMPRIN87/147 (59.2%)*,**77/112 (68.8%)***43/112 (38.4%)32/117 (26.5%)VEGF95/151 (62.9%)*,**90/114 (78.9%)***61/114 (53.5%)*31/121 (25.6%)P-GSK3β-ser988/144 (61.1%)*74/113 (65.5%)67/113 (59.3%)*48/116 (41.4%)MUC-255/145 (37.9%)*,**31/118 (26.3%)27/118 (22.9%)23/117 (19.7%)MUC-458/153 (37.9%)*39/118 (33.1%)***21/118 (17.8%)18/118 (15.3%)MUC-5AC84/148 (56.8%)57/106 (53.8%)50/106 (47.2%)54/113 (47.8%)MUC-660/146 (41.1%)*,**26/117 (22.2%)18/117 (15.4%)9/123 (7.3%)CD4464/152 (42.1%)*,**32/114 (28.1%)31/114 (27.2%)32/120 (26.7%)E-cadherin91/142 (64.1%)*58/112 (51.8%)***38/112 (33.9%)43/119 (36.1%)Membrane β-catenin57/144 (39.6%)*36/117 (30.8%)26/117 (22.2%)26/118 (22.0%)*p < 0.05 compared with the diffuse-type carcinomas**p < 0.05 compared with intestinal-part carcinomas***p < 0.05 compared with diffuse-part carcinomas
Patients' outcome with different gastric carcinomas
Follow-up information for 500 carcinoma patients was used before 2002 for a period ranging from 5 days to 9.15 years (mean = 49.4 months). Figure 3 shows the survival curves stratified according to Lauren's classification. Kaplan–Meier analysis indicated that the patients with IT carcinoma had a higher cumulative survival rate than those with the DT and MT lesions (p < 0.05). Although the MT carcinoma patients' survival rate was comparatively lower than the DT ones, there was no statistical significance (p > 0.05). Multivariate analysis demonstrated that invasive depth, venous invasion, lymph node, peritoneal or liver metastasis, and Lauren's classification were independent factors for the poor prognosis of gastric carcinoma patients (p < 0.05; Table 5).
Fig. 3Relationship between prognosis of carcinoma patients' and Lauren's classification Kaplan–Meier curves of cumulative survival rate for the patients with gastric carcinoma according to Lauren's classification. IT Intestinal type, DT diffuse type, MT mixed typeTable 5 Multivariate analysis of the clinicopathological variables of gastric carcinomasClinicopathological parametersRelative risk (95% CI)p ValueAge (≥65 years)1.224 (0.849–1.764)>0.05Sex (male:female)1.255 (0.827–1.904)>0.05Tumor size (≥4 cm)1.466 (0.862–2.493)>0.05Invasive depth (Tis,1/T2,3)2.945 (1.533–5.657)<0.05Lymphatic invasion (−/+)1.394 (0.865–2.247)>0.05Venous invasion (−/+)1.692 (1.097–2.608)<0.05Lymph node metastasis (−/+)1.983 (1.111–3.537)<0.05Peritoneal spread (−/+)3.197 (2.031–5.031)<0.05Liver metastasis (−/+)5.248 (2.575–10.697)<0.05Lauren's classification (intestinal/diffuse/mixed)1.351 (1.084–1.683)<0.05CI Confidence interval
Discussion
Histologically, Lauren's classification of gastric carcinomas is based on the morphological appearances. IT carcinomas are characterized by cohesive carcinoma cells forming gland-like tubular structures with an expanding or infiltrative growth pattern, like a papillary and well-differentiated adenocarcinoma. However, the cell cohesion is less apparent or absent in DT carcinomas, and cancer cells diffusely spread in the gastric wall as poorly differentiated adenocarcinomas, SRC, or undifferentiated carcinomas [23, 30]. However, there are two types of histological distribution in the MT carcinoma, as described by Lauren [16]. In this study, carcinoma diagnosed as MT made up 21.9% of the whole series in this population, more than that reported by Borch et al. (13%), Lauren (14%), and Stelzner and Emmrich (14.1%), and less than that reported by Carneiro et al. (38.5%) [2, 5, 16, 26]. This discrepancy might be due to the differences in diagnosis criteria, geographical regions, human races, living habits, and so forth. The age and sex distribution of the patients with MT carcinoma was intermediate although IT carcinomas frequently occurred in old men, compared with DT carcinomas.
Morphologically, there are both intestinal and diffuse components in MT carcinomas, and they were found to be significantly larger, more advanced, and more frequently spread into local lymphatic and venous vessels or regional lymph nodes, compared with IT and DT carcinomas in the present study, consistent with other reports [6, 16, 30]. Kozuki et al. [14] found that prominent lymphatic permeation and lymph node metastasis were more frequently observed in MT than in the pure type of gastric carcinomas. It was previously documented that the MT carcinomas showed a deeper infiltration of the gastric wall, a higher metastatic rate to the regional lymph nodes, and the need for higher staging with the TNM system of UICC classification, in contrast to IT and MT carcinomas [26]. These findings suggest that the pathobiological characteristics of the MT carcinomas were more aggressive than the other two types, which accounted for their comparatively poor prognosis observed in the present study although there was no statistical significance. Among the MT carcinomas, five groups were classified according to their growth pattern. Half of the present cases belonged to group I, which means there was diffuse distribution of the intestinal and diffuse component without clear borders in most of the MT cases. In general, groups III and IV of MT carcinomas did not exhibit deep invasion, comparatively high staging or frequent spreading, compared with other types. However, group IV of carcinomas with mucin and signet cells was more advanced with wide spreading, including local vessels and lymph node metastasis. Therefore, surgeons should carefully treat the patients with MT carcinomas in clinical practice because of their aggressive characteristics and poor prognosis, especially those classified as group IV. Furthermore, it was found that the major component in the MT primary foci morphologically paralleled to that in the corresponding lymph node metastasis although the converse situation also appeared. This inconsistent phenomenon could be explained by sampling bias or the metastatic potential of individual carcinoma cells.
Compared with IT carcinomas, the MT counterpart always exhibited more severe characteristics, including invasion into muscularis propria, lymphatic invasion and lymph node metastasis, and high TNM staging in the present study, as observed by Lauren and other studies [16, 26, 30]. To clarify the pheno/genotypes of gastric carcinoma, the expression of Ki-67, caspase-3, p53, FHIT, maspin, VEGF, EMMPRIN, P-GSK3β-ser9, MUC-2, MUC-4, MUC-5AC, MUC-6, E-cadherin, and β-catenin was examined by immunostaining. p53, FHIT, and maspin as tumor suppressor genes play important roles in regulating the balance between the proliferation and apoptosis of cancer cells [25, 30, 31, 34, 35]. Ki-67 antigen is present in the nuclei of cells undergoing proliferation and should be regarded as a good marker for cell proliferation [34]. Caspase-3 is responsible for the cleavage of poly (ADP-ribose) polymerase, actin, and sterol regulatory element binding protein and reflects the apoptotic level as a key protease in the cascade reaction of the apoptotic pathway [28]. In vivo and vitro evidences indicated that EMMPRIN and VEGF are involved in angiogenic processes in malignancies [29, 32]. Glycogen synthase kinase-3β (GSK3β) belongs to the serine/threonine protein kinase family and is also involved in regulating the balance between proliferation and apoptosis and can be inactivated via ser-9 phosphorylation by p70 S6 kinase, p90Rsk, Akt, certain isoforms of proteins kinase, and cyclic AMP-dependent protein kinase [1]. MUC-2, MUC-5AC, and MUC-6 are markers for intestinal goblet cells, superficial epithelium, and gastric pyloric gland cells, respectively, which can reflect the mucin secretion and variations in gastrointestinal malignancies. MUC-4 is a heterodimeric glycoprotein complex and expressed in several human epithelial carcinomas [3, 33]. CD44 is a cell surface glycoprotein involved in cell/cell and cell/matrix interactions. CD44 overexpression has been linked to the growth and spread of a range of different types of malignancies [8]. The E-cadherin can interact with β-catenin to form a complex, which is closely linked to cell adhesion and differentiation [19, 27, 36]. In the present study, all of the markers except MUC-5AC showed greater expression in the IT than in the DT carcinomas. Although some evidences indicate that MUC-5AC was strongly expressed in gastric DT carcinoma, in comparison with IT ones, there was no statistical difference in line with our finding [20, 22]. These data suggested that there were significant differences in the proliferation, apoptosis, angiogenesis, mucin production, and cell adhesion between the IT and DT carcinomas and these molecules mechanistically contributed to the molecular distinction in the morphological, behavioral, and histogenic aspects between both types of gastric carcinomas.
In general, DT carcinoma is believed to derive de novo from the peripheral stem cells of gastric gland neck proliferative zone without a recognizable precursor lesion except hereditary diffuse gastric cancer because in situ carcinoma or globoid dysplasia is its precancerous lesion, respectively, according to the Chinese and Western pathologists' opinions [4, 5, 9, 18,30]. When long-standing gastric inflammation causes tissue injury and stem cell failure with time, BMDCs are recruited and engrafted into the tissue stem cell niche, where BMDCs can behave in a way indistinguishable from endogenous tissue stem cells. With continued inflammation and injury, they can undergo sustained proliferation and malignant transformation into IT carcinomas, passing through precancerous stages of metaplasia and dysplasia when genetic defects, such as mutation, deletion, or rearrangement, are accumulated resulting in corruption of the balance between proliferation and apoptosis [17]. Likewise, dedifferentiation of intestinal to diffuse carcinoma had been identified as another histogenic pathway according to the histological heterogeneity of tumor cells, especially in the carcinomas containing moderately and poorly differentiated components with similar morphological appearance and diffuse distribution. If so, it is possible that the distinct components in MT carcinomas arise from the stem cells with common genetic traits and follow different carcinogenic pathways. Conversely, diffuse carcinoma could be derived from heterogeneity of the intestinal counterpart. According to our present data, no differences in the expression of Ki-67, caspase-3, FHIT, CD44, P-GSK3β-ser9, MUC-2, MUC-5AC, MUC-6, and membrane β-catenin were observed between the intestinal and diffuse components of MT carcinomas, which supports the possibility of similar origin or dedifferentiation. However, increased expression of EMMPRIN, VEGF, MUC-4, and E-cadherin in the intestinal component compared to the diffuse counterpart also provided evidence that the original carcinoma cells might undergo distinct carcinogenic routes resulting in the morphological distinction of both components. Both the intestinal and diffuse components in MT carcinomas had increased expression of Ki-67, EMMPRIN, and VEGF and reduced E-cadherin. Serum VEGF levels were found to be significantly higher in patients with MT gastric carcinomas than those with pure lesions [10]. It was proposed that MT carcinomas biologically displayed more aggressive behaviors than other types, including decreased cell adhesion, increased proliferation, and angiogenesis.
Most of the evidences in the present data indicated that the pathological behaviors of MT carcinomas were more aggressive than the other types, which was closely linked to the prognosis. It was previously documented that the MT patients' survival is significantly worse than those with IT or DT carcinomas regardless of their location [5]. However, this study demonstrated that there was no significance between the MT and DT carcinoma patients' survival rates although the former was lower than the latter. Additionally, both survival rates were lower than that with IT carcinomas. To avoid sampling bias from a shorter follow-up time, the older cases were chosen for the survival analysis, but they yielded consistent results. Furthermore, Cox's hazard proportional analysis indicated that the invasive depth, venous invasion, lymph node, peritoneal or liver metastasis, and Lauren's classification were independent prognostic factors for gastric carcinomas.
In summary, IT carcinoma, which is positively correlated with favorable prognosis, frequently displayed high levels of proliferation, apoptosis, angiogenesis, mucin production, and cell adhesion. Gastric MT carcinoma showed more aggressive behaviors than IT and DT ones. There was a significant difference in the proliferation, apoptosis, angiogenesis, mucin secretion, and cell adhesion between the IT and DT carcinomas, whereas only a few characteristics were differentially detected in the intestinal and diffuse component of the mixed-type carcinoma, suggesting that different components of MT carcinoma might originate from common stem cells, but follow distinct histogenic pathways. Furthermore, these results confirm that Lauren's classification is significant regarding the histopathogenesis and differentiation and considered as a guide to the clinical treatment of gastric carcinoma.
|
[
"gastric carcinoma",
"pathobiological behaviors",
"carcinogenesis",
"lauren's classification",
"mixed type"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Biochim_Biophys_Acta-2-1-2323027
|
Translational control by cytoplasmic polyadenylation in Xenopus oocytes
|
Elongation of the poly(A) tails of specific mRNAs in the cytoplasm is a crucial regulatory step in oogenesis and early development of many animal species. The best studied example is the regulation of translation by cytoplasmic polyadenylation elements (CPEs) in the 3′ untranslated region of mRNAs involved in Xenopus oocyte maturation. In this review we discuss the mechanism of translational control by the CPE binding protein (CPEB) in Xenopus oocytes as follows:1.The cytoplasmic polyadenylation machinery such as CPEB, the subunits of cleavage and polyadenylation specificity factor (CPSF), symplekin, Gld-2 and poly(A) polymerase (PAP).2.The signal transduction that leads to the activation of CPE-mediated polyadenylation during oocyte maturation, including the potential roles of kinases such as MAPK, Aurora A, CamKII, cdk1/Ringo and cdk1/cyclin B.3.The role of deadenylation and translational repression, including the potential involvement of PARN, CCR4/NOT, maskin, pumilio, Xp54 (Ddx6, Rck), other P-body components and isoforms of the cap binding initiation factor eIF4E.
1
Introduction
The regulation of translation of mRNAs by cytoplasmic elongation of the poly(A) tail was discovered some twenty years ago in the oocytes and early embryos of clams, worms, frogs and mice [1–6]. These maternal mRNAs are stored in the growing oocyte with a short poly(A) tail of 20 to 40 nucleotides and are translationally repressed (masked). Upon oocyte maturation or after fertilisation, the poly(A) tail of masked mRNAs is elongated to 80–250 residues and the mRNAs are translationally activated. A long poly(A) tail is thought to be stimulatory to translation through the binding of cytoplasmic poly(A) binding proteins, which recruit initiation factors and form a closed-loop complex through their association with the translation initiation factor eIF-4G [7]. Cytoplasmic polyadenylation is essential for the meiotic maturation of the oocyte in both Xenopus and mouse, as it mediates the translational activation of the mRNAs encoding c-Mos kinase and mitotic cyclins [8]. In addition, cytoplasmic polyadenylation has been implicated in the cell cycle, in cellular senescence and in the formation of memory through synaptic plasticity [9]. In this review we will discuss the current knowledge of the mechanism of translational control by cytoplasmic polyadenylation, with an emphasis on Xenopus oocyte maturation and CPEB, the cytoplasmic polyadenylation element binding protein. A review of the role of cytoplasmic polyadenylation in mammalian gametogenesis is provided elsewhere in this volume (Tadashi Baba).
2
mRNA elements and RNA binding proteins
In order for mRNAs to be singled out for cytoplasmic polyadenylation, they have to be recognised by the polyadenylation machinery. In oocytes and early embryos of the African clawed frog, Xenopus, cytoplasmic polyadenylation can be conferred by several discrete elements in the 3′ untranslated region (3′ UTR) of the mRNA which are recognised by specific RNA binding proteins. At least 4 different cytoplasmic polyadenylation elements have been identified in Xenopus. The C-rich cytoplasmic polyadenylation element and the U-rich embryonic cytoplasmic polyadenylation element normally mediate cytoplasmic polyadenylation in the zygote and early embryo [10,11]. The putative cytoplasmic polyadenylation factors binding to these elements are poly(rC) binding protein 2 (PCBP2) and Elav related protein A, the ortholog of HuR, respectively [10,12–14]. A polyadenylation element that functions early in oocyte maturation, the polyadenylation response element (PRE), was predicted computationally [15,16]. The RNA binding protein Musashi was shown to bind to some, but not all PREs and mutating Musashi or its known consensus binding site blocked cytoplasmic polyadenylation, indicating that it is likely to be a cytoplasmic polyadenylation factor [17]. These data suggest that the collection of cytoplasmic polyadenylation elements that have been called PREs may represent a mixture of Musashi binding sites and at least one more uncharacterised cytoplasmic polyadenylation element.
By far the best characterised cytoplasmic polyadenylation element is the CPE, which is required for the cytoplasmic polyadenylation of a number of mRNAs, including cyclin B1 mRNA, during oocyte maturation and during the embryonic cell cycle [18–20]. The consensus CPE is often described as U5AU, but variations such as U4AU (c-mos), and U4–5A2U (cyclin B1) have also been shown to be active [5,6,21,22]. The CPEs in the mRNA encoding mouse c-Mos are active in mature Xenopus oocytes, but only conform to U5A [23,24]. Strongly polyadenylating mRNAs, however, generally adhere to the U5A1–2U consensus. In addition, examination of the cases in the existing literature suggests that CPEs that mediate significant polyadenylation tend to be close to the polyadenylation signal, from downstream to overlapping and up to approximately 60 nt upstream [21,25–29].
The CPE binds cytoplasmic polyadenylation element binding protein (CPEB), an RNA binding protein containing two RNA recognition motifs and a zinc finger region, all of which are required for recognition of the CPE [21,30,31]. In addition to the fact that many mRNAs bearing a CPE undergo cytoplasmic polyadenylation, a wealth of evidence indicates that CPEB is required for cytoplasmic polyadenylation. Depletion of CPEB from an egg extract abolishes polyadenylation and injection of a CPEB antibody blocks it in oocytes and embryos [20,21,31,32]. N-terminal deletions and phosphorylation site mutants of CPEB act as dominant negative blockers of cytoplasmic polyadenylation [20,21,31,32]. In mice, knock out or knock down of Cpeb1 causes poly(A) tail changes in target mRNAs in early oogenesis and oocyte maturation [33,34]. CPEB is therefore the only well-established mRNA specificity factor for cytoplasmic polyadenylation in vertebrates.
In addition to an mRNA specific element, the nuclear poly(A) signal (AAUAAA or AUUAAA) is absolutely required for cytoplasmic polyadenylation [5,6,11,22,28]. The poly(A) signal is thought to recruit a variant of the cleavage and polyadenylation specificity factor (CPSF), a four subunit complex that binds this element and mediates polyadenylation in the nucleus [35,36]. The 160 kDa subunit is the RNA binding protein that recognises the poly(A) signal [37], while the 73 kDa subunit is the endonuclease that mediates the formation of the 3′ ends of all mRNAs [38–41]. The 100 kDa subunit resembles the 73 kDa subunit but does not appear to have nuclease activity and its precise function in 3′ end processing is unknown. The 30 kDa subunit is a zinc finger protein that has been reported to have some affinity for RNA and it potentially has endonuclease activity. It was recently shown to mediate interactions between CPSF and the body of RNA polymerase II [42–45]. An additional CPSF associated factor, Fip1, has been shown to mediate the interaction with poly(A) polymerase and have U-rich RNA binding activity [46]. It has probably been present in most of the previously utilised preparations of CPSF and is thought of as a fifth subunit by many in the field [40,46].
The 100 and 30 kDa CPSF subunits are present in the Xenopus oocyte cytoplasm, and the 160 kDa subunit has been inferred to be present from its binding activity, while the 73 kDa endonuclease is absent from the cytoplasmic CPSF complex [47]. It is unknown whether Fip1 is present in the oocyte cytoplasm. Immunodepletion of the 100 kDa subunit from egg extracts or expression of a viral protein that blocks nuclear polyadenylation by binding to the 30 kDa subunit also abolishes cytoplasmic polyadenylation, demonstrating that CPSF is required for this process [47]. In addition, CPEB is coimmunoprecipitated with the 100 kDa CPSF subunit from oocyte extracts (but not with the 73 kDa subunit) and it binds to the 160 kDa subunit in reticulocyte lysate [47,48]. Whether these interactions are direct or mediated by other proteins present in both oocyte extract and reticulocyte lysate is not yet clear. Curiously, the binding of CPSF to the poly(A) signal is stimulated by a CPE even in the absence of CPEB, a phenomenon that could be mediated by the RNA binding activity of the 30 kDa subunit or by Fip1 [46,49]. However, CPEB strongly stimulates polyadenylation of a CPE containing RNA in a pure reaction system containing nuclear CPSF and poly(A) polymerase [48]. It is therefore likely that CPEB improves the recruitment of CPSF to the mRNA by binding directly to both.
3
Symplekin
Symplekin is a protein found in nuclear complexes containing CPSF and other processing factors and is thought of as a scaffold protein involved in 3′ end RNA processing [50–52]. In yeast and plants, homologues of symplekin are required for proper cleavage and polyadenylation, and this is possibly also the case in vertebrates [53–55]. In vertebrate somatic cells symplekin is predominantly nuclear and in epithelial cells also localised at tight junctions [50,56]. In the Xenopus oocyte, however, symplekin is also found in the cytoplasm in complexes with the 100 kDa subunit of CPSF [50]. A symplekin antibody also precipitates CPEB from oocyte extracts, indicating that the protein is in cytoplasmic polyadenylation complexes [57]. Moreover, the symplekin antibody inhibits cytoplasmic polyadenylation in oocytes and extracts [57]. These data strongly suggest that symplekin is involved in cytoplasmic polyadenylation.
4
The cytoplasmic poly(A) polymerase
Initially, a form of one of the classical nuclear poly(A) polymerases was thought to be responsible for cytoplasmic polyadenylation too. In vitro reconstitution experiments indicated that the poly(A) polymerase (PAP) isolated from bovine thymus can mediate CPE and poly(A) signal dependent polyadenylation [47–49]. Cytoplasmic poly(A) polymerase species were detected in Xenopus oocytes and an isoform was cloned that was 87% identical to bovine PAP but lacked the nuclear localisation signal [58,59]. Three different polyclonal antibodies against this protein inhibited polyadenylation in egg extracts [58]. However, after these findings the work on the cytoplasmic poly(A) polymerase stagnated. No native complex containing CPEB and a PAP isoform was ever reported and studies on the regulation of PAP activity indicated that the enzyme is inactivated by phosphorylation during oocyte maturation, when cytoplasmic polyadenylation is actually induced [60].
In 2002, a Caenorhabditis elegans germline determinant called gld-2 was cloned and shown to encode a cytoplasmic poly(A) polymerase that is recruited to certain mRNAs by an RNA binding protein [61]. The Gld-2 protein belongs to the same large family of DNA polymerase β nucleotidyl transferases, but only has limited additional homology to the classical poly(A) polymerases and it lacks the RNA binding domain. It soon transpired that this protein is widely conserved, having homologues in fission yeast and mammalians [62,63].
The Xenopus Gld-2 protein (xGld-2) coimmunoprecipitated very efficiently with symplekin in both mature and immature oocytes. In addition, tagged xGld-2 was shown to bind the 160 kDa CPSF subunit as well as CPEB synthesised in reticulocyte lysate [57]. This indicates that xGld-2 can form complexes with symplekin, CPSF and CPEB, although the direct interactions are as yet unclear. Gld-2 overexpressing oocytes and extracts supplemented with Gld-2 displayed an increase in their CPE-dependent polyadenylation and symplekin depleted extracts regained some polyadenylation activity if they were replenished with recombinant CPEB, symplekin and Gld-2 [57].
The less than optimal reconstitution of cytoplasmic polyadenylation by Gld-2 may indicate that cytoplasmic PAP or an unknown additional cytoplasmic polyadenylation factor is being depleted with symplekin. However, it is also very plausible that not all the supplemented proteins are completely active. Especially CPEB is notoriously difficult to obtain in a soluble form and isolation procedures usually include a denaturing step [30]. On the whole, the evidence seems to indicate that xGld-2 is an authentic cytoplasmic poly(A) polymerase in Xenopus oocytes, but involvement of the cytoplasmic PAP can as yet not be excluded. In Drosophila, PAP has been implicated in cytoplasmic polyadenylation in embryos [64], and although initial experiments implicated mouse Gld-2 in cytoplasmic polyadenylation in oocytes, polyadenylation appeared unimpaired in a knock out mouse [65,66]. The cytoplasmic polyadenylation mechanisms in mammalian germ cells are discussed in more detail by T. Baba elsewhere in this volume.
What part of the cytoplasmic polyadenylation complex recruits either PAP or Gld-2 is also still an open question. It is possible that CPEB replaces Fip1 in cytoplasmic polyadenylation, or that another specific recruitment factor exists for Gld-2. Also, the nuclear cleavage stimulation factor CstF77 has been reported to be present in the cytoplasmic polyadenylation complex in Xenopus oocytes, but appears to have an as yet undefined role in translational repression, rather than in cytoplasmic polyadenylation [67].
In nuclear polyadenylation, binding of the nuclear poly(A) binding protein, PABP2, is required to limit the length of the poly(A) tail. It has been suggested that such a function is required in cytoplasmic polyadenylation, as symplekin immunoprecipitates from polyadenylating egg extracts add abnormally long poly(A) tails [57]. It is unknown if the length of the poly(A) tail added is controlled in cytoplasmic polyadenylation, but it is worth noting in this context that the main poly(A) binding protein in oocytes is not the somatic cytoplasmic PABP, but an embryo specific protein, ePABP, which may have a specific function in this context [68,69].
Recently, it was shown that ePAB and somatic PABP both can bind to CPEB, and that this interaction is reduced by phosphorylation of CPEB during oocyte maturation [65,70]. The ePABP–CPEB interaction has been proposed to stimulate cytoplasmic polyadenylation and translational activation, as demonstrated by depletion experiments [65]. ePABP may therefore be a cytoplasmic polyadenylation factor. However, the data presented appear to show that symplekin is partially depleted with ePABP, which could explain the reduction in polyadenylation. The add-back experiment does unfortunately only show a minimal restoration of polyadenylation. Depletion of CPEB with ePABP is described for experiments described in the discussion section of this paper, adding to these concerns. In addition, because most panels in this paper do not show direct comparisons between CPE containing and CPE lacking RNA substrates, the intriguing effects of ePABP depletion could perhaps also be explained by a general effect of enhanced deadenylation and reduced translation of all polyadenylated mRNAs, which are predicted effects of low ePABP levels [68,69]. For the moment, it therefore appears prudent to reserve judgement on the function of the CPEB–ePABP interaction.
5
The activation of cytoplasmic polyadenylation
In Xenopus oocytes, meiotic maturation is induced by steroid hormones, probably through the action of both classical nuclear steroid receptors and a G protein coupled transmembrane receptor [71–73]. A rapid drop in cyclic AMP leads to the inactivation of protein kinase A, which is necessary for the induction of cytoplasmic polyadenylation. In recent years, progress has been made in various aspects of the activation mechanisms of cytoplasmic polyadenylation, although the picture is far from complete.
Tethering of mammalian Gld-2 to an mRNA using a fusion with an RNA binding domain and an mRNA with a binding site causes strong polyadenylation in immature Xenopus oocytes, indicating that the protein is constitutively active and only needs to be recruited to the mRNA to elongate the poly(A) tail [62]. However, as described above, Gld-2 is already associated with the polyadenylation factors in immature oocytes [57]. In addition, Gld-2 overexpression also leads to robust polyadenylation of CPE containing RNA in immature oocytes, which implies that the protein is probably inactivated by factors that are titrated by the overexpressed protein [57]. Since the action of Gld-2 requires recruitment to the RNA for activity, the titrated factor is unlikely to be CPEB itself. An intriguing possibility is that the protein depleted by Gld-2 is the cytoplasmic isoform of poly(A) ribonuclease (PARN). A 64 kDa catalytically active fragment of PARN has been found to coimmunoprecipitate with symplekin, CPEB and Gld-2 in immature but not in mature oocytes and binds to CPEB and Gld-2 synthesised in reticulocyte lysate [74]. The full length PARN is not abundantly found in this complex. The 64 kDa PARN isoform is thought to be responsible for the short poly(A) tails of CPE containing mRNAs, as a PARN antibody inhibited CPE-mediated deadenylation in immature oocytes [74]. In addition, catalytically inactive PARN can induce polyadenylation of CPE containing RNAs in immature oocytes and overexpression of wild type PARN represses the polyadenylation induced by Gld-2 overexpression [74]. These data indicate that PARN is the repressive factor depleted by Gld-2 injection and that it is the ejection of PARN from the polyadenylation complex that induces cytoplasmic polyadenylation [74]. There are two possibilities to explain the action of PARN on CPE containing mRNAs, either Gld-2 continuously adenylates while PARN takes these nucleotides off, or the presence of PARN blocks Gld-2 activity, for instance by preventing access to the 3′ end of the mRNA.
Since not all mRNAs are activated by cytoplasmic polyadenylation at the same time, the activation of the complexes must depend at least in part on mRNA specificity factors. In maturing Xenopus oocytes, cytoplasmic polyadenylation has been classed into 2 types, early (class I) and late (class II). Early polyadenylating mRNAs gain a poly(A) tail well before the activation of cdk1/cyclin B and the breakdown of the nuclear envelope (germinal vesicle breakdown, GVBD). Early polyadenylation is independent of the synthesis of the kinase c-Mos (which is itself encoded by a class I mRNA), while late polyadenylating mRNAs are dependent on c-Mos synthesis and the subsequent cdk1/cyclin B activation and gain their tails after GVBD [16,75,76]. The less well characterised polyadenylation response element (PRE, see above) and its binding factor Musashi have been shown to be responsible for some of the early polyadenylation events, despite the fact that these mRNAs also contain CPEs [15,17]. However, there is a clear correlation between the placement of the CPEs in an mRNA and the timing of its polyadenylation [15,77,78]. In addition, some late polyadenylating mRNAs also contain PREs/Musashi binding sites and the histone B4 mRNA, which contains both a PRE and a CPE, has been reported to be polyadenylated early, but has been shown to be Mos-dependent, indicating that the current classification in early and late perhaps oversimplifies the situation [15,17,75,78].
The evidence for which factor mediates early polyadenylation rests primarily on the use of dominant negative Musashi and CPEB mutants, which block translation and polyadenylation of c-Mos mRNA and GVBD [15,17,48]. Charlesworth et al. compared Musashi and CPEB mutants, and came to the conclusion that Musashi alone mediates the early polyadenylation, in contrast to earlier findings by Mendez et al., which implicated CPEB in the early polyadenylation [17,32]. The CPEB phosphorylation mutant used by Charlesworth et al. is a rather weak dominant negative (L.E. Hake, Boston, USA, unpublished observation), so it is possible that CPEB-dependent early polyadenylation is not completely repressed by it in all cases. Differences in expression levels and examination of endogenous or injected RNAs could also explain the discrepancies between the two studies. In addition, either or both dominant negative proteins could be titrating other cytoplasmic polyadenylation factors and inhibit polyadenylation on mRNAs to which they do not normally control, or they could fail to enter a specific mRNA–protein complex due to low exchange rates. Importantly, injection of a CPEB antibody prevented GVBD and gave strong repression of polyadenylation of injected early RNAs, including a Mos 3′ UTR construct, indicating that CPEB-mediated early polyadenylation exists [21]. In addition, interference with pathways that are known to affect CPEB phosphorylation and activation leads to a reduction in pre-GVBD c-Mos mRNA polyadenylation and a delay or block in oocyte maturation, including in those papers that claim that early polyadenylation is CPEB-independent [15,16,32,79,80]. However, there remains the possibility that the same signal transduction pathways ultimately activate Musashi and CPEB, which would make it impossible to distinguish between the effects of the two proteins in these experiments, with the possible exception of the CPEB antibody injection. With the limitations of the currently available evidence, it seems safest to assume that Musashi and CPEB both contribute to the polyadenylation of early mRNAs and that some mRNAs may be exclusively dependent on one or the other. Further work is required to resolve this issue definitively.
The induction of Xenopus oocyte maturation and early cytoplasmic polyadenylation by progesterone is dependent on a drop in cAMP levels and the resulting inhibition of protein kinase A [17,81]. Downstream of this event is the phosphorylation of CPEB on Ser174, which happens early in oogenesis. Although CPEB, CPSF and Gld-2 are already in a complex in immature oocytes, this phosphorylation appears to induce a stronger association of CPEB with CPSF and Gld-2 [32,48,82]. Importantly, Ser174 phosphorylation also induces the ejection of PARN from the polyadenylation complex [74]. Mutation of Ser174 to alanine results in a dominant negative protein that inhibits oocyte maturation and early polyadenylation partially or fully and late polyadenylation completely [15,32,83]. In addition, injection of a peptide containing the phosphorylation site delays GVBD considerably, suggesting that it is inhibiting an early polyadenylation event [32]. All evidence therefore indicates that phosphorylation of Ser174 is a crucial event in the activation of cytoplasmic polyadenylation.
There is some controversy surrounding the kinase mediating the early phosphorylation of CPEB on Ser174. Initially, it was demonstrated that Aurora A kinase (Eg2) can phosphorylate this site and that this requires the N-terminus of CPEB, in addition to the phosphorylation site itself [32]. This suggested that the N-terminally deleted CPEB may owe its dominant negative properties to the lack of a binding site for Aurora A. In addition, in vitro polyadenylation reactions containing recombinant CPEB, purified CPSF and nuclear PAP are strongly stimulated by the addition of Aurora A [48]. Glycogen synthase kinase 3 (GSK3) is thought to negatively regulate Aurora A by phosphorylation and this block appears to be relieved during oocyte maturation [83]. Consistent with a role in early oocyte maturation, GSK3 overexpression inhibited GVBD in Xenopus oocytes [83]. However, studies from three independent groups have failed to detect Aurora A activation early in oocyte maturation [79,84–86]. In addition, Aurora A inhibitors do not affect oocyte maturation or CPEB phosphorylation and depletion of Aurora A from oocyte extracts does not inhibit early CPEB phosphorylation either [79]. There is therefore sufficient evidence to throw serious doubt on the role of Aurora A in the activation of CPE-mediated cytoplasmic polyadenylation early in oogenesis. In mouse neurons, calmodulin dependent kinase II (CamKII) can phosphorylate the corresponding site on mouse CPEB, Thr171 [87]. CamKII plays a role in the induction of mouse oocyte maturation and is known to be present in Xenopus eggs, and could therefore be an alternative candidate kinase for mediating the early phosphorylation [88–91]. Because CamKII can be activated independently of calcium, the absence of a calcium flux in early oocyte maturation is not necessarily an obstacle for this hypothesis [92]. Evidence for a requirement of CamKII activity early in oocyte maturation would be necessary to make this more than speculation. However, Aurora A is very active later in oogenesis and it is undeniably capable of activating CPEB. It therefore appears likely that at the very least the maintenance of CPEB phosphorylation on Ser174 is carried out by this kinase.
Mitogen activated protein kinase (MAP kinase) has been implicated in the activation of early CPE-mediated, but not PRE-mediated, cytoplasmic polyadenylation [16,17,79,93]. An early, c-Mos independent activation of MAP kinase has been detected during oocyte maturation and CPEB is phosphorylated by MAP kinase on multiple sites [32,79,94]. MAP kinase does not phosphorylate Ser174, but has been suggested to either prime CPEB for Ser174 phosphorylation or to activate the Ser174 kinase [79]. As the early MAPK activation is dependent on protein synthesis, translational activation of another mRNA is probably required [94]. One potential candidate is RINGO/Speedy, a Cdk1 interacting and activating protein that is transiently expressed after progesterone treatment and required for oocyte maturation and CPEB phosphorylation [65,95–98].
A guanine nucleotide exchange factor for the Rho family of GTPases was found to interact with the N-terminal half of CPEB in a yeast 2 hybrid screen and called XGef [80]. This interaction was confirmed in oocytes and has been shown to be direct [80,99]. There appear to be 2 binding sites for CPEB on XGef [99]. Antibodies against XGef block oocyte maturation and prevent the polyadenylation and translation of c-Mos mRNA [80,99]. Conversely, overexpression of XGef accelerates oocyte maturation and c-Mos polyadenylation, independent of the production of c-Mos protein, indicating that it is upstream of this kinase in the signal cascade of meiotic maturation [80]. XGef enhances the early phosphorylation of CPEB and the DH domain associated with nucleotide exchange activity is required for its effects of on cytoplasmic polyadenylation [80,99]. In addition, a mutant of XGef that retains nucleotide exchange activity but has impaired binding to CPEB reduced early CPEB phosphorylation and delayed oocyte maturation [99]. However, a broad spectrum Rho GTPase inhibitor did not affect oocyte maturation or CPEB phosphorylation, indicating that the DH domain may not function to activate a Rho family GTPase, but is required in another capacity [79]. Strikingly, XGef immunoprecipitates were found to contain MAPK in both immature and mature oocytes and it therefore may be required to bring CPEB to the signalling complexes involved in its phosphorylation [79].
In another yeast 2 hybrid study, mouse CPEB was found to bind the small intracellular domain (ICD) of the transmembrane protein amyloid precursor like protein 1 (APLP1) and its relatives [100]. In Xenopus oocytes, Gld-2, the CPSF 100 kDa subunit and symplekin were predominantly associated with the plasma membrane in the same fractions as overexpressed APLP1 and immunoprecipitation of CPEB from these fractions indicates that it can associate with APLP1 [100]. Immunogold electron microscopy confirmed that CPEB and CPE containing RNA is associated with membranes. Overexpression of full length APLP1 induced some cytoplasmic polyadenylation in untreated oocytes and enhanced the effect of low concentrations of progesterone. The ICD alone had even stronger effects, indicating that the action of APLP1 is not dependent on membrane localisation. After treatment with low concentrations of progesterone, APLP1 stimulated CPEB phosphorylation on Ser174 [100]. While the association with amyloid precursor proteins may have great significance for the role of CPEB and cytoplasmic polyadenylation in neurons, it is as yet unclear whether APLP1 or its relatives are required for cytoplasmic polyadenylation in oocytes and whether it mediates the membrane association of the cytoplasmic polyadenylation complexes.
Late cytoplasmic polyadenylation is required for progression from meiosis I to meiosis II during oocyte maturation [101]. mRNAs that undergo late polyadenylation often have a CPE overlapping with their poly(A) signal, e.g. UUUUAAUAAA [15,77,78]. The late polyadenylation during oocyte maturation is dependent on the activation of the mitotic kinase cdk1 and its regulatory cyclin subunits [16,77,78,93]. Cdk1 does phosphorylate CPEB leading to degradation of most of the oocyte CPEB via ubiquitin-mediated degradation [101–103]. This degradation is required for late polyadenylation to occur. Remarkably, symplekin bound CPEB appears to represent the stable fraction of CPEB, as symplekin immunoprecipitates do not show a difference in CPEB content between immature and mature oocytes [57]. This implicates that CPEB in polyadenylation complexes is stable during oocyte maturation, consistent with the polyadenylation activity observed. A model for late cytoplasmic polyadenylation can be devised in which the abundant ‘free’ CPEB (i.e. CPEB in a different complex than symplekin, CPSF and the poly(A) polymerase) binds to the CPE overlapping the poly(A) signal in late polyadenylating mRNAs and prevents the recruitment of CPSF and symplekin by the polyadenylation signal early in oocyte maturation. After cdk1 activation, most free CPEB is depleted and a complex containing CPEB, symplekin and CPSF binds both the CPE and the poly(A) signal [101].
A current view of CPE-mediated cytoplasmic polyadenylation in oocyte maturation is depicted in Figs. 1 and 2 and can be summarised as follows:1.The activation of CPE-mediated cytoplasmic polyadenylation during progesterone induced oocyte maturation is induced by a drop in protein kinase A activity and requires an early translation event, perhaps translation of RINGO/Speedy.2.This induces the early activation of MAP kinase, which is associated with the polyadenylation complex through XGef and phosphorylates CPEB on multiple sites, but not on Ser174.3.CPEB is phosphorylated on Ser174 by an as yet not fully confirmed kinase, possibly Aurora A or CamKII, which is required for the induction of CPE-mediated polyadenylation and causes an increase in the binding between Gld-2, CPSF and CPEB, causing the ejection of PARN from the complex and allowing Gld-2 to elongate the poly(A) tail of the mRNA.4.After GVBD, CPEB is phosphorylated by cdk1 and the free CPEB is mostly degraded, allowing the CPEB in polyadenylation complexes (now probably phosphorylated on Ser174 by Aurora A) to activate cytoplasmic polyadenylation on mRNAs which contain a CPE overlapping with the poly(A) signal.
A surprising number of proteins involved in the activation of polyadenylation are themselves upregulated by this process. This should create positive feedback loops which amplify the signal and contribute to the progression of meiotic maturation. As discussed above, the mRNA for the serine/threonine kinase c-Mos is one of the early targets of cytoplasmic polyadenylation [16,75,76]. It is an activator of the MAPK kinase MEK, and its synthesis leads to further activation of MAP kinase, activation of cdk1 and phosphorylation of CPEB [71]. The mitotic cyclins, activators of cdk1, are induced by cytoplasmic polyadenylation, as is Aurora A [19,26]. Similarly, CamKII is regulated by cytoplasmic polyadenylation in neurons [104]. Finally, Gld-2 targets its own mRNA, which contains CPEs in its 3′ UTR[105]. A system with so much positive feedback will require strong brakes, and it appears that at least in some cases this is provided by deadenylation factors, as discussed below.
6
Deadenylation and translational repression
The maternal mRNAs that are stored in the oocyte in an untranslated state have short poly(A) tails, but have been reported to have normal poly(A) tail addition in the nucleus in both mouse and frog [74,106]. In both organisms, the deadenylation of RNA injected into the cytoplasm requires CPEs, indicating that the CPEB–PARN complex, discussed above, mediates this process [74,106]. However, RNA substrates containing CPEs alone have not been tested in these experiments and it is therefore not clear if a CPE is sufficient for deadenylation or if other sequence elements are required. In fact almost all CPE containing mRNAs have putative binding sites for other deadenylation factors.
Firstly, the deadenylation factor Pumilio binds directly to CPEB, and therefore could contribute to the deadenylation of all CPE containing mRNAs [107,108]. Many CPE containing mRNAs also contain Pumilio binding sites, including cyclin B1 and Gld-2 [97,107–109], indicating that the recruitment of Pumilio maybe both by protein–protein and RNA–protein associations. Pumilio is a member of a highly conserved family of RNA binding proteins called the Puf family. Members of this family can mediate translational repression and mRNA destabilisation in organisms from yeast to vertebrates [110,111]. In yeast, the Puf family protein Mpt5 interacts with the conserved deadenylase complex CCR4–Pop2–Not, by binding to Pop2/Caf1 [112–114]. This association was also confirmed for the human and worm proteins [113]. In addition, Pumilio is known to aid the recruitment of the translational repressor Nanos, both in fly and frog [107,115]. Nanos can recruit the CCR4–Pop2–Not complex by binding to Not4 and contributes to the translational repression of fly cyclin B [116]. So far, the role of the CCR4–Pop2–Not complex in poly(A) tail metabolism has not been studied in Xenopus oocytes and embryos, but the conservation of both the recruitment proteins and one of the target mRNAs (cyclin B) seems to indicate that it is very likely to play a role.
A subset of mRNAs that are activated by CPE-mediated cytoplasmic polyadenylation during oocyte maturation loose their poly(A) tails after fertilisation [19,117,118]. In most cases, this is mediated by an embryonic deadenylation element (EDEN) in the 3′ UTR of the mRNA, which binds the deadenylation factor EDEN-BP [119]. Aurora A, cyclin B1 and c-Mos encoding mRNAs all bind EDEN-BP and the 3′ UTR of Gld-2 mRNA contains good consensus binding sites, making EDEN-BP a good candidate repressor of the cytoplasmic polyadenylation-mediated positive feedback loops during early embryogenesis [105,119–121]. The mammalian ortholog of EDEN-BP, CUGBP1, can recruit PARN [122]. A similar association of EDEN-BP and PARN in the Xenopus embryo could explain its role in deadenylation, but so far this complex has not been reported. It can therefore not be excluded that EDEN-BP mediates its deadenylation activity in Xenopus embryos through another deadenylase, for instance the CCR4–Pop2–Not complex. No direct contacts between CPEB containing and EDEN containing complexes has been described so far. Because CPEs, EDEN sequences, Musashi and Pumilio binding sites often occur in the same 3′ UTR, their associated complexes can be expected to compete for the end of the mRNA to mediate deadenylation or polyadenylation.
Although a short poly(A) tail will lead to inefficient translation of an mRNA, it is not sufficient for complete translational repression, as illustrated by the partial polysomal association of histone B4 mRNA and the accumulation of B4 protein in Xenopus oocytes [78,123,124]. Many of the other stored maternal mRNAs have to be strongly translationally repressed to enable the subsequent stages of growth, maturation and fertilisation of the oocyte [71]. CPEs, Musashi and Pumilio binding sites can repress translation of specific mRNAs, often in the absence of a poly(A) tail or in the absence of deadenylation, indicating that they have an additional function in blocking translation [75,105,107–109,125–127]. Several detailed models exist to explain the translational repression by CPEs, some of which are presented in Fig. 3.
In full grown oocytes, a CPEB binding factor called maskin was found to associate with the cap binding initiation factor eIF4E in a manner that should preclude recruitment of eIF4G and thus inhibit translation [128]. Maskin is a homologue of the transforming acidic coiled-coil domain protein 3 (TACC3) and these proteins play a vital role in the formation of the mitotic spindle in multiple organisms [129–134]. The binding between maskin and eIF4E is weak but detectable and this association is abrogated by binding of PABP to the poly(A) tail elongation and by phosphorylation of maskin in both oocytes and embryos [129,135–137]. As the eIF4E binding site in maskin appears not to be present in other organisms, it is not clear how widespread this regulation is [138]. A more conserved CPEB associated eIF4E binding protein is the neuronal protein neuroguidin, which has been proposed to function in a similar manner to maskin in repressing translation in the nervous system [139].
The deadenylase PARN is a well characterised cap binding protein and a 5′ cap structure is required for deadenylation by PARN [140–142]. CPEB can recruit PARN to the mRNA, but deadenylation still requires a cap structure [74]. This leads to a conflict between the maskin model and the opposing polymerase-deadenylase model [74,128]. Either the maskin model is true and eIF4E/maskin is stably bound to the cap and PARN is not continuously active, or the opposing polymerase-deadenylase model is correct and CPEB bound PARN is stably bound to the cap and continuously active, excluding eIF4E from the cap. In fact, PARN could very well mediate both the deadenylation and the translational repression of mRNAs to which it is recruited by excluding eIF4E from the mRNA [120].
Both maskin and PARN are absent in early oocytes, so other repression mechanisms must exist to mediate translational repression in these cells. Recently, a large CPEB containing RNP complex was found in early oocytes [143]. This complex contains several proteins found in P-bodies and related RNP granules that have been implicated in both translational control and mRNA degradation [144]. In mammalian cells, CPEB can also be found in large RNP granules including P-bodies [145]. The early oocyte CPEB complex contains the eIF4E variant eIF4E1b, the eIF4E binding protein 4E-T, the RNA helicase Xp54 (Ddx6, Rck, Me31B, Dhh1) and the P-body components p100 (Pat1) and Rap55 (Scd6, CAR-1). It does not contain PARN, maskin or the canonical cap binding initiation factor eIF4E1a. Rap55 and Xp54 can mediate translational repression in Xenopus oocytes when tethered to an mRNA [146–148]. In addition, the yeast homologues of both proteins, as well as of p100, are involved in general repression of translation [149]. Strikingly, the yeast Pumilio homologue Mpt5 recruits Dhh1, a homologue of the Xp54 RNA helicase, indicating that Pumilio-mediated translational repression may involve a similar complex [113]. The mechanism of this highly conserved type of translational repression is not entirely clear, but it is likely to involve assembly of large RNP particles that exclude translation initiation factors and/or ribosomes. In addition, eIF4E1b was shown to be defective in eIF4G binding and to bind 4E-T independently of the canonical eIF4E binding site. This implies that the mRNAs in this complex are translationally repressed through sequestration of the cap by eIF4E1b, which would exclude eIF4E1a/eIF4G from the mRNA. Indeed, tethering of 4E-T causes translational repression and injection of an eIF41b antibody enhances oocyte maturation [143]. It is as yet unclear which proteins in the early complex bind directly to CPEB and therefore are likely to confer specific repression of CPE containing mRNAs. As eIF4E1b is also the predominant CPEB associated eIF4E in fully grown oocytes, this protein could in principle convey CPE-mediated translational repression throughout oogenesis [67,143].
At present it is difficult to choose amongst the multitude of models for translational repression by CPEs. Some of the complexes detected by pulldown and immunoprecipitation may not contain the majority of the repressed mRNA, even though presence of some mRNA was demonstrated by RT-PCR. Alternatively, it may well be that every one of these models is correct at a particular stage of oogenesis or embryogenesis or that the repression is different for specific mRNAs, depending on binding sites for other proteins such as Pumilio and EDEN-BP. An attractive option is that sequestration to P-body like large complexes is the result of translational repression and acts as an enhancer of translational repression, while smaller, more mRNA specific translational repression complexes are formed during the movement of mRNAs in and out of P-bodies. To distinguish between these models, it will be important to characterise the cap binding proteins and P-body components present on specific mRNAs at different developmental stages.
7
Cytoplasmic polyadenylation and translational activation
The simplest explanation for why cytoplasmic polyadenylation leads to translational activation is that the poly(A) tail recruits translation initiation factors through its association with PABP and that these mediate release of the mRNA from the repression complexes, either by exchanging the cap binding complex for eIF4E1a/eIF4G or by somehow extracting the mRNA from a P-body-like RNP complex, a function which has been ascribed to PABP in yeast [150]. In support of this hypothesis, the translational repression of the histone B4 and cyclin B1 3′ UTRs on mRNAs injected into Xenopus oocytes can be abolished by the addition of a poly(A) tail, at least until that tail is deadenylated [29,74]. However other mRNAs (G10, Cl2) appear to be still repressed even if a long poly(A) tail is added and only become translated if the mRNA is actively polyadenylated in the maturing oocyte [6,13]. In a seminal experiment, Barkoff et al. cleaved the endogenous c-Mos mRNA with an oligonucleotide, removing the polyadenylation signal and CPE. As expected, this abrogated c-Mos mRNA polyadenylation and translation in oocytes exposed to progesterone. A ‘prosthetic’ poly(A) tail was tethered to the remaining 3′ UTR using basepairing. This did not lead to c-Mos synthesis in unstimulated oocytes, but restored the synthesis of c-Mos in response to progesterone [76]. This suggests that, at least for the translation of c-Mos mRNA, both the presence of a poly(A) tail and an additional, polyadenylation independent, modification of the mRNP are required for translational activation.
8
Discussion
Since the discovery of CPEB ten years ago, our knowledge of the mechanisms of CPE-mediated cytoplasmic polyadenylation has made great strides, especially in Xenopus oocytes. However, there are still quite a few questions that need to be resolved, for instance:1.A more systematic investigation of the consensus CPE sequence and its maximum distance to the poly(A) signal would enable a more reliable bioinformatic prediction of the targets of CPE-mediated cytoplasmic polyadenylation. Are all mRNAs containing these sequences polyadenylated during oocyte maturation?2.The relative roles of Gld-2 and PAP in cytoplasmic polyadenylation need to be further evaluated. Can they substitute for each other? What are their contact points in the cytoplasmic polyadenylation complex? A cytoplasmic polyadenylation system reconstituted from pure components would be ideal to resolve these questions.3.The kinase(s) responsible for the early activating phosphorylation on Ser174 of CPEB should be identified unequivocally. If Aurora A is responsible, why do inhibitors of its activity not block oocyte maturation and can no early activation of this kinase be detected? Do inhibitors of CamKII affect early polyadenylation? What is the signal transduction cascade leading to induction of the activating kinase?4.What are the relative roles of Musashi and CPEB in the polyadenylation of c-Mos mRNA and other substrates? How do dominant negative mutants of CPEB and Musashi achieve their repressive functions? May there be off target effects through titration of common polyadenylation factors or signal transduction machinery?5.What are the direct interaction between proteins and RNA elements in the cytoplasmic polyadenylation complex? Much of the work discussed here has been performed either in oocyte or reticulocyte lysates, where other interacting proteins may be present. It appears necessary that these interactions are mapped in detail in vitro or in a 2 hybrid system, as this could lead to predictions for the functions of dominant negative mutants and potentially clarify how different complexes are assembled.6.Are the complexes on cytoplasmic polyadenylation sequences other than the CPE similar to the CPEB associated complexes? For instance, do Gld2 or PAP associate with the RNA binding proteins that recognise them?7.Which models for translational repression and deadenylation of CPE containing mRNAs are correct in what stage of oogenesis? Again, a more intimate knowledge of the direct interactions involved in assembling the repression complexes is likely to yield important new investigative tools for resolving this question. In addition, the study of mRNP complexes assembled in vivo could yield some more conclusive answers.8.Are the polyadenylation and translation repression complexes identical for each CPE containing mRNA or are there mRNA specific differences in the complexes? Immunoprecipitation followed by RT-PCR and RNA affinity chromatography can partially answer this question, but affinity purification methods that target specific mRNPs are likely to be crucial to obtain a full answer.
Some of the controversies discussed here are a result of the experimental restrictions of the Xenopus oocyte and egg systems. These systems are very good for biochemical assays both in injected oocytes and egg extracts and they provide ample material for purification and identification. However, the high levels of stable maternal proteins make clean knock down or knock out experiments impossible in most cases, which has led to a heavy reliance on dominant negative approaches. To resolve some of the outstanding issues we are therefore likely to have to look to vertebrate genetic systems such as mouse and/or to develop tissue culture based cytoplasmic polyadenylation systems that are amenable to siRNA knock down, for instance using neuronal cells or cells synchronised in the cell cycle.
The advent of high throughput screens will strongly impact the field of cytoplasmic polyadenylation in the near future as the results of RNP immunoprecipitation microarray experiments and poly(A) tail profiling studies will increase our knowledge of the range of mRNA substrates [9,151–153]. Such screens may also identify new cytoplasmic polyadenylation elements as well as novel cell types and biological processes in which this form of regulation is involved. Only one cytoplasmic polyadenylation element and its binding protein have so far been studied in detail, predominantly in oocyte maturation. As outlined above, a large number of important questions are still unresolved, even in this one system. With at least three other cytoplasmic polyadenylation sequences in existence in oocytes alone and the evidence that cytoplasmic polyadenylation also is involved in neuronal events and the mitotic cell cycle [9], there appears ample scope for expansion of the field in the near future.
|
[
"translational control",
"cytoplasmic polyadenylation",
"oocyte",
"deadenylation",
"meiotic maturation"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Acta_Neuropathol_(Berl)-3-1-1794627
|
Hippocampal sclerosis dementia differs from hippocampal sclerosis in frontal lobe degeneration
|
Hippocampal sclerosis (HS) is characterized by selective neuronal loss and gliosis in CA1 and the subiculum and has been associated with several disorders, including Alzheimer’s disease, frontotemporal lobar degeneration with ubiquitin immunoreactive inclusions (FTLD-U), vascular dementia and some tauopathies. In some cases, HS is not associated with other degenerative pathologies. Such cases are sometimes referred to as HS dementia (HSD). Differences between HSD and HS in the setting of FTLD-U have not been systematically investigated. To this end, eight cases of HSD and ten cases of HS associated with FTLD-U were studied with Nissl and periodic acid-Schiff stains to assess neuronal loss and corpora amylacea, respectively. Sections were immunostained with antibodies to glial fibrillary acidic protein, HLA-DR and synaptophysin and immunoreactivity was measured with image analysis in CA1 and the subiculum of each case. Additionally, sections were immunostained with antibodies to 4-R tau to determine the presence of argyrophilic grains. HSD was different from HS associated with FTLD-U. Specifically, it was more common in the elderly, and it was associated with more marked neuronal and synaptic loss and with greater reactive gliosis. Corpora amylacea tended to be more frequent in HSD than in FTLD-U, but there was no difference in frequency of argyrophilic grains.
Introduction
Hippocampal sclerosis (HS) is defined as selective neuronal loss and gliosis of CA1 and the subiculum of the hippocampus. Its prevalence in demented patients ranges from 2.8 to 13% [4–6, 15, 25, 26]. It is associated with several other disorders, including Alzheimer’s disease (AD), frontotemporal lobar degeneration with ubiquitin immunoreactive inclusions (FTLD-U), vascular dementia, dementia with Lewy bodies and some tauopathies. Occasionally, HS has been reported as an independent pathologic explanation for dementia, the so-called HS dementia (HSD). The reported prevalence of HSD ranges from 0.4 to 2% [2, 4, 15, 27].
The etiology of HS is uncertain in most cases, but it is hypothesized to be due to either hypoxic-ischemic injury or to neurodegeneration of selectively vulnerable neuronal populations in CA1 and the subiculum [4, 36]. An hypoxic-ischemic origin is supported by the known deficient hippocampal vascular supply [8] and the fact that CA1 hippocampal neurons have a higher susceptibility to hypoxia due to their concentration of glutamate receptors, which may be involved in excitotoxic cell death [30]. Further arguments in favor of this theory are the association of hypoxic-ischemic episodes in patients with HS [12, 34, 37] and the increased prevalence of cardiac [3, 10, 29] and cerebrovascular disease [11, 12, 26, 35] in cases of HS of presumed hypoxic-ischemic origin.
Since many cases of HS do not have histories of hypoxia or ischemia [2, 10, 12, 15, 27], cardiac disease [2, 11, 12, 27] or cerebrovascular accidents [10, 27], a degenerative cause is also proposed. This hypothesis is supported by the frequent association of HS with degenerative dementias like AD [3, 4, 12, 15, 16, 27, 29] and FTLD-U [3, 4, 13, 15, 19, 22, 24, 29] and the less frequently with tauopathies, mainly argyrophilic grain disease (AGD) and FTDP-17 [5, 12, 17].
The aim of this study was to compare the pathological characteristics of HSD versus HS in the setting of FTD-U.
Materials and methods
The Mayo Clinic Jacksonville brain bank database between 1998 and 2005 of individuals with dementia or degenerative neuropathology (N = 1,487) was screened for specimens with HS that had complete neuropathologic evaluations. This evaluation included description of gross and microscopic findings, as well as quantitative information about Alzheimer type pathology. Specifically, the reports include counts of senile plaques and neurofibrillary tangles in six cortical sections, four sectors of the hippocampus, two regions of the amygdala, as well as subcortical regions with thioflavin-S fluorescent microscopy. It also involves immunohistochemistry for tau, α-synuclein and ubiquitin, as appropriate, as previously described [4]. A Braak neurofibrillary tangle stage was assigned to all cases based upon the distribution of neurofibrillary tangles with thioflavin-S fluorescent microscopy, as previously described [4, 14, 21, 31, 33].
HS was detected in 103 cases; in 95 cases (92%) it was associated with a primary degenerative disease process, mostly AD (36; 18 men and 18 women; 85 ± 5 years of age) and FTLD-U (44; 23 men and 21 women; 75 ± 12 years of age; P < 0.01 compared to AD). HS was detected in about 5% of AD cases (36 of 637) and 75% of FTLD-U cases (44 of 59). The remainder of cases included HS in the setting of a wide range of pathologies, including Lewy body disease, progressive supranuclear palsy, corticobasal degeneration and multiple system atrophy. In eight cases, HS was not associated with degenerative pathology.
For this study, 18 cases were selected for further study and placed in one of two groups. The first group of eight cases was operationally termed HSD; most HSD cases (75%) were associated with evidence of cerebrovascular disease. When compared to a control group of 30 cases from the same brain bank matched for age, sex and Braak stage, cerebrovascular disease was detected in 80% of the controls, a frequency similar to that in HSD. Therefore, we could not necessarily assume that the HS was due to hypoxic-ischemic causes. In all HSD cases, FTLD-U was ruled out with ubiquitin immunohistochemistry. The second group (“FTLD-U HS”) included 10 cases of HS associated to FTLD-U. None of the FTLD-U HS cases had pathologic evidence of cerebrovascular disease or other pathologic processes, and HS was assumed to be related to FTLD.
In all 18 cases other pathologic processes, such as progressive supranuclear palsy, corticobasal degeneration or Lewy body disease were absent. Alzheimer type pathology was minimal. None of the cases in either the HSD or FTLD-U HS group had a Braak neurofibrillary tangle stage greater than IV. In none of the cases could neuronal loss in the hippocampus be attributed to neurofibrillary tangles, since none had many extracellular (“ghost”) neurofibrillary tangles. Medial temporal tauopathy due to AGD was not a basis for exclusion, and all cases were screened for this age-related pathology with immunohistochemistry using isotype-specific tau antibodies, as previously reported [17].
Coronal sections of the posterior hippocampus at the level of the lateral geniculate nucleus were cut at a thickness of 5 μm. Sections were stained with Nissl and periodic acid-Schiff (PAS) stains. Sections were immunostained on a DAKO Autostainer with antibodies to HLA-DR, glial fibrillary acidic protein (GFAP), 4R tau and synaptophysin using procedures similar to those reported previously [23]. The following primary antibodies were used: anti-GFAP (GA-5, monoclonal, 1:1,000, Biogenex, San Ramon, CA), anti-HLA-DR (LN3, monoclonal, 1:100, eBioscience, San Diego, CA), anti-4R tau (ET3, monoclonal, 1:25, Dr. Peter Davies, Albert Einstein College of Medicine, Bronx, NY) and anti-synaptophysin (EP10, monoclonal, 1: 10, Dr. Peter Davies, Albert Einstein College of Medicine, Bronx, NY). For antigen retrieval, slides were steamed in distilled water for 30 min. For ET3 immunostaining, which has been shown to be a sensitive and specific method for detecting AGD [17, 32], slides were also pretreated in 99% formic acid for 30 min.
The degree of neuronal loss in the subiculum, CA1, CA2/3, CA4 (end plate) and dentate gyrus was evaluated on Nissl stained sections using a three-point rating scale: 0, absent; 1, mild-to-moderate; and 2, marked. PAS staining was used to assess the density of corpora amylacea. Corpora amylacea were defined as round, PAS-positive structures in the neuropil, perivascular, subpial and subependymal areas [7]. Presence of corpora amylacea throughout the hippocampus was graded on a four-point scale: 0, absent; 1, sparse; 2, moderate; and 3, many.
To assess astrocytic, microglial and synaptic immunoreactivity, sections immunostained for GFAP, HLA-DR and synaptophysin were used. Standardized, non-overlapping, digital images of the pyramidal layer of CA1 and subiculum were taken under a 40× objective on a light microscope. Immunoreactivity was quantified as a percentage area using MetaMorph software, version 6.3r0 (Molecular Devices, Sunnyvale, CA).
Four-R (4R) tau immunostained slides were used to determine the presence of AGD. Cases were considered to have AGD if they had small dot- or comma-like argyrophilic lesions in neuronal processes of the pyramidal layer and the entorhinal cortex, as well as coiled bodies in temporal white matter, which are the same criteria used in other studies [31].
Clinical and neuropathologic records were reviewed to collect data on age, sex, Braak neurofibrillary tangle stage, brain weight and macroscopic and microscopic neuropathology findings at autopsy. The recorded gross findings at autopsy were lobar, hippocampal and mammillary body atrophy. The remarkable microscopic findings were vascular pathology, ischemic brain injury and white matter disease.
Differences between the two groups in age, Braak stage, brain weight, corpora amylacea, neuronal loss and glial and synaptic immunoreactivity pathology were analyzed with unpaired t test and Mann–Whitney U tests depending upon the variable. Fisher’s exact test was used to compare sex, presence of AGD and remarkable pathologic findings. The statistical analyses were performed using Sigma Stat for Windows, version 3.11 (Systat Software, Richmond, CA) and the significance levels were set at P < 0.05.
Results
A summary of the demographic and neuropathologic findings of the 18 cases included in this study is presented in Table 1. A comparison of the two groups is presented in Table 2. The cases with HSD had an older age of death than those with FTLD-U HS (84 vs. 65, P = 0.001). In the HSD group, six of eight cases were men, while four of six cases in the FTLD-U HS group were men, but this difference was not significant.
Table 1Demographic and neuropathologic features of HS casesAgeSexNFT stageNeuronal lossVascular pathologyCerebrovascular pathologyWMSubCA1CA2CA4DGAtherosclerosisArteriosclerosisAmyloid angiopathyLacunar infarctsMicroscopic infarctsIschemic gliosisP187 MI2200050–75% (PCA)PV WM, BG, Th––F, M, WSAThPVP284 FII22000MinimalBG–BG, ThO–PV, DCP386 M02200050–75% (ICA, MCA)WM, BGF, P, OBG, pons–WM, BG, ThPV, DCP476 MIV02202Minimal–F, O––––P579 MIII1100050–75% (MCA, PCA, ACA)BG, CBO–BGMPVP679 FIII02111––––Cortical WSACortical WSA, ThPVP787 MII22002Minimal–O–––PVP890 MIII21000<50%WM, BG–PV WM, BG, amygdala–ThPV, DCF183 FI22NA00––––––F, T subcorticalF256 M022000––––––F, P subcorticalF356 F021000––––––F, T, P subcorticalF441 FI22002––––––F, T subcorticalF570 MI22000Minimal–––––F, subcorticalF670 M022000––––––F, T subcorticalF762 F010000Minimal–––––F, T subcortical and PVF872 FIII10002––––––F, T subcortical and PVF966 F022002Minimal–––––F, P subcorticalF1073 M022002NA–––––F, T, P subcorticalSub subiculum, CA cornu ammonis, EP endplate, DG dentate gyrus, F frontal, T temporal, P parietal, O occipital, M motor, ICA internal carotid artery, ACA anterior cerebral artery, MCA middle cerebral artery, PCA posterior cerebral artery, VA vertebral artery, BA basilar artery, PV periventricular, DC deep cerebral, WM white matter, BG basal ganglia, Th thalamus, CB cerebellum, WSA watershed area, NA not available, P # represent HSD cases; F# represent FTLD-U HS casesTable 2Comparison of HSD and FTLD-U HSHSD (n = 8)FTLD-U HS (n = 10)Age at death (years), mean ± Std. Dev.84 ± 565 ± 12‡Sex (% of cases)Male7540Female2560Braak NFT stage, median (25%-tile, 75%-tile)2.25 (1.5, 3)0 (0, 1)†Brain weight (g), mean ± Std. Dev.1130 ± 100868 ± 130‡Gross atrophy (% of cases)Frontal lobe25100†Medial temporal lobe1378†Lateral temporal lobe3367Parietal lobe1767Occipital lobe1311Mammillary bodies100100Hippocampal formation8990Neuronal loss (score), median (25%-tile, 75%-tile)Subiculum1 (0.25, 1)1 (1, 1)CA11 (0.75, 1)1 (0.5, 1)CA2/30 (0, 0.25)0 (0, 0)EP0 (0, 0)0 (0, 0)DG0 (0, 1)0 (0, 1)Synaptophysin (% area), mean ± SEMSubiculum 23.4 ± 12.336.4 ± 21.3CA126.5 ± 24.451.7 ± 20.8†GFAP (% area), mean ± SEMSubiculum 9.5 ± 7.56.2 ± 2.9 CA114.9 ± 12.59.2 ± 7.8HLA-DR (%area), mean ± SEMSubiculum0.80 ± 0.770.56 ± 0.50CA10.59 ± 0.750.19 ± 0.18Corpora amylacea (score), median (25%-tile, 75%-tile)1 (0.5, 1.75)0.5 (0, 2)Argyrophilic grain disease (% of cases)1320Fisher exact test was used for categorical variables, t test for continuous variables and Mann–Whitney Rank Sum test for discontinuous variables†P < 0.05‡P < 0.001
Neuropathologic findings
The Braak neurofibrillary tangle stage was significantly greater in HSD than in FTLD-U HS. The brain weight was also greater in HSD than in FTLD-U HS. Cortical atrophy was more frequent in FTLD-U HS than HSD in frontal (100% vs. 25%), medial temporal (78% vs. 13%), lateral temporal (67% vs. 33%) and parietal lobes (67% vs. 17%). The difference reached statistical significance in frontal and medial temporal regions. Both groups showed marked gross atrophy of the hippocampal formation and mammillary body.
In this comparative study, cerebrovascular pathology was virtually limited to the HSD group. Atherosclerosis of large vessels at the base of the brain was present in seven of eight HSD cases. Three had minimal, 1 had mild (occluding less than 50% of the lumen) and 3 had moderate atherosclerosis (occluding between 50% and 75% of the lumen). Arteriosclerosis defined by presence of hyalinosis of the media and adventitial fibrosis of small (<100 μm diameter) vessels was present in 5 (63%) cases and cerebral amyloid angiopathy (CAA) was present in 4 (50%) of 8 HSD. Three (30%) cases of FTLD-U HS had minimal atherosclerosis. None of the cases of this group had significant arteriolosclerosis or evidence of CAA.
Ischemic brain injury was present in six of eight HSD cases, but in none of the FTLD-U HS cases. The eight HSD cases included 3 (38%) with lacunar infarcts, 4 (50%) with microscopic infarcts and 5 (63%) with microscopic foci of ischemic gliosis.
Most cases, independent of their group, showed white matter pathology defined by rarefaction and attenuation of white matter, associated with dilated perivascular spaces, loss of myelinated fibers and gliosis. These findings were present in the periventricular white matter in seven of eight HSD cases. In three of these cases more diffuse white matter pathology was also present, involving the centrum semiovale. All of the cases with FTLD-U HS showed white matter pathology that followed the distribution of the cortical atrophy.
Two of the HSD cases (P4 and P7) had only minimal atherosclerosis and CAA, which was probably insufficient to cause hippocampal hypoperfusion, but both cases had a history of coronary artery disease and congestive heart failure, which might account for hypoxic-ischemic injury to the hippocampus, although this remains purely speculative, since these risk factors are common in individuals of this age without HS [12].
Another case of HSD (P6) with minimal atherosclerosis, arteriosclerosis and CAA had a history of epilepsy, as well as laminar ischemic gliosis in the watershed cortices and cortical cystic infarcts and thalamic ischemic gliosis. Epilepsy in this patient was probably secondary to a porencephalic cyst that was present in the temporal pole. Given the absence of degenerative pathology, this case was included in the HSD group, but presence of extensive cerebrovascular disease raises the possibility that HS in this case was related to hypoxic-ischemic injury.
Quantitative and semi-quantitative analyses
Both HSD and FTLD-U HS had severe neuronal loss in the pyramidal layer of CA1 and the subiculum. No significant differences were found in the severity of neuronal loss between the two regions; nevertheless, differences in the distribution of neuronal loss were noted. Most cases had neuronal loss in both CA1 and the subiculum, except for two cases of HSD that had isolated CA1 involvement and two cases of FTLD-U HS that had isolated subicular involvement.
The CA2/3 region and the endplate were not affected in most cases. Only cases of HSD occasionally had neuronal loss in these regions, with involvement of CA2/3 in P4 and involvement of CA2/3 and endplate in P6. The dentate was involved in half of the cases of both groups.
Synaptic loss assessed by image analysis of density of synaptophysin immunoreactivity was significantly less in CA1 region in HSD compared to FTLD-U HS (Fig. 1). A similar trend was noted in the subiculum. Gliosis as assessed by image analysis with GFAP and HLA-DR immunostains tended to be greater in HSD compared with FTLD-U HS (Fig. 1), although it did not reach statistical significance probably because of small sample size. Both GFAP and HLA-DR immunoreactivity were greater in HSD than in FTLD-U HS in both CA1 and the subiculum.
Fig. 1Comparison of CA1 region of hippocampus in HS associated with FTLD-U (a, c, e) and HSD (b, d, f) immunostained for synaptophysin (a, b), GFAP (c, d) and HLA-DR (e, f). Note better preservation of neuropil and less gliosis in FTLD-U HS compared to HSD (All figures are ×200)
There tended to be more corpora amylacea in HSD than in FTLD-U HS, but this did not reach statistical significance. A concomitant diagnosis of AGD was made in two cases of FTLD HS and in one case of pure HS (P6).
Discussion
The present study confirmed the characteristic findings of neuronal loss and gliosis in CA1 and the subiculum in HS regardless of etiology; however, there were some significant differences between HSD and HS that occurs in FTLD-U. In addition, there were subjective differences that did not meet statistical significance probably related to the small sample size in this study. The results support the hypothesis that there are differences in underlying disease mechanisms in HSD compared to HS in FTLD-U.
From a demographic perspective, demented patients with HS have been shown to be older than demented patients without HS [11, 12, 27, 29], although there are exceptions [2, 10]. Data concerning the age distribution among patients with HS has to our knowledge not been reported. In the present study, we show HSD to mainly affect the elderly (>80 years of age) as opposed to HS in FTLD-U. The younger age of the FTLD-U may account for the lower frequency of concurrent cerebrovascular disease, since an age-, sex- and Braak neurofibrillary tangle stage-matched cohort had comparable frequency of cerebrovascular disease.
Microscopically, there were three important differences in the pathology in HSD compared with HS in FTLD-U. First, HSD had lower synaptophysin immunoreactivity, suggesting a more destructive process resulting in more severe synaptic loss, while in FTLD-U there is selective neuronal loss, but relative preservation of the neuropil, presumably from preserved afferent and efferent projections to sectors that show neuronal loss. Second, the difference in the distribution of neuronal loss suggested subicular involvement to be characteristic of HS in FTLD-U. Even though the differences in distribution of neuronal loss did not reach statistical significance, other observations support this idea [20]. Third, HSD was associated with greater astrocytic and microglial reaction compared to HS in FTLD-U. Glial activation has been proposed to be caused by neuronal damage, which is present in both, ischemic and degenerative pathology, but may be exacerbated in HSD due to more extensive tissue damage affecting not only neurons, but also the neuropil.
Macroscopically, FTLD-U HS had more severe cortical atrophy, a not surprising finding, since it correlates with FTLD [25]. HSD on the other hand showed a pattern of atrophy often localized to the hippocampal formation, arguing against this pathology being associated with more widespread cortical degeneration. This fits with results from previous studies of HSD that showed that loss of synaptophysin immunoreactivity in the hippocampus was not accompanied with similar loss in the cortex or basal ganglia [12].
Corpora amylacea have been proposed as a possible marker for hypoxic-ischemic HS, since they may be numerous in brains of patients exposed to repetitive hypoxic episodes [1, 6]. They have also been proposed as a useful marker for evaluating HS in surgical specimens removed for treatment of temporal lobe epilepsy [9]. Although there was a trend for more corpora amylacea in HSD than in FTLD-U HS, it did not reach statistical significance, which may be further evidence against the argument that HSD is due to hypoxic-ischemic injury.
Even though we did not find a difference in the frequency of AGD in HSD and HS associated with FTLD-U, the presence of AGD in 3 of the 18 cases, may suggest an association of HS with AGD as reported in previous studies [5, 17].
This study included two cases of HSD (P4 and P7) in which there was not only no evidence of degenerative pathology, but also no evidence of cerebrovascular pathology. Cases in which HS is the only pathologic explanation for dementia are rare. Hattanpa found some cases of HS to be false negative cases for FTLD-U after further studying them with ubiquitin immunohistochemistry [13]. Since we analyzed our cases with ubiquitin, false negatives were excluded, giving a probable explanation for the lower prevalence compared to the 0.48, 0.53 and 2% previously reported by others [2, 15, 27].
The two cases of HSD with neither degenerative nor cerebrovascular pathology had cardiovascular risk factors that could link HS to an ischemic condition of systemic origin. They are similar to the ones that have been reported by others [34, 37]. On the other hand, in a systematic survey of cardiovascular risk factors, including electrocardiograph abnormalities, congestive heart failure and cardiomegaly by imaging and autopsy studies, there was no increased frequency of these findings in HSD compared to age-matched controls without HS [12].
The present study confirms the finding that HS in the setting of dementia in most cases should be regarded as a feature linked to a concurrent degenerative disease process, particularly FTLD-U [13, 18, 19, 28]. In other cases, HS may be linked to a systemic condition (e.g., cardiorespiratory failure) or cerebrovascular disease [8, 36]. In rare cases, the etiology is unknown. The present study suggests that HSD is a disorder of the elderly that is associated with more marked neuronal and synaptic loss and greater reactive gliosis than that seen in HS associated with FTLD-U. This argues for a different pathogenesis
|
[
"hippocampal sclerosis dementia",
"image analysis",
"immunohistochemistry",
"frontotemporal degeneration"
] |
[
"P",
"P",
"P",
"R"
] |
Surg_Endosc-4-1-2292804
|
Retracting and seeking movements during laparoscopic goal-oriented movements. Is the shortest path length optimal?
|
Aims Minimally invasive surgery (MIS) requires a high degree of eye–hand coordination from the surgeon. To facilitate the learning process, objective assessment systems based on analysis of the instruments’ motion are being developed. To investigate the influence of performance on motion characteristics, we examined goal-oriented movements in a box trainer. In general, goal-oriented movements consist of a retracting and a seeking phase, and are, however, not performed via the shortest path length. Therefore, we hypothesized that the shortest path is not an optimal concept in MIS.
Currently, there are various virtual-reality (VR) trainers available to train MIS skills [1, 2]. These trainers have the advantage over box trainers of providing objective feedback about the performance of the user (performance parameters), which motivates residents to learn [3–5]. The most common parameter used to measure basic MIS skills objectively is task completion time. However, this parameter alone does not adequately evaluate the task performance [6]. Therefore, the time for completion is often used in conjunction with other parameters [1, 7–12]. One of the task-independent parameters that is used in each VR trainer is path length – the length of the curve described by the tip of the instrument over time [13]. Intuitively, a longer path length indicates a less-efficient movement and is often interpreted as lack of experience.
In MIS, goal-oriented movements (point-to-point movements) are very common (e.g., during grasping, placing a clip on the vessel, or while using diathermy). In general, a simple strategy to perform such a movement is to move along a straight line between an initial position and a specified target position [14, 15]. However, it is hardly possible to make such a movement during MIS. Movements of MIS instruments are done through the incision point, which acts as a pivot point. Therefore, point-to-point aiming movements in MIS may simply result in a curved workspace path (Fig. 1). Both these ways of performing point-to-point movements are very efficient. In general, however, goal-oriented movements are not performed via the shortest path; a pull-back movement is performed before the specified target position is approached [16, 17]. Therefore, we hypothesized that the shortest path, as presently used to indicate more-efficient performance in MIS, is not a suitable reference.
Fig. 1Three strategies of performing a goal-oriented movement in MIS. (a) a movement along the straight line between the initial and the end position; (b) a movement along a curved path between the initial and the end position, which is the result of the movement of the MIS instrument through the incision point; and (c) a realistic movement performed by the surgeon during the goal-oriented task. A – the initial position; B – the end position; P – incision (pivoting) point
The objective of this study was to examine goal-oriented movements during training of the basic MIS skills. The study consisted of two parts: first, we investigate whether goal-oriented movements can be split into two phases: a retracting phase and a seeking phase; and, subsequently, we investigate whether movements of the MIS instrument in these two phases depend on the surgeon’s experience. The performance of the participants was analyzed and compared using normalized path lengths.
Methods
Participants
Experienced gynaecologists (with the experience of more than 100 MIS procedures), their residents (experience of 10–100 MIS procedures), and novices (medical students with no previous experience in MIS) were invited to participate in this study. Each participant, voluntarily enrolled in this study, was asked to complete a short questionnaire detailing demographic information and prior experience in laparoscopy.
Task
The hypothesis was tested using a simple one-hand positioning task, which requires touching the top of eight cylinders with the tip of the MIS instrument (Fig. 2). All cylinders were situated in a box trainer in various three-dimensional (3D) positions. Every correctly touched cylinder resulted in the lighting of the lamp corresponding to this cylinder. The start/end point and the order of touching cylinders (indicated by the numbers located next to the cylinders) were the same for each participant.
Fig. 2The positioning task. The task requires touching the top of the eight cylinders (of varying 3D position) with the tip of the MIS instrument. A correctly touched cylinder results in lighting up a lamp (above on the picture) corresponding to this cylinder. The start/end point and the order of touching the cylinders are indicated by the numbers located next to the cylinders
To mimic the in vivo gynaecological MIS situation, all participants performed the task with their left hand, while the camera was held in the right hand. To provide the same conditions for each participant, the position of the task and the incision points for the camera and the MIS instrument were standardized in the box trainer. The image of a 0° laparoscope was presented on a monitor.
Before the test, all participants were instructed how to perform the positioning task. Furthermore, they were allowed to make one trial before testing.
Data analysis
Movements of the MIS instrument were recorded with our newly developed TrEndo tracking system with a sample frequency of 100 Hz [18]. The data of the eight simple goal-oriented movements performed by each participant during the task were analyzed. The first goal-oriented movement was defined by the movement between the start point and first cylinder. Each successive goal-oriented movement was defined by the movement between two consecutive cylinders (Fig. 2).
The retracting phase describes the pull-back movement of the MIS instrument in the direction of the pivoting point. Therefore, we analyzed the projected movements of the instrument in the ABP plane (Fig. 3), which passes through the begin point of the movement (point A), the end point of the movement (point B), and the pivoting point (point P). To examine the deviation of the movement from the plane ABP, we also analyzed the movements projected in the plane AB, which is perpendicular to the plane ABP, and which passes through the points A and B (Fig. 3). Point M, defined as the point of the movement most distant from the line AB, was used to make a distinction between the retracting phase AM (from point A to point M) and the seeking phase MB (from point M to point B). The analysis of each of the two phases was done using a normalized path length, which was derived as follows: where PLreal is the real path length (between A and M for retracting, and M and B for seeking phase), and PL0 is the distance (length of the straight line) between A and M and M and B (for retracting, and seeking phases, respectively). The normalized path lengths of all eight successive goal-oriented movements were averaged per participant and analyzed.
Fig. 3The ABP (described by the points A, B, and P) and AB (that passes through points A and B, and is perpendicular to the plane ABP) planes presented in an orthographic projection: (a) the ABP and AB planes, and the goal-oriented movement presented in the three-dimensional space; (b) front view: projection of the goal-oriented movement in the ABP plane; (c) top view: projection of the goal-oriented movement in the AB plane; and (d) side view: the ABP and AB planes, and the goal-oriented movement presented in a side view. A – the initial position; B – the end position; M – a point, which is used to make a distinction between the retracting and the seeking phases; P – incision (pivoting) point
Statistics
Data were analyzed using the Statistics Toolbox of MATLAB 7. Statistical analysis was performed using one-way analysis of variance (ANOVA), and Wilcoxon tests. A probability p < 0.05 was considered to be statistically significant.
Results
Participants
Thirty-eight participants completed the task and the questionnaire. The group of experts consisted of 9 experts (age 38–59 years). The group of residents consisted of 17 gynaecological residents (age 29–41 years). The group of novices consisted of 12 medical students (age 23–28 years). All participants were right-handed.
Retracting and seeking phases
Typical trajectories for an expert and a novice performing the goal-oriented movement are presented in Fig. 4. The figure shows that the retracting and seeking phases can easily be distinguished from each other. No significant difference between the experts, residents, and novices was observed in the distance between point M and the line AB.
Fig. 4The typical trajectories for an expert (left) and a novice (right) performing a goal-oriented movement. Top: projection of the goal-oriented movement in the plane ABP. Bottom: projection of the goal-oriented movement in the plane AB. A – the initial position; B – the end position; M – a point, which is used to make a distinction between the retracting and the seeking phases
Normalized path lengths of the two phases are presented in Fig. 5. In 3D space, the experts’ normalized path length was shorter during the retracting phase [median (range) %: 152 (129–178)] than during the seeking phase [180 (172–247)]. Similarly, the residents’ and novices’ path lengths were shorter during the retracting phase [164 (126–250), and 168 (136–268)] than during the seeking phase [201 (163–287), and 290 (244–469), respectively].
Fig. 5Normalized path lengths of the retracting and the seeking phases for experts, residents, and novices. Upper: normalized path length in 3D space. Middle: normalized path length in the ABP plane. Lower: normalized path length in the AB plane. The results are presented as notched box-and-whisker plots, where every box has a line at every quartile, median, and upper quartile values. The whiskers are presented as lines that extend from each end of the box in order to show the extent of the rest of the data. The notches represent the 95% confidence interval for the median. Boxes whose notches do not overlap are significantly different (p < 0.05). **p < 0.01; ***p < 0.001; E – experts; R – residents; N – novices; 1 – retracting phase; 2 – seeking phase
In the ABP plane, the experts’ normalized path length in the retracting phase [132 (120–194)] was significantly shorter than in the seeking phase [199 (154–234)]. The normalized path lengths of residents and novices were also significantly shorter in the retracting phase [133 (108–492), and 180 (118–287)] than in the seeking phase [208 (162–286), and 310 (252–469)].
In the AB plane, the experts’ normalized path length was significantly shorter in the retracting phase [122 (100–200)] than in the seeking phase [183 (141–209)]. Normalized path lengths of residents and novices were also significantly shorter during the retracting phase [147 (107–264), and 207 (128–271)] than during the seeking phase [199 (162–285), and 310 (252–469)].
Influence of the experience
The experts’, residents’, and novices’ normalized path lengths during the retracting and the seeking phases are presented in Fig. 6. The data in Fig. 6 are the same as the data in Fig. 5, but presented per phase for all groups. In 3D space, there was no significant difference between experts, residents, and novices for the retracting phase. The normalized path lengths of experts and residents were significantly shorter than the path length of the novices during the seeking phase. No significant difference between experts and residents was observed in that phase.
Fig. 6The influence of the experience on the performance of the retracting and the seeking phases. The data presented in this figure are the same as the data in Fig. 7, but presented per phase. *p < 0.05; ***p < 0.001; E – experts; R – residents; N – novices; 1 – retracting phase; 2 – seeking phase
In the ABP plane, we found no significant difference between experts, residents, and novices for the retracting phase. In the seeking phase, the normalized path lengths of experts and residents were significantly shorter than the path length of novices. No significant difference between experts and residents was observed in that phase.
In the AB plane, a significant difference between experts, residents, and novices was found for both phases. In both the retracting and the seeking phases, the normalized path lengths of the experts and residents were significantly shorter than the path length of the novices. No significant difference between experts and residents was observed for both phases.
Discussion
Our findings indicate that goal-oriented movements are not performed via the shortest path. The movements clearly distinguish a retracting and a seeking phase. The results show that the normalized path length during the retracting phase is significantly shorter than during the seeking phase. Furthermore, experience in MIS does only influence the seeking phase. Experts’ and residents’ normalized path lengths were significantly shorter than the path lengths of the novices. The shorter path length in the seeking phase implies better performance; therefore, the seeking phase is characteristic of the differences in performance. This finding confirms that the path length is an important measure in the assessment of basic MIS skills, in which the seeking phase is the only component that can distinguish between novice and expert.
The retracting phase is essential in MIS, because it improves safety of the patient by avoiding unpredicted contact with the tissue. It seems that the best strategy to perform this phase safely is to pull back the instrument along its axis (in the direction of the pivoting point) and avoid any movements in the AB plane, which can cause unpredicted contacts with the tissue. The results of this study confirm this assumption; experts’ and residents’ normalized path lengths are smaller in the AB plane than the path length of the novices. Movements in the AB plane can result in unintended tissue contact and be dangerous for the patient; therefore, it would be beneficial to let novices learn how to perform a more-precise retracting movement.
The experience in MIS does not influence the retracting phase in the ABP plane. Moreover, a longer retraction of the MIS instrument in the direction of the pivoting point does not denote worse or less-efficient performance. In contrast, it may be a sign of the safer (for the patient) use of the MIS instruments. This finding is important, because it shows that the shorter path length, as presently used to indicate more-efficient performance in MIS, is not a correct concept for analyzing optimal movements in laparoscopy.
To investigate whether the retracting phase really represents safety, it is necessary to design a different study. For example, one possibility would be to include obstacles which, upon touching represent an error. With such a study design, it would be feasible to determine how much retraction is necessary and whether the experts have learnt the optimal retraction amount. The present study cannot be used to decide on these aspects.
The results of this study indicate that the comparison of the novices’ path length to the experts’ path length is an important and valid component of the overall criterion-based assessment of basic MIS skills. It is necessary to realize that this study was not designed to determine whether expert performance actually results in the optimal path length and, therefore, the results of this study cannot be used to indicate that experts’ movements are optimal.
In the literature, there is a tendency to assess MIS performance using very basic terms (e.g., time, path length, penalty points), which do not show “the size and the nature of the gap between expert and novice performance” [13]. The results of this study show that the analysis of the separate phases of the movement in various planes can give such details. For example, movements of the MIS instrument along its axis (in and out the trocar) during the seeking phase characterize the surgeon’s level of depth perception. Movements of the instrument in the AB plane may indicate eye–hand coordination problems (e.g., orientation). Since every phase of the movement analyzed in different planes provides details about different limitations, it is crucial to analyze the movements in MIS separately for each phase and in both planes. Only then will the feedback about the performance and the nature of the limitations lead to goal-oriented training curricula.
This study shows for the first time the importance of proper analysis of the instruments’ movements during training of MIS skills. The clinical impact of such extended analysis is that only in this way is it possible to implement a correct objective score that will measure and certify the competence of surgeons’ basic psychomotor MIS skills in addition to the existing criteria for the assessment of MIS performance. Moreover, this extended motion analysis can result in improvement of the training of basic MIS skills, since it will identify the differences between the experts’ and novices’ performance, and which areas require more training.
Conclusions
Movements during laparoscopic tasks are not performed via the shortest path. Therefore, the shortest path length, as presently used for the assessment of basic MIS skills, may be not a proper concept for analyzing optimal movements and therefore needs to be revised. Goal-oriented movements should be split into two phases: a retracting and a seeking phase. Novices are less efficient in the seeking phase. This finding confirms that path length is an important measure in the assessment of basic MIS skills, but that the seeking phase is the only component that distinguishes between novice and expert. Furthermore, the retracting phase is very important in MIS, since it improves safety by avoiding intermediate tissue contact. Analyzing motions in the separate phases provides greater insight into the nature of the gap between expert and novice performance. Motion analysis in MIS should be seen as an addition to the existing criteria in assessing performance.
|
[
"goal-oriented movement",
"path length",
"minimally invasive surgery",
"objective assessment",
"training",
"motion analysis"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
J_Med_Internet_Res-5-4-1550578
|
Searching for Cancer Information on the Internet: Analyzing Natural Language Search Queries
|
Background Searching for health information is one of the most-common tasks performed by Internet users. Many users begin searching on popular search engines rather than on prominent health information sites. We know that many visitors to our (National Cancer Institute) Web site, cancer.gov, arrive via links in search engine result.
Introduction
For members of the general public who use the Internet, many seek medical information [1- 6]. According to a recent systematic review of 24 peer-reviewed publications describing the proportions of Internet users among various populations of cancer patients in the developed world, about 39% of cancer patients are using the Internet directly, and in addition, 15% to 20% of persons with cancer use the Internet "indirectly" through family and friends [7]. Studies have evaluated information-seeking behavior on the Internet by cancer patients generally [8- 10], their companions [11,12], and patients with the following common specific cancer diagnoses: breast [13- 16], prostate [17,18], lung [19], and gastrointestinal cancers [20]. Studies have also evaluated information gathering by cancer patients undergoing radiotherapy [21] and chemotherapy [22], and those from centers outside of North America [23,24]. Individuals from certain disadvantaged groups have been shown to seek medical information online less frequently and with more difficulty [7,25,26].
Eysenbach and Kohler found that general consumers search for medical content using search engines rather than medical portals or sites of medical societies or libraries [27]. Newly-diagnosed cancer patients and their families often start their searches as users less sophisticated in Web and medical terminology. They too commonly begin searching on popular search engines rather than on prominent cancer-information sites. We know that many visitors to our own Web site [28] arrive via search engine result links.
To better understand users' needs this research aimed to establish what lay users really want to know when they search online for cancer information. To do this we evaluated data from Ask.com [29], a popular natural-language-processing (NLP) search engine. Natural-language-processing search engines allow users to create queries using whole phrases and sentences of any length, rather than just key words.
Earlier reports of this project have been published in abstract form only. The abstracts reported a brief project summary [30], and data specific for breast cancer [31] and gastrointestinal cancer [20]. This is the first comprehensive report of the entire project.
Methods
The National Cancer Institute (NCI) partnered with AskJeeves, Inc to develop a methodology to capture, sample, and analyze 3 months of cancer-related queries on the Ask.com Web site, a prominent US natural-language-processing consumer search engine. At the time of the project, Ask.com was receiving over 35 million queries per month.
Search Terms
An NCI oncologist (JLB) developed a benchmark set of 500 terms and word roots that were matched against actual AskJeeves user queries. Most terms and word roots were from the NCI dictionary on the NCI Web site [32]. NCI also suggested additional terms not included in the dictionary. These terms related to anatomy, organ systems, treatments, pharmaceuticals, treatment and diagnostic procedures, genetics, epidemiology, and pathology.
Table 1
Top 37 search terms and roots with ≥ 5 queries per week during test week
Term
Actual Queries During Test Week
% of Total Queries
*cancer*
9765
56.75
*tumor*
1396
8.11
*carcino*
656
3.81
*leukemia*
635
3.69
lymphom*
419
2.43
chemotherapy
378
2.20
biopsy/biopsies
375
2.18
*melano*
348
2.02
*sarcoma*
294
1.71
*dysplasia*
255
1.48
hodgkin*
245
1.42
MRI
214
1.24
clinical trial
187
1.09
mammogr
175
1.02
maligna*
170
0.99
*metasta*
155
0.90
"*" is a placeholder for the part of the search term before or after the root.
The test sample of these 500 words and roots was used to filter cancer queries from the Ask.com Web site for 1 week in August 2001. From these 500 terms, only 37 appeared ≥ 5 times per day over the trial week. The list of 37 terms (plus common misspellings) yielded 17208 queries for the test week. The frequency of each term is shown in Table 1. Queries with common misspellings, (eg, prostate and prostrate, biopsy and biopsey, leukemia and lukemia, chemotherapy and chemothereapy) were captured and analyzed. It was felt that the cut off of 5 times per day (≥ 35 times per week) would capture the key queries and include any common query topic, since even with a frequency of 35 queries a week, the majority of these terms accounted for less than 1% of the total population. Of the original 500 terms supplied by NCI, only 7% (35/500) appeared in the logs at a high frequency, but this 7% accounts for over 37% of user queries identified as cancer related on Ask.com during the study period.
Collecting Queries and Sampling
The process used for collecting and sampling queries is outlined in Figure 1. Using the 37 terms to search the Ask.com query logs, 204165 instances of cancer-related queries were found for June, July, and August 2001. Of these queries, 7500 individual user questions were randomly selected by AskJeeves for detailed analysis (see Appendix 1).
Figure 1
Processing of cancer queries on Ask.com
Very often there were multiples of the same questions. Thus, these 7500 queries actually represented 76077 queries that were entered into Ask.com, about 37% (76077/204164) of all queries identified as cancer-related from the 3 month period of log analysis. For example, a user question might be "Where can I find information about breast cancer?" This individual example represents 1 user question, but might have been queried by more than 100 people on any given day. Each query was counted only once.
Sampling Issues
The random sample of 7500 individual queries provides a confidence interval of 1.11% at a confidence level of 95%. This means that even if more samples were taken from 204165 queries, 95% of those samples should not be off by more than 1.1%. While this means that the samples themselves would not vary more than 1.1% over 95% of the samples taken, as the data are categorized and classified, in effect smaller and smaller samples are taken. Therefore, to offset this problem additional queries were examined, even though a smaller sample would still provide a high degree of confidence in the results.
In other words, although broad generalizations—such as "breast cancer accounts for 25% of all cancer queries"—can be easily presented, a large sample size is required to break down data far enough to conclude that when users ask about breast cancer, they are most often asking about specific types of treatments.
Highest-Level Categories for Queries
User queries were assigned to a set of 6 highest-level categories (as shown in Table 2):
Cancer (ie, specifically mentioning a cancer type)
General Research
Treatment
Diagnosis and Testing
Cause/Risk/Link
Coping
Table 2
Highest-level categories for queries
Highest-Level Category
Number of Queries
Percent of All Sampled Queries
*
Cancer †
59619
78.37
General Research
7808
10.26
Treatment
3832
5.04
Diagnosis and Testing
3315
4.36
Cause/Risk/Link
1249
1.64
Coping
254
0.33
Total
76077
* Percentages do not sum to 100% due to rounding.
† ie, specifically mentioning a cancer type.
Highest-level categories were created in a collaborative effort between the AskJeeves data-analysis team and NCI staff before the study period began, but the final category titles were revised as the actual queries were analyzed. The initial categories were based on user queries entered into Ask.com and a variety of online sources, such as NCI's online dictionary [32] and NCI's Physician Data Query (PDQ) [33].
The highest-level categories were populated using proprietary AskJeeves filters and automated-analysis tools that sorted queries according to specific types of cancers, or—in the absence of mentioning a specific cancer type—whether the query asked about other areas such as Treatment or Coping. (AskJeeves did not share the filters and automated analysis tools with the authors.) Queries that could not be sorted by the filters and automated-analysis tools were placed in a temporarily-uncategorized category; they were categorized during the next step (reading and analysis).
Reading and analyzing each individual query not only verified the automated process, but also helped to refine existing categories and create new categories and subcategories, as appropriate. For example, without this type of analysis, the query "Where can I find a Web site with information on using high protein food to fight Breast cancer?" might have been left under Breast Cancer > Media and Organizations > Web sites (where ">" indicates a change in category level). This would not be correct, as the true user intent was to inquire about Alternative Treatments. As a result, under the category Breast Cancer > Treatment, "Alternative" was added to the Breast Cancer > Treatment category analysis as a subtopic. (Treatment—without a specific cancer site designated—is both a highest-level category and a subcategory under Breast Cancer and under most cancer types.)
Approximately 78% of all categorized queries from the sample referenced a particular type of Cancer, and were placed in the highest-level category Cancer. An example of this kind of query would be "Where can I find information about Breast Cancer?" (This query would be classified as Cancer > Breast Cancer > General Information.) Any query that did not mention a specific kind of Cancer, even though the question was about cancer, was placed on 1 of the 5 other highest-level categories. An example of this type of query would be "Where can I find information on cancer treatment with radiation?" This query was assigned to the Radiation subcategory in the highest-level category Treatment (ie, it was classified as Treatment > Radiation).
Queries that did not relate to a specific Cancer type were placed in 1 of the 5 other highest-level categories: General Research, Treatment, Diagnosis and Testing, Cause/Risk/Link, or Coping. For example the query "How does smoking cause cancer?" would be placed in the Cause/Risk/Link category, as it did not refer to any specific type of cancer.
"Cancer" Queries (Related to Specific Cancer Types)
As shown in Table 3, there were 14 cancer types (N = 59619 queries) selected as subcategories of the Cancer highest-level category. For cancer types with the most-frequent queries, like Digestive/Gastrointestinal/Bowel (D/G/B), Breast, and Genitourinary, there were enough queries to populate standard subcategories like General Information, Treatment, Symptoms, Diagnosis and Testing, and Cause/Risk/Link. These common cancer types often warranted the creation of customized subcategories, like Breast > Media and Organizations > Web sites. For the less-common cancer type queries, like Bile (duct) in D/G/B, few queries were received and only those in General Information are shown.
Table 3
Cancer types
Type within Top-Level Cancer Category
Number of Queries
% Queries in Cancer Category
*
†
% Queries in This Report
*
‡
Digestive/Gastrointestinal/Bowel (D/G/B)
8959
15.0
11.8
Breast
6953
11.7
9.1
Skin
6709
11.3
8.8
Genitourinary
6250
10.5
8.2
Hematologic/Blood
5448
9.2
7.2
Gynecological
5344
9.0
7.0
Lung
4630
7.8
6.1
Soft Tissue/Muscle
3954
6.6
5.2
Lymphoma
3333
5.6
4.4
Head and Neck
2522
4.2
3.3
Brain/Neurological
1852
3.1
2.4
Miscellaneous Cancer
1633
2.7
2.1
Bone
1429
2.4
1.9
Pediatric
603
1.0
0.8
* Percentages do not sum to 100% due to rounding.
† Denominator (N = 59619) was the total number of queries about specific types in the Cancer category.
‡ Denominator (N = 76077) was the total number of queries analyzed in this report.
Privacy Issues
Although NCI helped create the search terms and the categories into which the analyzed data was placed, NCI did not have access to: the raw query logs at AskJeeves, any information about what AskJeeves users did with the searches generated on the AskJeeves Web site (ie, what links they picked), or the identities of any users of the Ask.com Web site. NCI did not require permission from the Institutional Review Board.
Results
Frequency of Top-Level Categories
As shown in Table 2, The 6 highest-level categories in order of decreasing frequency of queries were:
Cancer (N = 59619, 78.37%)
General Research (N = 7808, 10.26%)
Treatment (N = 3832, 5.04%)
Diagnosis and Testing (N = 3315, 4.36%)
Cause/Risk/Link (N = 1249, 1.64%)
Coping (N = 254, 0.33%)
The data in Table 2 indicate that the great majority of users asked for information about specific types of cancers, but rarely asked about a Treatment option or Diagnosis and Testing procedure without specifying the particular cancer about which they were concerned. Similarly, users asked few queries about general Symptoms of cancer unrelated to a specific type of cancer (see Diagnosis and Testing > Symptoms, N = 473, 14.27%). An example would be "what are some symptoms of cancer?"
Subdividing Cancer Queries
Table 3 breaks down the highest-level category Cancer queries (N= 59619) into more specific cancer types. In order of decreasing frequency within the Cancer category, the 14 subcategories were:
Digestive/Gastrointestinal/Bowel (D/G/B) (N = 8959, 15.0%)
Breast (N = 6953, 11.7%)
Skin (N = 6709, 11.3%)
Genitourinary (N = 6250, 10.5%)
Hematologic/Blood (N = 5448, 9.2%)
Gynecological (N = 5344, 9.0%)
Lung (N = 4630, 7.8%)
Soft Tissue/Muscle (N = 3954, 6.6%)
Lymphoma (N = 3333, 5.6%)
Head and Neck (N = 2522, 4.2%)
Brain and Neurological (N = 1852, 3.1%)
Miscellaneous (N = 1633, 2.7%)
Bone (N = 1429, 2.4%)
Pediatric (N = 603, 1.0%)
Any query specifically mentioning a cancer type by name, was assigned to that subcategory. For example, questions about Breast-Cancer-specific Treatment, Diagnosis and Testing, Causes, and Coping are found in the Cancer > Breast Cancer category, within 1 of the 10 subcategories displaying Breast Cancer information. All questions about Leukemia or Myeloma would be found in Hematologic/Blood, Hodgkin's Disease queries in Lymphoma, and Esophageal cancer questions in D/G/B.
The number of subcategories assigned to each of the 14 different cancer types varied somewhat and was driven by the nature and number of the specific queries in those cancer types.
Detailed Analysis of Queries
The detailed categorizations and verbatim display of examples of sampled queries are shown in Appendix 1. There is a breakdown of all the 14 cancer types within the highest-level category Cancer as well as a breakdown of queries within the 5 other highest-level categories not referencing any particular cancer type. These 19 are arranged alphabetically in the Appendix.
Major observations about the 19 categories and subcategories are noted below, in the order they appear in the Appendix. Our comments emphasize issues related to requested cancer content more than technology issues related to the natural language processing.
1.0 Bone Cancer
As shown in Appendix 1, there were 1429 queries about Bone Cancer. The vast majority of Bone Cancer queries asked for General Information (N = 1107, 78%). An example of this category would be: "Where is information on bone cancer?" Users asked questions about Bone Cancers linked to various sites of Anatomy as well as certain Histologies. There were some questions related to Bone Cancers in teenagers that were assigned to this category, rather than the Pediatric category. There were more questions about Diagnosis and Testing (N = 64, 4.48%) and Symptoms (N = 135, 9.45%) than Treatment (N = 26, 1.82%).
2.0 Brain and Neurological Cancer
Of the 1852 Brain and Neurological Cancers queries, General Information accounted for the vast majority (N = 1323, 71.44%). There were 427 (23.1%) questions about specific cancer types in this category. Some cancer types queries asked about Medulloblastoma, which is typically but not always a Pediatric tumor. As with Bone Cancer above, some questions could have been meaningfully assigned to more than 1 top-level Cancer site category. In this category there were more queries about Symptoms (N = 259, 13.98%) than Treatment (N = 112, 6.05%).
3.0 Breast Cancer
As shown in Appendix 1, Breast Cancer was one of the simpler cancer types, from a data-display standpoint. There was only 1 anatomic-cancer type and all of the individual queries for that cancer type were assigned into 1 of 10 subcategories.
The 10 top-level Breast Cancer subcategories were:
General Information (N = 3423, 49.23%)
Symptoms (N = 889, 12.79%)
Treatment (N = 570, 8.20%)
Media/Organization (N = 428, 6.16%)
Cause/Risk/Link (N = 393, 5.65%)
Diagnosis and Testing (N = 376, 5.41%)
Statistics (N = 274, 3.94%)
Pictures (N = 225, 3.24%)
Type (N = 217, 3.12%)
Definition (N = 158, 2.27%)
Nine of the 10 Breast Cancer subcategories were analyzed in detail in Appendix 1. The tenth, Pictures, did not require further analysis. Most queries asked for General Information.
There were more frequent queries about Breast Cancer (N = 6953) than any other cancer type. This may not be apparent from Table 3, which appears to show more D/G/B cancers (N = 8959). However, D/G/B overall is actually composed of 10 cancer types. The most frequently queried cancer type in D/G/B was Colorectal (N = 4,801) which had fewer queries than Breast.
Even though other cancer types may have been assigned more subcategories than the 10 for Breast, the detail and the medical specificity and technical vocabulary of Breast queries appear to be the most complex than other Cancer sites, probably reflecting the sophistication of basic research and clinical data on this topic and the relative sophistication of the breast cancer information seekers.
4.0 Cause and Risk
There were 1249 queries in this highest-level category. Without mentioning a specific cancer by name, there were N = 1115 (89.27%) queries about Causes and Links but only N = 134 (10.73%) about Prevention. Among the 1115 queries in the Causes and Links subcategory, the following topics were noted:
Drugs (N = 287, 25.74%)
Unspecified (N = 247, 22.15%) (eg, "What is cause a cancer?" [sic])
Radiation (N = 247, 22.15%)
Personal (N = 116, 10.40) (eg, "Can anti-persperant [sic] deodorant cause cancer?")
Chemical/Plastics (N = 74, 6.64%)
Environmental (N = 70, 6.28%)
Food Supplement (N = 64, 5.74%)
Genetic Mutation/Virus (N = 10, 0.90%)
Smoking was not in this list, probably because most queries about smoking were included under a query about a specific type of cancer, like Lung or Head and Neck.
5.0 Coping
There were only 254 queries about Coping. The queries referenced Support Groups (N = 127, 50%), Pain (N= 98, 38.58%), and Depression (N = 29, 11.42%). Even though there were few questions in this highest-level category, the issue was of specific interest to NCI, which asked for this category to be created and analyzed separately.
6.0 Diagnosis and Testing
There were 3315 queries in this highest-level category, which did not mention a specific cancer by name. Most were queries about specific Testing (N = 2842, 85.73%). The others (N = 473, 14.27%) were queries about Symptoms. Among Testing queries, CAT/CT scan (Computerized Axial Tomography/Computed Tomography scan) (N = 1509, 53.10%) and MRI (N = 587, 20.65%) were the most-common Testing topics, followed by Biopsy (N = 502, 17.66%).
7.0 Digestive/Gastrointestinal/Bowel (D/G/B)
The presentation of data queries for D/G/B in Appendix 1 is complex because, there were 7 top-level subcategories, including General Information and 10 cancer types identified in the General Information subcategory
As shown in Appendix 1, 8959 queries for D/G/B sites were broken down into 7 subcategories:
General Information (N = 5568, 62.15%)
Symptoms (N = 1506, 16.81%)
Diagnosis and Testing (N = 1125, 12.56%)
Treatment (N = 294, 3.28%)
Statistics (N = 184, 2.05%)
Definition (N = 163, 1.82%)
Cause/Risk/Link (N = 119, 1.33%)
Most queries asked for General Information. Examples of General Information queries would be "Where can I learn about the cancer esophageal cancer?"? and "Where can I find information on Stomach cancer"?
A breakdown of all D/G/B queries by cancer type is shown in the list below. The absolute numbers and percentages (of all D/G/B queries) in the list below differ from the pie diagram in Appendix 1 because the list below includes organ-type queries from General Information plus the 6 other subcategories in D/G/B.
Colorectal (N = 4801, 53.59)
Liver (N = 1413, 15.77%)
Gastrointestinal (stomach) (N = 1094, 12.21%)
Pancreas (N = 965, 10.77%)
Bowel (N = 273, 3.05%)
Esophagus (N = 260, 2.90%)
Other (N = 153, 1.7%)
The organ subsites in Other include Gall Bladder, Bile Duct, Anal, and Abdominal.
As noted in Appendix 1, for D/G/B there were far more questions about Symptoms (N = 1506, 16.81%) than Treatment (N = 294, 3.28%) possibly reflecting the fact that (1) users of Ask.com were just beginning their D/G/B information seeking and (2) there is less complexity in the published Treatment data for D/G/B compared to some other cancer types, like Breast Cancer.
The terms Bowel, Gastrointestinal, Stomach, and Abdominal may have been used interchangeably by users. They appear not to recognize that queries for sigmoid, rectum, cecum, appendix, transverse colon, small bowel, and stomach (gastric) cancer would provide much more useful information.
For D/G/B, some queries about Liver Metastases were included with queries about primary Liver Cancers.
8.0 General Research
There were 7808 queries assigned to the highest-level category General Research, a topic not linked to a specific cancer type. In this category the 5 most-common subcategories were:
Research (N = 2819, 36.10%)
Organization (N = 1656, 21.21%)
Clinical Trials (N = 1272, 16.29%)
Concerns (N = 1201, 15.38%)
Pictures (N = 559, 7.16%)
Among the queries about Organization, there were 1065 queries about the American Cancer Society (ACS) and 223 about the National Cancer Institute (NCI).
Among the 1272 queries about Clinical Trials, the most-common 3 questions/topics were:
What are ... (N = 634, 49.84%) eg, "What are clinical trials?"
Latest ... (N = 260, 20.44%) eg, "latest cancer clinical trial research"
Types of ... (N = 111, 8.73%) eg, "types of cancer trials"
9.0 Genitourinary Cancers
In decreasing order, the frequency of Genitourinary organ-type queries (N = 6250) in all 12 Genitourinary subcategories including General Information was:
Prostate (N = 3141, 50.26%)
Testicular (N = 1772, 28.35%)
Bladder (N = 708, 11.33%)
Kidney (N = 496, 7.94%)
Other (N = 133, 2.12%)
Although it has been estimated that there were 198100 new cases of Prostate Cancer diagnosed in the US in 2001 and only 7200 cases of Testis Cancer [34], the relative frequency of Testis Cancer queries was quite high. One possible reason might be that males diagnosed with Testis Cancer are generally much younger than those diagnosed with Prostate Cancer, and those younger individuals might be more-frequent information seekers on the Internet. It may also reflect the fact that the 2001 Tour de France bicycle race won by Lance Armstrong, a Testis Cancer survivor, was held during July, coinciding with the study period for this project.
As with most sites, the most-common Prostate Cancer questions were General Information (N = 1715, 54.6%). For Prostate Cancer, there were more questions about Treatment (N = 460, 14.65%) than Symptoms (N = 364, 11.59%). This may reflect major medical controversies about treatment options and the typically asymptomatic presentation of the disease.
For the Genitourinary category as a whole, there were more questions about Symptoms (N = 854, 13.66%) than Treatment (N = 604, 9.66%).
Expected misspellings of prostate (prostrate) were noted.
10.0 Gynecological Cancers
There were 5344 queries overall. The breakdown of subcategories in decreasing frequency was:
General Information (N = 3409, 63.79%)
Symptoms (N = 939, 17.57%)
Diagnosis and Testing (N = 452, 8.46%)
Treatment (N = 247, 4.62%)
Definition (N = 158, 2.96%)
Cause/Risk (N = 83, 1.55%)
Statistics (N = 42, 0.79%)
Prevention (N = 14, 0.26%)
In decreasing order of frequency, the cancer types queried in all 8 Gynecological subcategories included the following:
Ovarian (N = 2031, 38.00%)
Cervical (N = 1924, 36.00%)
Uterine (N = 606, 11.34%)
Endometrial (N = 225, 4.21%)
Vulvar (N = 166, 3.11%)
Vaginal (N = 219, 4.09%)
Other or not specified (N = 173, 3.24%)
There were nearly as many questions about Cervical Cancer as Ovarian Cancer despite the fact that in the United States in 2001 the estimated incidence of new Ovarian Cancers was about twice that of invasive Cervical Cancer [34].
There were questions about Endometrial cancer as well as Uterine cancer. These data suggest that Web site information needs to be provided using both labels.
11.0 Head and Neck
There were 2522 queries overall. Most queries asked for General Information (N = 1485, 58.88%). The vocabulary used to ask about specific cancer types within General Information was:
Throat
Mouth
Oral
Tongue
Head
Neck
The vocabulary confirms the need to offer health information with words that are not technical like larynx, glottis, pharynx, or nasopharynx. There were 59 questions asking about Definitions of Head and Neck cancer. Specifics about cancer anatomy of this cancer type may be less familiar to the general public than other sites.
There were 422 queries asking for Pictures of Head and Neck Cancer. There were only 47 questions (1.86%) asking about Cause/Risk/Link issues, despite the fact that there is a great deal known about the Causes and Prevention of Head and Neck Cancer. There were 418 questions (16.57%) about Symptoms and but only 52 (2.06%) about Treatment.
12.0 Hematologic and Blood Cancers
Among 5448 queries in this category, the 5 most common of the 12 subcategories were: General Information (N = 3781, 69.40%), Definition (N = 701, 12.96%), Symptoms (N = 539, 9.89%), Treatments (N = 175, 3.21%), and Organizations (N = 102, 187%). Within General Information users asked about Leukemia (N = 2895, 76.57%), Myeloma (N = 592, 15.66%), Bone Marrow (N = 148, 3.91%), and Blood Cancers (N = 146, 3.86%). Various misspellings of Leukemia were noted and nontechnical terms such as Blood Cancer and Bone Marrow Cancer were frequent.
13.0 Lung Cancer
Lung Cancer (N = 4630) accounted for 8% of organ-type specific queries within the highest-level Cancer category. This is a disproportionately-low percentage given the relative incidence of Lung Cancer in the United States in 2001 [32]. There were more queries about Gynecological and Hematologic/Blood cancers, even though the US incidence for these is far lower.
Among Lung Cancer queries, the queries were classified as follows:
General Information (N = 3223, 69.61%)
Symptoms (N = 530, 11.45%)
Cause/Risk/Link (N = 305, 6.59%)
Treatment (N = 219, 4.73%)
Definition (N = 150, 3.24%)
Statistics (N = 113, 2.44%)
Diagnosis and Testing (N = 90, 1.94%)
In the Cause/Risk/Link category of Lung Cancer, there were only N = 180 queries (59.02%) that asked generally about Causes of Lung Cancer and N = 102 queries (33.44%) that asked specifically about Smoking. There were N = 23 queries (7.54%) asking if Marijuana caused Lung Cancer.
Only N = 255 (7.91%) queries within General Information asked about Lung Cancer by (histologic cell) Type, despite the fact that this is a major determinant of triage for treatment.
For Lung Cancer > Treatment, there were 219 queries (4.73%). Most Treatment queries were Unspecified (N = 118, 53.88%), eg, "What are treatments for lung cancer?" There were 26 Treatment questions about Cure (11.87%). There were few specific questions about Medications (chemotherapy) (N = 21, 9.59%), Radiation (N = 19, 8.68%), or Surgery (N = 10, 4.57%). Although all numbers were small, there were more questions about Alternative Treatment (N = 13, 5.94%) than Surgery (N = 10, 4.57%). There were only 4 Treatment questions (1.83%) about palliative care, despite the grave prognosis for most Lung Cancers. Clearly the questions about Lung Cancer, the most-common lethal cancer, were far less sophisticated than the questions about either Breast Cancer or Prostate Cancer.
14.0 Lymphomas
Among the 3333 queries about Lymphoma (including both Hodgkin's Disease and Non-Hodgkin's Lymphoma), General Information (N = 2391, 71.74%) questions were the most common. Unlike many cancer types, there was frequent mention of histologic types, as is appropriate, given the wide variety of clinically-different prognoses and treatments in this subcategory. There were many different spellings of Hodgkin's Disease.
15.0 Miscellaneous Cancers
There were 1633 queries assigned to this Cancer subcategory. The Miscellaneous Cancers were:
Endocrine (N = 901, 55.17%)
Neoplasm (N = 272, 16.66%)
Kaposis (N = 262, 16.04%)
Ocular (N = 179, 10.96%)
Germ Cell (N = 19, 1.16%)
Several of the Ocular queries, eg, Ocular Melanoma and Retinoblastoma, could have been considered for other subcategories, such as Skin and Pediatric respectively. Germ cell tumors could also have been placed in either Genitourinary or Gynecological subcategories. These ambiguities illustrate the difficulty in categorizing precise user information needs despite the use of natural language processing.
16.0 Pediatric
There were only 603 Pediatric queries, and most asked about a specific cancer type (N= 403, 66.83%). There were relatively few General Information queries (N = 81, 13.43%) eg, "where can I find information on children's cancers?" Since patients with Pediatric cancers in the US are usually managed generally by pediatric oncology specialists at major regional medical centers, those seeking Pediatric cancer information are probably directed to specialized Web sites rather than general sites like Ask.com.
Of 403 queries for cancer types, the most common were Hematologic/Blood (N = 137, 34%), Neuroblastoma (N = 133, 33%), and Rhabdomyosarcoma (N = 68, 16.87%). There were only 4 questions referring to pediatric Brain and Neurological cancers. Since this is such a common Pediatric tumor type, it is possible that some Pediatric neurological tumor questions were assigned to the Brain and Neurologic Cancer category even though the questions were really meant to target a Pediatric issue.
17.0 Skin Cancers
Among 6709 queries in this Cancer subcategory, 3596 (53.60%) asked for General Information. Like Lymphoma, there was frequent mention of specific Skin Cancer types (N = 2157, 32.15%), probably because of the significantly-different clinical prognoses and treatments.
Only 169 queries (2.52%) asked about Cause/Risk/Link, and 60 queries (0.89%) asked about Prevention despite the fact that so much is known about these topics and Skin Cancer.
Among Skin Cancers queried by histologic cancer type (N = 2157, 32.15%), Melanoma was the most common (N = 1707, 79.14%), even though it is far-less common than Basal Cell Skin Cancers (N = 322, 14.93%) [10]. Frequent mention of Melanoma probably reflects its more-serious prognosis and more-complicated clinical triage.
18.0 Soft Tissue Cancers
There were 3954 queries in this Cancer subcategory. Although most appropriately refer to sarcomas of various types, there was a minority of misplaced queries. Some queries appear to reference conditions that are probably benign (Ganglion, Fibroid, Dysplasia, and Lipoma) and others should have been placed in different Cancer subcategories eg, Brain and Neurological (Oligodendroglioma and Glioma) These will be corrected on later analyses.
19.0 Treatment
In the 3832 highest-level category queries about Treatment, most questions were about a specific Treatment Type (N = 3223. 84.11%), even though no specific cancer was mentioned. Within Treatment > Treatment Type there were many general queries about Chemotherapy (N = 2275, 70.59%). There were questions about general Radiation Therapy (N = 534, 16.57%), and few about specialized Radiation Therapy treatments like Gamma Knife, Laser, and Protons. There were more general questions about Alternative Therapies (N = 239, 7.42%) than Surgery (N = 127, 3.94%) Many Alternative Therapy questions also appear in specific organ-type subcategories, particularly Breast.
Query Frequency Relative to US Incidence of Cancer Types
Table 4 compares the incidence of selected cancers in the United States (US) in the year 2001 with the frequency of selected site-specific cancer queries in this report. It has been estimated that there were 1268000 new cancer cases in the US in 2001 [34]. The sites in Table 4 were selected specifically because they were easiest to compare directly.
The relative percentage of specific organ-type queries exceeds the percentage of annual incidence only for rarer cancers. The difficulty of finding useful information on prominent cancer portals or with standard search engines may be one explanation, although there are others. The comparison is not meant to be definitive as there are clearly issues with validity of this comparison:
Cancer prevalencemight be a better benchmark than incidence
US incidence data exclude cases of in situ breast and cervix cancers as well as the very-common basal cell and squamous cell skin cancers
Queries could have come from anywhere in the world, not just the United States
Query total may include those who accessed the site more than once
Queries could have come from individuals who are not newly-diagnosed patients
Table 4
Comparing relative annual US incidence of selected cancers and query frequency
Cancer Site
Estimated Number of New US Cancers Diagnosed in 2001
*
% of Estimated New US Cancers in 2001
*
†
Number of Cancer Site-Specific Queries in This Report
% Queries in Cancer Category
†
‡
Digestive (D/G/B)
235700
18.6
8959
15.0
Prostate
198000
15.6
3141
5.3
Breast
193700
15.3
6953
11.7
Lung
169000
13.3
4630
7.8
Lymphoma
63600
5.0
3333
5.6
Bladder
54300
4.3
708
1.2
Uterus/Endometrial
38300
3.0
931
1.6
Head and Neck
30100
2.3
2522
4.2
Ovary
23400
1.9
2031
3.4
Brain and Neurological
17200
1.4
1852
3.1
Cervix
12900
1.0
1924
3.2
Soft Tissue
8 700
0.69
3954
6.6
Testis
7200
0.57
1772
3.0
* Data from 2001 Estimated Annual US Cancer Incidence Figures (N = 1268000) [10].
† Percentages in columns 3 and 5 do not add up to 100% because only selected cancers were included in this chart.
‡ Only selected Cancers were included in this chart. Denominator (N = 59619) was the total number of queries about specific subsites in the Cancer category.
Other Observations
The query analysis reveals that online users generally seek information about Symptoms and Treatment for specific cancers, rather than about cancers generally. In addition, Symptom queries showed a frequency between 2 and 5 times that of Treatment queries, for most cancers.
For this study we did not specifically target queries about Adult Immune Deficiency Syndrome (AIDS), even though AIDS can often be associated with Cancer. There were 262 questions about Kaposi's Sarcoma in the Miscellaneous Cancers category.
Discussion
General Information was the largest category for almost all cancers, probably reflecting the nature of the Ask.com consumer search engine. It is a consumer-oriented Web-wide search engine where users tend to seek general information that can help them learn either how or where they should further pursue their inquiries. It is likely the users are just starting their Web searches on Ask.com and they are not yet interested in or they do not yet know enough information to ask more-sophisticated questions. This behavior may not reflect that of users who go directly to a known cancer-information portal with a predetermined need for detailed information.
We attempted to capture and analyze all cancer-related queries, including those with correct and incorrect spellings. Misspellings were noted relatively frequently, but we have no data on the number of misspellings, as we did not target this in advance as an endpoint, and we did not have direct access to the raw data logs. Appendix 1 shows verbatim queries with examples of the misspellings. Automating help for users who enter misspelled words is a major issue for search engines in order to optimize query results. Other researchers have noted the search difficulties related to spelling of cancer search terms correctly [35].
Ask.com users entered both keyword searches and sentence-style queries, despite the fact that this is a natural-language-processing search engine. We recognize that even if users typed in a long query it was still sometimes difficult to discern absolutely what specific information the user needed, particularly since we did not have access to the links users picked.
The vocabulary employed by users of Ask.com ranged from unsophisticated to very sophisticated. This suggests that allowing users to employ less-technical language on cancer Web sites would significantly help them find the information they seek.
The queries captured for this study undoubtedly reflect the news and research studies in the public arena during the time period from June to August 2001. A different time period would certainly reflect a different distribution. Examples of the kinds of events that could affect the results include the diagnosis or death of a celebrity with cancer, the publication of a major trial about bone marrow transplantation for breast cancer, or the Food and Drug Administration approval of an important new drug.
The presence of a search engine with natural language processing on a Web site, while potentially valuable to users, does not obviate the need for good user-centered Web site design and information architecture [36]. It has been shown that searching via search engine can be minimized and user satisfaction maximized if information architecture and link titles follow appropriate guidelines [37]. Nevertheless, for less-sophisticated users, a natural-language-processing search engine can be helpful in finding the information users seek and provide enhanced success in searching.
An October 30, 2003 search of the PubMed Web site [38] of the National Library of Medicine [39] yielded 458 search results from a query for "Natural Language Processing." Most citations were from publications within the last 3 years, attesting to the currency of natural language processing as an important research topic cutting across a wide variety of research disciplines. Potential data-mining applications of this tool in medicine extend far beyond the use described in this paper.
Eysenbach and Kohler have recently developed a novel methodology, similar to the method used in this study, to estimate the actual volume and prevalence of health-related searches on the Web in relation to the total number of searches conducted daily on the Internet [40]. They collected queries from 2 search engines, Metacrawler (a search engine of search engines) [41] and Ask.com [29] (the same natural-language-processing search engine used for this report). These 2 search engines were selected because they allowed "peeking" at actual user search-query topics. They concluded that 4.5% of all searches on the Web might be health related. The queries were collected from Metacrawler between February 2001 and April 2002, and from Ask.com between February 2001 and April 2001. The first date range overlapped our study dates and the second occurred just before data collection for our study.
In summary, natural-language-processing tools such as the one used for this study are able to filter and subset raw query data into useful analysis categories. Retrieval and analysis of these data can be used to better understand the actual content users want and the level of understanding and sophistication they have when they come to the Web site. Using the information on a continuing basis can form the basis for updating content on Web sites based on the most-current user needs. If a natural-language search engine were offered on a health-information portal, for example, it could improve customer access to desired information, particularly for those users with less sophistication about content or language. Additional analyses of query results are planned for the future. Consideration has been given to piloting the use of natural language processing on subsites of our Web portal.
|
[
"cancer",
"internet",
"search engines",
"natural language processing"
] |
[
"P",
"P",
"P",
"P"
] |
Intensive_Care_Med-3-1-1915613
|
Pediatric delirium in critical illness: phenomenology, clinical correlates and treatment response in 40 cases in the pediatric intensive care unit
|
Objective To study the phenomenology, clinical correlates, and response to treatment of delirium in critically ill children in the pediatric intensive care unit (PICU).
Introduction
Delirium is a neuropsychiatric disorder secondary to a general medical condition, and must be considered a serious complication of the underlying disease or its treatment. In the revised fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR®), delirium is defined by four concurrent diagnostic criteria: (1) acute onset and fluctuation of (2) a disturbance of consciousness with reduced ability to focus, shift or maintain attention and (3) a change of cognition with memory deficit, disorientation, language disturbance, perceptual disturbances or hallucinations, (4) caused by the direct physiological consequences of a general medical condition [1]. It is frequently seen in critically ill adult and geriatric patients [2–5] and is associated with a poor prognosis, reflected by longer hospital stay, worse functional and cognitive outcome, and a higher mortality rate after discharge from hospital [3]. In mechanically ventilated critically ill adults, delirium is an independent predictor of elevated 6-month mortality and a longer hospital stay [6]. If appropriate diagnostic tools validated for bedside use by non-psychiatrists [e.g. Delirium Rating Scale (DRS), Confusion Assessment Instrument for the Intensive Care Unit (CAM-ICU)] are used, delirium is diagnosed in over 80% of critically ill adult patients [7]. Thus, systematic monitoring for delirium and appropriate treatment with haloperidol in critically ill adult patients were included in the recently published clinical practice guidelines for sedatives and analgesia of the Society of Critical Care Medicine [8]. However, the optimal management of patients with delirium and the effects of the pharmacological treatment on the outcome are still key concerns for today [9]. Given lack of age-appropriate diagnostic criteria and assessment tools in children, even less is known about the incidence, clinical presentation, response to treatment and consequences of childhood delirium in general, and in critically ill children in particular [10–12]. The few available published data on childhood delirium suggest that morbidity and mortality are higher in children with than in children without delirium [13]. Therefore, delirium in children should be considered a serious complication and be treated accordingly. Unfortunately, while there are comprehensive guidelines on the diagnosis and treatment of delirium in adults, clinical guidelines for delirium in children are nonexistent.
The aim of this study was to investigate the incidence, patient and population characteristics, clinical presentation and response to treatment of delirium in a cohort of critically ill children admitted to a tertiary pediatric intensive care unit (PICU). Given the necessarily multidisciplinary approach to assessment and treatment of these children, the input of four disciplines – child psychiatry, pediatric intensive care medicine, child neurology and adult neuropsychiatry – was used.
Methods
Design, setting and patients
A descriptive study was carried out over a 4-year period (January 2002 to December 2005) in an eight-bed tertiary PICU. This PICU is a tertiary referral center for both general and surgically critically ill children in the southeastern region of the Netherlands (population 1.4 million, 350 annual admissions). Critically ill children, acutely, non-electively and consecutively admitted, were prospectively sampled. Both mechanically ventilated and non-ventilated patients were included.
Diagnostic approach
All children with (1) confusion, agitation, anxiety, moaning, discomfort, or behavioral disturbances with no acceptable medical explanation or (2) failure of standard analgosedative treatment were systematically assessed for the presence of delirium in a two-step diagnostic approach. The standard analgosedative treatment can be summarized as follows: children who required analgosedation because of obvious or expected pain or because of stress related to their underlying disease or treatment received adequate doses of opioids and/or benzodiazepines according to internationally published guidelines for analgesia and sedation in critically ill children [14]. Drug doses were individually tailored to achieve optimal patient comfort and were slowly reduced in order to avoid a withdrawal syndrome. In the event that a withdrawal syndrome was suspected, based on clinical observation or the revised Finnegan score, specific treatment with long-acting benzodiazepines (e.g. lorazepam) or opioids (e.g. methadone) was started according to internationally published guidelines [15, 16].
At the time the assessment for delirium was initiated, none of the patients had signs of imminent life-threatening respiratory, circulatory or neurological failure, while ongoing asphyxia, respiratory acidosis, metabolic disturbances, fighting the ventilator due to inappropriate ventilator settings and withdrawal syndrome all had been excluded systematically as an explanation for the observed behavior.
The first step of the diagnostic approach was a systematic assessment by a child neuropsychiatrist (J.S.) using DSM-IV criteria for delirium. Criteria were evaluated on the basis of (1) hetero-anamnestic information from parents, nurses, intensivists, and child neurologists about behavior and behavioral changes and (2) child psychiatric examination. Based on the findings, patients were categorized as having a (probable) delirium or not. In a second step, the provisional diagnosis of delirium was further tested in a daily multidisciplinary consensus meeting. The team consisted of the child neuropsychiatrist, the attending pediatric intensivist, and occasionally a geriatric neuropsychiatrist specialized in delirium in geriatric patients and/or a child neurologist. If this team agreed that alternative explanations for a child's behavior were unlikely, the consensus diagnosis was delirium.
Based on the dominant clinical presentation, cases of delirium were classified as “hyperactive” when psychomotor agitation was present, and as “hypoactive” when retardation and/or inhibition was present. A number of children presented with cognitive and/or attentional disturbances in the context of severe anxiety states, often accompanied with moaning and restlessness, but without clear agitation or retardation. This latter group was classified as “emerging” or “veiled” delirium [17], in reference to the way delirious syndromes have been described as “partial delirium” in adult ICU patients or as “subsyndromal delirium” in elderly medical patients [7, 18, 19]. The different presentation forms were not always clear-cut, and some fluctuated dramatically over time. The severity of illness was scored according to the Pediatric Index of Mortality (PIM) and Pediatric Risk of Mortality (PRISM) [20].
Therapeutic approach
Whenever delirium was identified or suspected, a two-track treatment approach consisting of both psychosocial and pharmacological interventions was implemented. Psychosocial interventions – the parents' presence and comforting throughout the day (and night), familiar music, favorite toys, pictures of home and pets, friends, school, lighting schedules, sometimes even fragrances – are standard in the PICU. The parents also received an information leaflet on childhood delirium [21]. All patients were also treated with antipsychotic medication after the referring pediatric intensivist had agreed and the parents, because of the off-label use, had given informed consent, which was never refused. In children with psychomotor agitation that was acutely threatening to their health status, haloperidol at a loading dosage of 0.15–0.25 mg i.v. was used, given slowly over a period of 30–45 min, followed by a maintenance dose of 0.05–0.5 mg/kg/24 h i.v. [11, 22, 23]. In less acute situations, and when oral medication was possible, risperidone at a loading dose of 0.1–0.2 mg p.o. was used, followed by a maintenance dose of 0.2–2.0 mg/24 h p.o. as the treatment of choice. Clinical response and side effects were recorded by the child neuropsychiatrist and the pediatric intensivists. In order to tailor the treatment for delirium, daily discussions were held with the multidisciplinary team. Adjustment of treatment was based on the clinical observations and judgements of the parents, nurses, intensivists and child psychiatrist.
Children were followed up for 6 weeks after discharge from the hospital either at the outpatient clinic or by contacting the parents by telephone.
As the study solely involved the structured recording of routine clinical practice, under Dutch law no institutional review board approval was required.
Results
From January 2002 to December 2005, there were 877 acute, non-elective admissions to the PICU. Distribution of age and gender are shown in Table 1. In 61 cases (7%), a systematic assessment by a child neuropsychiatrist was requested, usually for agitation, anxiety, moaning, discomfort, behavioral disturbance or problematic analgosedation.
Table 1Number and incidence of delirium in the total sample* by age and genderAgeTotal sample*Patients with deliriumIncidence (%)0–2.99 years51314 2.7Male310 9 2.9Female203 5 2.53–5.99 years106 4 3.8Male 56 3 5.4Female 50 1 2.06–8.99 years 80 6 7.5Male 46 1 2.2Female 34 514.79–11.99 years 77 3 3.9Male 61 3 4.9Female 16 0 012–14.99 years 70 710Male 35 514.3Female 35 2 5.715–18 years 31 619.4Male 13 430.8Female 18 211.1Total87740 4.6Male52125 4.8Female35615 4.2*Critically ill children, acutely, non-electively and consecutively admitted
during a 4-year period
Of these 61 patients, 40 (61%) were diagnosed with delirium, yielding a cumulative incidence of 5% (boys 5%; girls 4%). Age-specific incidences increased from 3% in the 0–3 years age group (boys 3%; girls 3%), to 19% in the 16–18 years age group (boys 31%; girls 11%) (Table 1, Fig. 1).
Fig. 1Incidence of delirium in the sub-groups by age and gender
The child psychiatric disorders diagnosed in the 61 referrals are summarized in Table 2.
Table 2Child psychiatric diagnosis at the first consultation (n = 61)nDiagnosis40delirium 5adjustment disorders with anxiety and depressed mood, post operative 4psychological–psychiatric factors affecting the medical condition 3anxiety disorder 3emotional and behavioral problems during chronic ventilation 2adjustment disorders with depressed mood 1mood disorder 1adjustment disorder with anxiety 1sleeping problem 1feeding problem
Table 3 summarizes the population characteristics of the sample diagnosed with delirium, while Table 4 lists the underlying somatic disease or causative pharmacological treatment.
Table 3Population characteristics of the 40 PICU cases with delirium, 2002–2005CharacteristicsFrequency (total n = 40)Age (mean ± SD) 7.6 ± 5.9Malen25 age (mean ± SD) 7.6 ± 6.1Femalen15 age (mean ± SD) 7.6 ± 5.8Ethnicity White36 (90%) African 3 (7.5%) Asian 1 (2.5%)Mechanical ventilation34 (85%)PIM score (mean ± SD)9.96 ± 16.20PRISM score (mean ± SD)23.54 ± 24.94Major somatic pharmacological features Recent increase or decrease of analgosedative medication22 (55%) Neurological disorders21 (52%) Infectious disorders20 (50%) Respiratory disorders12 (30%)PIM, Pediatric Index of Mortality; PRISM, Pediatric Risk of MortalityTable 4Patient characteristics of the 40 PICU cases with delirium 2002–2005No.SexAgePrimary diagnosis on admission PICUMechanical ventilationDelirium typeTreatment 1M 3 monthsMultiple congenital malformations+EmergingHaloperidol 2F 4 monthsMeningococcal septic shock+EmergingRisperidone 3M 4.5 monthsSevere CLD+EmergingRisperidone 4F10 monthsNear drowning+EmergingHaloperidol 5F 1 yearPneumonia+EmergingHaloperidol 6M 1 yearSepsis due to perforated appendicitis+EmergingHaloperidol 7M 1 yearSubarachnoid bleeding+EmergingHaloperidol 8M 1 yearPericardial effusion with pretamponade+HyperactiveHaloperidol 9M 1 yearMultiple dysmorphia, upper airway obstruction–EmergingHaloperidol10F 2 yearsMeningococcal septic shock+HyperactiveHaloperidol11M 2 yearsADEM+HyperactiveRisperidone12M 2 yearsCervical mass, upper airway obstruction+EmergingHaloperidol13M 2 yearsAspiration and pneumothorax+Hypoactive–14F 2 yearsMeningococcal meningitis with sepsis and DIC+HyperactiveHaloperidol15F 3 yearsCystic fibrosis and pneumonia+HypoactiveHaloperidol16M 4 yearsIntracerebral hemorrhage, Marfan syndrome+HypoactiveHaloperidol17M 5 yearsMedulloblastoma post surgery+EmergingRisperidone18M 5 yearsUpper respiratory tract infection+HyperactiveHaloperidol19M 6 yearsMultiple trauma+EmergingRisperidone20F 8 yearsMeningo-encephalitis+HyperactiveRisperidone21F 8 yearsViral encephalitis+Hyperactive–22F 9 yearsStatus asthmaticus+HyperactiveRisperidone23F 9 yearsTBI, gunshot wound+HypoactiveRisperidone24M 9 yearsStatus asthmaticus+HyperactiveHaloperidol25M 9 yearsNeural tube defect and drain dysfunction–Emergingfirst Haloperidol, then Risperidone26M11 yearsHypovolemic shock, typhus abdominalis–HyperactiveHaloperidol27F12 yearsTBI+HypoactiveHaloperidol28M12 yearsTBI and fracture of lower leg+EmergingHaloperidol29M13 yearsSepsis, paronychia–HyperactiveHaloperidol30M13 yearsStatus epilepticus+EmergingHaloperidol31F14 yearsTBI+HyperactiveHaloperidol32F15 yearsPost TBI+HypoactiveHaloperidol33M15 yearsPostoperative state+HypoactiveRisperidone34M15 yearsAcute lymphoblastic leukemia–HypoactiveHaloperidol35M15 yearsTBI+EmergingHaloperidol36F15 yearsStatus asthmaticus+HyperactiveHaloperidol37M16 yearsMultiple trauma–HypoactiveHaloperidol38F16 yearsBacterial meningitis+HyperactiveRisperidone39M16 yearsRespiratory failure, Duchenne disease+EmergingHaloperidol40M17 yearsSeptic shock+EmergingHaloperidolCLD, chronic lung disease; ADEM, acute disseminated encephalomyelitis; DIC, diffuse intravascular coagulation; TBI, traumatic brain injury
The underlying features were: recent increase or decrease in analgosedative medication (n = 22), neurological disorders (n = 21), infections (n = 20) and respiratory disorders (n = 12). Usually, a combination of these existed.
All but two patients were treated with an antipsychotic drug. Twenty-seven children were given haloperidol, 10 risperidone, and 1 child received both drugs in succession. In most cases, the beneficial results were observed rapidly, especially in the hyperactive forms, sometimes even after a single dose [11]. Sometimes it took a few hours or days. Two patients experienced acute dystonia as a likely side effect of the haloperidol, responding well to biperidene. Two children received no medication: one because of lack of consensus in our expert team, and one because of an endotracheal intubation at the time that medication was indicated. In most cases, the medication was stopped or tapered off successfully during hospitalization or afterwards in an outpatient setting. Five children (12.5%) with delirium died of their underlying disease; the mean PIM was 10% and the mean PRISM 24% (Table 3).
Discussion
This is the first systematic multidisciplinary study of the phenomenology and treatment of delirium in 40 critically ill children in a PICU context. The low cumulative incidence of 5% is mainly the result of the low incidence in the younger age groups (< 9 years old), this segment constituting the majority of the total sample (80%). A clearly higher incidence is seen in the older age groups. However, in critically ill adult patients substantially higher incidences have been reported, ranging from 10–30% in general hospital settings to 50% in postoperative patients and up to 80% in the terminally ill [2].
There are several possible explanations for this difference. First, the incidence of delirium in young critically ill children may be truly low: differences in age-related resilience and underlying conditions may contribute to true differences in the incidence between the very young and the very old. However this explanation seems unlikely, given the tendency in the very young to develop delirium under even much less severe circumstances [1, 11, 24]. A second explanation may relate to the fact that extensive psychosocial interventions are provided as a routine in Dutch PICU settings, with a possibly preventive effect on delirium in much the same way as observed in geriatric patients [25]. A third factor may be an anti-delirium effect of the routinely used analgosedative medication, although especially benzodiazepines may have excitatory side effects in children. A fourth factor may relate to reluctance on the part of the intensivists and/or child neurologists to request psychiatric evaluation in the case of behavioral changes for fear of adding stigmatization to an already burdened system. Perhaps the fifth, most likely and important explanation is that parents, nurses, pediatric intensivists and child neurologists do not easily recognize delirium, because the medical condition of these critically ill PICU children is so complex and changeable. It is possible that a psychiatric consultation was readily requested only in cases of anxiety and/or hyperactive delirium, not in the hypoactive and or veiled ones.
The differential diagnosis of pediatric delirium consists of acute stress reactions, acute anxiety states, adjustment disorders with mixed emotions, dissociative and/or regressive states and childhood-onset psychosis (see also Table 2). However, differentiating delirium from extreme stress and agitation due to acute and life-threatening conditions is not only impossible, it is also unwanted, because it is usually irrelevant at that point of time in the process of medical care. Causative treatment, if possible, is always the first step to be taken. Furthermore, the diagnosis of delirium in children is complicated by the fact that the criteria for adult delirium are not always easily applicable to children because of important differences in age and developmental levels. The DSM-IV criteria for delirium are not always useful in the case of pediatric delirium, especially in a PICU context. This is a source of concern. Moreover, delirium is not mentioned in the DSM-IV section on child psychiatry. DSM-IV describes as an essential feature of delirium the “disturbance of consciousness” leading to “impairment of the ability to focus, shift and sustain attention”. This, however, is of little relevance in the critically ill in a PICU context, where disturbance of attention is routinely present due to the disorder(s) itself. In fact, attention is often the first “to go” [17, 26]. In addition, patients almost always require treatment with opioids and benzodiazepines, which also have a strong impact on attention. Furthermore, it has been hypothesized that a disturbance of consciousness is not a discriminating characteristic of delirium in an ICU setting [27].
Our case series suggests that in addition to the hyper- and hypoactive forms of delirium, a third form may be characterized by anxiety, moaning, and/or restlessness. This was referred to as an “emerging” or veiled delirium, as described previously [17]. In the PICU population this state did not develop into a frank hyper- or hypoactive form, but appeared to exist in its own right, accompanied by disturbances of consciousness and cognition.
Although atypical presentations of disorders are often not captured in classification systems, the high prevalence of “emerging” delirium in our sample (17/40) stresses the importance of further phenomenological study. Adhering too strongly to DSM-IV criteria for adult delirium, for clinical use in a PICU context, may result in persistent under-diagnosis. On the other hand, the incidence of delirium may be overestimated by using the CAM-ICU, so most important is the issue of what constitutes delirium in critical illness [28].
There are accumulating indications that pediatric delirium can be subtle and accompanied or even dominated by other neuropsychiatric signs such as: reduction of awareness of the caregiver and/or the surrounding environment, purposeless actions, restlessness, inconsolability, signs of autonomic dysregulation and other subtle higher cortical dysfunctions [29–32]. Parents and nurses have a way of discriminating patterns in their children that may be diagnostically important. Thus, parents sometimes state: “This is not my child anymore.” Pediatric delirium therefore may have various subtle presentations and can be considered a spectrum disorder, which makes it sometimes difficult to diagnose [33]. Neither haloperidol nor risperidone is registered for the treatment of childhood delirium, although both are mentioned as the treatment of choice for adults [1]. Moreover, haloperidol is not registered for i.v. administration, even though it is used in this way in many places. We prefer risperidone in non-acute situations because of the theoretically lower risk of extrapyramidal side effects. Haloperidol and risperidone have been used for other indications in young children as well, such as childhood psychosis [34] or aggression in autism [35]. There are two limitations regarding our observations of treatment response. First, no severity scale for pediatric delirium was used, because none exists for this PICU population. Second, because no studies on the natural course of childhood delirium exist that have established the rate of spontaneous remission, spontaneous remissions may have been misclassified as response to treatment. In our opinion, however, the time frame of response points towards a medication effect.
Most medications employed in pediatrics and child psychiatry are used off-label [36], which means that special attention should be paid to information and informed consent procedures. Given the relatively high incidence of extrapyramidal symptoms with haloperidol, the “off-label” use needs further study [37].
This study has several limitations. First, it was a study of routine clinical practice. Observations were based on referrals emerging from care as usual. Although the focus on delirium may have altered referral paths and rates, we did not actively advocate any change, nor did we screen all admissions for delirium. In the absence of a hard clinical indication, no routine blood level measurements were performed to rule out persistent high levels of sedatives as a possible explanation for any neuropsychiatric symptom. Next, the lack of DSM-IV criteria for pediatric delirium and the disputed relevance of its main criterion in a (P)ICU setting make a standardized diagnosis difficult. Finally, treatment was provided in an open setting and based on consensus, especially among child psychiatrist, pediatric intensivist and child neurologist.
In conclusion, critically ill children in a PICU can develop delirium, with a hyperactive, hypoactive or veiled presentation, despite adequate analgosedation and intensive psychosocial interventions. The approach using an algorithmic structuring and an intensifying of daily clinical care, including the use of multidisciplinary daily consensus meetings, appeared effective in assessing, diagnosing and treating childhood delirium at the PICU. The findings suggest that the incidence is much lower than in adults, but a likely explanation is that delirious states requiring child psychiatric referral are still frequently under-diagnosed. There is also still a great need for developing delirium criteria in critically ill patients, children and adults alike, in a (P)ICU setting. Treatment with haloperidol or risperidone was successful in all patients. Future research is necessary to identify the risk factors for pediatric delirium in a multivariate prospective approach, to develop “easy to use” bedside tools for non-psychiatric trained team members for the early detection of delirium in all pediatric PICU patients, and to study the effects of interventions in a double-blind and ideally placebo-controlled fashion.
Disclosure
Jim van Os is a speaker or member of the advisory board for Lilly, BMS, Janssen-Cilag and AstraZeneca. He received grant or research support from Lilly, GSK, BMS and AstraZeneca. Albert Leentjens participated in research for Boehringer Ingelheim and is on the advisory board of a study related to Parkinson's disease by the same company. The other authors have no financial relationships to disclose.
|
[
"delirium",
"critically ill children",
"pediatric intensive care unit (picu)",
"haloperidol",
"consensus meetings",
"risperidone"
] |
[
"P",
"P",
"P",
"P",
"P",
"P"
] |
Diabetologia-3-1-1849422
|
Impact of the population at risk of diabetes on projections of diabetes burden in the United States: an epidemic on the way
|
Aims/hypothesis The aim of this study was to make projections of the future diabetes burden for the adult US population based in part on the prevalence of individuals at high risk of developing diabetes.
Introduction
Considerable evidence has been presented on the rise in diabetes prevalence in the United States and the United Kingdom [1, 2]. The prevalence of diabetes has become so large that it has been termed an epidemic [1, 3]. This rise is particularly important for healthcare needs in the US because almost 30% of individuals with diabetes are currently undiagnosed and diabetes is disproportionately represented in minority populations [4].
Projection of future disease prevalence helps to plan for healthcare needs. An understanding of the population at risk of developing the disease is critical when projecting future disease burden. Several studies have projected future diagnosed diabetes prevalence for the US and other countries [2, 5–9]. These studies, however, did not consider that not all individuals are equally at risk of developing diabetes, thereby possibly distorting estimates of downstream prevalence. For example, some risk factors for diabetes, e.g. obesity, have increased substantially in the population [10–12]. Moreover, many of these projections are based on estimates of diagnosed diabetes and exclude estimates of total diabetes (diagnosed and undiagnosed), which may lead to serious underestimation of the diabetes burden in the population.
Major risk factors for diabetes have been identified and are currently used by the American Diabetes Association to guide screening strategies. Although there are various measures for assessing the risk of having undiagnosed diabetes [13–17], few measures are available for assessing the risk of developing diabetes [18, 19]. Moreover, accounting for changes in the proportion of high-risk individuals, particularly as assessed through clinical indicators, has not been incorporated into previous projections of future diabetes burden.
The purpose of this study was to project the prevalence of diabetes for the adult US population up to 2031, using models based on data contained in the nationally representative National Health and Nutrition Examination Survey (NHANES) II mortality survey (1976–1992), NHANES III (1988–1994) and NHANES 1999–2002.
Materials and methods
Diabetes prevalence model
The model for diabetes prevalence used in this study was created using data from the NHANES III (1988–1994), and then fitted to data from the NHANES 1999–2002 as a validity check of the accuracy of the model’s projections. The resulting model was then used to project the number of individuals with diabetes in the US in 10-year increments into the future. We evaluated 10-year age classes at each 10-year interval. Our model has the following components:
Number of individuals with diabetesTime 2 = ∑ (number of individuals with diabetesTime 1i + incident casesi − mortalityi), where i equals each 10-year age group, and incident cases consist of: (1) persons converting from a disease-free state to having diabetes; (2) diabetic patients immigrating to the United States; and (3) persons with diabetes moving into the 20 to 29-year-old age class.The percentage of persons with diabetes, which is calculated thus:percentage diabetesTime 2 = (number of individuals with diabetesTime 2 / total populationTime 2) × 100.The estimate of future diabetes is therefore based on this equation, including the number of individuals with diabetes in the previous time period, conversion to diabetes, migration, and mortality, rather than being a linear extrapolation of the change in diabetes prevalence from the known values of 1991 and 2001.
Data sets
The NHANES is a programme of surveys conducted by the National Center for Health Statistics and designed to assess the health and nutritional status of adults and children in the United States. The survey is unique on a national level in that it combines interviews and physical examinations. The NHANES uses a complex multistage sampling design, making it representative of the non-institutionalised US population and allowing weighted estimates to be computed.
For this study we used several of the NHANES data sets. Specifically, we used the NHANES III (1988–1994) (unweighted n = 4,950) and the NHANES 1999–2002 (unweighted n = 3,804) to estimate among individuals of 20 years of age and older the prevalence of diagnosed diabetes, the total diabetes burden (diagnosed and undiagnosed diabetes), and the proportion of the population at risk of developing diabetes. Since mortality within the population affects future prevalence [20], we also used the cohort from the NHANES II mortality survey (1976–1992) (unweighted n = 3,916) to provide estimates of diabetes mortality. Computation of all analyses using the NHANES data sets to provide nationally representative estimates for the models was designed to account for the complex survey design and the appropriate sample weights. All analyses were conducted using SUDAAN software (Research Triangle Institute, Research Triangle Park, NC, USA).
Variables used in models
Prevalence of diabetes
Diabetes burden was assessed as diagnosed diabetes plus undiagnosed diabetes. Because of the substantial proportion of people with undetected diabetes, we focused on this formula for total diabetes, rather than using diagnosed diabetes to indicate diabetes burden in the population. Moreover, by focusing on diabetes as diagnosed and undiagnosed disease, we minimised the possible impact on future diabetes prevalence of changes in screening practices for diagnosing diabetes during an ensuing time period.
Diagnosed diabetes was assessed as individuals who answered yes to a question of whether a doctor had told them they had diabetes. Undiagnosed diabetes was estimated on the basis of individuals who said they had not had a previous diagnosis of diabetes, but who had fasting plasma glucose (FPG) > 7.0 mmol/l. Although, the diagnostic criteria for diabetes during the time between the NHANES II and the NHANES III changed from FPG > 7.8 mmol/l to FPG > 7.0 mmol/l, we used the newer criteria to gain an awareness of the total diabetes burden at each point in time using the same criteria [21].
Persons converting from a disease-free state to having diabetes
Although a variety of diabetes risk scores exist, most have been created from cross-sectional studies and have as their aim the identification of individuals with undiagnosed diabetes. Their ability, therefore, to make predictions on development of diabetes is unknown [13, 14, 16]. The risk score used in this study is based on one developed for the Atherosclerosis Risk in Communities (ARIC) cohort study [18]. Among individuals without diagnosed diabetes or FPG > 7.0 mmol/l, we used a scoring strategy which includes: high waist circumference (>102 cm in men, >88 cm for women), raised blood pressure (>130/85 mmHg or antihypertensive medications), low HDL-cholesterol (<1.03 mmol/l for men, <1.29 mmol/l for women), high triacylglycerol (>1.7 mmol/l), BMI > 30 kg/m2, and hyperglycaemia. Each of the characteristics is worth 1 point except for hyperglycaemia, which can be worth 2 points if FPG is > 5.6 mmol/l or 5 points when FPG is >6.1 mmol/l. A score of >4 puts an individual at high risk of developing diabetes, whether diagnosed or undiagnosed. A score of <4 indicates that a person has a low risk of developing diabetes.
This particular risk score was chosen for several reasons. First, it has moderate sensitivity (68%) and specificity (75%). Second, it is computed in a reasonably straightforward manner without having to use coefficients from the ARIC cohort that may be specific to that cohort. Third, data and results provided in the study by Schmidt et al. [18] allowed for computation of the rate of development of diabetes in both the high-risk group and the low-risk group. The ratio of development of diabetes in the high-risk group versus the low-risk group was 4.5:1. Variables needed to compute this diabetes risk score are available only in the NHANES III and the NHANES 1999–2002.
Although the ARIC diabetes risk score did not specifically consider race or age in the computation [18], we computed conversion rates for 10-year age classes for three race/ethnic groups (non-Hispanic Whites, non-Hispanic Blacks and Hispanic individuals) by fitting age categories for the data from 1991 to 2001 and then fitting race/ethnicity on to the same time change. We did not compute specific sex-specific conversion rates because sex was already differentiated in several of the variables in the ARIC diabetes risk score [18].
Migration of persons with diabetes
Migration of individuals with or without diabetes into the population can also affect future diabetes prevalence. Recent projections have included migration within their models [5]. Because we are looking at changes in diabetes prevalence among adults, migration of adults, particularly from ethnic minorities, could substantially affect the 10-year projections. We used data from the NHANES III to estimate migration of persons with diabetes in the 20 years and older age groups. The NHANES III measured how many years foreign-born immigrants had been in the US. Thus, we estimated the number of foreign-born individuals who had been in the country for 9 years or less for the total population as well as for different racial/ethnic groups. The NHANES III data allowed us to make estimates for non-Hispanic Whites, non-Hispanic Blacks and Hispanic individuals.
Persons with diabetes moving into the 20 to 29-year-old age class
For 2011, 2021 and 2031 the total number of persons with diabetes in the 20 to 29-year-old age class was estimated using a linear projection of the NHANES III and NHANES 1999–2002 data. The proportion of 20 to 29-year-olds with diabetes in each race/ethnic group was held constant at the proportions found in the NHANES 1999–2002 data at the later time intervals.
Mortality among individuals with diabetes
Diabetes mortality for the total population was based on data from the NHANES II mortality survey (1976–1992). This population-based cohort study was used to provide estimates of diabetes mortality, since mortality within the population affects future prevalence [20]. Diabetes mortality was estimated as all-cause mortality among individuals with diabetes (either diagnosed or undiagnosed) at baseline, rather than as mortality with diabetes listed as the cause of death. This definition is more consistent with the potential impact of diabetes on future prevalence. Mortality estimates were computed separately for the total population by age classes.
The NHANES II mortality cohort is based on a sample of individuals aged 30 to 75, whereas we made diabetes estimates on individuals aged 20 years and older. Consequently, we assumed no deaths due to diabetes in the 20 to 29-year-old age group over the 10-year period.
Population estimates
Total population of 10-year age classes was estimated using data from NHANES III for 1991, NHANES 1999–2002 for 2001, and US Census Bureau, Middle Series projections for 2011, 2021 and 2031 [22]. Total population of race/ethnic groups was also determined by 10-year age classes using the same sources of information.
Analysis
In an effort to provide an estimate of future trends in diabetes and the population at high risk of developing diabetes, we employed the following procedure. We used the NHANES III data to fit a model to predict total diabetes in the NHANES 1999–2002. We used this strategy prior to making future projections, because it allowed us to develop and fit the model to an existing national estimate of diabetes prevalence. Because both the NHANES III and the NHANES 1999–2002 are based on multi-year data collection, we estimated a mid-point of 1991 and 2001 for the two surveys.
The number of persons with diabetes 10 years post-baseline was calculated for 10-year age classes by first adding baseline prevalence and incidence (the number of low-risk and number of high-risk persons who developed diabetes over the 10-year interval), then adding persons with diabetes who immigrated to the United States, and persons with diabetes who moved into the 20 to 29-year-old age class, and finally subtracting the number of diabetic subjects who died. Percentage of persons with diabetes was estimated for each time period by taking the total number of persons with diabetes and dividing by the expected total population, then multiplying by 100.
Varying model assumptions
Our initial predictions of future diabetes burden were based on the assumption of a constant proportion of individuals at high risk of diabetes at the levels present in the NHANES 1999–2002. To account for potential changes in the proportion of persons at high risk of diabetes, we also evaluated increases in the proportion of persons at high risk by 10, 20 and 30%, as well as estimates based on decreases in the proportion of persons at high risk by 10, 20 and 30%. Theoretically, it is unlikely that the proportion of persons at high risk will remain stable, because from NHANES III to NHANES 1999–2002 the proportion at high risk was seen to increase. Also, a major risk factor for diabetes, obesity, has increased substantially over a 40-year time period [10, 12]. We evaluated the effect of decreasing proportions at high risk, to account for the possibility that interventions to improve lifestyle of adults in the US may be effective.
In addition, to address the potential impact on mortality of healthcare interventions in management of diabetes, we examined potential reductions of 10, 20 and 30% in mortality among individuals with diabetes. Finally, we computed a model examining a combination of effects, assuming that lifestyle interventions would yield a 10% decrease of persons at high risk and healthcare interventions would yield a 10% decrease in mortality of persons with diabetes.
Results
Table 1 shows estimates of the total diabetes burden from the NHANES III and the NHANES 1999–2002 and the future 10-year projections for 2011 through to 2031. The number of individuals at high risk of diabetes based on the multivariable diabetes risk score was 38.4 million in 1991 and 49.9 million in 2001. Using our model to predict the known diabetes prevalence in 2001 from the 1991 data, results were satisfactory and within 0.2% of the actual population prevalence of total diabetes. If the proportion of individuals at high risk within the adult population remains stable at 2001 levels, we could expect 55.8 million in 2011, 60.9 million in 2021, and 66.1 million in 2031. As can be seen, the prevalence of diabetes is projected to increase. The diabetes prevalence of 6.3% in 1991 and 8.8% in 2001 is projected to increase to 14.5% in 2031 with 37.7 million adults having diagnosed or undiagnosed diabetes. Assuming stability in the population proportion of individuals at high risk of developing diabetes, the rate of increase in the number of individuals with diabetes and the proportion with diabetes tends to slow over time. Among individuals aged 30 to 39 years who are not currently targeted for screening according to age, the prevalence of diabetes is expected to rise from 3.7% in 2001 to 5.2% in 2031.
Table 1Number of people (in millions of persons) with and prevalence of diabetes by year and age categoryAge (in years)NHANES IIINHANES 1999–200220112021203120–29Number0.20.40.50.70.9Percentage0.51.01.31.72.030–39Number0.61.61.82.12.3Percentage1.43.74.75.05.240–49Number1.53.54.34.34.9Percentage4.58.110.311.111.250–59Number2.53.96.67.26.9Percentage11.412.215.617.718.260–69Number3.14.16.59.810.2Percentage15.319.822.025.126.6>70Number3.24.15.78.412.4Percentage16.217.420.522.623.9TotalNumber11.117.525.432.637.7Percentage6.38.811.513.514.5NHANES National Health and Nutrition Examination Survey
The results shown in Electronic supplementary material (ESM) Table 1 show the projected prevalence of diabetes according to different racial/ethnic groups. Non-Hispanic White adults are projected to continue to have a lower prevalence of diabetes than both non-Hispanic Black and Hispanic individuals. By 2031, the Hispanic community will have an overwhelming diabetes burden, with more than 20 percent of the adult population having diabetes.
The projections in ESM Table 2 are based on different assumptions regarding changes in the number of individuals who are at high risk of developing diabetes and changes in mortality among individuals with diabetes. As might be expected, as mortality decreases the prevalence of diabetes increases in the subsequent 10 years. The estimate for 2031 indicates that potential decreases in mortality and a potential decrease in individuals at high risk of developing diabetes yields a prevalence similar to that achieved if the proportion at high risk is kept stable from 2001. All of these estimates indicate a larger diabetes burden among Hispanics.
Discussion
This national projection of diabetes prevalence for the US is the first to model the projection on the number of individuals at high risk of developing diabetes using a multivariable risk assessment. Projections suggest a rising and substantial diabetes burden for the population. Hispanic adults will be most affected, with estimates suggesting that by 2031 more than 20% of the adult Hispanic community will have diabetes. These results are particularly worrisome for this community in light of recent evidence that the gap in healthcare quality between Hispanic and non-Hispanic White individuals has continued to widen [23].
Many previous diabetes projections have been limited to estimates of diagnosed diabetes and thus have lower estimates of projected diabetes burden, and have not incorporated an evaluation of the population at high risk of diabetes, with clinical indicators, into their models [5, 9]. Our estimates will be less likely to be affected by changes in screening strategies. Additionally, they incorporate potential changes in the level of risk for diabetes in the US population, a change which is likely given national trends in obesity [3]. Moreover, recent data have suggested that individuals with undiagnosed diabetes are similar to those with diagnosed diabetes with regard to the development of complications; thus our estimates are more robust in describing the burden of disease in the population [24].
Comparing our projections with those from other studies, we note that an estimate, published in 2006, for diagnosed diabetes in the US among individuals aged 20 to 64 years in 2030 is 16.8 million [25]. Our estimates are based both on diagnosed and undiagnosed diabetes, and our projection of total diabetes among that age group for 2031 is higher, namely 19 million. It is possible that estimates based solely on diagnosed diabetes could become more consistent with our estimates, if greater vigilance were shown for screening for undiagnosed diabetes. However, not accounting for the at-risk population in the estimates is likely to lead to inaccurate estimates. A comparison of our estimates of total diabetes with those of another study [7], which projected total diabetes but did not account for the population at high risk of developing diabetes, reveals that the latter’s projections are most probably underestimates. Using data from 1993, the investigators projected a population prevalence estimate of total diabetes in the US among individuals aged 20 years and older for the year 2000 to be 7.6%, while the NHANES 1999–2002 yielded a prevalence of 8.8%. For 2025 the same team [7] projected a prevalence of 8.9% versus 13.5% for 2021 in our study.
The results have several implications for the delivery of healthcare and healthcare financing.
First, we estimated our models under several assumptions for the number of individuals at high risk of diabetes in the population. Regardless of these assumptions, the US will have a substantial number of individuals at high risk of diabetes in 2011, 2021 and 2031. Interventions to modify lifestyle are critical to decrease the number of individuals at high risk, and consequently to lower the expected increase in diabetes in the future. Although some of the diabetes estimates suggest seemingly small decreases in future prevalence, based on decreases in the population at risk, the actual numbers are substantial. For example, a one-percentage point drop in the US population estimate of diabetes among individuals aged 20 and older in 2031 is quite substantial and would account for a decrease in prevalence of diabetes equivalent to 2,600,000 people.
Second, the projection that a substantial proportion of the population will have diabetes indicates greater spending will be necessary to manage the disease. This will include spending on drugs, ongoing monitoring, and treating of complications including nephropathy, retinopathy, and cardiovascular disease.
Third, the disproportionate impact of diabetes on minorities, particularly Hispanics, demands new intervention strategies to decrease the number of individuals at high risk and to deliver care to individuals who have historically had poor access to care. Additionally, with the projected increase in diabetes prevalence among 30 to 39-year-olds, a population not currently targeted for screening, a re-examination of current public health policy and screening strategies may be warranted [26].
There are several strengths to the design of this study. One is that the study utilised multiple NHANES data sets, which have the advantage of allowing for nationally representative population estimates. Thus, the initial data used to fit the model as well as to make mortality estimates of diabetes, both diagnosed and undiagnosed, are nationally representative. Another strength is that this study is the first to make a nationally representative assessment of the at-risk population for development of diabetes and then use that assessment to model the future prevalence of diabetes. The assessment of risk used, moreover, is based on the ARIC diabetes risk score [18], a multivariable risk score that used clinical indicators.
When interpreting our results, however, several limitations need to be considered. Thus, although this is the first study to use a validated diabetes risk score to assess the high-risk population for the development of diabetes for the entire US population, potential limitations exist with regard to the diabetes risk score. The ARIC diabetes risk score [18] was based on a cohort of individuals aged 45 to 64 years at baseline and may therefore be limited when estimating diabetes development among individuals aged 20 years and older. However, we estimated diabetes prevalence in 10-year age increments. Moreover, the risk score’s moderate sensitivity and specificity may cause the model to under- or potentially overestimate future prevalence projections. Another possible limitation is that estimates of future disease burden are based on assumptions about the number at risk of disease and about mortality within the population. We have attempted to address this limitation by presenting the results of a sensitivity analysis, which includes variations in the proportion of the population at risk and in mortality. The third limitation is the diagnosis of diabetes in the NHANES data on the basis of a single FPG value. This strategy, although common in epidemiological studies, could potentially underestimate the prevalence of diabetes associated with isolated post-challenge hyperglycaemia, which occurs more commonly in women, the elderly, and in lean populations. It could also overestimate diabetes prevalence, because a clinical diagnosis of diabetes in asymptomatic patients requires two abnormal fasting glucose levels.
In summary, a continued focus on effective interventions for lifestyle modifications to decrease diabetes risk, as well as vigilant ascertainment of diabetes, appears crucial if the future prevalence and burden of diabetes in the US population are to be adequately addressed. This is especially important for minority populations, particularly the Hispanic community, which is projected to have an overwhelming future diabetes burden. Considering that minorities have historically had limited access to healthcare, these findings emphasise the importance of interventions targeting these populations.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Table 1
Projections of non-Hispanic White, non-Hispanic Black and Hispanic persons with diabetesa and prevalence in 2011, 2021, and 2031 (43 kb)
Table 2
Projections, using different assumptions, of diabetes prevalence among individuals aged >20 years in 2031 (32 kb)
|
[
"diabetes",
"projection",
"epidemiology"
] |
[
"P",
"P",
"P"
] |
Skeletal_Radiol-4-1-2257990
|
Distinguishing benign notochordal cell tumors from vertebral chordoma
|
Objective The objective was to characterize imaging findings of benign notochordal cell tumors (BNCTs).
Introduction
The giant notochordal rest and notochordal hamartoma are terms that have been used to describe a benign notochordal cell tumor (BNCT) that may be mistaken for a chordoma [1, 2]. This recently discovered intraosseous benign lesion of notochordal cell origin may be a potential precursor of chordoma [3–6], but based on the few cases with long-term follow-up (those previously published and our present series) none has clinically grown to become a chordoma. The anatomical predilection is identical to that of chordoma. The most common locations are the sacro-coccygeal region and the base of the skull, followed by cervical and lumbar vertebrae. Most tumors are usually small and asymptomatic and are often found at autopsy. The occasional lesion may become large and can be detected on advanced imaging. Our working group collected 7 cases of BNCT that were detected on imaging. We report on their management, outcomes, and histology. Five patients underwent surgical management based on a misdiagnosis of chordoma and failure to recognize the entity of BNCT. This paper focuses on the clinical, radiological, and histological features of these cases, which get overtreated when misdiagnosed.
Materials and methods
The data came from 7 patients with BNCTs that were found clinically. Clinical records included patient sex, age, presenting symptoms, and anatomic site involved. All formalin-fixed specimens were decalcified with formic acid solution and then routinely processed for hematoxylin-eosin and periodic acid Schiff (PAS) staining. For immunohistochemical study, paraffin-embedded specimens were cut into 3- to 5-μm-thick sections and then mounted on glass slides coated with 3-aminopropyltriethoxysilane. The sections were stained by the streptavidin-biotin-peroxidase method (Nichirei, Tokyo, Japan). The antibodies used are listed in Table 1. Microwave antigen retrieval was carried out prior to immunostaining with vimentin, AE1/AE3, CAM 5.2, and cytokeratin (CK) 18. To determine the specificity of immunostaining, we included known positive and negative tissues as controls.
Table 1Antibodies used in the studyAntibodiesSourceClonalityDilutionEpithelial membrane antigenDAKO Cytomation, Glostrup, DenmarkM1:100Cytokeratin (AE1/AE3)DAKO Cytomation, Glostrup, DenmarkM1:100Cytokeratin (CAM5.2)Becton Dickinson, San Jose, CA, USAM1:1Cytokeratin 18DAKO Cytomation, Glostrup, DenmarkM1:20VimentinDAKO Cytomation, Glostrup, DenmarkM1:100S-100 proteinNichirei, Tokyo, JapanP1:200M, monoclonal; P, polyclonal
Results
Table 2 summarizes ages, symptoms, treatment, and outcomes of the 7 patients with BNCTs that were found clinically (Figs. 1, 2). Two lesions were documented as possible vertebral osteonecrosis without collapse; however, recent histological re-evaluation indicated BNCTs [7, 8]. The other cases were originally diagnosed as chordoma histologically. The patient group consisted of 3 men and 4 women ranging in age from 22 to 55 years (mean, 44 years). One patient was found to have two lesions at different sites, evident in the fifth lumbar vertebra and the second sacral vertebra (Fig. 3). The latter was not examined microscopically. A small lesion was found incidentally in another patient in the far distal coccyx and was resected. Of the 9 lesions, 4 were in the cervical vertebra; 2 each in the lumbar vertebra and sacrum; and 1 in the coccyx. The most common symptom was pain in 5 patients; however, 1 patient complained of pain radiating to the fingers and numbness that possibly was caused by known ossification of the posterior longitudinal ligament (OPLL) at the sixth to seventh cervical vertebrae. Another patient was found to have cervical OPLL on radiographs. The pain was usually indolent except for that caused by OPLL. Lesions were found incidentally in 2 patients on MR images when they were examined for unrelated symptoms due to traffic accidents. A needle or open biopsy was done for 5 lesions. Four lesions were diagnosed as chordoma and 1 as osteonecrosis. The remaining 2 lesions were clinically diagnosed as osteonecrosis and an intraosseous benign condition respectively. The patients whose lesions were diagnosed as osteonecrosis and chordoma respectively underwent curettage. Three patients underwent vertebrectomy after the diagnosis of chordoma. One patient underwent wide resection following a preceding curettage. One patient who did not undergo a surgical procedure was followed because the diagnosis of osteonecrosis was made at biopsy. Another patient did not undergo any surgical management in spite of the histological diagnosis of chordoma at biopsy and was followed because the clinician doubted the histological diagnosis of chordoma based on unusual radiological findings for chordoma. The prognosis of all patients but 1 was excellent. Four patients who underwent surgical management were well without recurrent or metastatic disease for follow-up periods between 17 and 84 months (mean 54.5 months). Two patients who did not undergo surgery were alive without progressive disease, as judged by imaging, for 13 months and 44 months after biopsy respectively (Fig. 4). One patient who underwent vertebrectomy died of hemostasis after surgery because she would not accept a blood transfusion on religious grounds.
Table 2Summary of benign notochordal cell tumors found clinicallyPatientAge (years)GenderLocationSymptomsTreatmentOriginal diagnosisFollow-up information130MaleL4Low back painVertebrectomy following biopsy and curettageChordomaAlive with no evidence of disease, 60 months253FemaleC6Numbness and pain of the fingers bilaterally, probably due to OPLL at C6, 7Vertebrectomy following biopsyChordomaAlive with no evidence of disease, 57 months355FemaleC5Neck pain, shoulder stiffnessBiopsyChordomaAlive with no progressive disease, 13 months450MaleC3Car accident injury (incidentally found on MRI)CurettageOsteonecrosisAlive with no evidence of disease, 84 months552MaleC5Mild upper back pain OPLL, focalNeedle biopsyOsteonecrosisAlive with no progressive disease, 44 months622FemaleS4, Cox (microscopic)Buttock painWide resection following curettageChordomaAlive with no evidence of disease, 17 months743FemaleL5, S2Car accident injury (incidentally found on MRI)Vertebrectomy following biopsy at L5ChordomaDied of postoperative complicationsL, lumbar vertebra; C, cervical vertebra; S, sacrum; Cox, coccyx; OPLL, ossification of posterior longitudinal ligamentFig. 1A 30-year-old man who complained of low back pain (patient 1). a Lateral radiograph of the lumbar spine reveals vague sclerosis in the fourth lumbar vertebra. b Axial computed tomography (CT) scan of the fourth lumbar vertebra demonstrates significant sclerosis in the center of the body, partly extending to the cortex. c Sagittal T1-weighted spin echo magnetic resonance (MR) image reveals a large intraosseous lesion with low signal intensity. The normal bone marrow signal is preserved in the anterior and posterior portions of the body. d Sagittal T2-weighted MR image shows slightly bright signal intensity intermingled with intermediate signal. e Sagittal gadolinium-DTPA-enhanced MR image does not demonstrate any enhancement. No soft tissue mass is recognizedFig. 2A 53-year-old woman who was examined for ossification of the posterior longitudinal ligament at the sixth to seventh cervical vertebrae (patient 2). A lesion was found incidentally. a Lateral radiograph reveals faint or vague osteosclerosis in the 6th cervical vertebral body. The physiological lordosis of the cervical spine is decreased. b Axial CT scan reveals diffuse sclerosis within the vertebral body. c Sagittal T1-weighted spin echo MR image reveals low signal intensity in almost the entire marrow space of the sixth cervical vertebra. d Sagittal T2-weighted MR image reveals intensely high signal replacing the entire marrow space. No soft tissue mass is evidentFig. 3A 43-year-old woman who was found to have two separate abnormalities in the lower spine during an imaging study for a traffic accident (patient 7). a Lateral radiograph of the lumbar spine and sacrum reveals intense sclerosis of the entire fifth lumbar vertebral body and mild sclerosis in the cephalad portion of the sacrum. b Sagittal T1-weighted MR image reveals homogeneous low signal intensity in both the fifth lumbar and second sacral vertebraeFig. 4A 52-year-old man who complained of mild upper back pain (patient 5). a Sagittal T2-weighted MR image reveals an intraosseous lesion with high signal intensity in the fifth cervical vertebra at initial presentation. b Sagittal T2-weighted MR image demonstrates no progressive disease 14 months after needle biopsy. No extraosseous tumor extension or enlargement is recognized
Imaging findings
Clinical images reviewed included conventional radiographs (n = 9), bone scintigrams (n = 2), CT scan (n = 7), and MR imaging (n = 8). Conventional radiographs revealed ill-defined, vague sclerosis within the vertebral body in 5 lesions (Figs. 1, 2) and diffuse prominent sclerosis presenting as ivory vertebra in 1 (Fig. 3). The remaining 3 lesions were not identified on radiographs. In all 4 lesions involving the cervical spine, physiological lordosis was decreased. Bone scintigrams did not reveal any abnormal uptake.
All lesions were sclerotic on CT. One was mildly sclerotic in the center of the vertebral body. The others were moderately to markedly sclerotic (Figs. 1, 2). Among them, 5 lesions that were located in the central part of the vertebral body extended partly to the cortex. Another lesion had diffuse sclerosis replacing the entire bone marrow space. No cortical disruption or bone destruction was evident in any of the lesions.
T1-weighted MR images revealed an intraosseous lesion with homogeneous low signal intensity (Figs. 1–3). T2-weighted MR images revealed homogeneous intermediate to bright signal intensity. Gadolinium-DTPA-enhanced T1-weighted MR images were available for 3 lesions and did not show any significantly increased intensity (Fig. 1). Three lesions occupied the entire marrow space and the other 5 lesions were surrounded by residual bone marrow tissue. No extraosseous disease was demonstrated.
Histological findings
Eight lesions were histologically examined. All lesions were located within the vertebral bodies. The lesions were unencapsulated and composed of solid sheets of adipocyte-like vacuolated chordoid cells mixed with eosinophilic, less vacuolated cells of various degrees (Fig. 5a). The adipocyte-like cells had clear cytoplasm and eccentrically or centrally located round or polygonal nuclei (Fig. 5b). Some tumor cells were multi-vacuolated and their nuclei were mildly polymorphic or hyperchromatic (Fig. 5c). No mitotic figures were recognized in any of the lesions. No myxoid background was evident. In some cases, cystic spaces were found in the lesions that contained eosinophilic colloid-like material. The lesions lacked any apparent vascular network. Islands of hematopoietic bone marrow were seen within the lesions. The entrapped bone trabeculae were sclerotic because of appositional or reactive new bone formation (Fig. 5a,b). PAS staining demonstrated intracytoplasmic glycogen granules. In all 7 lesions examined immunohistochemically, the tumor cells stained intensely positive for vimentin, S-100 protein, EMA, CAM5.2, AE1/AE3, and CK18 (Fig. 5d,e). The appearance in each of the cases was consistent with benign notochordal cell tumor [4, 5]. There were no features to support the diagnosis of chordoma. Patient 7 was presented at the Members Meeting of the International Skeletal Society in 1999.
Fig. 5Photomicrographs of benign notochordal cell tumors. a Low power magnification (patient 1) reveals solid sheets of adipocyte-like vacuolated cells. The affected bone trabeculae are sclerotic. Some islands of non-neoplastic hematopoietic bone marrow are seen in the lesion (hematoxylin-eosin [HE] stain). b Higher power magnification (patient 1) reveals a solid sheet of vacuolated tumor cells of various sizes. The nuclei are oval and appear bland (HE stain). c Some tumor cells (patient 1) are multivacuolated and centrally located nuclei are mildly polymorphic. They are reminiscent of lipoblasts (HE stain). d Biopsy specimen (patient 5) reveals sheets of vacuolated cells with pyknotic nuclei between bone trabeculae. The tumor cells may be mistaken for degenerative fatty marrow (HE stain). e The tumor cells (patient 5) stain immunohistochemically positive for cytokeratin (AE1/AE3) indicating notochordal cell origin (SAB immunohistochemical stain)
Discussion
Six lesions consistent with BNCT have been documented previously by others [1, 2, 9–11]. Darby and coauthors [1] first described a case in 1999, which was surgically excised. A 39-year-old man who complained of a 7-week history of persistent low back pain was found to have an intraosseous lesion replacing almost the entire fifth lumbar vertebra. Radiographs and scintigraphy appeared normal. The lesion was well visualized by MRI, but showed only mild sclerosis on a CT scan. Histological findings were identical to those of our cases presented here. They were not typical for chordoma; however, the authors suspected that the lesion might be an intraosseous chordoma rather than a giant notochordal rest. Mirra and Brien [2] reported 2 cases under the name of giant notochordal hamartoma of intraosseous origin. A 42-year-old woman complained of low back pain after falling down the stairs and was found to have an intraosseous lesion involving almost the entire sixth thoracic vertebral body on MRI. The other case was a 41-year-old man who had suffered from intermittent stiffness and mild pain in the neck. MR images demonstrated an intraosseous lesion involving about 98% of the fifth cervical vertebral body. Neither of the lesions was visible on radiographs. CT scans revealed osteosclerosis in the center of the affected vertebral bodies. The lesions showed low signal intensity on T1-weighted MR images and intermediate to high signal intensity on T2-weighted images and no extraosseous disease. No abnormal uptake was detected on bone scintigrams in either case. The patients did not undergo a surgical procedure after the biopsies. The follow-up examinations after 10 years and 38 months respectively demonstrated no progressive disease and indicated their benign nature. Kyriakos et al. [9] reported a case that involved the fourth lumbar vertebra. An 11-year-old boy with Tourette’s syndrome had a 2.5-year history of repeated low back pain after a playground injury. Radiographs, CT scan, and single photon emission computed tomographic bone scan (SPECT) were normal. MRI revealed an intraosseous lesion with low T1- and high T2-weighted signal intensities within the fourth lumbar vertebral body. Cell block material obtained by a fine needle aspiration biopsy led to the diagnosis of chordoma, and the patient underwent L4 vertebrectomy. On subsequent evaluation, they concluded that the lesion was likely a giant notochordal rest rather than chordoma. The most recently documented case reported by Chauvel et al. [10] was an intraosseous lesion in the fifth lumbar vertebral body. The patient was a 45-year-old woman who had complained of low back pain for 2 years. Radiographs and bone scan were normal; however, CT scan revealed fine sclerosis in the L5 vertebra. MRI demonstrated an intraosseous lesion with homogenous low signal on T1-weighted image and high on T2-weighted image. The patient underwent L5 vertebrectomy because the lesion was histologically diagnosed as a chordoma on biopsy. The histological features from the L5 vertebra were similar to those of previously documented cases reported by Darby et al. [1], Mirra and Brien [2], and Kyriakos et al. [9]. A further case was illustrated in the textbook by Dorfman and Czerniak [11]. The lesion entirely occupied the third lumbar vertebra.
Recent studies have discovered unique benign notochordal cell lesions within axial bone that may be precursors of chordoma [3–6, 12]. The new term “benign notochordal cell tumor” was proposed for these lesions [4, 5]. Careful autopsy examination can identify the lesions in approximately 20% of adult cadavers [4, 5]. They are usually tiny or small and found most frequently in both ends of the axial skeleton followed by the mobile spine. Their histological features are different from those of classic chordoma or notochordal vestiges in fetal intervertebral disks (Table 3). BNCTs are characterized histologically by intraosseous sheets of adipocyte-like vacuolated chordoid cells intermingled with less vacuolated eosinophilic cells of various degrees. The lesions were not lobulated. Some colloid-like material containing cystic spaces are seen in the lesions; however, the lesions lack any myxoid background. The nuclei are usually bland, but sometimes appear mildly atypical. No mitotic figures are recognized. The affected bone trabeculae are sclerotic because of appositional or reactive new bone formation. These histological features are completely identical to those of the lesions reported by Darby et al. [1], Mirra and Brien [2], Kyriakos et al. [9], Chauvel et al. [10], and Dorfman and Czerniak [11]. These reported cases are considered by us to represent BNCTs that grew and replaced almost the entire vertebrae.
Table 3Imaging and microscopic features of benign notochordal cell tumorsEvaluation method FeaturesRadiograph (n = 9)Vague sclerosis (n = 5)Marked sclerosis (n = 1)Invisible (n = 3)Bone scintigram (n = 2)No abnormal uptake (n = 2)CT scan (n = 7)Osteosclerosis in vertebral body (n = 7)No bone destruction or cortical disruption (n = 7)MRI (n = 8)T1-WI (n = 8): low signal intensityT2-WI (n = 8): intermediate–high signal intensityGd-DTPA T1-WI (n = 3): no enhancementNo invasive soft tissue mass (n = 8)Histology (n = 8)Solid sheets of adipocyte-like vacuolated cells combined with less vacuolated eosinophilic cells of various degreesRound or polygonal pyknotic nuclei occasionally with polymorphismNo lobular configurationNo extracellular myxoid matrixSome cystic spaces containing eosinophilic colloid-like materialNo mitotic figuresPoor vascular networkEntrapped islands of hematopoietic bone marrowSclerotic bone trabeculae affectedPositive immunostaining for vimentin, cytokeratin (AE1/AE3, CK18), EMA, and S-100 proteinCT, computed tomography; MR, magnetic resonance; WI, weighted image; Gd; gadolinium; CK, cytokeratin; EMa, epithelial membrane antigen.
Benign notochordal cell tumors are usually asymptomatic or indolent. Most lesions are overlooked, even at autopsy, because of their anatomical location or size. Only one such lesion had been documented with a name of ecchordosis physaliphora vertebralis in 1982 [13]. Enlarged BNCTs may be detected on CT or MR images, or rarely by conventional radiography. The imaging findings of our cases are somewhat variable, but distinctive in the absence of trabecular destruction on CT or of a soft tissue mass on either CT or MR images (Table 3). Gadolinium enhancement was absent in the 3 patients in whom it was administered. Radiographs may reveal ill-defined, vague sclerosis in the vertebral body and occasionally show diffuse sclerosis presenting as an ivory vertebra (Fig. 3). Often the lesions are invisible on radiographs. CT scans may reveal unequivocal sclerosis, which is much more obvious than on radiographs. No bone destruction or cortical disruption is recognized. T1-weighted MR images reveal relatively well-demarcated intraosseous lesions with homogeneous low signal intensity. T2-weighted MR images show homogeneous high signal intensity. Gadolinium-DTPA enhanced T1-weighted MR images do not show enhancement. Bone scintigraphy did not show any abnormal uptake.
Diagnosis of the cases presented was based upon biopsy or preoperative imaging; the diagnosis was chordoma in 5 cases and osteonecrosis in 2. Hence, 5 patients underwent vertebrectomy, wide resection, or curettage. All patients but 1 are well without recurrent or metastatic disease. Two patients who underwent only biopsy had no progressive disease, as judged by radiography, CT, or MR images, for 13 months and 44 months respectively. As described by Mirra and Brien, the follow-up information was consistent with the benign nature of the tumors.
The terminology of the lesions is controversial. Dorfman and Czerniak [11], Kyriakos et al. [9] and Chauvel and coauthors [10] used the term giant notochordal rest. Mirra and Brien [2] proposed the term giant notochordal hamartoma of intraosseous origin. In addition, the International Skeletal Society has used the term benign chordoma on the brochure of the Pathology Refresher Course since 2003. However, we prefer the term “BNCT” to the others. BNCTs show histological and immunohistochemical features that are different from those of notochordal vestiges in fetal intervertebral disks. BNCTs lack any myxoid background and show positive immunoreactivity for cytokeratin 18 although notochordal vestiges in the fetal intervertebral disks are associated with myxoid matrix and are negative for cytokeratin 18. Notochordal vestiges are often found in the intervertebral disks of neonates and finally disappear by the age of 1 to 3 years. In our personal autopsy experience, we have never seen any notochordal tissue in the vertebral bodies in fetuses and babies who had notochordal vestiges in their intervertebral disks. In contrast, BNCTs are found in 20% of the axial skeleton dissected from adult cadavers. Therefore, BNCTs are considered to develop after birth. The term of hamartoma is not proper for the lesions, as mentioned by Kyriakos et al. [9] and one of the present authors [3]. The term of benign chordoma is oxymoronic because “chordoma” has been used to describe a malignant notochordal cell tumor [14].
In 15 clinically identified BNCTs, including our cases and those reported by others, 6 tumors were found in the lumbar spine, 5 in the cervical spine, 2 in the sacrum, and 1 each in the thoracic spine and coccyx. This anatomical distribution is different from that found at autopsy [2]. We believe the reason for this discrepancy is that the cervical and lumbar regions often are examined with MRI for neck pain and low back pain. Lesions in the sacrum and coccyx may be overlooked because they are less frequently imaged.
The most important disease that should be distinguished from BNCT is chordoma. Microscopically, chordoma is composed of cords or strands of atypical chordoid cells with a myxoid matrix of various degrees [11, 15, 16]. Chordomas are associated with a large soft tissue component. However, differential diagnosis may be difficult because cellular atypia of BNCT and chordoma may overlap. They also share the same immunohistochemical profile [3]. Imaging is critical in separating the benign lesion from the chordoma. Chordomas are osteolytic tumors. In contrast, BNCTs may show osteosclerotic reaction and do not exhibit osteolytic change. Chordomas are not intracompartmental lesions and are associated with a soft tissue mass. In those cases with histological overlap, with our present understanding of these diseases, we would suggest that the absence of a soft tissue mass should favor a BNCT and the presence of a soft tissue mass should indicate a chordoma.
Biopsy specimens of BNCTs may be easily mistaken for fatty marrow because of their morphological similarity to that tissue. Indeed, 2 lesions among our cases were overlooked as normal fat marrow on biopsy sections. Careful observation can distinguish BNCTs from fatty marrow. Furthermore, immunohistochemical study can help to distinguish BNCTs from fatty marrow. BNCTs stain positive for epithelial markers whereas fat cells do not (Fig. 5). Differential diagnosis also includes metastatic carcinoma, particularly originating from clear cell carcinoma of the kidney. Metastatic clear cell carcinoma usually is osteolytic on radiographs and CT scans. Gadolinium-DTPA enhanced T1-weighted MR images have increased signal intensity. Histologically, clear cell carcinoma has an alveolar pattern and vascular fibrous septa. Clear cell carcinoma rarely exhibits positive immunoreaction for S-100 protein [17].
Large BNCTs may be found during routine clinical examination. They may have the potential of malignant transformation to chordoma; however, the transformation, as we know from the collected cases, is a rare happening [6]. We, however, are not aware of any cases of what we term BNCT growing beyond the confines of bone and displaying features of chordoma, and therefore they should be managed as benign lesions and carefully followed by thin section CT and MRI for signs of extraosseous disease. More extensive follow-up by cross-sectional imaging over a long period of time would be required before firmly determining the biological behavior of BNCTs. BNCTs do not require any surgical management until they undergo malignant transformation. BNCTs should be recognized by radiologists, pathologists, and orthopedic surgeons to prevent unnecessary radical surgery and carefully followed by advanced imaging.
|
[
"benign notochordal cell tumor",
"chordoma",
"notochordal rest",
"notochordal hamartoma",
"vertebra",
"mri",
"spine"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Ann_Surg_Oncol-4-1-2190341
|
Second-Look Operation for Unresectable Pancreatic Ductal Adenocarcinoma at a High-Volume Center
|
Background The value of re-exploration for pancreatic ductal adenocarcinoma after the initial diagnosis of unresectability is unclear.
Surgical resection is the only curative approach available for patients suffering from pancreatic cancer.1 However, resection rates remain low, not only because advanced disease is present in the majority of patients, but also because nonspecialized centers often have insufficient experience with radical surgery.2–4 Resection rates in tertiary referral centers are increasing due to technical progress, improved perioperative management, increasing emphasis on a more-radical approach, such as resection of tumor-infiltrated portal vein, and improved experience in judging resectability.3,5–8 This holds particularly true for evaluating vascular infiltration, which may preclude resection. In these cases, extensive mobilization and careful dissection are necessary to distinguish accompanying inflammatory adhesions from tumor infiltration of retroperitoneal vessels. However, determination of resectability remains challenging, and thus in selected cases less-experienced surgeons are potentially more prone to mistakenly regard a tumor as unresectable.
Since randomized trials on the value of neoadjuvant therapy have not yet been published, it also remains unclear whether—particularly in younger patients—neoadjuvant treatment (or reoperation after initially palliative chemotherapy) may be beneficial, at least in some patients, and may thus lead to resectability of an initially unresectable tumor.
Reoperation of patients who were deemed unresectable may be a further option to increase resection rates.9–18 The value of re-exploration has only been analyzed by a few groups: 10 studies have addressed this question, with the first study concluding that an “appreciable salvage rate” is feasible on re-evaluation.9 Subsequent studies confirmed these results and demonstrated that a second-look operation can be performed with low morbidity and mortality rates, and that this approach confers a survival benefit for the then resectable patients.10–18 Calculations of statistically significant differences in survival of resected and unresected patients at reoperation are hampered by insufficient power of the individual studies due to small patient groups. However, there is a general trend showing that re-exploration may be an option in carefully selected patients at specialized centers.
In this study, we retrospectively analyzed our single-center experience with a group of patients who had initially been diagnosed as unresectable and who were subsequently referred to us for re-exploration. In addition, a literature review of the available (retrospective) studies was performed to more precisely define the value of re-exploration.
PATIENTS AND METHODS
From our database we identified 33 patients with ductal adenocarcinoma of the pancreas who had initially been diagnosed as unresectable (at other institutions) but were re-explored in the course of their disease. Patient characteristics, initial procedures, subsequent palliative or neoadjuvant therapies, operative details of the re-exploration, morbidity, and mortality (death within 30 days after surgery) were prospectively recorded. Patients were followed up until February 2007. At the time of analysis, two out of 14 patients in the group of unresectable patients and 11 out of 19 patients in the resection group were alive.
During the second-look operation, resectability was defined using previously reported criteria.1 If resection was impossible, a tumor biopsy was taken from those patients in whom the initially performed palliative operation was sufficient. In the case of an obstruction, a gastroenterostomy, hepaticojejunostomy or a double bypass procedure was performed. In resectable patients, a pylorus-preserving pancreaticoduodenectomy (ppPD) was the standard operation. However, if the tumor involved the pyloric region, a classical PD was carried out. To achieve tumor-free resection margins, a total pancreatectomy had to be performed in two cases. A distal pancreatectomy was performed for one tumor of the tail of the pancreas.
Pancreatic tumors other than ductal pancreatic adenocarcinoma were excluded. The specimens from resectable (at reoperation) patients were classified according to the tumor, node, metastasis (TNM) classification.19
For the preoperative imaging of operable patients, computed tomography of the abdomen or magnetic resonance imaging/magnetic resonance cholangiopancreaticography (MRI/MRCP) was the standard.
Statistical Analysis
The Kaplan–Meier method was used for presentation of the survival curve, and differences in survival were evaluated with the log-rank test. A P value of < 0.05 was regarded as statistically significant.
Literature Review
A literature search was performed independently by CWM, JK, and JB as described previously.20 The last search was carried out on 30 January 2007. Available manuscripts were cross-searched to identify studies that had not initially been found. The search strategy revealed 10 retrospective studies on re-exploration after initial diagnosis of unresectability. No stratification was performed because only retrospective analyses have been published. Data collection was conducted as previously described.20
RESULTS
In this study, 33 patients (median age 60 years) from our prospectively maintained database who had been re-explored after having initially been defined as unresectable (operation between 2001 and 2005) were included. At the time of the initial operation, an exploratory laparotomy was performed in 51% of the patients, while in the other 49% a bypass operation or a biopsy was carried out (Table 1). The main reason for the definition of unresectability at the initial operation was vascular infiltration (six arterial infiltrations, nine venous infiltrations). In some patients, metastatic disease was present (Table 2). After the first operation, most patients received palliative/neoadjuvant therapy. The types of treatment varied considerably: 13 patients received chemoradiation, 12 patients were treated with chemotherapy, and eight patients received no therapy. There were no standardized treatment protocols for palliative/neoadjuvant therapy, but patients were treated according to local guidelines at the institution where the initial diagnosis had been made.
Table 1.Patient characteristicsUnresectable patients (n = 15)Resectable patients (n = 18)Age60 (41–68)61 (33–71)Female34Male1214Initial operation Explorative laparotomy48 Gastroenterostomy50 Double bypass35 Lymph node biopsy10 Resection of abdominal wall tumor10 Laparoscopy02 Hepaticojejunostomy13Interval between surgeries (median)88 days101 daysPreoperative tumor markers CA19-9 (median)608 U/l117 U/l CEA (median)39143 U/l136 U/lDiabetes mellitus36Neoadjuvant therapy1213Table 2.Criteria for initial inoperabilityUnresectable patients (n = 15)Resectable patients (n = 18)Peritoneal metastasis40Vascular infiltration69Liver metastasis31Duodenal infiltration11Lymph node infiltration14Other/unknown03
Upon referral to our department, all patients presented in adequate general condition (ASA score ≤ 3; American Society of Anesthesiologists). At the second operation, 18 patients were found to have a resectable tumor, while in 15 patients the primary tumor could not be resected. The median time from initial surgery to re-exploration was 88 days in the unresectable group and 101 days in the group of patients in whom a resection could be performed. Except for one patient with a large tumor infiltrating the retroperitoneal vessels, metastatic disease (not detectable in the abdominal CT scan) was present in all the unresectable patients, precluding a surgical resection (Table 4).
In the group of patients who were found to be resectable, 15 pancreaticoduodenectomies, two total pancreatectomies and one distal pancreatectomy were performed (Table 3). Portal vein/superior mesenteric vein (SMV) resections had to be carried out in four patients. There was one patient with a liver metastasis at the initial operation (Table 2) who was resected at the second-look surgery. The patient had no liver metastases at the second look and was therefore considered resectable. This patient is alive at 1241 days after the initial surgery. Four out of five patients without neoadjuvant treatment who were resected (Table 1) received adjuvant chemotherapy. One patient refused adjuvant treatment.
Table 3.Surgical procedures at second operation, morbidity and mortalityUnresectable patients (n = 15)Resectable patients (n = 18)Exploration with biopsy7–Double bypass3–Gastroenterostomy3–Hepaticojejunostomy2–Pancreaticoduodenectomy–5Pylorus-preserving pancreaticoduodenectomy–10Total pancreatectomy–2Left resection–1Duration of operation (min; median)105440Blood loss (ml; median)100500Morbidity23Delayed gastric emptying10Bilioma10Wound dehiscence01Cholangitis01Lymph fistula01Mortality01Length of stay (days; median)1112Table 4.Criteria for unresectability at reoperationUnresectable patients (n = 15)Peritoneal metastasis6Liver metastasis5Omental metastasis1Peritoneal and liver metastasis2Retroperitoneal infiltration1Table 5.Review of studies on re-exploration for pancreatic cancerAuthorYearInstitutionPatientsResection rateMoosa1979University of Chicago2417/24 (71%)Jones1985University of Toronto50N/AHashimi1989Bradford Royal Infirmary2611/26 (42%)McGuire1991Johns Hopkins Medical Institutions5533/55 (60%)Tyler1994M.D. Anderson Cancer Center1914/19 (74%)Robinson1996M.D. Anderson Cancer Center2929/29 (100%)Johnstone1996Naval Medical Center, San Diego2916/29 (55%)Sohn1999Johns Hopkins Medical Institutions7852/78 (67%)Chao2000Fox Chase Cancer Center4022/40 (55%)Shukla2005Tata Memorial Hospital, India1515/15 (100%)This series2007University of Heidelberg3318/33 (55%)
In the group of unresectable patients, bypass procedures were carried out in eight patients, while seven patients underwent exploration with tumor biopsy (Table 3). The postoperative course of the patients was mostly uneventful, with two morbidities in the inoperable group and five in the resected patients (Table 3). One patient who was resected died in the immediate postoperative period (mortality 1/18). The duration of the operation, amount of blood loss, and length of stay were increased in the resection group (Table 3).
Pathology reporting revealed that a T3 tumor was present in most of the cases. Only one patient was found with a T2 tumor. Nine pancreatic specimens (50%) were node positive, with an average of 21.5 harvested lymph nodes. A grade 1 tumor was found in only two patients, whereas a grade 2 tumor was present in nine patients and a grade 3 pancreatic cancer was present in seven patients.
The Kaplan–Meier survival curves revealed a median survival of 439 days for unresectable patients versus 934 days for resected patients (after the second operation). A log-rank test demonstrated that this difference was statistically significant (P = 0.013; data not shown). To better judge the influence of neoadjuvant/palliative treatment on survival, a Kaplan–Meier curve was calculated for the survival after the initial operation. This analysis showed a significantly different median survival of 547 versus 1078 days for unresectable versus resectable patients (P = 0.018; Fig. 1).
FIG. 1.Kaplan–Meier survival curve of re-explored patients after initial operation. Comparison of the survival curves of re-explored, resected patients (n = 17; one excluded due to postoperative mortality; red) and re-explored, unresectable patients (n = 15; black). A log-rank test demonstrated significantly increased survival in the resected patients (P = 0.018).
To judge whether resection at reoperation influences the outcome, we compared re-resected (18) and non-re-resected (15) patients with a cohort of 554 patients who were resected at the initial operation at our center and with a cohort of 250 patients who could not be resected. The Kaplan–Meier survival curve followed by a log-rank test revealed no differences in survival (P = 0.2514; Fig. 2) when comparing initially resected patients and patients resected at the second operation. Furthermore, there were also no differences in comparison of the unresectable patients, irrespective of the second surgery (P = 0.45; Fig. 3).
FIG 2.Survival curves of re-explored, resected patients and 572 primarily resected pancreatic ductal adenocarcinoma (PDAC) patients. Comparison of the survival curves of re-explored, resected patients (n = 18; red) and a control cohort of 572 patients (black) who were initially resected revealed no differences regarding survival (P = 0.2514).FIG. 3.Overall survival of re-explored, unresectable patients and 256 unresectable patients. The survival of a control cohort of 256 PDAC patients who were unresectable (n = 265) was compared with re-explored, unresectable patients (n = 15; red), revealing no differences in survival (P = 0.45).
Subsequently, we performed a retrospective review of the causes of initial unresectability. This revealed that only 11 of 33 patients had initially been treated at a high-volume center (defined as hospitals with a case load of >16 pancreatic resections per year7). Except for four cases with metastasized disease, infiltration of local structures was the main reason for initial unresectability. A detailed analysis of these causes of unresectability showed that 14 of 33 patients would probably have been resected if they had initially been referred to our center. In these cases, mainly venous (SMV/PV) infiltration was present, which may not be considered a contraindication for a resection.
Review of Retrospective Studies
Ten studies on the value of re-exploration to potentially achieve resection have been published so far.9–18 In the first study, by Moosa,9 17 patients were re-evaluated after having initially been diagnosed as unresectable. Of these, nine patients underwent total pancreatectomy, while in two patients no tumor was detectable. The survival of the resected cohort was 1.5–6 years, which was higher than the mean survival in the group of patients with inoperable disease (7–12 months). Similar results were reported by Hashimi et al., who performed a second-look operation on 30 patients, revealing that nine were resectable, while in two patients no tumor was found despite several biopsies.11 In the study by Jones and co-authors, preliminary surgery (nonresective) had been carried out before referral in 51% of the patients, leading to the correct diagnosis in only 44% of these.10 However, outcomes for this patient cohort are not specifically described. Two publications from the Johns Hopkins Institutions also extensively describe results obtained in re-exploratory surgery of two patient cohorts (recruited at different time periods: from 1979 to 1990 and from 1991 to 1997). In the first publication,12 the resection rate for pancreatic cancer at re-exploration was 58%, with subsequent mean survival periods of 20.5 months (pancreatic adenocarcinoma) and 33 months (nonpancreatic periampullary malignancies), which were comparable to the cohort of initially resectable patients. Sohn et al. subsequently described results obtained from re-exploration of a second cohort of patients that included all periampullary tumors.16 The resection rate at reoperation was 67%. There was a striking survival difference when comparing the unresectable and the resectable groups (at re-exploration), with a median survival of 7 versus 23 months, respectively. Comparison of the patients who were resected at re-exploration with a control group who had initially been resected revealed significantly increased survival rates in the reoperative group (Kaplan–Meier survival curve generated from the time of initial operation: 20 months in the control group versus 33 months in the reoperative group, P = 0.02). In the study by Tyler et al. from 1994, reoperation rates were comparable with the earlier published studies (14 out of 19 patients who underwent re-laparotomy were resected), as were the (low) complication rates (three out 14 resected patients).13 At the time of median follow-up of 26 months, four of ten patients with pancreatic adenocarcinoma were alive, suggesting survival rates comparable to previously published results. Similarly, Johnstone and Sindelar showed a resection rate of 55% at re-exploration and significantly improved disease-specific survival rates in the resected group (P < 0.01) 14. Robinson and co-workers demonstrated an exceptionally high resection rate at reoperation of 100%.15 In a study by Chao and co-workers, a resection rate at the second look operation of 55% was shown.18 Furthermore, they could demonstrate that resection at the second operation led to “a median survival comparable with that of patients who did not undergo previous exploration”. In the most recently published study, by Shukla et al., all patients who were reoperated could be resected.17 Confirming the previously published studies, low morbidity and mortality rates were demonstrated.
DISCUSSION
This single-center experience shows that re-exploration of patients initially deemed unresectable may be an option for the treatment of pancreatic cancer. In concordance with previously published studies (which were all retrospectively conducted), re-exploration can yield resectability rates greater than 50% and can be performed with low morbidity and mortality. However, whether reoperation confers a survival benefit is still a matter of discussion.
Although diagnosis of resectability has been facilitated by the use of multidetector, high-resolution computed-tomography exploration often remains necessary.21,22 Particularly when neoadjuvant therapy has been applied, fibrosis and an inflammatory reaction frequently hamper exact assessment of tumor size and (potential) metastatic disease. A number of studies have demonstrated that both the complication rate and mortality after pancreatic surgery are significantly higher in hospitals that perform few operations.3,7,8 We found that, upon re-exploration, more than half of the tumors were resectable, and that 42% of the patients initially deemed unresectable could have been resected at high-volume centers. These results support the concept that patients with pancreatic cancer should be treated with an interdisciplinary approach at large tertiary referral centers.
Furthermore, resection in a second-look operation was shown to potentially confer a survival benefit, although it remains unclear whether this is due to resection or to favorable tumor biology in the patients in whom a regression of the initially metastasized or infiltrating tumor was achieved. Altogether, there was significantly increased survival in the reoperated group of patients compared with a control cohort of patients who were not reoperated (resected and unresectable). These results suggest that there might be a generally favorable tumor biology in this group of reoperated patients. However, the differences in survival conferred by resection at reoperation confirm the notion that in a subgroup of patients in good condition, re-laparotomy (and potentially resection) should be attempted. Furthermore, reoperated but not resected patients do not fare worse than those who were not resected initially and who were not explored again after chemotherapy or chemoradiation. Therefore, we again conclude that in a selected cohort of patients with a tendency towards remission, re-laparotomy may be a valid option. This holds particularly true when patients have initially been considered unresectable at smaller hospitals with less experience in pancreatic surgery and especially with less experience in judging whether a tumor is resectable. This is supported by many studies on reoperative pancreaticoduodenectomy for pancreatic cancer.12,13,15–17 As stated by Robinson and co-authors, “detailed preoperative imaging and a clearly defined operative plan would have allowed successful resection at the initial operation” in the majority of patients. However, accompanying inflammatory changes allow judgment of resectability only during the surgical procedure itself in many cases. Therefore, not only preoperative assessment but also the extension of resectability criteria (as present in high-volume centers) is a key factor for improving the outcome in patients with pancreatic cancer. We strongly suggest that all patients suffering from pancreatic cancer should be referred to high-volume centers. Furthermore, patients in generally good condition but locally unresectable disease should receive neoadjuvant chemoradiation, preferably within randomized controlled clinical trials.23 However, we do not consider infiltration of the SMV/portal vein (in contrast to superior mesenteric artery (SMA)/celiac trunk involvement) a contraindication for resection. Following the last cycle of chemoradiation, high-resolution CT will allow judging resectability in many cases and in the absence of metastatic disease, surgical exploration remains the only option to determine resectability.
In conclusion, reoperation for pancreatic cancer after initial classification of unresectability revealed resectability in half of patients. Since the majority had received chemotherapy/chemoradiation after the initial operation, the concept of selecting patients by neoadjuvant therapy may be supported. Furthermore, a subgroup analysis revealed that a large number of patients who were initially deemed unresectable at smaller hospitals would have been resectable at a large tertiary referral center, which again promotes the concept of patient centralization in pancreatic surgery. A survival benefit for the resected patients underlines the efficacy of a surgical resection even in a situation in which the initial findings preclude a potentially curative approach.
|
[
"unresectability",
"pancreatic cancer",
"reoperation",
"second look"
] |
[
"P",
"P",
"P",
"P"
] |
Ann_Biomed_Eng-2-2-1705516
|
An Ex Vivo Study of the Biological Properties of Porcine Aortic Valves in Response to Circumferential Cyclic Stretch
|
Normal physiological mechanical forces cause constant tissue renewal in aortic valve leaflets (AVL) while altered mechanical forces incite changes in their structural and biological properties. The current study aims at characterizing the remodeling properties of AVL subjected to cyclic circumferential stretch in a sterile ex vivo bioreactor. The leaflets cultured were stretched at a maximum rate of 300%s−1 corresponding to a 15% strain for 48 h. Collagen, sulfated glycosaminoglycan (sGAG), and elastin contents of the stretched, fresh, and statically incubated leaflets were measured. Cusp morphology and cell phenotype were also examined. AVLs exposed to cyclic stretch showed a significant increase in collagen content (p < 0.05) when compared to fresh and statically incubated AVLs. sGAG content was significantly reduced in the stretched AVLs (p < 0.05) when compared to the fresh leaflets and was comparable between stretched and statically incubated AVLs. There was no statistically significant change in elastin content in all the three groups of AVLs (p > 0.05). Native aortic valve morphology was well preserved in stretched leaflets. Immunohistochemistry and immunoblotting studies showed an increased expression of α-smooth muscle actin (α-SMA) in stretched leaflets while α-SMA expression was reduced in statically incubated AVLs when compared to the fresh leaflets. To conclude, circumferential cyclic stretch altered the extracellular matrix remodeling activity of valvular cells, and consequently the extracellular matrix composition of the AVLs. Most interestingly, the contractile and fibrotic phenotypic expression of valve interstitial cells was enhanced. These results show that circumferential cyclic stretch is a possible mediator for AVL remodeling activity.
Introduction
Cardiac valves play a crucial role in controlling blood flow through the body.6 Of particular interest in this study is the aortic valve, which at its most basic level controls blood flow between the left ventricle and the systemic circulation. In the United States, aortic valve disease has emerged as the third most common cardiovascular disease,17 with the number of aortic valve replacement surgeries steadily increasing over the past 20 years.33
The aortic valve is subjected to 30–40 million cycles of opening and closing a year, which translates to approximately 3 billion cardiac cycles in a lifetime.26 During each cardiac cycle, the normal functional aortic valve interacts closely with the surrounding hemodynamic environment and is exposed to a myriad of mechanical forces such as transvalvular pressure, axial, shear and bending stresses, and cyclic flexure.26 Normal hemodynamic forces have been shown to cause constant tissue renewal in aortic valves,19 while altered mechanical forces are believed to induce changes in aortic valve biology possibly leading to valvular disease.16,30 However, the cellular and molecular events involved in these processes are not well characterized. Studies indicate that abnormal hemodynamics, especially hypertension, experienced by the valve leaflets cause tissue inflammation, which can lead to calcification, stenosis and ultimately valve failure1,15,23Knowledge of the adaptive mechanisms of the aortic valves to altered hemodynamics will therefore improve our understanding on mechanotransduction of valvular disease progression and valve failure.
Mechanical forces modulate cell physiology, and have been shown to affect the biosynthetic activity of cells in tissue matrices. Previous studies on aortic valve mechanobiology have shown that porcine aortic valves cultured under normal hemodynamic conditions in a sterile ex vivo flow loop undergo constant renewal to maintain their native phenotypes.13 Porcine aortic valve endothelial cells exposed to 20 dyn cm−2 of steady laminar shear stress showed an organization of focal adhesion complexes and a cell alignment different than in static endothelial cell culture.4 Aortic valve leaflets exposed to hypertensive cyclic pressure showed increased collagen and sulfated glycosaminoglycan (sGAG) synthesis which varied in a magnitude and frequency-dependent manner.31 Studies exposing native porcine aortic valve leaflets to the isolated effects of flow, laminar steady shear stress, and constant static and cyclic pressure resulted in a reduced level of expression of α-smooth muscle actin (α-SMA).29,31,32 This last finding might indicate the involvement of other forces in maintaining and regulating the contractile phenotype of aortic valve interstitial cells. To date the effect of cyclic stretch on aortic valves has not been studied.
The mechanical mode of cyclic stretching is important for normal aortic valve functionality as it allows, among other things, the valve leaflets to extend and form a coaptive seal with each other. However, it was seen that the valve leaflet tissue becomes less extensible and loses its ability to stretch rapidly with increasing age.15 This is mainly because collagen fibrillogenesis proceeds to maturity by increasing the diameter of some of the constituent collagen fibrils that require greater force to produce the same leaflet extension.8 Although the valve leaflet undergoes biaxial stretching in vivo, studies that investigate the effects of separate uniaxial (circumferential and radial) as well as biaxial cyclic stretch on the aortic leaflets would provide useful information on valve functionality or failure.
We hypothesize that the regulation of extracellular matrix remodeling as well as the expression of the contractile phenotype of aortic valve cells depend on the loading forces due to cyclic stretch. Here, we examined this hypothesis with a focus on the effect of physiological uniaxial circumferential cyclic stretch on the biosynthetic response of aortic valve leaflets using porcine aortic valve leaflets and an ex vivo tensile stretch bioreactor. Results from the current study will improve our understanding of the biological responses of native valves to circumferential stretching and will provide a good basis of comparison for conducting studies on radial cyclic stretching.
Material and Methods
Tissue Harvest
Fresh porcine aortic valves were obtained from a local abattoir (Holifield Farms, Covington, GA) following on-site dissection of the hearts within 30 min of slaughter. The valves were then transported to the laboratory in sterile, ice-cold Dulbecco’s Phosphate Buffered Saline (DPBS; Sigma, St. Louis, MO).
Upon arrival to the laboratory, the three leaflets were excised from each valve and a rectangular section of tissue was isolated from the center of each leaflet. This tissue section was further divided into three rectangular sections having an aspect ratio of 12 × 5 mm and oriented in the circumferential direction from the basal region. These three sections from each leaflet were randomly selected as fresh control, static control, and stretched leaflets, respectively (Fig. 1). Leaflet sections from the fresh control group were analyzed within 4 h following valve dissection. Leaflet sections from the static control group were statically incubated in Dulbecco’s Modified Eagle Medium (DMEM; Sigma) at 37°C for the duration of the experiment (48 h). Leaflet sections from the stretched group had loops sutured to the 5 mm ends and were suspended in DMEM in a tensile stretch bioreactor.10 These samples were then cyclically stretched at a frequency of 1.167 Hz (equivalent to 70 beats per minute) to 15% strain at a maximum strain rate of 300%s−1 at 37°C. The temporal strain waveform imposed (Fig. 2a) in the circumferential direction was derived from the variations of the leaflet length in the circumferential direction over one cardiac cycle.25,26 The extension of the leaflet corresponded to diastole (2/3 of the cardiac cycle), while the unstretched state of the leaflet corresponded to systole (1/3 of the cardiac cycle). The gradients of the extension and relaxation approximated those experienced in vivo.25 The above waveform resulted in a tissue extension of 1.8 mm at a maximum rate of extension of 36 mm s−1. Each experiment was run for 48 h, with 12–15 leaflet sections for each experimental run.
Figure 1.Preparation of aortic valve leaflets for the experiment. Three sections were cut out in the circumferential direction from the base, belly and tip regions of the aortic valve leaflet. These sections were randomly chosen as stretched, fresh and static leaflet groups. Fresh sections were processed immediately. Static control sections were incubated for 48 h in DMEM. Stretched sections were stretched for 48 h.Figure 2.(a) Loading curve used in this study. A near physiologic loading curve 25 was input to the actuator to stretch the leaflets. (b) Ex vivo tensile stretch bioreactor used in this study. A magnified image of the tissue chamber is shown on the right showing eight tissue wells, with aortic valve leaflets in four of the wells.
Stretch Bioreactor System and Validation
The stretch bioreactor used in this study (Fig. 2b) was developed by Engelmayr et al. 10The device consists of two tissue chambers coupled via a cross-arm to a centrally positioned linear actuator (Ultra Motion LLC, Mattituck, NY). Each chamber consists of eight culture wells within which two “stationary posts” are positioned. The linear motion of the actuator is translated into uniaxial motion of two “moving posts” via the cross-arm which bifurcates to two moving arms. These moving arms then enter each individual bioreactor chamber and terminate at each culture well with the two “moving posts” mentioned above. The interior of the chamber is protected from external contamination by a neoprene-lined lid. The bioreactor was constructed from polysulfone, which is thermally and chemically stable, while the lids were constructed using acrylic. The device was assembled using 18-8 stainless steel screws (McMaster Carr, Cleveland, OH) and fabricated by Apollo Precision Inc. (Plymouth, MN). The motor input waveform (Fig. 2a) was downloaded to the actuator controller card using the Si-Programmer software (Ultra Motion LLC).
After assembly, the bioreactor was validated for uniformity of strains on the leaflet sections in all the culture wells. This was achieved by placing dye markers (Shandon tissue marking dye, Thermo Electron Corp, Pittsburgh, PA) on aortic valve leaflets and capturing the distance between the markers under uniaxial cyclic stretch using a high-speed camera capture technique developed in our laboratory.12The bioreactor was validated at input strains of 15 and 25%.
In order to determine if the stretch bioreactor induced a shear stress on the leaflet surface due to the relative motion of the leaflet with respect to the surrounding media, a computational fluid dynamics (CFD) model was designed, validated and used to predict the leaflet wall shear stress under the imposed loading curve in the stretch bioreactor.
For the experiment, the bioreactor was sterilized using hydrogen peroxide in a gas-plasma sterilizer (Advanced Sterilization Products, Irvine, CA), and assembled aseptically in a sterile laminar flow hood (Fisher Hamilton, Two Rivers, WI). The culture wells were filled with 8 ml of serum-free DMEM supplemented with 50 mg l−1 ascorbic acid, 3.7 g l−1 sodium bicarbonate, 1% (v/v) nonessential amino acid solution, and 1% (v/v) antimycotic-antibiotic solution (all reagents from Sigma). Each tissue section was suspended in the medium by the two sutured loops – one end was looped over the stationary posts while the other end was looped over the moving post. The bioreactor was then placed in an incubator (Fisher Scientific, Hampton, NH) maintained at 37°C, 5% CO2 atmosphere at normal atmospheric pressure and the leaflets were stretched for 48 h.
Tissue Processing
At the end of the experiment, each leaflet section was further cut into three equal rectangular sections. These resulting samples were dehydrated for 48 h at room temperature in a vacuum oven (Yamato Scientific America, San Francisco, CA) prior to digestion in pepsin (Sigma), protease (Sigma) and proteinase K (Sigma), for analyzing collagen, sulfated glycosaminoglycan (sGAG), and elastin contents, respectively. All digestions were carried out in a water bath maintained at 37°C. The total amount of collagen, sGAG and elastin was determined using the SircolTM Collagen assay kit (Biocolor, UK), BlyscanTM sGAG assay kit (Biocolor), and FastinTM Elastin assay kit (Biocolor), respectively. These kits used quantitative dye-binding methods to determine the total quantity of the respective ECM component in the sample.
Valve leaflet samples from additional experiments were isolated for analysis of tissue morphology and cell phenotype. For histological and immunohistochemical analyses, leaflet sections were fixed in 10% neutral buffered formalin (Fisher Scientific, Suwanee, GA) for 24 h, saturated in 70% ethanol, processed in a tissue processor (Shandon Pathcenter enclosed Tissue Processor) in descending grades of ethanol, embedded in paraffin and cut into 5 μm sections.
Morphological Characterization
Routine Hematoxylin and Eosin (H & E) stain was used to examine gross tissue structure and indicate any changes in the structural compositions of the leaflets. H & E staining was done on an autostainer (Leica, Vashaw Scientific, Norcross, GA). Picrosirius red staining was performed to examine collagen fiber structure and morphology. The protocol was as follows. Deparaffinized slides were incubated in Weigert’s hematoxylin (Sigma) followed by picrosirius red solution for one hour. The slides were then washed in acidified water, three changes of 100% ethanol, and xylene before being coverslipped. These slides were viewed at 100× using a Nikon imaging microscope (Nikon Inc., Melville, NY). A Q-imaging camera, Retiga 1300C (I-Cube, Glen Burnie, MD) and its bundled software Q-Capture Pro was used to acquire images of the tissue sections. H & E images were recorded using normal white light while picrosirius red images were recorded with circularly polarized light.
Characterization of Cell Phenotype
α-smooth muscle cell actin (α-SMA) immunohistochemistry (IHC) was used to detect myofibroblasts and smooth muscle cells. The staining provided information on changes in cell phenotype between the three experimental treatments. The protocol was as follows. Deparaffinized slides were blocked using 1% gelatin/phosphate buffered saline (PBS; Sigma) for 30 min. The slides were then incubated in mouse monoclonal anti α-SMA primary antibody (Sigma) in 1% Bovine Serum Albumin (BSA)/PBS (Sigma) for one hour. The sections were saturated in biotinylated horse anti-mouse IgG (Vector laboratories) in 1% BSA/PBS, and 2% normal horse serum (Vector laboratories) for 30 min. Avidin-D Texas red (Vector Laboratories) fluorochrome was applied to the sections after the secondary antibody was washed off. The sections were then counter stained with 0.25 μg ml−1 4′, 6-Diamidino-2-phenylindole (DAPI; Sigma), coverslipped and stored at 4°C. Slides were subsequently imaged at 200× using a TR-FITC-DAPI triple filter under a mercury lamp.
Image Analysis
H & E images were analyzed by measuring the fibrosa: spongiosa length ratio, and the fibrosa: ventricularis length ratio. α-SMA immunopositive cells appear red with cell nuclei counterstained blue, while the color of collagen fibers changes from green to yellow to orange to red as the fiber thickness (and consequently maturity) increases. The proportion of α-SMA positive cells were therefore assessed by using the Image J program (NIH, Bethesda, MD) to determine the relative area coverage of α-SMA positive red staining compared to the area coverage of entire leaflet section. The relative proportion of the different maturity of the collagen fibers in the leaflet section were calculated using published methods.2 Basically, picrosirius red images were first imported into Matlab 7.01 (The Mathworks, Natick, MA) and converted into their hue, saturation and value components. Only the hue components were retained, and a histogram of hue frequency was obtained from the 8-bit hue images that contained 256 possible colors. The colors were defined as follows: red 2–9 and 230–256; orange 10–38; yellow 39–51; and green 52–128. All other hue values corresponded to interstitial space and were confirmed by visual inspection. The number of pixels within each hue range were determined and expressed as a percentage of the total number of collagen pixels and plotted.
Immunoblotting Analysis
The rest of the leaflet sections were processed for analysis by immunoblotting. Leaflets were homogenized and sonicated in ice-cold lysis buffer (in mmol l−1: Tris–HCl 20 (pH 7.5), EGTA 5, NaCl 150, glycerol-phosphate 20, NaF 10, sodium orthovanadate 1, 1% Triton X−100, 0.1% Tween-20 and protease inhibitors) to extract proteins from the tissue, and then centrifuged at 14,000 rpm to pellet cellular debris. The supernatant was assayed for protein concentration using the DC Protein Assay (Bio-Rad, Hercules, CA) and equal aliquots of tissue lysates were resolved on a 10 % SDS-PAGE gel and subsequently blotted to a polyvinyldifluoride (PVDF) membrane (Millipore, Billerica, MA). Membranes were incubated with a mouse anti α-SMA primary antibody (Sigma) at 4°C overnight (1:1000), and then with a goat anti-mouse alkaline phosphatase-conjugated secondary antibody (Bio-Rad). CDP-Star reagent (NE Biolabs, Ipswich, MA) was added and chemiluminescence of the resulting product was detected using photographic paper. GAPDH antibody (Santa Cruz Biotech, Santa Cruz, CA) was used as a loading control (1:5000). Pixel densities of α-SMA immunoreactive bands were then analyzed using the Image J program and normalized by the intensities of the GAPDH immunoreactive bands.
Analysis of Tissue Culture Medium
As a follow-up to determine if sGAGs were leaching out from the leaflet in the DMEM, six samples of the DMEM were analyzed before and after the experiment. The BlyscanTM sGAG assay kit (Biocolor) used in the ECM analysis was used to detect sGAGs in the medium.
Statistical Analysis
The data for collagen, sGAG and elastin contents were expressed as mean plus one standard error of the mean. A total of 27 leaflets were analyzed for the ECM assays, 8 leaflets for the histological and immunohistochemical studies, and 10 leaflets for the Western blotting studies. The quantitative data from the ECM assays, image analysis and immunoblotting were analyzed first using ANOVA followed by pairwise comparisons with the Tukey post-hoc test. A p-value of less than 0.05 was used as a measure of statistical significance. When only two groups were being compared, data were analyzed using student t-tests with a p-value of less than 0.05 indicating statistical significance. All statistical analyses were performed using Minitab (Minitab R14, Minitab Inc. State College, PA).
Results
Validation of Stretch Bioreactor
The stretch bioreactor produced a uniform output strain across the leaflet of 14.80 ± 0.53% for 15% input strain, and 25.48 ± 0.53% for 25% input strain. Performing a student t-test between input and output signals revealed no significant difference (p > 0.05, n = 16), indicating that the bioreactor could accurately and uniformly reproduce the specified input strain across the entire leaflet section.
The peak steady-state wall shear stress experienced by the leaflet as calculated from CFD was 2.5 dyn cm−2. This is small compared to the peak 79 dyn cm−2 experienced by the leaflet under physiological flow conditions during systole as measured in our laboratory,28 and it can be assumed that stretch is the dominant force acting on the leaflet.
Extracellular Matrix Content1
A significant increase in collagen content was observed in stretched leaflets when compared to fresh and static control leaflets (p < 0.05, n = 27). Stretched leaflets showed a 90% and 161.5% increase in collagen content in comparison to fresh and static controls, respectively. There was no significant difference in collagen content between fresh and static control leaflets (p > 0.05, n = 27). sGAG content was found to be decreased by 40% in stretched leaflets when compared to fresh valve leaflets (p < 0.05, n = 27), while sGAG content was comparable between stretched and static control leaflets (p > 0.05, n = 27). ANOVA showed that there was no significant difference (p > 0.05, n = 27) between treatment groups for elastin content. The results for the ECM assays are graphically depicted in Fig. 3.
Figure 3.Cyclic stretch increased collagen, decreased sGAG, while not affecting elastin contents in aortic valve leaflets. The leaflets were exposed to physiologic circumferential cyclic stretch (15% at heart rate of 70 bpm) or static control conditions for 48 h. The amounts of collagen, sGAG and elastin in the leaflets exposed to static or stretching were compared to those of fresh leaflets. The data were normalized by tissue dry weight and expressed as a mean value plus one standard error of the mean (* p < 0.05, n = 27).
Analysis of the conditioned medium at the end of experimental duration revealed a statistically significant increase (p < 0.05, n = 6) in sGAG secreted in to the media (5.83 ± 0.56 μg ml−1 of DMEM) compared to the fresh media (1.08 ± 0.30).
Cell Morphology and Phenotype
H & E staining (Fig. 4a) of the leaflets showed preserved three-layered leaflet architecture in stretched leaflets consistent with fresh and static control leaflets. The cell nuclei were found to be intact with no tears in the tissue section. Image analysis showed no significant difference (p > 0.05, n = 8) in the fibrosa-spongiosa length ratio as well as the fibrosa-ventricularis length ratio (Fig. 5a) indicating that there was no gross damage to leaflets structure.
Figure 4.(a) Cyclic stretch did not damage tissue structure and morphology. H & E stained images of fresh, static and stretched aortic valve leaflets are depicted here. Cytoplasm was stained pink and cell nuclei were stained blue. The representative images show the three-layered morphology of the leaflets was intact in all three groups (F – fibrosa, S – spongiosa, V – ventricularis). (b) Cyclic stretch maintained native collagen architecture. Picrosirius red images of fresh, static and stretched aortic valve leaflets are shown. Mature collagen fibers were stained orange-red. Layered collagen morphology of leaflets was observed and crimp was preserved in stretched leaflets (F – fibrosa). (c) Cyclic stretch increased α-smooth muscle actin (α-SMA)-positive cells in the ventricularis side of the aortic valve leaflets. Fresh, static and stretched leaflets were examined by α-SMA IHC. Actin was stained red and cell nuclei were counterstained blue. Increased expression of α-SMA was observed in stretched leaflets, and α-SMA was reduced in static leaflets (V – ventricularis).Figure 5.(a) Fibrosa-to-spongiosa ratio and fibrosa-to-ventricularis ratio remained unchanged. There was no significant difference (p > 0.05, n = 8) in the relative thicknesses of the fibrosa, spongiosa and ventricularis layers between fresh, static and stretched leaflets. (b) Static leaflets exhibited reduced levels of newly synthesized collagen fibers. Newly synthesized fiber proportion was reduced in static leaflets signifying reduced levels of collagen synthesis. Newly synthesized fiber proportion was comparable between fresh and stretched leaflets, but the proportion of mature fibers was greater in stretched leaflets. (c) Cyclic stretch increases α-smooth muscle actin in the aortic valve leaflets. α-SMA immunopositive staining coverage was increased in stretched leaflets and reduced in static leaflets when compared with fresh leaflets.
Picrosirius red staining (Fig. 4b) of the leaflets showed that collagen architecture was preserved in fresh, static as well as stretched leaflets. The layered structure of collagen along with the crimp in the collagen fibers could clearly be seen in the stretched leaflets, demonstrating that the stretching motion did not damage valvular collagen. These findings demonstrated that uniaxial cyclic stretch of the leaflets did not damage leaflet morphology and maintained native ECM structure and cellular composition.
Additionally, hue analysis of the picrosirius red images (Fig. 5b) revealed that the proportion of newly synthesized (green) fibers in statically incubated leaflet sections was significantly decreased by 36.5% (p < 0.05, n = 8) compared to fresh leaflets. There was no significant difference in green fibers (p > 0.05, n = 8) between fresh and stretched leaflets, however, mature (red) collagen fibers significantly increased by 101.4% in stretched leaflets compared to fresh leaflets.
The immunohistochemical study with α-SMA (Fig. 4c), a marker of myofibroblast-like or smooth muscle cell-like valvular interstitial cells2, revealed that cyclic stretch up-regulated the expression of α-SMA, while static incubation down-regulated the expression of α-SMA. This was corroborated by image analysis (Fig. 5c) of the images, which showed significantly increased (p < 0.05, n = 8) α-SMA immunopositive staining in stretched leaflets compared to fresh leaflets, and significantly reduced (p < 0.05, n = 8) α-SMA staining in static leaflets compared to fresh leaflets. Another interesting observation was that the α-SMA positive cells in the stretched leaflets were predominantly found in the ventricularis region of the aortic valve leaflet while there was little or no α-SMA expression in the fibrosa. In contrast, α-SMA positive cells were randomly distributed in all the three regions of the fresh and static leaflets.
The above finding was further reinforced by the Western blotting results (Fig. 6). There was a 54.5% increase (p < 0.05, n = 10) in α-SMA content of stretched leaflets and 34.9% decrease (p < 0.05, n = 10) in statically incubated leaflets when compared to fresh leaflets. These results together suggest that cyclic stretching induces phenotypic change of valve interstitial cells toward myofibroblast-like cells in a side-dependent manner.
Figure 6.Cyclic stretch increased α-smooth muscle actin in the aortic valve leaflets. Fresh, static and stretched aortic valve leaflets were analyzed by Western blot with α-SMA antibody and analyzed by densitometry. When compared to fresh leaflets, α-SMA expression decreased in static leaflets, but increased in stretched leaflets (*p < 0.05, n = 10). Anti-GAPDH antibody was used as a loading control.
Discussion
Effect of Cyclic Stretch on ECM Components
The current study investigated the effects of isolated circumferential cyclic stretch on the biological responses of aortic valve leaflets using a sterile ex vivo stretch bioreactor. Initial validation of the bioreactor showed that strains were uniform across the leaflet section. It was also validated that there was negligible wall shear stress on the leaflet, and therefore cyclic stretch was the only force acting on the leaflet.
It was revealed that the collagen content of the aortic valve leaflets stretched to 15% was increased when compared to fresh and static control leaflets, while sGAG content was decreased in stretched leaflets compared to fresh leaflets. Furthermore, elastin content was comparable between the stretched and fresh leaflets. This suggests that exposure to isolated effects of circumferential cyclic stretch at physiological levels has altered the extracellular matrix composition, consequently, altering the valve remodeling.
Collagen fibers, densely populated in parallel alignment along the circumferential direction of the leaflet, form the load bearing structure of the leaflet during each cardiac cycle.18,26 The increase in collagen suggests that the leaflets adapt to altered mechanical loading due to the effects of isolated cyclic stretch, by either increasing synthesis, or decreasing degradation of collagen. The image analysis of the picrosirius red staining (Fig. 5b) showed comparable proportions of newly synthesized fibers, and hence collagen synthesis in both fresh and stretched leaflets. The increased proportion of mature (red) fibers in stretched leaflets therefore leads us to believe that it is the reduced degradation of collagen that causes the overall total increase in collagen content. The preserved crimp in collagen fibers in the stretched leaflet, as evidenced by picrosirius red staining also implies that the native structure of collagen was not damaged by the stretching motion.
This data is in agreement with a study conducted by Ku et al.14 where porcine aortic valve interstitial cells were cultured on a collagen substrate and stretched 48 h. For cells subjected to 14% cyclic stretch, collagen incorporation in the cell layer was approximately 170% of unstretched cells. No significant changes were observed at time points less than 48 h. This provides further evidence that 48 h of culture is a necessary and sufficient time point to observe changes in the valvular ECM.
The semi-fluidic nature of the sGAG in the spongiosa layer of the leaflet gives the leaflet considerable plasticity.26 It is therefore believed that the purpose of sGAG might be (i) to help reduce leaflet stresses by lubricating the relative motion of the other two layers as the leaflet deforms, and (ii) to dampen compressive forces during diastole.7 The absence of compressive stresses in this study, with the stretch being carried out at atmospheric pressure, might therefore explain the reduced levels of sGAG observed in the stretched leaflet compared to fresh leaflets. Additionally, analysis of the DMEM at the end of experimental duration showed a statistically significant increase in sGAG content in the media. This suggests that cyclic stretching of the aortic valve leaflets resulted in secreting or leaching of the sGAG from the leaflets into the media, as the various layers of the leaflet rub against each other. A study conducted by Grande-Allen et al.11 also showed a loss of GAG from explanted porcine bioprosthetic valves due to leaching even under normal storage conditions. This suggests that there is a high propensity for sGAGs to be forced out of the native aortic valve either during static incubation or leaflet motion. In contrast, aortic valve leaflets when exposed to normal physiological forces undergo constant sGAG renewal by the interstitial cells.13
In the ventricularis layer of the leaflet, the major extracellular component is elastin in the form of fibers that are oriented primarily along the radial direction. Their primary role of elastin fibers is to maintain a circumferential collagen fiber orientation and recoil the collagen fibers to their initial state after the external load has been released.27 Since the valve leaflets were cyclically stretched along the circumferential direction at physiological rates in the current study, it is possible that the elastin fibers were not “activated” beyond their normal levels and consequently additional synthesis of elastin was not required, explaining why elastin content remained unchanged between fresh and stretched leaflets.
The above findings reinforce the fact that valve cells require mechanical stimulation to maintain a balance between synthesis and degradation of ECM components. This occurred without undue damage to the normal tri-layered structure of the aortic valve leaflet as the relative thicknesses of the fibrosa, spongiosa, and ventricularis were maintained (Fig. 5a).
Effect on Cell Phenotype
Valvular interstitial cells are mainly comprised of fibroblasts, myofibroblasts, and smooth muscle cells that exhibit different phenotypes depending on the microenvironment surrounding the cells. Expression of α-SMA was associated with the contractile phenotype of smooth muscle cells24 and myofibroblast-like cells,2 and hence was chosen as a marker for examining the interstitial cell phenotype in the present study. Studies have shown that the expression of α-SMA was decreased by changes in culture conditions and mechanical and biochemical stimulation in cultured vascular cells.21,22 Studies in our laboratory have shown that α-SMA expression decreased in response to isolated effects of pressure and shear forces.13,29,31 Similar behavior was observed in valvular interstitial cells in 3D static gel culture.5
The current study revealed via IHC (Fig. 5c) and Western blotting (Fig. 6) that circumferential cyclic stretch increases the expression of α-SMA in aortic valve leaflets. It is therefore speculated that cyclic stretch is a regulator of the contractile and fibrotic phenotypes of aortic valve interstitial cells. A reversible equilibrium between F-actin depolymerization and G-actin polymerization exists in the cells of aortic valves. The contractile motion of the leaflets due to stretch may have caused a shift in this equilibrium towards polymerization of G-actin to form F-actin fibrils, which is detected as α-SMA immunopositive cells.
These findings also lead to the possibility that in the normal hemodynamic environment, the up-regulating effect of stretch on α-SMA expression may be counterbalanced by the down-regulating effects on α-SMA expression of pressure and shear stress. This provides an additional possible explanation for the fact that aortic valves cultured in normal hemodynamic forces maintain a constant α-SMA expression.13
Another interesting observation was that α-SMA immunopositive cells occurred preferentially in the ventricularis of the stretched leaflet. This is in contrast to fresh and static leaflets which exhibited an apparent random distribution of α-SMA through all the three layers of the valve leaflet. This may provide insights into specific regions or “hot spots” on the aortic valve leaflet where there is a focal response to mechanical forces. In a study by Simmons et al.,20 spatial heterogeneity of aortic valve endothelial phenotypes was speculated to contribute to the focal susceptibility for calcific aortic valve lesion development, implicating the endothelium as a regulator of valvular calcification. The current study shows the potential focal and side-specific response of the interstitial cells of the aortic valve to mechanical forces. In addition, α-SMA is a known marker for mesenchymal cells,3, 9 and the expression of α-SMA could also indicate regions of endothelial cell to mesenchymal cell transdifferentiation. It is clear that further side-specific study is needed in this area as it may provide insights into how valves sense and transmit mechanical forces.
Conclusion
The current study demonstrated that circumferential cyclic stretch altered the extracellular matrix remodeling activity of native porcine aortic valve leaflets. These results combined with results from previous ex vivo studies show that the changes in aortic valve leaflet biology depend intimately on the influence of local hemodynamic environmental factors. Further research is needed to understand the mechanistic pathways behind these biological changes. The knowledge gleaned from such studies will ultimately allow for elucidation of mechanisms of aortic valve pathogenesis leading to valve failure.
|
[
"aortic valve leaflets",
"circumferential stretch",
"contractile phenotype",
"uniaxial stretch bioreactor"
] |
[
"P",
"P",
"P",
"R"
] |
Virchows_Arch-3-1-2151131
|
MicroRNA—implications for cancer
|
MicroRNAs (miRNAs) are small RNA molecules that regulate gene expression post-transcriptionally. After the discovery of the first miRNA in the roundworm Caenorhabditis elegans, these short regulatory RNAs have been found to be an abundant class of RNAs in plants, animals, and DNA viruses. About 3% of human genes encode for miRNAs, and up to 30% of human protein coding genes may be regulated by miRNAs. MicroRNAs play a key role in diverse biological processes, including development, cell proliferation, differentiation, and apoptosis. Accordingly, altered miRNA expression is likely to contribute to human disease, including cancer. This review will summarize the emerging knowledge of the connections between human miRNA biology and different aspects of carcinogenesis. Various techniques available to investigate miRNAs will also be discussed.
The “small RNA revolution”
Small ribonucleic acid (RNA) can act as a specific regulator of gene expression. This discovery has been an exciting breakthrough in Biological Sciences of the past decade, culminating in last year’s Nobel Prize in Physiology or Medicine awarded to Andrew Fire and Craig Mello. Building on previous work mainly in plants [50], Fire et al. [23] discovered that exogenous double-stranded RNA can be used to specifically interfere with gene function. This phenomenon was called RNA interference (RNAi). They also speculated that organisms might use double-stranded RNA naturally as a way of silencing genes. It was then shown that RNA interference was mediated by 22 nucleotide single-stranded RNAs termed small interfering RNAs (siRNAs) derived from the longer double-stranded RNA precursors [87]. The small interfering RNAs were found to repress genes by eliminating the corresponding messenger RNA transcripts, and thus, preventing protein synthesis.
Over the following years, many new small functional RNAs have been found. RNA is usually thought of as messenger RNA that serves as the template for translation of genes into proteins. In contrast, functional or non-coding RNA molecules are transcribed from a DNA sequence, but not translated into protein. The encoding DNA sequence is often referred to as an RNA gene. Functional RNA genes in the human genome include transfer RNA (tRNA), ribosomal RNA (rRNA), and various other small non-coding RNAs. Several hundred genes in our genome encode small functional RNA molecules collectively called microRNAs (miRNAs). Precursors of these miRNA molecules form structures of double-stranded RNA that can activate the RNA interference machinery. MicroRNAs downregulate gene expression either by degradation of messenger RNA through the RNA interference pathway or by inhibiting protein translation.
The first miRNA was discovered in 1993 by Victor Ambros and colleagues Rosalind Lee and Rhonda Feinbaum [42]. A genetic screen in the roundworm Caenorhabditis elegans, a millimeter-long animal used as a model organism in biological research, identified genes involved in developmental timing [42]. Surprisingly, one of the genes, termed lin-4, did not encode a protein but instead a novel 22-nucleotide small RNA. Seven years later, Reinhart et al. [70] discovered a second 22-nucleotide small RNA of this type, let-7, a gene also involved in C. elegans developmental timing. The lin-4 and let-7 small regulatory RNAs soon became very exciting for two reasons. Firstly, homologs of the let-7 gene were identified in other animals including humans [65]. The conservation of let-7 across species suggested an important and fundamental biological role for this small RNA. Secondly, the mechanism of RNA interference (RNAi) was discovered at that time, and it became clear that miRNA and RNAi pathways were intricately linked and shared common components. Within the following year, more than 100 additional small regulatory RNAs similar to lin-4 and let-7 were identified in worms, the fruit fly Drosophila, and in humans [38, 40, 41]. These small non-coding RNAs were named microRNAs (miRNAs) [38, 40, 41].
Subsequently, many more short regulatory RNAs were identified in almost all multicellular organisms, including flowering plants, worms, flies, fish, frogs, mammals [38, 40, 41, 48, 71], and in single cellular algae and DNA viruses [66, 75]. To date, more than 500 human miRNAs have been experimentally identified. Computational predictions of miRNA targets suggest that up to 30% of human protein coding genes may be regulated by miRNAs [46, 68]. This makes miRNAs one of the most abundant classes of regulatory genes in humans. MicroRNAs are now perceived as a key layer of post-transcriptional control within the networks of gene regulation.
MicroRNAs are sequentially processed from longer precursor molecules that are encoded by the miRNA genes [1] (Fig. 1). MiRNA genes are referred to by the same name (termed mir) written in italics to distinguish them from the corresponding mature miRNA (termed miR) followed by a number, e.g., mir-1 or miR-1. The encoding DNA sequence is much longer than the mature miRNA. Two ribonuclease enzymes, Drosha and Dicer, subsequently process the primary transcripts (or pri-miRNA) to generate mature miRNAs. The primary transcripts contain one or more stem-loop structures of about 70 bases. Stem-loops are double-stranded RNA structures consisting of a nucleotide sequence that can fold back on itself to form a double helix with a region of imperfect base pairing that forms an open loop at the end (Fig. 1a). The ribonuclease Drosha excises the stem-loop structure to form the precursor miRNA (or pre-miRNA) [43]. After export into the cytoplasm, the pre-miRNA is cleaved by the ribonuclease Dicer to generate a short RNA duplex [6, 28]. After untwisting, one RNA strand becomes the mature single-stranded miRNA, while the complementary strand, termed miRNA*, is usually rapidly degraded (Fig. 1b).
Fig. 1The biogenesis and function of miRNAs. a Primary miRNAs (pri-miRNA) are transcribed from longer encoding DNA sequences (miRNA genes). The pri-miRNA contains one or more stem-loop structures of about 70 bases. In the nucleus, the ribonuclease enzyme Drosha excises the stem-loop structure to form the precursor miRNA (pre-miRNA). b After export into the cytoplasm, the pre-miRNA is cleaved by the ribonuclease Dicer to generate a short RNA duplex (miRNA:miRNA*). c The mature single-stranded miRNA is incorporated into the RNA-induced silencing complex (RISC), while the complementary strand (miRNA*) is usually rapidly degraded. The miRNA incorporated into the silencing complex can bind to the target messenger RNA by base pairing, causing inhibition of protein translation and/or degradation of the target messenger RNA
MicroRNAs recognize their targets based on sequence complementarity [10]. The mature miRNA is partially complementary to one or more messenger RNAs. In humans, the complementary sites are usually within the 3′-untranslated region of the target messenger RNA. To become effective, the mature miRNA forms a complex with proteins, termed the RNA-induced silencing complex. The miRNA incorporated into the silencing complex can bind to the target messenger RNA by base pairing. This base pairing subsequently causes inhibition of protein translation and/or degradation of the messenger RNA (Fig. 1c). The potential mechanisms underlying this process were recently reviewed [30, 67]. Protein levels of the target gene are consequently reduced, whereas messenger RNA levels may or may not be decreased. In humans, miRNAs mainly inhibit protein translation of their target genes and only infrequently cause degradation or cleavage of the messenger RNA [1].
The biological role and in vivo functions of most mammalian miRNAs are still poorly understood. In invertebrates, miRNAs regulate developmental timing (e.g., lin-4), neuronal differentiation, cell proliferation, growth control, and programmed cell death [9, 33, 42]. In mammals, miRNAs have been found to play a role in embryogenesis and stem cell maintenance [7], hematopoietic cell differentiation [17], and brain development [59, 60]. To date, knowledge of human miRNAs has been primarily descriptive. MicroRNA expression has been found to be deregulated in a wide range of human diseases including cancer. However, it remains uncertain whether altered miRNA expression is a cause or consequence of pathological processes. The underlying mechanisms of why and how miRNAs become deregulated are largely unknown. Although bioinformatics approaches can predict thousands of genes that are potentially targeted and regulated by miRNAs based on sequence complementarity, only very few miRNA target genes have been functionally validated. Our group is currently investigating the role of miRNAs in mammary gland development and breast cancer pathogenesis. A comparison of miRNA and gene expression identified miRNAs that classify molecular breast cancer subtypes [8]. As cancer is ultimately a consequence of disordered gene expression, miRNAs have been suggested to contribute to the development of cancer [11]. This review will focus on the connection between human miRNA biology and different aspects of carcinogenesis. Various techniques available to investigate miRNAs will also be discussed.
MicroRNAs and cancer
Three important observations early in the history of miRNAs suggested a potential role in human cancer. Firstly, the earliest miRNAs discovered in the roundworm C. elegans and the fruit fly Drosophila were shown to control cell proliferation and apoptosis [9, 42]. Their deregulation may therefore contribute to proliferative diseases such as cancer. Secondly, when human miRNAs were discovered, it was noticed that many miRNA genes were located at fragile sites in the genome or regions that are commonly amplified or deleted in human cancer [14]. Thirdly, malignant tumors and tumor cell lines were found to have widespread deregulated miRNA expression compared to normal tissues [12, 24, 52]. The question remained whether the altered miRNA expression observed in cancer is a cause or consequence of malignant transformation.
MicroRNAs as causal cancer genes at genomic breakpoints
Five years ago, the first direct evidence for an involvement of miRNAs in cancer was reported [13]. Calin et al. studied a well-known deletion on chromosome 13, which is the most frequent chromosomal abnormality in chronic lymphocytic leukemia (CLL). This deletion had long been suspected to contribute to leukemogenesis. However, extensive studies had failed to identify a causal gene. Calin et al. [13] found that two miRNA genes, mir-15 and mir-16, were located within this 30-kb deletion. They subsequently analyzed the expression of miR-15 and miR-16 in blood samples from patients with CLL. Both miRNAs were absent or downregulated in the majority (68%) of cases when compared to normal tissue or lymphocytes. This finding suggested that these two miRNAs were causally involved in the pathogenesis of chronic lymphocytic leukemia.
In 2005, three reports provided the first mechanistic insight into how miRNAs might contribute to carcinogenesis. Two independent studies described the relationship between a miRNA cluster, mir-17-92, and the Myc oncogenic pathway [27, 63]. A third report demonstrated an interaction between let-7 miRNA and the RAS proto-oncogene [32].
The mir-17-92 cluster—small RNAs with oncogenic potential
A cluster of six miRNAs, the mir-17-92 cluster, was found to be located within a region on chromosome 13 that is commonly amplified in human B-cell lymphomas [64]. He et al. [27] demonstrated that the miRNAs from the mir-17-92 cluster were overexpressed in lymphoma cell lines carrying this amplification, and expression levels correlated with gene copy number of the mir-17-92 locus [27]. Further, the miR-17-92 primary transcript was found to be overexpressed in tumor samples from lymphoma patients. To test their hypothesis that mir-17-92 actively contributes to lymphomagenesis, the authors took advantage of a mouse model of human B-cell lymphoma. These mice develop lymphomas due to an overexpression of the Myc oncogene. The Myc oncogene encodes the transcription factor c-Myc that regulates cell proliferation, growth, and apoptosis, and overexpression of c-Myc is common in cancer. He et al. [27] demonstrated that additional expression of the mir-17-92 cluster accelerated c-Myc-induced tumorigenesis in mice. The authors therefore suggested that mir-17-92 was the first potential non-coding oncogene, referred to as oncomir-1.
The cellular function of miR-17-92 was not identified in these experiments. Nevertheless, the pathology of the tumors indicated lower rates of apoptosis as compared to tumors with Myc overexpression alone. Three recent studies contributed towards our understanding of the oncogenic potential of miR-17-92. Two reports demonstrated an anti-apoptotic effect of miR-17-92 through various pathways that promote cell proliferation and growth [55, 76]. A third study identified mir-17-92 as a mediator of angiogenesis in tumors induced by the oncogene c-Myc [19].
O’Donnell et al. [63] independently identified the same cluster of miRNAs, mir-17-92, to be regulated by the transcription factor c-Myc. The transcription factor Myc induces expression of E2F1 growth factor. The mir-17-92 cluster which is also induced by c-Myc does, in contrast, inhibit E2F1 expression. The authors therefore suggested a novel regulatory mechanism by which c-Myc fine-tunes gene expression by activating the transcription of target genes and by simultaneously inducing inhibitory miRNAs that reduce their translation.
The example of the mir-17-92 cluster highlights that a distinction between oncogenic and tumor suppressor miRNAs is likely to be an oversimplification. The same miRNAs may have oncogenic or tumor suppressor activity depending on the context and the cell type they are expressed in. A single miRNA may regulate various unrelated target genes and thereby control opposing activities such as cellular proliferation and apoptosis. The ultimate function of a miRNA may depend on the tissue type they are expressed in and what target genes are present.
MicroRNAs with tumor suppressor potential
The let-7 family of miRNAs was the first group of miRNAs shown to regulate expression of a proto-oncogene, the RAS protein. RAS proteins are membrane-associated signaling proteins that regulate cell growth and differentiation. A miRNA that controls expression of these potentially oncogenic proteins would be predicted to possess tumor suppressor activity.
Mutations in the RAS oncogene are present in approximately 15–30% of all human cancers, and overexpression of the RAS oncogene is common in lung cancer. Johnson et al. [32] showed that overexpression of RAS protein in lung cancer tissue correlated with reduced expression of let-7 miRNA. They experimentally confirmed that let-7 can inhibit RAS expression in human cancer cell lines. Loss or reduction of let-7 in lung cancer leads to RAS overexpression, thus, promoting cellular growth and contributing to tumorigenesis. The authors therefore suggested that let-7 acts as tumor suppressor [32]. Another group independently reported reduced expression of let-7 in lung cancers and found that this correlated with a poor prognosis [77].
Global loss of miRNA expression in cancer
A global decrease in miRNA levels has been observed in human cancers, indicating that small RNAs may have an intrinsic function in tumor suppression. Lu et al. [52] were the first to show that the expression levels of many miRNAs were significantly reduced in cancers compared to the corresponding normal tissues. They analyzed a total of 217 human and mouse miRNAs across 334 human cancers, cancer cell lines, and normal tissues. Cancers had significantly reduced global miRNA expression. Poorly differentiated tumors had lower miRNA levels compared with more-differentiated tumors. The authors hypothesized that miRNAs can function to drive terminal differentiation and prevent cell division. Global changes in miRNA expression may reflect the degree of cell differentiation [52]. A recent study examined the expression of 241 human miRNAs in a comprehensive panel of human cancer cell lines, the NCI-60 panel, and in normal tissues [24]. The authors confirmed the finding that most miRNAs were expressed at lower levels in human tumor-derived cell lines compared with the corresponding normal tissue [24].
Until recently, considerable uncertainty remained as to whether the altered miRNA expression observed in cancer was a cause or consequence of malignant transformation. Earlier this year, a study by Kumar et al. [37] proved for the first time that widespread reduction in miRNA expression does, indeed, promote tumorigenesis. The authors globally reduced the production of mature miRNAs through a knockdown of the miRNA-processing enzymes Drosha and Dicer in cell lines. The mouse and human cancer cells consequently showed decreased steady-state miRNA levels. These cells with global miRNA loss showed enhanced cellular growth in vitro [37]. When injected into nude mice, these cells generated faster growing and more invasive tumors compared to controls. To assess the effect of global miRNA loss in vivo, the authors deleted the miRNA-processing enzyme Dicer in a mouse model of lung cancer. The Dicer mutant mice who had impaired miRNA processing developed an increased tumor burden, with an expansion in tumor number and tumor size, as well as tumors which were less well differentiated compared to controls [37]. Overall, these data clearly suggest that global miRNA loss enhances tumorigenesis. Kumar et al. demonstrated that loss of miRNAs leads to upregulation of proto-oncogenes such as RAS and c-Myc. However, it remains to be elucidated whether loss of all miRNAs is necessary or whether reduction of a subgroup of key tumor suppressor miRNAs, such as let-7, is the event that promotes malignant transformation.
MicroRNAs in the p53 tumor suppressor network
Transcriptional networks are often deregulated in cancer cells and may lead to altered transcription of miRNA genes. Two recent studies identified a miRNA, miR-34, to be regulated by the p53 transcription factor [16, 26]. The p53 protein, also called “the guardian of the genome”, regulates the cellular response to stress and cancer-initiating events such as DNA damage. He et al. [26] found that a miRNA, miR-34, is directly activated by the transcription factor p53 after DNA damage. Expression of miR-34 induces cell cycle arrest and thereby acts together with other effectors of the p53 tumor suppressor network to inhibit inappropriate cell proliferation. Another group independently demonstrated that miR-34 is upregulated by p53 upon DNA damage and promotes apoptosis [16].
Together, these data indicate that altered expression of miRNAs is not simply a secondary event that reflects the less differentiated state of cancer cells. In contrast, at least in some cases, miRNA expression is specifically driven by tumor suppressors and oncogenes.
MicroRNAs with a role in tumor invasion and metastasis
Transcriptional networks may drive miRNA expression in cancers. Recent work from Ma et al. [54] suggested a model by which a pleiotropic transcription factor, Twist, induces expression of a specific miRNA, which suppresses its direct target and in turn activates a pro-metastatic gene, leading to tumor cell invasion and metastasis. The expression of miR-10b induced by the transcription factor Twist promoted cell migration and invasion in mouse and human breast cancer cells. Furthermore, the expression level of miR-10b in primary human breast carcinomas correlated with clinical progression [54]. These findings, if confirmed, suggest that specific miRNAs may have a role beyond the tumor-initiating event and directly participate in tumor progression and metastasis.
Regulation of miRNAs in cancer—who regulates the regulators?
In few cases, the underlying cause of miRNA deregulation in cancer is clear. As discussed above, the overexpression of miR-17-92 correlates with amplification of its gene locus [27]. Similarly, decreased expression of miR-15 and miR-16 is associated with a corresponding chromosomal deletion [13].
Transcriptional or epigenetic regulation of miRNAs has been recently reported [53, 73]. The transcription of a miRNA gene, mir-124a, was shown to be inactivated by hypermethylation of its promoter in various human tumors. This process of epigenetic silencing is a well-known mechanism to inactivate protein-coding genes in cancer cells and may similarly apply to miRNAs. The miRNA gene mir-127 is usually expressed in normal cells but not in cancer cells. Saito et al. [73] demonstrated that miR-127 was highly induced in cultured human cancer cells after treatment with demethylating drugs, suggesting that it is subject to epigenetic silencing through promoter hypermethylation.
A novel mechanism of miRNA regulation was suggested by Mayr et al. [56] and Lee and Dutta [44]. They demonstrated that miRNA function could be regulated through loss of miRNA binding sites in the target gene. Both groups independently demonstrated that chromosomal translocations in a known oncogene, high mobility group A2 (Hmga2), led to loss of the let-7 miRNA binding sites in its messenger RNA. Disrupted repression of Hmga2 by let-7 promoted oncogenic transformation and growth in mammalian cells. These two studies provide the first evidence that disrupting the interaction of a single miRNA and its target can produce an abnormal phenotype in mammalian cells [44, 56].
In addition, there is evidence that miRNAs are regulated indirectly through control of their processing enzymes. Thomson et al. [81] showed that a downregulation of miRNAs in human cancer was not associated with reduced levels of the primary miRNA transcripts. The authors therefore suggested regulation of miRNAs during subsequent processing steps, e.g., through altered function of the enzyme Drosha [81].
MicroRNA profiling—implications for cancer diagnosis
Lu et al. [52] asked the question whether global miRNA expression profiles could classify human cancer. MicroRNA expression profiles clearly differentiated human cancers according to their developmental origin.
Cancers of epithelial and hematopoietic origin had distinct miRNA profiles. A subgroup of gastrointestinal tumors, which arise from endoderm, was distinguished by miRNA expression patterns. Furthermore, tumors within a single cell lineage such as acute lymphoblastic leukemia were further differentiated according to their underlying genetic abnormality into BCR/ABL-positive tumors, T-cell tumors, and those with MLL gene rearrangement [52]. Finally, the authors applied the miRNA expression profiles they had established to an independent series of 17 poorly differentiated tumors of unknown origin. Based on the differential expression of 217 miRNAs, a correct diagnosis could be established in 12 out of 17 of the tumors. In contrast, gene expression profiling based on ∼16,000 messenger RNAs did not accurately classify the tumors [52]. This has potential important clinical implications. If miRNAs prove useful for clinical diagnosis, their key advantage might be their high stability. In contrast to most messenger RNAs, they are long-lived in vivo [49] and very stable in vitro [78], which might allow analysis of paraffin-embedded samples for routine diagnostic applications.
MicroRNAs—novel therapeutic targets?
Regulatory RNAs may also have therapeutic applications by which disease-causing miRNAs could be antagonized or functional miRNAs restored. The most intuitive choice of molecules to correct altered miRNA–messenger RNA interactions are RNA oligonucleotides. These oligonucleotides need to be chemically modified to allow for stability in serum and cellular uptake. Modified antisense oligonucleotides are already being developed to utilize the intrinsic RNAi pathway for delivery of gene therapy. If the delivery problem can be overcome, then miRNA therapies may also be possible.
Two studies have successfully applied 2′-O-Methyl-modified antisense RNAs to inhibit miRNA function in cultured cells [29, 57]. Recent work by Krutzfeldt et al. [36] demonstrated that modified cholesterol-conjugated antisense RNAs designated “antagomirs” could effectively inhibit miRNA function in vivo in the adult mouse. The authors applied three daily intravenous injections of antagomirs and achieved effective inhibition of four miRNAs over a period of weeks in most tissues except brain [36]. A novel approach was recently reported by Ebert et al. [21]. They developed miRNA inhibitors that can be transiently expressed in cultured mammalian cells. These competitive inhibitors termed “miRNA sponges” derepressed miRNA targets at least as strongly as chemically modified antisense oligonucleotides [21]. A different approach was taken by Tsuda et al. [83]. The authors designed synthetic miRNAs to target overexpressed tumor proteins, such as HER-2 protein. A synthetic miRNA targeting HER-2 messenger RNA successfully inhibited HER-2 protein expression in ovarian cancer cells [83]. Together, these studies hold some promise of miRNAs as future therapeutic targets.
One limitation of antisense RNA therapies is the restricted number of cells that can be targeted. Any approach to knock down a particular miRNA with antisense oligonucleotides will only result in partial knockdown. This may represent a limitation for cancer therapies. It remains to be seen whether indirectly mediated bystander effects on cancer cells that have not been directly targeted may partly overcome this limitation. In contrast, a partial effect on function may be of therapeutic value in neurodegenerative diseases, such as Parkinson’s or Alzheimer’s disease. A partial restoration of dopamine production by antisense therapy might result in a significant clinical improvement in Parkinson patients. Similarly, a partial reduction of the disease-causing proteins in Alzheimer’s disease may lead to a clinical improvement and might be achievable by RNA based or miRNA gene therapy.
Techniques and approaches to study miRNAs
All known miRNAs are registered in a public web-based registry, the “miRBase” database that provides up-to-date information on all published miRNAs [25]. Novel miRNA genes can be discovered by bioinformatics approaches searching for evolutionary conserved stem-loop structures in the genome (reviewed in [3, 5]). Experimentally, miRNAs are discovered by cloning all small RNAs from a certain tissue type or developmental stage and subsequent sequencing to identify the subgroup of small RNAs that fulfill the criteria for miRNAs [38, 51]. Both computational and experimental approaches indicate that many more miRNAs are likely to be identified [4, 5], which is reflected by the rapidly increasing number of annotated miRNAs which increased from less than 300 to more than 4,000 over the past 4 years [58].
MicroRNA expression studies
Northern blot analysis is a well-established technique for studying messenger RNA expression and was soon adapted to detect miRNAs in cells or tissues [42, 84]. Subsequently, conventional DNA microarray technology was modified to form miRNA microarrays, allowing for the detection of multiple miRNAs simultaneously across various samples [15, 60, 62, 82].
Lu et al. [52] developed a novel microarray strategy to improve probe specificity, which is critical due to the short nature of mature miRNAs. They performed hybridization in solution using polystyrene capture beads that are coupled to oligonucleotide probes complementary to the miRNAs of interest. The solution hybrids are then analyzed using a multicolor flow cytometer measuring bead color, denoting miRNA identity, and labeling intensity, denoting miRNA abundance [52].
In parallel to microarray platforms, commercial assays for quantitative reverse transcriptase polymerase chain reaction (RT-PCR) have become available. These allow for the analysis of miRNAs in small tissue samples or even single cells [79], as well as validation of microarray data. In addition to mature miRNAs, these quantitative RT-PCR assays can be applied to analyze miRNA precursors and primary transcripts [31]. In situ hybridization for the detection of mature miRNAs has recently become possible by using special high-affinity locked nucleic acid (LNA)-modified DNA oligonucleotide probes and holds promise for the application on human formalin-fixed and paraffin embedded tissue [35, 61].
Functional characterization of miRNAs
Various strategies have been used to investigate the function of specific miRNAs. In worms and flies, loss-of-function mutants for specific miRNAs or miRNA families allow us to draw conclusions regarding possible physiological functions of miRNAs from the resulting abnormal phenotype [34, 39, 86]. The knockdown of miRNAs or pre-miRNAs using modified antisense oligonucleotides has proven particularly useful in cell lines [29, 45, 57]. LNA-modified antisense oligonucleotides have been successfully utilized to knock down specific miRNAs in cultured cells [22]. This approach allowed identification of a crucial role for a miRNA, miR-223, in granulocytic differentiation [22]. In addition, the modified antisense RNAs (antagomirs) described by Krutzfeldt et al. [36], which inhibit miRNA function in the adult mouse, may provide a potential research tool to study miRNA function in vivo.
In mammals, induced defects in miRNA biogenesis are a useful tool for investigating the biological roles of miRNAs, as loss-of-function mutants are not available for most miRNA genes. Dicer knockout mouse models have revealed essential roles for miRNAs in murine organogenesis [88]. A recent study utilized a combined knockdown of the miRNA-processing enzymes Drosha, Dicer1, and DGCR8 to study the consequences of a global decrease in mature miRNAs in cancer cell lines and in a mouse model for lung cancer [37].
Earlier this year, four independent groups have, for the first time, deleted genes for single miRNAs in mice [72, 80, 85, 88]. Two of the groups deleted the same DNA sequence for mir-155 and described severe immune defects [72, 80]. Mice lacking miR-155 showed impaired function of B and T lymphocytes and dendritic cells [72]. In particular, T helper cell differentiation and the germinal center reaction to produce a T-cell-dependent antibody response were defective [80]. Together, these two studies demonstrated a key role for miR-155 in normal immune function. The two other groups deleted different miRNAs, miR-1-2 and miR-208, and reported cardiac defects. Mice lacking miR-208 showed inadequate cardiac growth in response to stress [85], while mice lacking miR-1-2 had defects in cardiac morphogenesis and electrical conduction [88].
MicroRNA target sites
A validated biochemical strategy for identifying miRNA targets would be highly desirable. Two groups have recently reported promising approaches to experimentally identify miRNA targets. Both approaches apply biochemical methods to purify the effector complexes of miRNAs associated with proteins and bound messenger RNA targets [2, 20].
An increasing number of sophisticated bioinformatics approaches are being developed to predict putative miRNA target genes [3, 10, 69, 74]. This is based on the fact that miRNA target recognition is at least partly based on simple sequence complementarity. Interestingly, exact base pairing between miRNAs and their targets commonly appears to be required only in the first six to eight bases from the 5′ end of the miRNA. The short nature of this designated “seed region” allows a single miRNA to act on up to a hundred different target sites, and all human miRNAs together may regulate up to one third of protein coding genes [10, 47]. A different approach to discovering miRNA target genes is to knock out or overexpress a particular miRNA and use conventional microarrays to identify genes that show changes in expression. This approach is based on the observation that some miRNAs can also downregulate messenger RNA levels in addition to downregulating protein levels of their target genes [49]. Experimental validation of miRNA target sites has been limited to date. A common approach has been to express a miRNA in vivo while simultaneously expressing and monitoring the target messenger RNA linked to a reporter gene, i.e., Luciferase [10, 18, 55, 56, 76]. The fact that a single miRNA can regulate multiple targets and a particular target may be regulated by various miRNAs suggests a highly complex network of miRNA-target interactions, which is only beginning to be unraveled.
Conclusions
Over recent years, miRNAs have emerged as major players in the complex networks of gene regulation and have been implicated in various aspects of human disease. Only 5 years after the first study reported a direct involvement of miRNAs in cancer, these small RNAs have already significantly improved our understanding of carcinogenesis. In addition to protein-coding oncogenes and tumor suppressor genes, we will have to take into account miRNAs and their regulatory networks if we aim to understand the complex processes underlying malignant transformation.
|
[
"microrna",
"cancer",
"mirna",
"development",
"pathogenesis"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
J_Chem_Ecol-4-1-2292484
|
Increased Terpenoid Accumulation in Cotton (Gossypium hirsutum) Foliage is a General Wound Response
|
The subepidermal pigment glands of cotton accumulate a variety of terpenoid products, including monoterpenes, sesquiterpenes, and terpenoid aldehydes that can act as feeding deterrents against a number of insect herbivore species. We compared the effect of herbivory by Spodoptera littoralis caterpillars, mechanical damage by a fabric pattern wheel, and the application of jasmonic acid on levels of the major representatives of the three structural classes of terpenoids in the leaf foliage of 4-week-old Gossypium hirsutum plants. Terpenoid levels increased successively from control to mechanical damage, herbivory, and jasmonic acid treatments, with E-β-ocimene and heliocide H1 and H4 showing the highest increases, up to 15-fold. Herbivory or mechanical damage to older leaves led to terpenoid increases in younger leaves. Leaf-by-leaf analysis of terpenes and gland density revealed that higher levels of terpenoids were achieved by two mechanisms: (1) increased filling of existing glands with terpenoids and (2) the production of additional glands, which were found to be dependent on damage intensity. As the relative response of individual terpenoids did not differ substantially among herbivore, mechanical damage, and jasmonic acid treatments, the induction of terpenoids in cotton foliage appears to represent a non-specific wound response mediated by jasmonic acid.
Introduction
Plants produce a broad range of defense chemicals that act as deterrents or toxins against herbivores and pathogens. Several different strategies have evolved concerning the deployment of such compounds. When the synthesis of defense chemicals occurs solely after initial attack, this reduces metabolic costs (Gershenzon 1994), as defenses are produced only when needed. Such induced defenses have ecological advantages (Agrawal and Karban 1999) and might be a useful strategy for plants that are sporadically attacked. However, plants that risk frequent and heavy damage may be better protected by investing in constitutive defense compounds (Wittstock and Gershenzon 2002). The accumulation of constitutive defenses within plants usually conforms to expectations of the ‘optimal defense theory’ (ODT), which predicts that the highest protection level will be found in parts with the highest fitness value such as young tissues and reproductive organs (McKey 1979; Rhoades 1979). This has been shown for a variety of plant secondary metabolites from different classes such as alkaloids (Hartmann and Zimmer 1986; Baldwin 2001), glucosinolates (Porter et al. 1991. Brown et al. 2003), furanocoumarins (Berenbaum and Zangerl 1999), phenylpropanoids (Harborne 1991; Opitz and Schneider 2002), and terpenes (Gershenzon and Croteau 1991). Under abiotic or biotic stress, plants often exhibit induced responses by enhancing the accumulation of constitutive compounds in certain tissues that increases their deterrence or toxicity to enemies (Baldwin 1988; Wittstock and Gershenzon 2002; Vazquez-Flota et al. 2004). These changes also may be rationalized as a way for plants to reduce the costs of defense by producing high concentrations of constitutive defenses only when and where these are needed. However, more information is required to understand how the levels of constitutive defenses change under different stresses in plants on an organ-by-organ basis.
Among the vast number of defensive compounds in plants, terpenoids form the largest group with over 25,000 known structures (Buckingham 1994). Some of the best-studied examples for constitutive plant defenses are terpenoid volatiles such as monoterpenes (C10), which are considered defensive against herbivores and pathogens (Gershenzon and Dudareva 2007). Monoterpenes are frequently constituents of oils and resins that are accumulated in complex secretory structures such as glandular trichomes, secretory cavities, or resin ducts (Fahn 1979). Among the most comprehensively studied systems are monoterpenes in the glandular trichomes of peppermint (Gershenzon et al. 2000; McConkey et al. 2000). However, less is known about the accumulation of other classes of terpenes.
Cotton (Gossypium hirsutum) is particularly suitable for the study of terpenoid accumulation because a large variety of terpenes are produced constitutively in all photosynthetically active parts of the plant and stored in subepidermal glands. Beside monoterpenes, G. hirsutum produces sesquiterpenes (C15; Elzen et al. 1985) and terpenoid aldehydes, such as hemigossypolone (C15) and the heliocides H1 to H4 (C25; Fig. 1; Hedin et al. 1992). All three classes of terpenoids in cotton are biosynthetically related (Stipanovic et al. 1977, 1978a; Davis and Essenberg 1995). The precursor of sesquiterpenes, farnesyl diphosphate, is formed by addition of a C5 isopentenyl diphosphate unit to geranyl diphosphate, the precursor of all monoterpenes. The C25 heliocides are formed from Diels–Alder-type reactions of the C15 hemigossypolone (derived from δ-cadinene) to the monoterpene myrcene for the formation of heliocides H2 and H3 or the monoterpene E-β-ocimene for the formation of heliocides H1 and H4 (Fig. 2).
Fig. 1Terpenoid products of three structural classes, which are accumulated in glands of the foliage of Gossypium hirsutum. 1 α-pinene, 2 (E)-β-ocimene, 3 myrcene, 4 (E)-β-caryophyllene, 5 γ-bisabolene, 6 α-humulene, 7 hemigossypolone, 8 heliocide H1, 9 heliocide H2Fig. 2Proposed scheme for the biosynthesis of heliocide H1 in cotton
The accumulation of terpenoid aldehydes in cotton leaves was shown to be increased after herbivory (Bezemer et al. 2004), and these compounds have been considered to function as feeding deterrents on generalist herbivores such as Spodoptera exigua (McAuslane et al. 1997). The heliocides specifically have been correlated with antifeedant effects and are locally and systematically induced in response to herbivore attack (Karban and Carey 1984; Croxford et al. 1989; Alborn et al. 1996). However, besides the terpenoid aldehydes, little information is available on the accumulation of other terpenes found in subepidermal glands. In studying the control of terpenoid accumulation in cotton, it is of interest to determine if higher concentrations arise from increases in gland number or from increases in the amount of terpenoids per gland (McAuslane et al. 1997). Increasing numbers of subepidermal glands in cotton leaves were observed after attacks of spider mites and caterpillars (McAuslane et al. 1997; Agrawal and Karban 2000).
To learn more about the control of terpenoid accumulation in cotton, we compared the response of plants to herbivore damage, mechanical damage, and the application of jasmonic acid (JA), a ubiquitous mediator of defense responses in the plant kingdom (Browse 2005). Differences in the reaction of plants to these treatments were investigated with respect to plant growth, gland production, and accumulation of major compounds from three classes of terpenoids: terpenoid aldehydes, monoterpenes, and sesquiterpenes at the level of the individual leaf. We also examined how the accumulation was influenced by the number of glands in the total foliage or the relative filling of glands with terpenoids.
Methods and Material
Plants G. hirsutum L. of a single variety, ‘Deltapine acala 90’ (Delta and Pine Company—Hollandala, MS, USA), were reared in a growth chamber under high light intensity [1 mmol (m2)−1 s−1 of photosynthetically active radiation], a 15-h light period, 28°C/23°C (day/night), and 65% relative humidity. Plants were established in pots containing ‘Seramis’-clay granulate (Masterfood GmbH, Verden/Germany), which were watered daily and fertilized weekly with a complete fertilizer [1 g/l Flory 3 (N/P/K, 15:10:15), Euflor/Germany]. Under these conditions, optimal growth and production of the various terpenoids was obtained.For all experiments, leaves were numbered sequentially such that number 1 refers to the first true leaf initiated right after the cotyledons (Fig. 3). In this study, leaves 1–7 were investigated and collectively referred to as total foliage.
Fig. 3Sketch of a Gossypium hirsutum plant at the end of experiment 1, approximately 1-month-old. Leaves are numbered according to the order of development. Grid pattern indicate leaves, which were damaged either mechanically or by caterpillars of Spodoptera littoralis
Herbivores Eggs of the generalist herbivore, Spodoptera littoralis (Boisd.; Lepidoptera, Noctuidae), Egyptian cotton worm, were obtained from Syngenta (Basel, Switzerland). After hatching, larvae were reared in 2-l plastic boxes containing artificial diet, which consisted of 500 g of bean flour, 500 ml water, 9 g ascorbic acid, 5 g ethyl-4-benzoic acid, 0.6 ml α-tocopherol, 9 ml Mazola germ oil, 4 ml 3.7% formaldehyde, and 650 ml 7.5% agar solution. Fresh diet was applied every third day. Larvae were reared at constant 20°C, 50–60% relative humidity, and ambient light conditions. Two weeks after hatching, larvae reached the third instar and were used for the study.
Experiment 1
Treatment of Cotton Plants Three weeks after germination, when plants had developed four to five true leaves, the length of the midrib of leaf 4 was determined as a good measure for growth stage. Next, 24 plants at comparable growth stage were transferred into a second growth chamber (Vötsch, Germany) with abiotic conditions identical to the chamber where plants were reared. Here, plants were maintained separately in 250-ml plastic pots and watered twice a day to ensure adequate water supply. After 24 h, plants were divided into four treatment groups of six individuals each. Plants of one group were damaged mechanically three times on leaf 2 on the first day and three times on leaf 3 on the fourth day by using a fabric pattern wheel. In every case, leaves were scratched four times in parallel to the midrib (two times on each side). For a second group, three caterpillars of Spodoptera littoralis were caged on each plant on leaf 2 during the first day and on leaf 3 during the fourth day until larvae consumed at least one third of the available leaf material. Plants of the third group were watered with a 100-μM JA solution in water (Sigma) during the first day. JA concentrations in this range are commonly used to induce the biosynthesis of secondary products in plants (Koch et al. 1999; van Poecke and Dicke 2002). Afterward, this solution was replaced by pure tap water. A control group of plants was not treated at all. One week after the start of treatments, all true leaves were harvested separately, frozen in liquid nitrogen, and stored at −20°C until further analyses.
Foliar Glands Just after removal, all leaves were scanned to calculate leaf areas by using the program ImageJ (Wayne Rasband, National Institutes of Health, USA). In this procedure, herbivore-damaged leaves were used as templates to reconstruct the areas of undamaged leaves. In addition, surface view pictures were taken of two circular 24-mm2 sections from the center of each leaf half with a stereomicroscope connected to a digital camera. In doing so, leaves were illuminated from below with a lamp to visualize filled subepidermal glands as pigmented spots. After the number of filled glands was counted, gland densities for 24-mm2 sections were calculated. Based on these sections, gland numbers and gland densities per leaf and per total foliage were extrapolated. These values were slightly underestimated because of higher gland densities at leaf edges (personal observations).
Analysis of Monoterpenes and Sesquiterpenes True leaves were separately analyzed for their content of constitutive terpenes. Depending on available material, 30–100 mg of ground frozen leaf were extracted with 300–500 μl pentane/n-hexane 1:1. Additionally, 2 μg of nonylacetate were added as internal standard before extracts were shaken for 3 h at room temperature. After extracts were cooled to 4°C, they were filtered through cotton and Na2SO4 to remove remaining water. A subset of extracts was analyzed qualitatively by gas chromatography–mass spectrometry (GC–MS) on a Hewlett-Packard 6890 gas chromatograph (injector temperature 220°C; 1 μl splitless injection) coupled to a Hewlett-Packard 5973 quadrupole mass selective detector. Separations were performed on a DB-5MS column (30 m × 0.25 mm × 0.25 μm film; Agilent Technology) with 2 ml min−1 helium as the carrier gas. The following temperature gradient was used: 40°C for 3 min, increased to 90°C at 5°C/min, further increased to 140°C at 40°C/min, followed by an increase to 160°C at 4°C/min, maintained at 160°C for 3 min, and a final heating to 300°C for 3 min. Mass spectrometry was performed with an ionization potential of 70 electron volts and a scan range of m/z from 50 to 300. Compound identification was based on comparisons with mass spectra in the Wiley and National Institute of Standards and Technology (NIST) libraries or on direct comparison of mass spectra and retention times with available standards. For quantification, all samples were analyzed on a Hewlet-Packard 6890 gas chromatograph and a flame ionization detector (temperature 250°C). The separation procedure was identical to the method for GC–MS analysis. Terpenes were quantified by using nonylacetate and corrected with response factors according to Scanlon and Willis (1985). Concentrations and levels of terpenes per subepidermal gland were calculated for single leaves and for total foliage. Eight individual monoterpenes and sesquiterpenes were analyzed, which represent more than 90% of the total terpenes in G. hirsutum.
Analysis of Terpenoid Aldehydes True leaves were separately analyzed for their content of terpenoid aldehydes according to Stipanovic et al. (1988). Depending on available material, 30–100 mg of ground frozen material were extracted ×3 with 3 ml of ethyl acetate/n-hexane (1:3) and 50 μl of 10% HCL. Extracts were shaken at 5°C for 3 h. Afterwards, extracts were removed from the leaf material, evaporated with nitrogen, and redissolved in 40% H2O and 60% solvent B (see below) from the following high-performance liquid chromatography (HPLC) procedure. Reversed phase HPLC analysis was performed on a Chromolith LC-18 column (Merck; 5 μm; 100 × 4 mm) with single wavelength detection at 272 nm. The column was eluted with H2O + 0.05% trifluoracetic acid (solvent A) and EtOH/MeOH/MeCN/EtOAc/iPrOH, 26.6:7.4:40.5:6.1:19.4 (solvent B), with the following gradient: 50% B (0 min)–50% B (5 min)–61% B (5.1 min)–61% B (20 min). Flow rate was held constant at 2 ml min−1, and temperature was kept at 25°C. Five compounds, the hemigossypolone (C15) and the heliocides 1–4 (C25), were identified by their retention times according to the literature (Stipanovic et al. 1988) and additionally by liquid chromatography–mass spectrometry (LC–MS) measurements. Quantification was done by using standard curves of purified compounds. Concentrations and levels per subepidermal gland were calculated for single leaves and for the total foliage. Hemigossypolone and the four heliocides represent approximately 90% of the total terpenoid aldehyde content. As a sesquiterpene aldehyde, hemigossypolone was included in our analysis with other terpenoid aldehydes rather than with the sesquiterpenes. In comparison to other cotton varieties (Bezemer et al. 2004), the terpenoid aldehyde gossypol (C30) was found only in traces by LC–MS and was not quantified.
Experiment 2
Eighteen 2-wk-old plants at the growth stage of two true leaves were chosen and prepared for treatments in a way comparable to the first experiment. Plants were divided into three groups. Plants of one group were damaged mechanically nine times on leaves 1 and 2 during six consecutive days by using a mechanical fabric pattern wheel (procedure see experiment 1). For a second group, two caterpillars of Spodoptera littoralis were caged on each plant on one cotyledon during day 1, on leaf 1 during day 3, and on leaf 2 during day 5 until larvae consumed at least one third of the available leaf material. A control group was not treated at all. One week after the start of treatments, leaf 4 from each plant was harvested. Leaf areas and gland number per leaf were determined according to experiment 1. Afterwards, leaves were frozen in liquid nitrogen and stored at −20°C until further analyses of mono- and sesquiterpenes. Levels of terpene classes were calculated based on the levels of eight individual compounds that represent more than 90% of the accumulated monoterpenes and sesquiterpenes in G. hirsutum.
Statistics—Experiment 1
Total Foliage The effects of mechanical damage, herbivory, and JA application on foliage area, levels of total monoterpenes, sesquiterpenes, and terpenoid aldehydes, as well as on levels of individual major terpenoids, were analyzed with analyses of variance. Data were tested for normality and equal variances. For normally distributed data, one-way analyses of variance (ANOVAs) were used. Differences between treatments and controls were tested for significance by using Dunn’s post hoc tests. Data not normally distributed were analyzed by using Kruskall–Wallis one-way ANOVAs on ranks and Dunn’s post hoc tests. Analyses were performed with Sigma Stat 2.03. The effect of foliage area and treatment on the total number of glands was tested by analysis of codeviance. Generalized linear models (GLM) with the log link as link function were performed. To deal with overdispersion, a quasi-Poisson distribution was used in the models instead of Poisson distribution. The models were simplified by removing non-significant terms and by factor-level reduction (Crawley 2002). These analyses were performed in R, version 2.4.1. All data are presented as mean ± SE except when it is mentioned otherwise.
Single Leaves The effects of leaf position, treatment, and their interactions on levels of terpenoids per gland were tested by using nested two-way ANOVA. For this purpose, terpenoid data were root- or log-transformed to normalize them. For single leaf positions, the effects of the treatments on the leaf area were tested with one-way ANOVAs. The effect of leaf area and treatment on the total number of glands was tested by analysis of codeviance (GLM see above). To analyze the effect of leaf area and treatment on gland density, levels of terpenoids per gland and terpenoid concentrations per milligram fresh leaf material ANCOVAs were used. These analyses were performed in R, version 2.4.1. Changes in per gland concentrations of individual terpenoids at specific leaf positions after treatments were tested with Sigma Stat 2.03. For normally distributed data, one-way ANOVAs were used. Differences between treatments and controls were tested for significance by using Dunn’s post hoc tests. Data not normally distributed were analyzed using Kruskall–Wallis one-way ANOVAs on ranks and Dunn’s post hoc tests.
Statistics—Experiment 2
Single Leaves The effect of the treatments on the areas of leaf 4 was tested with one-way ANOVA. The effect of leaf area and treatment on the total number of glands was tested by analysis of codeviance (GLM see above). To analyze the influence of leaf area and treatments on the accumulation of mono- and sesquiterpenes an ANCOVA was achieved. All analyses were performed in R, version 2.4.1.
Results
Experiment 1
Terpenoid Accumulation in Total Cotton Foliage after Induction The accumulation of terpenoid classes among the four different treatments was first compared on the basis of total foliage (Fig. 4). For monoterpenes, sesquiterpenes, and terpenoid aldehydes, there was a gradual increase in the following order: control, mechanical damage, herbivory, and JA treatment (ANOVA on ranks: monoterpenes, H
= 9.420, P
= 0.024; sesquiterpenes, H
= 14.140, P
= 0.003; terpenoid aldehydes, H
= 19.760, P < 0.001). Levels of all three classes were elevated in herbivore-damaged (Dunn’s tests: monoterpenes, Q
= 2.572; sesquiterpenes, Q
= 2.490; aldehydes, Q = 2.939; all P < 0.05) and JA-treated plants (Dunn’s tests: monoterpenes, Q
= 2.735; sesquiterpenes, Q
= 3.674; aldehydes, Q
= 4.164; all P
< 0.05) in comparison to the controls.
Fig. 4Mean (±SE, N = 6) levels of terpenoids in the total foliage of 4-week-old cotton plants, which were either damaged at leaves 2 and 3 mechanically (Mech), fed upon by larvae of Spodoptera littoralis (Herb), treated with jasmonic acid (JA), or left as untreated controls (Ctrl) 7 days previously. Asterisks indicate significant differences between treatment and control plants based on Kruskall–Wallis one-way ANOVAs on ranks and Dunn’s post hoc tests. *P < 0.05
All individual compounds showed the same pattern that was seen for total terpenoid accumulation: a gradual increase from control to mechanical damage, herbivory, and JA treatment. In 11 out of 13 compounds, this increase was significant (Table 1) for herbivore and JA-treated plants compared to the control plants, while mechanically damaged plants showed tendencies for increased accumulation. Among monoterpenes, considerable differences in accumulation after treatment were shown. (E)-β-ocimene increased up to eightfold after both herbivory and JA treatment in comparison to controls, whereas myrcene levels slightly but significantly increased after herbivory and JA treatment, and the amounts of α- and β-pinene did not show significant differences between treatments and control. Little variation in the induction pattern occurred when comparing individual sesquiterpenes. Levels of all major sesquiterpenes, including β-caryophyllene, α-humulene, γ-bisabolene, and β-bisabolol doubled after herbivory and were 2.0–2.6-fold higher after JA treatment in comparison to control plants. Among terpenoid aldehydes, considerable differences appeared. Whereas the amounts of hemigossypolone were doubled, levels of heliocides, especially H1 and H4, increased much more after herbivory and JA treatment compared to the controls. Heliocide H4, for example, showed up to a ninefold increase after herbivory and up to a 15-fold increase after JA treatment in comparison to control plants. For H1, these increases were approximately five- and tenfold, respectively. Meanwhile, the highest levels of H2 and H3 were found in the JA-treated plants, reaching up to 3.3-fold that of control levels.
Table 1Mean levels (±SD, N = 6) of terpenoid products in total foliage of 1-month-old G. hirsutum plants that were either mechanically damaged with a fabric pattern wheel, damaged by larvae of Spodoptera littoralis (herbivory), treated with jasmonic acid (JA), or left as untreated controlsCompoundANOVA (df = 3)Levels of terpenoids (mg)F/HP ValueControlMechanical damageHerbivoryJAMonoterpenesα-Pinene7.2130.0651.11 ± 0.081.26 ± 0.111.31 ± 0.251.45 ± 0.44β-Pinene7.6470.0540.20 ± 0.020.23 ± 0.020.25 ± 0.050.26 ± 0.08Myrcene10.1670.0170.49 ± 0.060.68 ± 0.040.78** ± 0.150.80** ± 0.33(E)-β-Ocimene17.487<0.0010.06 ± 0.020.25 ± 0.060.47** ± 0.110.49** ± 0.26Sesquiterpenes(E)-β-Caryophyllene16.340<0.0010.50 ± 0.070.76 ± 0.120.95** ± 0.211.32** ± 0.50α-Humulene9.509<0.0010.14 ± 0.020.22 ± 0.030.27 * ± 0.060.36 * ± 0.13γ-Bisabolene18.007<0.0010.23 ± 0.040.34 ± 0.060.45** ± 0.100.60** ± 0.19β-Bisabolol14.3400.0020.46 ± 0.050.59 ± 0.090.82** ± 0.180.91** ± 0.26Terpenoid aldehydesHemigossypolone15.2870.0021.59 ± 0.302.16 ± 0.443.14** ± 0.403.10** ± 0.83Heliocide 120.487<0.0010.59 ± 0.141.65 ± 0.373.35** ± 1.115.95** ± 1.93Heliocide 221.031<0.0010.97 ± 0.211.45 ± 0.421.83 * ± 0.373.06 * ± 0.74Heliocide 323.938<0.0010.37 ± 0.070.56 ± 0.150.70 * ± 0.151.23 * ± 0.30Heliocide 420.247<0.0010.17 ± 0.080.59 ± 0.251.47** ± 0.392.60** ± 0.78The effect of treatment on terpenoid levels were tested using one-way ANOVAs. For normally distributed data, one-way ANOVAs were performed (F values). For not normally distributed data, Kruskall–Wallis one-way ANOVAs on ranks were performed (H values).*P < 0.05, significant differences from the control according to Bonferoni post hoc tests**P < 0.05, significant differences from the control according to Dunn’s post hoc tests
Growth and Subepidermal Gland Production of Cotton Foliage ANOVA showed an effect of the treatments on the total leaf area (F = 10.44, P < 0.001). Whereas controls, mechanically and herbivore-damaged plants were highly comparable in size (mean, 496–513 cm2), JA-treated plants (mean = 306 cm2) showed a reduction in the total area of foliage down to 60% that of the controls (Bonferoni post hoc test. P < 0.05).
Because of significantly smaller leaves in the JA-treated plants, the effect of treatment on the production of subepidermal glands was tested by using a GLM with leaf area as a covariable. For all treatments, this covariable showed an impact on the number of glands (t = 8.113, P < 0.001; Fig. 5). In comparison to the controls, an increase in the number of glands was found for herbivore-damaged plants (t = 4.147, P < 0.001) and for plants that were treated with JA (t = 5.585, P < 0.001). Because there were no differences between mechanically damaged and non-treated plants, the data of both groups were combined in this analysis.
Fig. 5Effect of leaf area and treatment on the number of glands in the total foliage of 4-week-old cotton plants, which were either damaged at leaves 2 and 3 mechanically (Mech), fed upon by larvae of Spodoptera littoralis (Herb), treated with jasmonic acid (JA), or left as untreated controls (Ctrl) 7 days previously. Regressions were drawn based on transformed data from a generalized linear model (GLM) with a quasi-Poisson error distribution. Data for control and mechanically damaged plants were combined due to factor level reduction
Single Leaf Analysis To assess more precisely plant response to the treatments, analyses were carried out at the level of individual leaves. For all treatments, the distribution of leaf areas followed a consistent pattern within plants with the biggest leaves at intermediate positions (Fig. 6a). Whereas no differences in leaf area were evident among mechanically damaged, herbivore-damaged, and control plants, plants treated with JA exhibited significant smaller areas for leaves 3–6 in comparison to corresponding control leaves (ESM Tables 3 and 4). Because of the negative impact of JA treatment on the leaf area, the following analyses of the effect of treatments on gland and terpenoid production were performed by using leaf area as covariable (statistical results see ESM Tables 3 and 4).
Fig. 6Mean (±SE, N = 3–6) leaf area, number of glands, gland density, terpenoid level per gland, and terpenoid concentration of 4-week-old cotton plants, which were either damaged at leaves 2 and 3 mechanically (Mech), fed upon by larvae of Spodoptera littoralis (Herb), or treated with jasmonic acid (JA) 7 days previously. Areas of herbivore-damaged leaves 2 and 3 were extrapolated from remaining leaf material. For each leaf position, asterisks indicate significant differences from corresponding control leaves based on one-way ANOVAs (a), on analyses of codeviance with leaf area as covariable (b), and ANCOVAs with leaf area as covariable (c–e). *P < 0.05
Gland number per leaf revealed a similar distribution pattern among leaves for all treatments: The oldest leaf 1 contained the lowest gland number, which increased gradually toward leaf 5 and again decreased in the expanding leaves 6 and 7 (Fig. 6b). Young leaves (5–7) of herbivore-treated plants showed significantly elevated numbers of glands in comparison to corresponding leaves of control plants. Lower but still significant increases in gland numbers were found for leaves 2, 3, and 7 (youngest leaf) of plants that were treated with JA compared to control leaves. Mechanical damage showed nearly no effect on gland number.
Changes in leaf area and gland numbers after treatment should have an impact on gland densities. Independent of treatment, gland densities increased gradually from the oldest leaf 1 toward the youngest leaf 7 (Fig. 6c). This pattern was enhanced in herbivore-damaged and JA-treated plants because of significantly higher gland densities in young leaves (5, 6, 7) compared to those leaves from the control plants. Mechanical damage did not cause an increased gland density in any leaf.
To analyze how terpenoid production was altered by treatment, total terpenoid levels per gland were calculated. Treatments had a considerable effect on the pattern of terpenoids per gland among leaves (Fig. 6d). Control plants showed highest levels in the oldest leaf and lowest levels in the youngest one. In contrast, mechanically and herbivore-damaged plants exhibited highest values in younger leaves, having significantly elevated total terpenoid levels in comparison to those of control leaves. Plants that were treated with JA showed significantly elevated levels of total terpenoids per gland in young and intermediate leaves in comparison to the controls.
The increase in total terpenoid and gland production caused by the treatments as well as altered leaf areas had a considerable impact on the pattern of the total terpenoid concentrations in leaves. Total terpenoid concentrations showed a consistent distribution pattern among leaves for all treatments (Fig. 6e). There was a gradual increase starting from the oldest leaf 1 toward the youngest leaves 6 and 7. All treatments enhanced this pattern by causing significantly higher concentrations in intermediate and young leaves in comparison to the leaves of corresponding control plants. The lowest effect was found for mechanically damaged plants, whereas plants treated with JA reached the highest concentration with mean levels of approximately 16 μg terpenoids per milligram fresh weight.
The analyses of individual terpenoids revealed a significant impact of treatment, leaf position, and interactions between these two factors on levels of all compounds per gland (Table 2). The lowest effects of treatment were found for the monoterpenes α- and β-pinene, whereas the biggest changes appeared for (E)-β-ocimene and heliocides 1 and 4. Differences of single terpenoid levels per gland among leaves appeared to be biggest for (E)-β-ocimene and heliocides 4, but in case of the latter, this was due to its unusual absence in leaves 2 and 3 (Fig. 7). Significant interactions between treatment and leaf position revealed that changes in levels of total and single terpenoids of induced plants were not consistent in all leaves, which is illustrated in Fig. 7. As with total terpenoid levels per gland (Fig. 6d), mechanical damage and herbivory caused significantly elevated accumulation of individual compounds per gland in the youngest leaves in comparison to control plants. In contrast, JA treatment induced elevated levels in young but also in intermediate leaves compared to control leaves. JA treatment led to the highest increases of most compounds, especially for heliocides, in comparison to those in the corresponding leaves of control plants.
Fig. 7Mean (±SE, N = 3–6) concentrations of terpenoids per gland among true leaves of 4-week-old Gossypium hirsutum plants, which were either damaged at leaves 2 and 3 mechanically (Mech), fed upon by larvae of Spodoptera littoralis (Herb), treated with jasmonic acid (JA), or left as untreated controls (Ctrl) 7 days previously. Asterisks indicate significant differences (P < 0.05) from corresponding control leaves based on one-way ANOVA’s and Bonferoni post hoc tests. Pound signs indicate significant differences (P < 0.05) from corresponding control leaves based on Kruskall–Wallis one-way ANOVAs on ranks and Dunn’s post hoc testsTable 2Results of nested two-way ANOVA for the effects of treatment (mechanical damage, herbivory, jasmonic acid application), leaf position, and interactions on levels of accumulated terpenoids per gland in plants of G. hirsutumCompoundTreatment (df = 3)Leaf Number (df = 6)Treatment × Leaf Number (df = 18)FPFPFPTotal terpenoids37.81<0.00117.45<0.0016.78<0.001Monoterpenesα-Pinene6.70<0.00139.98<0.0014.02<0.001β-Pinene4.340.00642.48<0.0013.67<0.001Myrcene10.25<0.00151.60<0.0014.06<0.001(E)-β-Ocimene106.72<0.00179.24<0.0016.50<0.001Sesquiterpenes(E)-β-Caryophyllene24.13<0.00113.30<0.0016.59<0.001α-Humulene20.56<0.0019.66<0.0013.21<0.001γ-Bisabolene37.08<0.00111.33<0.0015.74<0.001β-Bisabolol23.15<0.00112.48<0.0014.99<0.001Terpenoid aldehydesHemigossypolone10.16<0.00115.48<0.0016.47<0.001Heliocide 164.63<0.00113.78<0.0018.88<0.001Heliocide 227.13<0.00117.65<0.0014.94<0.001Heliocide 330.60<0.00117.66<0.0014.83<0.001Heliocide 455.70<0.001104.05<0.0018.59<0.001
Experiment 2
Morphological and Chemical Changes in Young Leaves after Induction In addition to herbivore-treated plants and untreated controls, this supplementary experiment also included mechanically damaged plants that were injured more frequently and over a longer period than in experiment 1. ANOVA showed an effect of the treatments on areas of young leaves (leaf 4; F = 5.97, P = 0.012; Fig. 8a, ESM Table 5). Herbivory reduced the areas of these leaves in comparison to those of the corresponding leaves of the control plants (t = −3.445, P = 0.004). Therefore, the effect of mechanical damage and herbivory on the production of glands were tested by using a GLM with leaf area as a covariable. For both treatments, leaf area showed an impact on the number of glands (t = 2.256, P = 0.041). In comparison to the control plants, a significant increase in the number of glands was found for mechanically damaged plants (t = 2.599, P < 0.021) and an even more pronounced increase for herbivore-treated plants (t = 3.938, P = 0.002; Fig. 8b).
Fig. 8Mean (±SE, N = 6) leaf area, number of glands, and monoterpene- and sesquiterpene levels of leaf 4 from 3-week-old Gossypium hirsutum plants, which were either damaged mechanically at leaves 1 and 2 (Mech), fed upon by larvae of Spodoptera littoralis (Herb) on one cotyledon, leaves 1 and 2, or left as untreated controls (Ctrl). Asterisks indicate significant differences from corresponding control leaves based on one-way ANOVAs (a), on analyses of codeviance with leaf area as covariable (b), and ANCOVAs with leaf area as covariable (c, d). *P < 0.05; **P < 0.01; ***P < 0.001
The accumulation of mono- and sesquiterpenes in leaf 4 after treatment followed a comparable pattern: a gradual increase from control to mechanical damage and herbivory (ANOVA: monoterpenes, F = 8.13, P = 0.004; sesquiterpenes, F = 27.43, P < 0.001; Fig. 8c,d). Levels of monoterpenes were elevated after herbivory in comparison to the controls (t = 4.029, P = 0.001). Levels of sesquiterpenes were elevated after mechanical damage (t = 5.115, P < 0.001) and herbivory (t = 7.197, P < 0.001) in comparison to the controls.
Discussion
Cotton (G. hirsutum) accumulates a large variety of terpenes constitutively, including monoterpenes (C10), sesquiterpenes (C15), and terpenoid aldehydes (C15, C25, and C30). This study is the first to describe the accumulation of all three of these classes after actual or simulated herbivory. Increased levels of terpenoids were observed in the total foliage 1 week after mechanical damage, feeding by Spodoptera littoralis caterpillars, or treatment with JA, in comparison to the untreated controls. A closer look revealed that, after damage to older leaves (leaves 2 and 3 numbered from the base), the increase was restricted mainly to young leaves (leaves 5–7). Similarly, studies on cotton that investigate the effects of herbivory by larvae of S. exigua and the wireworm Agriotes lineatus also showed elevated concentrations of terpenoid aldehydes in young cotton leaves (McAuslane et al. 1997; McAuslane and Alborn 1998; Bezemer et al. 2004). Such an induction pattern is in agreement with ODT, which predicts that an increase in the accumulation of defense chemicals occurs preferentially in plant parts with the highest fitness value such as young tissues or reproductive organs (McKey 1979; Frischknecht et al. 1987; Ohnmeiss and Baldwin 2000).
All three classes of cotton terpenoids are stored in subepidermal pigment glands found in leaves and other organs. After treatment, G. hirsutum displayed two ways to achieve elevated terpenoid accumulation: (1) production of additional glands and (2) increased filling of existing glands (Fig. 6). We showed that production of additional glands was restricted to leaves that were still under development or newly formed after treatment. This trend also has been observed after attack by spider mites or larvae of S. exigua on cotton (McAuslane et al. 1997; Agrawal and Karban 2000). In other plants, such as birch and tomato, an elevated number of defense structures like glandular trichomes is also produced in young leaves after induction by herbivores (Boughton et al. 2005; Valkama et al. 2005). Besides the fact that, according to ODT, plants have been selected to increase their resistance to herbivores especially in young tissue, the formation of subepidermal pigment glands in cotton might be necessarily restricted to newly developing leaves because of developmental constraints.
In counting the number of cotton leaf glands under a stereomicroscope, we may have overestimated the production of new glands after treatment if herbivory triggered the filling of pre-existing glands that were unpigmented (H. T. Alborn, personal communication). However, by using several microscopic techniques, we did not find any evidence for unfilled glands in the leaves analyzed. Therefore, we assume that actual or simulated herbivory does indeed induce increased numbers of glands in cotton.
The increased filling of existing glands with terpenoids was shown in the youngest leaves of induced plants for all major terpenoid classes. However, considerable differences in the magnitude of this additional accumulation were observed among individual terpenoid compounds. Among terpenoid aldehydes, levels of heliocides H1 and H4, showed the highest increase after treatments. Other studies on cotton species also have demonstrated that these two heliocides increase more than H2, H3, and hemigossypolone (HGQ) after herbivory by Spodoptera species (McAuslane et al. 1997; McAuslane and Alborn 1998; Agrell et al. 2004; Bezemer et al. 2004). Here, we showed that levels of (E)-β-ocimene increased in a magnitude similar to H1 and H4, and that this compound increased most among monoterpenes and sesquiterpenes. Interestingly, (E)-β-ocimene is a direct precursor in the formation of H1 and H4, combining with hemigossypolone in a Diels–Alder-type reaction (Stipanovic et al. 1978a; Fig. 2). Therefore, in non-treated plants, the biosynthesis of this monoterpene could represent a limiting step in the production of the appropriate heliocides. This proposition is supported by the fact that, in cotyledons of G. hirsutum, which do not contain any heliocides, hemigossypolone is present (S. Opitz, unpublished), but (E)-β-ocimene is absent. A comparable regulation mechanism that involves myrcene could determine the levels of H2 and H3 (Stipanovic et al. 1977, 1978b).
Our study compared the terpenoid content of total cotton foliage after three different treatments and showed a gradual increase in terpenoid levels in the following order: mechanical damage with a fabric pattern wheel, herbivory by S. littoralis caterpillars, and application of JA. However, a single incidence of mechanical injury is hard to compare with continuous feeding damage by an insect. Studies on lima bean have demonstrated that both the intensity and frequency of mechanical damage can alter the reaction of a plant (Mithöfer et al. 2005). This is consistent with the results of experiment 2 where more frequent mechanical damage caused significantly higher numbers of glands (Fig. 8), a response not detectable after the less frequent mechanical damage in experiment 1 (Fig. 5 and 6). However, in both experiments, herbivory induced stronger reactions of plants than mechanical damage. Elicitors found in the regurgitate of herbivores are known to trigger the biosynthesis of defense metabolites (Alborn et al. 1997). If such elicitors play a role in cotton defense reactions, mechanical injury may never result in terpenoid accumulation comparable to that caused by herbivory.
The strongest increase in both terpenoid and gland production appeared in plants that were treated with JA, a ubiquitous plant hormone known to mediate defense responses to biotic and abiotic stresses (Browse 2005). A previous study with cotton showed that the application of the methylated derivative, methyl jasmonate, induced the synthesis and emission of volatile terpenes, a response also observed for herbivore-damaged plants (Rodriguez-Saona et al. 2001). Obviously, JA plays an important role in mediating damage-induced signaling in G. hirsutum. Given the effects of exogenous application, herbivory can be assumed to trigger elevated internal levels of JA as in other species (Baldwin et al. 1997; Creelman and Mullet 1997). For example, in tobacco, lima bean, or maize, internal JA levels were determined in ranges of single nanograms per gram fresh weight, showing 20- to 40-fold increases within the first hours after induction (Baldwin et al. 1997; Koch et al. 1999; Schmelz et al. 2003). However, internal JA levels of cotton still need to be investigated. If the irrigation of plants with 100 μM JA in our study led to a non-physiological high internal concentration of JA, this could explain why JA treatment induced terpenoid accumulation to a greater extent than herbivory. We observed additionally that JA inhibited leaf development, a response also seen in other plants (Sembdner and Parthier 1993).
Studies on the induction of chemical defense in plants after herbivory always raise questions about the specificity of such reactions to particular enemies. Certain plants have been shown to induce different spectra of defenses in response to different herbivores (Turlings et al. 1998; Traw and Dawson 2002; Delphia et al. 2007). However, in this study, we did not find substantial differences in the pattern of induction among treatments for the major terpenoids measured (see Table 1). The amounts of single compounds varied in magnitude, leading to changes in the terpenoid profile, but these profile changes tended to be comparable among treatments. The induction of elevated terpenoid accumulation in leaves of G. hirsutum may thus be a non-specific reaction of plants to damage. Similar non-specific reactions to damage are known for terpenes in other species (Banchio et al. 2005), as well as for alkaloids (Frischknecht et al. 1987; Baldwin et al. 1997), phenolics (Cipollini 1997), and glucosinolates (Bodnaryk 1992).
The fact that cotton plants react to damage with elevated levels of terpenoids accumulated in the subepidermal pigment glands of their foliage suggests that these compounds function in plant defense. Caterpillar species such as S. exigua or Heliothis virescens prefer feeding on glandless instead of glanded cultivars of G. hirsutum, thus supporting the argument that the gland terpenes play a key role in defense of cotton against herbivores (Montandon et al. 1986; McAuslane and Alborn 2000). Leaf material from plants of glanded cultivar lines that had been induced by herbivory showed increased deterrent or toxic effects on Spodoptera species compared to material from uninduced plants, a finding thought to be due to their elevated contents of terpenoid aldehydes (Alborn et al. 1996; McAuslane et al. 1997; Anderson et al. 2001). Indeed, terpenoid aldehydes such as gossypol, hemigossypolone, and the heliocides H1 and H2 exhibit strong toxicity to caterpillars of H. virescens and Pectinophora gossypiella after addition to artificial diet (Elliger et al., 1978). While the feeding deterrent effects of induced cotton foliage are usually attributed to elevated terpenoid aldehydes contents (Hedin et al. 1992; McAuslane et al. 1997; McAuslane and Alborn 2000), we showed that cotton plants also increase their levels of monoterpenes and sesquiterpenes. Thus, these compounds may also participate in defense against herbivores. A sesquiterpene, caryophyllene oxide has been demonstrated to synergize the negative effect of gossypol, the dominant terpenoid aldehyde in roots and seeds of cotton, on larval development of H. virescens (Gunasena et al. 1988). Additionally, caryophyllene retarded the growth and delayed the time of development of these larvae. Other studies have shown that single sesquiterpenes such as β-bisabolene impair the development of insect herbivores and deter them from feeding (Bowers et al. 1976; Gonzalez-Coloma et al. 1995; Zipfel 2007).
Beside their function against herbivores, terpenoids are also considered to play a defensive role against fungi or pathogens. Because plants are exposed especially to infestation at wound sites, the elevated terpenoid levels in damaged leaves might have a critical role in helping plants to cope with pathogens. For cotton, studies have shown that certain terpenoid aldehydes possess antifungal and antipathogen activities (Zhang et al. 1993; Abraham et al. 1999). A good example for the antibiotic activity of terpenoids in other species comes from conifers where the growth and germination of bark beetle associated pathogens is inhibited (Keeling and Bohlmann 2006). In addition, analyses of terpene-rich essential oils from a variety of plant species have shown antibacterial and antifungal activities, indicating the potential role of these substances in plant defense (Oyedeji and Afolayan 2005; Ozer et al. 2007).
In summary, our results demonstrate that elevated levels of terpenoids in cotton leaves after real and simulated herbivory represent a general wound response that is mediated by JA. The increase in terpenoids is due to the production of additional glands, in which terpenoids are stored, as well as the increased filling of existing glands. As all three classes of terpenoids (monoterpenes, sesquiterpenes, and terpenoid aldehydes) were elevated after damage, these substances may act synergistically in defense against herbivores or pathogens (Stipanovic et al. 1988).
Electronic supplementary material
Below is the link to the electronic supplementary material.
ESM 1
(DOC 183 kb)
|
[
"cotton",
"gossypium hirsutum",
"pigment glands",
"monoterpenes",
"sesquiterpenes",
"terpenoid aldehydes",
"spodoptera littoralis",
"mechanical damage",
"jasmonic acid",
"induction",
"constitutive plant defense"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Appl_Microbiol_Biotechnol-4-1-2266784
|
Secretion of Streptomyces mobaraensis pro-transglutaminase by coryneform bacteria
|
We previously reported on the secretion of Streptomyces mobaraensis transglutaminase by Corynebacterium glutamicum ATCC13869 (formerly classified as Brevibacterium lactofermentum). In the present work, we investigated whether any other coryneform bacteria showed higher productivity than C. glutamicum ATCC13869. We found that most coryneform species secreted pro-transglutaminase efficiently. Moreover, we confirmed that Corynebacterium ammoniagenes ATCC6872 produced about 2.5 g/l pro-transglutaminase over a 71-h period in a jar fermentor. Our findings suggest that some other coryneform bacteria, especially C. ammoniagenes ATCC6872, are potential hosts for industrial scale protein production.
Introduction
The secretion of heterologous proteins into culture medium is a potentially useful method of production (Choi and Lee 2004), although the amounts obtained through this approach are generally small. Thus, there remains a need to develop an efficient secretion system for industrial proteins. Streptomyces mobaraensis transglutaminase (TG) has been used in the food industry to modify proteins (Yokoyama et al. 2004). It is utilized in binding meat and fish and in gelled food products such as jelly, yogurt, and cheese. Moreover, it has great potential for use in manufacturing the materials found in cosmetics, thermostable microcapsules, and carriers for immobilized enzymes (Yokoyama et al. 2004). However, so far, an efficient production system for TG has been lacking.
A large number of different coryneform bacteria have been isolated from soil, animals, and plants (Liebl 2005). Some are used in biotechnological production processes; for example, Corynebacterium glutamicum and Corynebacterium ammoniagenes are well-known industrial producers of amino acids and nucleotides, respectively (Liebl 2005). C. glutamicum is a Gram-positive, non-sporulating bacterium with about 53.8% guanine–cytosine (GC) DNA content and its genome has been sequenced (Ikeda and Nakagawa 2003; Kalinowski et al. 2003). This species is extensively used in the industrial production of amino acids, such as glutamate and lysine, which have been applied in human food, animal feed, and pharmaceutical products for several decades (Hermann 2003; Krämer 1994; Liebl 2005). However, there had been only few reports concerning heterologous protein secretion in C. glutamicum (Billman-Jacobe et al. 1995; Liebl et al. 1992; Salim et al. 1997). Recently, we demonstrated that TG could be efficiently secreted in an active form using a signal peptide derived from a cell-surface protein of coryneform bacteria in C. glutamicum ATCC13869 (Date et al. 2003, 2004; Kikuchi et al. 2003). We showed that C. glutamicum ATCC13869 could also secrete an active form of human epidermal growth factor, which consists of 53 amino acid residues, including six cysteine residues that form three disulfide bonds (Date et al. 2006). More recently, it has been shown that the twin-arginine translocation (Tat) pathway, which is a recently detected protein secretion pathway, is active in C. glutamicum ATCC13869 and that the Chryseobacterium proteolyticum pro-protein glutaminase, which is not secreted via the Sec pathway in C. glutamicum ATCC13869, is efficiently secreted by the Tat pathway (Kikuchi et al. 2006, 2007). These results demonstrate that protein production using C. glutamicum might be useful on an industrial scale. However, little is known about heterologous protein production by other species of coryneform bacteria.
In this report, we show that most other coryneform bacteria also secrete pro-TG efficiently and that C. ammoniagenes ATCC6872 is the best producer species, accumulating about 2.5 g/l pro-TG over a period of 71 h in a jar fermentor.
Materials and methods
Bacterial strains, plasmids, and culture medium
The bacterial strains used are shown in Table 1. The plasmids were as follows: pPSPTG1 (Kikuchi et al. 2003), which contains the cspB promoter derived from C. glutamicum ATCC13869; the CspA signal peptide sequence derived from a cell-surface protein of C. ammoniagenes ATCC6872; and the coding region of pro-TG, derived from S. mobaraensis, and pPKPTG1 (Kikuchi et al. 2003), in which the signal peptide for pro-TG secretion is replaced by the CspB signal peptide sequence from C. glutamicum ATCC13869. The coryneform bacteria were aerobically grown in CM2G medium (Kikuchi et al. 2003) at 30°C in a shaking test tube. Plasmids were introduced by electroporation with a Gene Pulser (Bio-Rad) using the standard protocol for Corynebacterium species. For pro-TG production in a test tube culture, we used MMTG medium containing 60 g glucose, 1 g MgSO4·7H2O, 30 g (NH4)2SO4, 1.5 g KH2PO4, 0.01 g FeSO4·7H2O, 0.01 g MnSO4·4H2O, 450 μg thiamine hydrochloride, 450 μg biotin, 0.15 g dl-methionine, and 50 g CaCO3 per l distilled water, adjusted to pH 7.5 (Kikuchi et al. 2003). Kanamycin (25 mg/l) was added to the culture medium when required.
Table 1Pro-TG accumulation by coryneform bacteria carrying pPSPTG1 in MMTG medium at 30°C for 72 hStrainPro-TG (mg/l)Corynebacterium ammoniagenes ATCC6872454Corynebacterium glutamicum ATCC13032147Brevibacterium taipei ATCC13744201Corynebacterium glutamicum (Micrococcus glutamicus) ATCC1376166Brevibacterium roseum ATCC13825269Corynebacterium glutamicum (Brevibacterium flavum) ATCC13826314Corynebacterium herculis ATCC1386854Corynebacterium glutamicum (Brevibacterium lactofermentum) ATCC13869118Corynebacterium acetoacidophilum ATCC13870369Corynebacterium glutamicum (Brevibacterium divaricatum) ATCC14020391Brevibacterium saccharolyticum ATCC14066238Brevibacterium immariophilium ATCC14068257Microbacterium ammoniaphilum ATCC15354423Corynebacterium glutamicum (Corynebacterium lilium) ATCC15990246Corynebacterium callunae ATCC15991759Brevibacterium thiogenitalis ATCC19240156
Jar fermentation
Fermentor growth was carried out in a 1-l jar fermentor (300 ml working capacity). Glycerol stock suspensions of bacteria were stored at −80°C in 20 v/v glycerol. Glycerol stock suspension (50 μl) was inoculated into 50 ml CM2G medium in a 500-ml Sakaguchi flask, and cultured at 27°C for 24 h. Then, 15 ml of the culture was inoculated into 300 ml MMTG-J medium containing 120 g glucose, 1 g MgSO4·7H2O, 30 g (NH4)2SO4, 1.5 g KH2PO4, 0.01 g FeSO4·7H2O, 0.01 g MnSO4·4H2O, 450 μg thiamine hydrochloride, 450 μg biotin, 0.15 g dl-methionine, and 0.2 g total nitrogen of soybean hydrolyzate per l distilled water, adjusted to pH 7.2. The jar fermentor operating conditions were as follows: agitation rate = 600 rpm, airflow rate = 1/2 vvm; temperature = 27°C. The culture was automatically maintained at pH 7.2 by the controlled addition of ammonia gas with airflow. The glucose concentrations were analyzed using an AS-210 glucose analyzer (Sakura Seiki, Tokyo, Japan), and the optical densities were measured at 625 nm with a DU640 photometer (Beckman Coulter).
Protein analysis
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in a 4–20% gradient polyacrylamide gel was carried out as described by Laemmli (1970), with Coomassie brilliant blue R-250 staining. The secreted pro-TG was measured by reverse-phase high-performance liquid chromatography, using purified native TG as a standard (Yokoyama et al. 2002). For processing the pro-TG to convert to an active form, the secreted pro-TG in a culture supernatant was incubated with purified SAM-P45 (Suzuki et al. 1997) at 30°C for 2 h, at a 100:1 ratio of pro-PG to SAM-P45. The purified SAM-P45 was a gift from Dr. S. Taguchi. The active-form TG was purified according to the method described previously (Yokoyama et al. 2002). TG activity was measured by the calorimetric hydroxamate procedure using N-carbobenzoxy-L-glutaminylglycine as described previously (Yokoyama et al. 2002). For determining the N-terminal amino acid sequence of the pro-TG, after separation by SDS-PAGE, proteins were transferred to polyvinylidene difluoride membrane by electroblotting. The membrane carrying the pro-TG was directly subjected to a gas phase protein sequencer (model PSQ; Shimadzu, Kyoto, Japan) equipped with an in-line amino acid analyzer (model RF-550; Shimadzu) as described previously (Kikuchi et al. 1997).
Results and discussion
Secretion of pro-TG by coryneform bacteria using C. ammoniagenes ATCC6872 CspA signal peptide
We transformed the bacteria listed in Table 1 with pPSPTG1, cultured each transformant in MMTG medium at 30°C for 72 h, and then subjected the supernatants to SDS-PAGE. As shown in Fig. 1, each transformant could stably secrete the pro-TG, and high levels of pro-TG with the anticipated molecular weight (42 kDa) were detected in the supernatants for all of the bacteria tested (Table 1). The pro-TG was most efficiently secreted by C. ammoniagenes ATCC6872, Corynebacterium acetoacidophilum ATCC13870, C. glutamicum ATCC14020 (formerly classified as Brevibacterium divaricatum), Microbacterium ammoniaphilum ATCC15354, and Corynebacterium callunae ATCC15991.
Fig. 1SDS-PAGE analysis of pro-TG secretion by coryneform bacteria carrying pPSPTG1. Each track contained a 10-μl aliquot of supernatant mixed with an equal volume of sample buffer. After electrophoresis, the gel was stained with Coomassie brilliant blue R-250. aLane 1, molecular weight markers; lanes 2 and 3, B. saccharolyticum ATCC14066; lanes 4 and 5, B. roseum ATCC13825; lanes 6 and 7, B. immariophilium ATCC14068; lanes 8 and 9, C. glutamicum ATCC13826; lanes 10 and 11, C. glutamicum ATCC15990; lanes 12 and 13, C. callunae ATCC15991; lanes 14 and 15, C. herculis ATCC13868; lanes 16 and 17, C. acetoacidophilum ATCC13870; lanes 18, C. ammoniagenes ATCC6872. bLanes 1 and 2, C. glutamicum ATCC13761; lane 3, molecular weight markers; lanes 4 and 5, C. glutamicum ATCC14020; lanes 6 and 7, M. ammoniaphilum ATCC15354; lanes 8 and 9, B. taipei ATCC13744; lanes 10 and 11, B. thiogenitalis ATCC19240; lanes 12 and 13, C. glutamicum ATCC13869; lanes 14 and 15, C. glutamicum ATCC13032; lane 16, C. ammoniagenes ATCC6872
We tested 16 strains of coryneform bacteria for the secretion of S. mobaraensis pro-TG. As shown in Fig. 1 and Table 1, all of the bacteria secreted pro-TG. We have previously shown that the C. glutamicum system has two advantages for heterologous protein secretion: a lack of proteolytic activity in C. glutamicum culture supernatants and a small amount of protein secretion by the bacteria themselves (Date et al. 2006; Kikuchi et al. 2003). As a result, the purification of secreted heterologous proteins is relatively easy. From the results shown in Fig. 1, it is clear that the other coryneform bacteria share these advantages for heterologous protein secretion: their culture supernatants contain little endogenous protein and no degradation of pro-TG is detected in the supernatants.
Effect of signal peptide sequence on pro-TG secretion
Next, to investigate whether the secretion efficiency depended on the signal peptide sequence, we examined the effect of using the CspA signal peptide from C. ammoniagenes ATCC6872 (Fig. 2a) or the CspB signal peptide from C. glutamicum ATCC13869 (Fig. 2a) for the pro-TG secretion. C. ammoniagenes ATCC6872, C. acetoacidophilum ATCC13870, C. glutamicum ATCC14020, M. ammoniaphilum ATCC15354, and C. callunae ATCC15991 could efficiently secrete the pro-TG using the CspA signal peptide from C. ammoniagenes ATCC6872 in this study. These strains and C. glutamicum ATCC13869 were transformed with pPKPTG1 harboring the CspB signal peptide from C. glutamicum ATCC13869. As shown in Fig. 2b, the substitution of the CspB signal peptide had little effect on pro-TG accumulation in C. acetoacidophilum ATCC13870, C. glutamicum ATCC14020, and M. ammoniaphilum ATCC15354. However, in C. callunae ATCC15991, the pro-TG accumulation (403 mg/l) was approximately halved (746 mg/l) while in C. ammoniagenes ATCC6872, it was about 2.8-fold higher.
Fig. 2Effect of signal peptide sequence on pro-TG secretion in coryneform bacteria. a Amino acid sequence alignment of C. ammoniagenes CspA signal peptide and C. glutamicum CspB signal peptide using Vector NTI software. Identical and similar residues are shown on black and gray backgrounds, respectively. b Accumulation of secreted pro-TG by coryneform bacteria carrying pPSPTG1 or pPKPTG1. The solid bars indicate pro-TG accumulation using the CspA signal peptide from C. ammoniagenes ATCC6872, and the open bars indicate pro-TG accumulation using the CspB signal peptide from C. glutamicum ATCC13869
The amount of pro-TG accumulation by C. ammoniagenes ATCC6872 using the CspB signal peptide from C. glutamicum ATCC13869 was about 2.8-fold higher than with its own CspA signal peptide (Fig. 2b). Interestingly, the difference between the effects of the two signal peptides was the opposite in C. glutamicum ATCC13869: the amount of pro-TG accumulation using the CspA signal peptide from C. ammoniagenes ATCC6872 was about 1.5-fold higher than with the CspB signal peptide from C. glutamicum ATCC13869 (Kikuchi et al. 2003). We found the same to be true for all of the heterologous proteins examined so far (data not shown). The reason for the differing effects of these signal peptides is unclear.
Cultivation of C. ammoniagenes ATCC6872 carrying pPKPTG1 in a jar fermentor
As C. ammoniagenes ATCC6872 carrying pPKPTG1 was the best producer strain in our study, we subjected it to a fermentation test in a jar fermentor under the conditions described in the “Materials and methods” section. The MMTG-J medium contained 120 g/l glucose; after the consumption of glucose for 71 h under these conditions, the optical density at 625 nm reached approximately 140, and the dry cell concentration was approximately 50 g/l. During the exponential growth phase, little pro-TG accumulated in the culture supernatant; however, the rate of accumulation increased linearly after entry into the stationary phase, and approximately 2.5 g/l pro-TG was produced (Fig. 3). The N-terminal amino acid of the secreted pro-TG was Asp, as in native pro-TG (Kikuchi et al. 2003), demonstrating that the CspB signal peptide had been correctly processed in C. ammoniagenes ATCC6872. We then incubated the culture supernatant with purified SAM-P45, which is a subtilisin-like protease from Streptomyces albogriseolus, for 2 h at a 100:1 ratio of pro-TG to SAM-P45 in order to process the pro-domain (Kikuchi et al. 2003). The pro-TG secreted by C. ammoniagenes ATCC6872 carrying pPKPTG1 was cleaved by purified SAM-P45 to the C-side of 41-Ser of the pro-domain and converted to an active form and the conversion yield of this enzymatic process was approximately 100%. The specific activity of the purified active-form TG, with additional Phe-Arg-Ala-Pro residues, was similar to that of the native TG (about 23 U/mg; Kikuchi et al. 2003).
Fig. 3Jar fermentation experiment employing C. ammoniagenes ATCC6872 carrying pPKPTG1. Growth (optical density at 625 nm), glucose concentration, and pro-TG accumulation are indicated by open circles, open squares, and closed triangles, respectively
The results of the fermentation experiments in the 1-l jar fermentors showed that the pro-TG secreted by C. ammoniagenes ATCC6872 was correctly folded. Thus, protein production using C. ammoniagenes ATCC6872 could be useful for industrial scale protein production.
|
[
"transglutaminase",
"corynebacterium glutamicum",
"corynebacterium ammoniagenes",
"protein secretion"
] |
[
"P",
"P",
"P",
"P"
] |
Bioinformation-1-10-1896059
|
Functional profiling and gene expression analysis of chromosomal copy number alterations
|
Contrarily to the traditional view in which only one or a few key genes were supposed to be the causative factors of diseases, we discuss the importance of considering groups of functionally related genes in the study of pathologies characterised by chromosomal copy number alterations. Recent observations have reported the existence of regions in higher eukaryotic chromosomes (including humans) containing genes of related function that show a high degree of coregulation. Copy number alterations will consequently affect to clusters of functionally related genes, which will be the final causative agents of the diseased phenotype, in many cases. Therefore, we propose that the functional profiling of the regions affected by copy number alterations must be an important aspect to take into account in the understanding of this type of pathologies. To illustrate this, we present an integrated study of DNA copy number variations, gene expression along with the functional profiling of chromosomal regions in a case of multiple myeloma.
Background
Genomic copy number alterations such as gains or losses of chromosomal regions have been shown to be on the basis of many human pathologies. Classical approaches
to characterize these genetic aberrations used comparative genomic hybridisation (CGH), in which genomic DNA was hybridised to metaphase
chromosomes. [1] Recently, the use of different types of microarrays to directly study genomic variations in DNA
copy number is becoming more and more popular. Such massive genomic approaches are known as array comparative genomic hybridisation, or Array CGH. [2]
These new technologies along with the use of expression arrays allow for a highly accurate characterisation of the dependence of gene expression on alterations in genomic
copy number. [3]
As in many genome-scale methodologies data analysis and, in particular, the biological interpretation of the results constitutes a well-known bottleneck. Specific
problems related to the analysis of Array CGH can be circumscribed mainly to two types: appropriate mapping and visualisation of the data onto the chromosomes, and
efficient copy number estimation. This last aspect has been the motivation for a number of analytical approaches recently proposed [4], that can be considered
the first generation of algorithms for Array CGH analysis. Obviously, copy number variations are expected to have a strong effect on gene expression. [5,6] Nevertheless, the ultimate aim of studies of copy
number chromosomal alterations is to understand what is the effect produced in functional terms. In the classical vision one or a few key genes are the causative factors
for the this type of pathologies, and the problem consisted in identifying such genes within the region amplified or deleted. The existence of regions in the
chromosomes containing coexpressing genes [7] which, in addition, are functionally related has recently been
reported even in higher eukaryotes. [8] Actually, regional arrangements of genes have found to be
regulated not only by copy number alterations but also by different mechanisms such as epigenetic modifications. [9] This reinforces
the functional role of chromosomal regions including groups of functionally related genes and its possible impact on diseases such as cancer. [10]
These observations give credence to a new vision in which chromosomal alterations can be causing effects not by altering single key genes but by acting on
complete molecular sub-systems such as pathways of functionally related genes. Recently, different approaches have focused on the functional aspects of the
results of microarray experiments. [11,12] Nevertheless,
the possible functional significance at regional level of copy number alterations has been largely ignored. Here we present a combined approach to the study of copynumber
alterations, gene expression and functional profiling, exemplified in a case of multiple myeloma. [13]
Methodology
Functional profiling of Array-CGH experiments under this new perspective would require of three steps: 1) detection of regions with copy number variations (the
origin of the disease), 2) detection of regional alterations in gene expression (the causes of the disease) and 3) analysis of enrichment in functional terms in the detected
regions (the consequences of the alteration or the functional basis of the disease). While copy number alterations can be detected by means of different
methods, alterations in the levels of gene expression are not always easy to be detected using the typical methods (t-test or similar) due several factors such as small
sample sizes. For this reason here we will only use plots to visualize the effect of one variable (copy number) into the other one (expression level). The third step, the functional profiling, becomes then the most important
aspect of the analysis given that it will provide a functional explanation of the molecular basis of the disease caused by copy number alterations.
Detection of copy number alterations
We have used a segmentation method which is a variant of the circular binary segmentation method [14], for copy number change detection (isowindow).
The isowindow method tries to identify boundaries between regions with a significant change in the values of intensity of hybridisation of the probes by some
consecutive steps. Firstly a t-test is used to determine differences between regions around all possible boundary points. Once all the candidate boundaries have
been selected (a liberal p-value is used at this stage) there are sorted from small to high minimum p-values. In a second step the boundary candidates in the list with
overlapping neighbourhoods are filtered to obtain a refined list of optimal non-overlapping boundary candidates. All the p-values are recalculated for the
redefined neighbourhoods and a more stringent threshold is applied here. Finally, regions at both sides of each boundary candidate are again compared with a t-test. If
they are not significantly different in their average hybridisation values, then they are merged as a unique region. Otherwise they define two regions with different
copy number value. This is a simple and quick procedure that allows for easily changing from fine to coarse resolution by modifying the thresholds for the p-values.
We have compared isowindow to other two methods for breakpoint detection, GLAD [15] and circular binary
segmentation (CBS) [14], which are among the best performers. [4]
In the GLAD method a likelihood function with weights determined adaptively is used to solve the copy number estimation problem locally based
on data smoothed. Then, the algorithm finds, for each probe, the maximal neighbourhood in which the local constant assumption holds. Each of the constant pieces
of the line define a block of probes with similar copy number among them and different copy number from that of the nearby regions. On the other hand, the CBS
method selects firstly a segment of the data (a group of probes that are all consecutively arranged in the genome or in a chromosome). The copy number measures of the
probes in that segment are compared to those in the reminder dataset using a t-statistic. Hence, the method can distinguish whether the segment chosen has a copy
number that is higher or lower than the overall copy number in the data, assumed to be the normal reference. This scheme is iterated exhaustively for all possible
segments in the dataset, spotting those that correspond to regions of altered copy number.
An approximation to the relative performances of the methods used was obtained by means of simulated data sets. Such datasets were generated by means of a
piecewise constant function plus random alterations normally distributed with mean value and three different levels for the standard deviation (corresponding to noise
levels 0.2, 0.5 and 1). A mean value of 0 would correspond to a normal region, without copy number alterations, while mean values lower and higher would
correspond to deletions or amplifications at different degrees, respectively. Amplified and deleted regions of different sizes are randomly situated within the simulated
normal chromosome and the methods have to locate them at different noise levels. The method proposed here performs at least as well as the GLAD and CBS (Table
1) while being more efficient in terms of runtimes. Isowindow shows a better performance in finding small amplicons.
Functional profiling of regions with copy number alterations
The final aim of a Array-CGH experiment is to find a molecular explanation for the effects of the detected copy umber alterations. The interpretation of genomescale
data is usually performed in two steps: in a first step genes of interest are selected in this case because they are located in the amplified (or lost) region detected.
In a second step, the selected genes of interest are compared to the background (here the rest of genes in the chromosome) in order to find enrichment in any
functional category (gene ontology, KEGG pathways, etc.) This comparison to the background is required because otherwise the significance of a proportion (even
if high) cannot be determined. Different approaches have been developed to this end. [11] Here we will use the
FatiGO+ (16) program, which uses a Fisher's exact test to determine the enrichment in different functional categories including gene ontology, KEEG pathways,
Interpro functional motifs, Swissprot keywords and some regulatory elements such as transcription factor binding sites or other regulatory motifs. [17]
Discussion
We have implemented all the described functionalities in a program, ISACGH (an acronym for In Silico Array CGH), which is used to illustrate the concept of
functional profiling of CGH arrays with an example of multiple myeloma (MM), an incurable form of haematological neoplasia.
Nine MM cell lines were obtained from the DSMZ (Deuche Sammlung von Mikroorganismen und Zelkuturen GmbH, Braunschweig, Germany) and were
cultured under recommended conditions. DNA and RNA were extracted using supplier's protocols. Microarray assays were performed using the CNIO OncoChip,
which contains 7657 different cDNA clones of cancer related genes. [18] CGH experiments onto cDNA arrays
and hybridisation were performed as described in [13] and quantified using the GenePix Pro 5.0 software (Axon
Instruments Inc., Union City, CA). Cy3/Cy5 ratio values were normalized using the DNMAD tool from the GEPAS [19,20
,21] and the resulting data were transformed to log2 ratios. Our purpose was to identify
any possible region that contained copy number gains (amplifications), to study the expression of the genes included in that particular region and to understand the
possible functional consequences of such alterations.
Using the segmentation method as implemented in the ISACGH we could detect a putative amplicon in the chromosome 18 (which remained undetected with both
GLAD and CBS, because of the low density of the array, although the effect would have been the same in a high density arrays with a small amplicon) The figure shows
the region (left) and the slight, although appreciable, differences in gene expression levels within the amplicon (right).
A unique feature offered by ISACGH is the possibility of obtaining a functional profile of the detected chromosomal regions. When the amplicon is analysed
through the FatiGO+ program [16,17] a number of GO
terms arise as over-represented in the genes contained in such region. Thus, the GO terms regulation of cellular process (GO:0050794) and regulation of physiological
process (GO:0050791) were significantly overrepresented in the amplicon (FDR adjusted p-value=0.0336). Genes annotated with these terms were: BCL2,
MALT1, NEDD4L, MBD2, TNFRSF11A and TCF4. Some of them have annotations at more detailed levels in GO, although the number of genes was too small as to
produce statistically significant results. For example BCL2 and MALT1 are annotated as negative regulation of programmed cell death (GO:0043069). These
observations suggest that some processes altered, that ultimately lead to diseases, are not produced by the deregulation of one unique gene, but are the combined
result of simultaneous deregulations of genes involved in a pathway or a particular biological function. In addition, these findings stress the importance of the use of
functional profiling methods for the proper understanding and interpretation of the results of the genome-scale experiments. This unique feature included
in ISACGH is of extreme importance since growing evidence suggests the existence of clusters of functionally related genes in the chromosomes [8]
and the possible impact on diseases such as cancer. [10]
Although ISACGH [22] can be used alone, it is tightly integrated in the GEPAS package. [19,
21,23] GEPAS, that stands for Gene Expression Profile Analysis Suite
(GEPAS), constitutes one of the most complete resources for microarray data analysis available over the web. GEPAS includes facilities for normalisations, clustering,
gene selection, predictors and functional profiling. Thus, different operations (including pre-processing or normalization) can directly be performed within the
same environment, without the necessity of any file reformatting step.
Conclusion
Despite a number of applications dealing with the estimation of genomic copy number have been recently published [4],
there are different aspects of the analysis of Array CGH data that have been poorly addressed or even ignored. Recent evidences strongly support the
existence of regional arrangements of functionally related genes [8], with obvious consequences for the understanding of diseases
characterised by copy number alterations, such as an important number of cancers. [10]
This fact reduces the validity to the classical vision, in which one or a few key genes would be the causative factors of the disease, and urges to take into
consideration the functional dimension in the interpretation of the effects of copy number alterations. In this new scenario, the deregulation of blocks of
functionally related genes located in the chromosomal regions with copy number alterations would be behind the disease phenotype.
The methods for functional profiling have proven in many scenarios its usefulness. An obvious challenge is to increase our knowledge in different aspects of function
and cooperation between genes in order to be able of applying this methods in a way that allows us to unravel new unknown functional aspects of the biology of the
cell and their connections to pathologies.
|
[
"function",
"profile",
"gene expression",
"chromosomal copy number"
] |
[
"P",
"P",
"P",
"P"
] |
Acta_Neuropathol-4-1-2270353
|
Neuronal pentraxin II is highly upregulated in Parkinson’s disease and a novel component of Lewy bodies
|
Neuronal pentraxin II (NPTX2) is the most highly upregulated gene in the Parkinsonian substantia nigra based on our whole genome expression profiling results. We show here that it is a novel component of Lewy bodies and Lewy neurites in sporadic Parkinson’s disease (PD). NPTX2 is also known as the neuronal activity-regulated protein (Narp), which is secreted and involved in long-term neuronal plasticity. Narp further regulates AMPA receptors which have been found to mediate highly selective non-apoptotic cell death of dopaminergic neurons. NPTX2/Narp is found in close association with alpha-synuclein aggregates in both substantia nigra and cerebral cortex in PD but unlike alpha-synuclein gene expression, which is down-regulated in the Parkinsonian nigra, NPTX2 could represent a driver of the disease process. In view of its profound (>800%) upregulation and its established role in synaptic plasticity as well as dopaminergic nerve cell death, NPTX2 is a very interesting novel player which is likely to be involved in the pathway dysregulation which underlies PD.
Introduction
Neuronal pentraxins constitute a family of proteins that are homologous to C-reactive and acute-phase proteins in the immune system and are thought to be involved in activity-dependent synaptic plasticity [1, 11, 32]. The neuronal pentraxins (NPTX1, NPTX2) and the neuronal pentraxin receptor were identified as synaptic proteins that bind to affinity columns of the snake venom toxin, taipoxin which presynaptically blocks neurotransmission, and the luminal calcium-binding protein TCBP49 [15]. NPTX2 [14] is also known as neuronal activity-regulated pentraxin (Narp), which was first identified as an immediate early gene responding to the induction of seizures in rat hippocampus [31]. The expression of this protein is regulated by synaptic activity and it induces the formation of new excitatory synapses and the regulation of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)-type glutamate receptor clustering at established synapses [7, 22, 23]. Importantly, dopaminergic neurons are selectively susceptible to excitotoxicity mediated via AMPA receptor activation with the resulting cell death being non-apoptotic [5].
A second line of research that makes NPTX2 a highly interesting gene in the context of Parkinson’s disease (PD) research concerns its expression in hypothalamic nerve cells that co-express hypocretin/orexin [3] and which are prodigiously affected by the neurodegeneration in PD. These nerve cells have been implicated in the sleep disturbances that are often found in PD patients [9, 30].
Using a whole genome expression dataset of the human Parkinsonian nigra [19], we have identified NPTX2 as the most highly upregulated gene in a cohort of sporadic PD cases. Here we demonstrate that NPTX2 labels Lewy bodies in paraffin sections and that there is dysregulation of NPTX2 also in PD frontal cortex.
Materials and methods
Tissue samples from human SN and frontal cortex
Some of the sporadic PD cases used in this study and several controls have been described previously [6, 19, 20]. Information on the cases employed for this project is provided in Supplemental Table. Brain tissue was obtained from the UK Parkinson’s Disease Society Brain Bank at Imperial College London. The diagnostic work was performed in the Department of Neuropathology following international neuropathological consensus criteria for PD (http://www.ICDNS.org) which require use of alpha-synuclein as a molecular diagnostic marker. Ethics committee approval was available for all human tissue work of this study. For the dissection of the substantia nigra tissue (medial and lateral nigra were prepared separately), the portion of the midbrain containing the substantia nigra was separated from the brainstem distal to the mamillary bodies (45° angle cut) at the level of the third cranial nerve and the superior colliculi. A slice approximately 5 mm in depth was prepared containing both SN and the cross-sectioned red nucleus at the level of its greatest diameter. Then, the SN was dissected from the surrounding nuclei and the pars reticulata. Once the SN was isolated, it was divided into two portions, the medial SN and the lateral SN. In all cases where frontal lobe was used, the superior frontal gyrus was sampled at the same level.
Quantitative real-time polymerase chain reaction
Total RNA was extracted from PD and control cases and transcribed using the RETROscript® kit (Applied Biosystems, Warrington, UK) following the manufacturer’s protocol. Primers for the internal control genes were as before [19]. Predesigned Taqman primers for NPTX2 (Hs00383983_m1) were purchased (Applied Biosystems) and Taqman mastermix (Applied Biosystems) was used. Each reaction was run in triplicate using an ABI PRISM™ 7700 Sequence Detector (Applied Biosystems). For every gene tested, a negative control was run without cDNA template. Data analysis for real-time PCR (DART-PCR Version 1.0; [25]) was used to determine relative differences in NPTX2 mRNA expression between PD cases and controls (also see Supplemental Figure 1).
In situ hybridization (ISH)
ISH was performed using radioactive oligonucleotide probes as described previously [18, 29] Oligonucleotides were designed using NetPrimer software (Premier, BioSoft International, http://www.premierbiosoft.com). Two oligonucleotides complementary to NPTX2 mRNA at nts 899–932 (AGCCACAGGCAGATGGTGAAGGCGTACAGCTCAG—probe A) and nts 1364–1397 (TCGACGTTATTGTCCACCCACGG-GATGATGTTGC—probe B) were synthesized (Sigma Genosys, Poole, UK). Oligonucleotides were radioactively end-labelled with 35S-dATP (Perkin-Elmer, UK) using terminal deoxynucleotidyl transferase (Promega, UK).
Snap frozen, unfixed post-mortem tissue blocks of the frontal cortex and SN were cut on a cryostat (12 μm), collected onto Superfrost Plus slides (VWR) and stored at −80°C. Before use, tissue sections were fixed for 5 min in fresh 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Pre-hybridisation treatment included acetylation in 0.25 M acetic anhydride/0.1 M triethanolamine, dehydration in graded alcohol (70–100% ethanols) and delipidation in 100% chloroform. Hybridisation with the oligonucleotide probe was performed overnight at 37°C. Post-hybridisation sections were washed in standard saline citrate solutions with increasing stringencies. The sections were dehydrated rapidly through graded alcohols (70–100% ethanol), air-dried, dipped in autoradiographic emulsion (LM1, GE Healthcare, Amersham, UK) and exposed for 5–6 weeks. The autoradiographic emulsion was developed, sections were counterstained with toluidine blue (VWR) for Nissl substance and mounted for visualisation using bright field microscopy for the counterstain and epi-polarized illumination for the silver grains of the autoradiographic emulsion. Control slides run to check for specificity of hybridization included competition with an excess of oligonucleotide (250 times) and the use of two probes to different portions of the mRNA of interest which produced identical patterns of signal. Black and white photographs of emulsion-coated sections were taken using a microscope provided with a Hamamatsu C4742–95 digital camera (Herrsching, Germany) and HiPic software (Herrsching) to capture images. Digital colour images were captured with a Retiga 1300 monochrome 12-bit camera using a color option provided by a RGB-HM-S filter (QImaging, Burnaby, BC, Canada) and QCapture 1.1.6 software (QImaging). Adobe Photoshop was used to assemble images.
Immunohistochemistry
Paraffin embedded tissue sections (8 μm) were used for all stainings. No antigen retrieval was required for NPTX2 whereas for alpha-synuclein (SNCA) antibody staining sections were treated with 80% formic acid for 60 min [24]. Adjacent sections from SN were stained with goat polyclonal anti-NPTX2 antibody (NP2 N-20, sc-12125; Santa Cruz Biotechnology Inc.; 1:25) and mouse monoclonal anti-SNCA (BD Transduction Laboratories, catalogue number 610787; 1:1,000) overnight (4°C). Secondary antibodies used were gG-HRP (sc-2020; Santa Cruz Biotechnology Inc.; 1:100) and biotinylated horse anti-mouse (B-2000; Vector Laboratories, 1:100), respectively. The ABC kit (Elite 6100, Vector Laboratories) was used for visualisation using 3,3′-diaminobenzidine tetrahydrochloride (0.02%; Sigma, UK). Sections were counter-stained in Mayer’s haemalum.
For fluorescence microscopy, anti-SNCA (1:1,000) and anti-NPTX2 (1:25) were added to the sections sequentially (each was incubated overnight at 4°C). For the detection of the primary antibodies, sections were rinsed in 0.1 M Tris buffered saline (TBS) and then incubated with either the goat anti-rabbit IgG tagged with Alexa Fluor 546 (red) for 1 h at room temperature or the biotinylated anti-mouse antibody (Vector Laboratories 1:100) for 1 hour followed by incubation with streptavidin Alex Fluor 488 (green) for 1 h at room temperature. Finally, tissue sections were rinsed in TBS and incubated in sudan black for 10 min before mounting in 80% Tris-buffered glycerol.
Western blotting
Snap frozen midbrain tissue samples containing SN from PD and control cases were homogenized with RIPA buffer (pH 8), heated to 70°C for 10 min and subsequently centrifuged (13,000 rpm, 10 min). The protein supernatant was removed and the concentration of protein determined (BIO-RAD protein reagent; BIO-RAD, UK) using a spectrophotometer. Denaturing gels were run using 30 μg of protein from each sample in 19.5 μl RIPA buffer, 3 μl β-mercaptoethanol and 7.5 μl lithium dodecyl sulphate sample buffer (Invitrogen, UK). NuPAGE Bis–Tris 4–12% polyacrylamide gels (Invitrogen) were run in NuPage SDS running buffer (Invitrogen). Gels were blotted onto nitrocellulose membranes (Invitrogen), immersed in NuPage LDS transfer buffer (Invitrogen) in a cold room (4°C) overnight applying 30 V. Successful protein transfer was verified by staining the membrane in Ponceau S (VWR, UK) for 10 min. Prior to overnight incubation in polyclonal goat antibody to NPTX2 (N-20), membranes were washed in phosphate buffered saline (PBS) containing 0.1% Tween 20 and subsequently rinsed in PBS/0.1% Tween 20 containing 5% skimmed milk (blocking step) for 30 min. Incubation with the secondary antibody coupled to horseradish peroxidase (1:5,000; Vector Labs, UK) was carried out for 1 h at room temperature. Visualization of antibody binding using enhanced chemilluminescence (ECL; Amersham Biosciences, UK) was performed according to the manufacturer’s instructions.
Semi-quantitative ratings
The NPTX2 labelled SN sections were examined by two independent observers blinded to the identity of the respective cases. The occurrence of NPTX2 immunoreactive profiles in the SN was scored for the overall burden of NPTX2 immunoreactive deposits using a seven-point scale ranging from 0–0.5 to 3 (absent to severe). Correlation analysis (Spearman; Sigmatstat 2.03) was conducted within the PD cohort for NPTX2 mRNA expression (log 2 microarray data) and the corresponding semi-quantitative immunocytochemical results.
Results
NPTX2 is a novel component of Lewy bodies
We demonstrate here that NPTX2 protein in the Parkinsonian substantia nigra localizes to Lewy bodies, Lewy neurites and some glial cells. NPTX2 is the most upregulated gene in our microarray data set [19]. It shows the greatest difference between disease and control nigrae (probe 213479_at, P = 4.60051E−07, differential expression, DElog 2, 3.0036217 which is equivalent to >800%). Many neurons show more than one NPTX2 immunoreactive LB (Fig. 1a). Western blotting confirmed specificity of the antibody revealing a band, which corresponds to the NPTX2 monomer (47 kDa; Fig. 1b). Very strong upregulation of NPTX2 mRNA expression in PD was confirmed by means of quantitative real time-PCR (Fig. 1c).
Fig. 1Neuronal pentraxin II (NPTX2) is strongly upregulated in substantia nigra and cerebral cortex of PD patients. a NPTX2 immunocytochemistry demonstrating specific staining of Lewy bodies in a dopaminergic neuron. b Western blotting confirms specificity of the anti-NPTX2 antibody. c qRT-PCR shows highly increased levels of NPTX2 mRNA in substantia nigra and, to a lesser extent, in frontal cortex. The arbitrary units along the ordinate represent relative fold changes [25]. The control value is 1; error bars indicate SEM. d, e In situ hybridisation demonstrates NPTX2 mRNA expression in nerve cells as well as glia in both substantia nigra and frontal cortex. Labelling is also found over non-pigmented neurons (arrow in d, SN). The arrows in e mark glial cells (cortex). In addition, there is autoradiographic signal in the neuropil (asterisks) which would be in keeping with the presumed dendritic expression of NPTX2 mRNA. f NPTX2 immunocytochemistry of cortical neuropil reveals occasional intercellular patches of “arborised” labelling also possibly suggestive of a dendritic localisation (asterisks). The arrow points to a NPTX2 immunoreactive glial cell. Counterstaining in a and f, haemalum, and toluidine blue in d. Scale bars: 20 μm (a, f), 50 μm (d), and 100 μm (e)
Localisation of NPTX2 and SNCA protein
In situ hybridisation revealed expression of NPTX2 mRNA in nerve cells as well as glia in both SN and frontal cortex, respectively (Fig. 1d, e). NPTX2 immunoreactive LBs were also observed in the frontal cortex and there were ramified (“arborised”) NPTX2 positive profiles of cell processes in the neuropil (Fig. 1f). Analysis of adjacent tissue sections demonstrated co-localisation of NPTX2 and SNCA in the same nerve cells (Fig. 2a, b). This was confirmed by dual immunofluorescence (Fig. 2c–e). Overall, NPTX2 immunoreactive neuronal deposits mapped very closely to alpha-synuclein aggregates, both cellularly and subcellulary, and NPTX2 was also observed in incidental Lewy bodies (for a summary of cases used see Supplemental Table of Electronic Supplementary Material). In contrast, no NPTX2 immunoreactive intracellular aggregates were observed in two cases of tauopathy, i.e. progressive supranuclear palsy, where alpha-synuclein deposits are also not found. However, not all SNCA immunoreactive profiles were NPTX2 positive, i.e. the number of SNCA labelled nerve cells outnumbered that of NPTX2 stained ones (about 2–3:1). In addition, in some nerve cells, SNCA immunopositivity was observed as small granular cytoplasmic deposits (“aggresomes”) that lacked NPTX2 immunoreactivity (Fig. 2a, b). However, NPTX2 immunoepitopes were found to decorate the halo of LBs more distinctly than alpha-synuclein (Fig. 3). Strongly NPTX2 immunoreactive Lewy neurites (Fig. 3) were also found. Cases with high levels of NPTX2 protein expression showed a larger alpha-synuclein burden. We estimate that about one-third of the alpha-synuclein immunoreactive protein deposits also stain for NPTX2. Labelling of normal appearing axons by NPTX2 was widespread in both PD (Fig. 3) and control brains. Glial cytoplasmic (Papp-Lantos) inclusions of multiple system atrophy (MSA) were NPTX2 positive (not shown) and in PD a sizeable number of glial cells expressed NPTX2 protein that appeared to be SNCA negative (Fig. 3). Correlation analysis for NPTX2 microarray expression data and the corresponding semi-quantitative immunocytochemical case ratings showed a significant correlation (r = 0.970, P < 0.001; Table 1). However, there was no association between NPTX2 expression and duration of the disease.
Fig. 2Co-localisation of NPTX2 and alpha-synuclein. a, b NPTX2 and alpha-synuclein immunolabelling of adjacent tissue sections demonstrates significant but not complete overlap between the two proteins. The large arrow in a denotes the same Lewy body-containing dopaminergic nigral neuron. Small granular cytoplasmic deposits showing alpha-synuclein immunoreactivity in some nerve cells (arrows in b) are NPTX2 negative. These aggresomes are sometimes difficult to distinguish from neuromelanin granules based on staining alone but their cytoplasmic distribution is also different from the latter. In contrast, Lewy neurites are double-labelled (arrow head in b). Neuromelanin (asterisks in a) is unstained. c, d, e Dual immunofluorecence demonstrates co-localisation of neuronal pentraxin II (red, c) and alpha-synuclein (green, d). Double-labelled structures (arrows in c) light up in yellow in the overlay (e). Scale bar: 40 μm (a, b), and 80 μm (c–e), respectivelyFig. 3Morphology of neuronal pentraxin II immunoreactive deposits in the Parkinsonian substantia nigra. a Lewy bodies are labelled which often show a crennelated surface (also see c). Arrows of the inset point to NPTX2 labelled connections/“bridges” between small Lewy bodies. Cross-sectioned axon bundles in white matter are numerous (longer arrows). b Arrows point to Lewy neurites in the substantia nigra. c NPTX2 is present in pre-Lewy bodies and glial cells. Labelling can be seen inside the cytoplasm of nerve cells in the neighbourhood and often directly attached to mature Lewy bodies (long arrows). Note the crenellated surface of compacted NPTX2 deposits (centre and Lewy body at the bottom of the figure) due to hair-like projections, which radiate into the cytoplasm. NPTX2 immunoreactivity is also present in glial cells. A subset of astrocytes as well as possibly microglia appear to be labelled (small arrows). The short arrow in the lower left points to what appears to be nuclear immunoreactivity. Scale bars: 40 (a), 50 (b) and 30 (c) μm, respectivelyTable 1Correlation analysis for NPTX2 microarray expression data and the corresponding semi-quantitative immunocytochemical ratings for eight of the PD casesCase numberExpression NPTX2 mRNA (log 2)Semi-quantitative ratings NPTX2 IR in SN27.278105587139.800657274368.7396793641.578.445081081.593.4182034440.5127.9801395781148.9652648212156.4144958231
Discussion
The most significant finding of this study is the striking presence of NPTX2 in classical Lewy bodies and in Lewy neurites. Experimental studies have demonstrated that Narp, the rat homologue of NPTX2, is important in neuronal development, promoting neuronal migration, synapse formation [1, 8] and neurite outgrowth [31]. Studies in neuronal pentraxin I and neuronal pentraxin II knockout mice have demonstrated a role in early synaptic refinements in the retina and dorsal lateral geniculate nucleus during development [1]. It has also been suggested that Narp’s role in synaptic plasticity may underlie the long-lasting aversive effects of drug withdrawal [10, 26]. These functions of the NPTX2 gene product and its dendritic expression in the cerebral cortex [17] are of special interest in relation to our finding that NPTX2 immunoreactivity was also occasionally present in the cortical neuropil where it had a ramified appearance reminiscent of dendritic arbors.
There is limited knowledge on the NPTX2 pathway (Fig. 4). Known physiological functions of NPTX2 include a role in the induction of AMPA receptor clustering [7]. NPTX2, NPTX1 and the GRIA3 subunit of AMPA receptors are all upregulated in our dataset. All other AMPA subunits and the receptor for NPTX2, NPTXR are downregulated. Two additional genes whose functions may be associated with AMPA receptors in dopaminergic neurons are also highly dysregulated in PD, NSF (N-ethylmaleimide sensitive fusion protein; 202395_at, P = 1.18252E−05, DElog 2 = −1.389396766) and Calcyon (dopamine receptor D1 interacting protein, 219896_at, P = 4.12902E−06, DElog 2 = −2.049585553). NSF binds the AMPA receptor GluR2 subunit and acts to disrupt GluR2-PICK1 interactions [28]; it may play a role in AMPA receptor trafficking and stabilization. Calcyon is co-expressed in a number of D1 dopamine receptor-positive neurons in brain and, like D1 dopamine receptors, is found in dendritic spines [16]. We also found NPTX2 in glial cells possibly including microglia. However, staining of glia was only present in a subpopulation of cells while neuronal cell processes appeared to be consistently labelled. Why exactly NPTX2 accumulates in Lewy bodies remains unclear at present. However, it is conceivable that NPTX2 is involved in the uptake of synaptic material. Synaptic dysfunction in PD is very significant as suggested by our microarray experiments (unpublished data). The presence of NPTX2 in only a subset of Lewy bodies and its association with their “crenellated surface” shown in Fig. 3 as well as NPTX2 immunoreactive “bridges” between Lewy bodies might indicate a role of NPTX2 in Lewy body growth.
Fig. 4NPTX2 pathway obtained by means of the Resnet 5.0 eukaryotic database of molecular interactions (Ariadne) overlayed with microarray expression values. Changes of PD relative to normal are indicated in red (increased in PD) and blue (downregulated), respectively. Abbreviations: AMPA alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; GRIA1 glutamate receptor, ionotropic, AMPA 1; GRIA2 glutamate receptor, ionotropic, AMPA 2; GRIA3 glutamate receptor, ionotrophic, AMPA 3; GRIA4 glutamate receptor, ionotrophic, AMPA 4; NPTX1 neuronal pentraxin I; NPTX2 neuronal pentraxin II; NPTXR neuronal pentraxin receptor. Green line, regulation; purple line, binding. The dotted grey lines connect to a cell process whereas the solid lines connect to a molecule. The grey colour indicates that the effect and/or mechanism are unknown. References for all protein interactions shown are provided in the Electronic Supplement to this figure
Many PD patients also show non-motor symptoms such as hallucinations, depression, autonomic dysfunction and sleep-disturbances. All these symptoms are present in narcolepsy where there is loss of orexin (hypocretin) producing nerve cells in the hypothalamus [2, 30]. It is noteworthy therefore that the same population of neurons shows massive cell loss in PD as shown recently [9, 30]. In addition, Thannickal et al. observed a significant correlation with the clinical stage of PD [30]. It is intriguing that orexin neurons of the hypothalamus demonstrate robust expression of Narp in the rat [27].
The relationship between NPTX2 and cell death deserves further study. All subunits of the NMDA, AMPA and kainate classes of glutamate receptors and a set of NMDA receptor-associated intracellular proteins are expressed in the human substantia nigra pars compacta, and may play a role in glutamate regulation of dopaminergic activity and/or release [21]. AMPA in turn has been discovered to be a mediator of selective non-apoptotic cell death in dopaminergic neurons [4]. The latter differs ultrastructurally from both neuronal apoptosis and typical excitotoxicity. Importantly, this mode of cell death is independent of caspase activity and does not have the morphological appearance of classical apoptosis. The findings are thus in keeping with our old postulate that nigral neurons are likely to die of a mechanism other than classical apoptosis and that aposklesis (“withering”) may be a more appropriate term [12, 13]. The involvement of NPTX2 in cell death mechanisms is additionally supported by studies on the effect of ceramide, which activates TNF-pathways [4]. Therefore, NPTX2/Narp has to be considered a new important gene in PD dysregulation which may be linked to both motor dysfunction caused by nigral dopaminergic nerve cell death in the midbrain and dementia due to synaptic changes in the cerebral cortex.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Supplemental Figure 1. Cluster analysis is a method to organize genes in a microarray dataset into groups (clusters) of similar expression values. We have used ArrayAssist 5.5 software (Stratagene) to produce this image which shows that NPTX2 is not the only gene that is significantly up-regulated in the Parkinsonian substantia nigra compared to control and whose expression varies within the cohort. The latter is important because unlike a number of other highly dysregulated genes which show greater similarity across cases and therefore higher p values, NPTX2 fits better with both the clinical and neuromorphological notion of different disease stages in different cases. There was a good correlation between expression values and immunocytochemical results (Table 1). Abbreviations: CON, control; LN, lateral nigra; MN medial nigra; PD, Parkinson’s disease. Codes CON10, CON2, CON3, CON9, PDC1, PD01- PD36 relate to study cases 19, 17, 18, 23, 16, 1-15, respectively (Supplemental Table). The colour range of expression values represents a logarithmic scale. The correlation of expression between different cases, or the distance between rows, i.e. their similarity measure, was determined as follows: a hierarchical clustering algorithm was applied with "Pearson centred" as the distance function, the "centroid" linkage rule was used, and clustering was performed on the whole nigral transcriptome. (JPG 3.35 mb)
Supplement to Fig. 4 (PDF 85.7 kb)
Supplemental Table 1. Cases used in the study (PDF 37.8 kb)
|
[
"ampa receptors",
"cell death",
"dementia",
"expresssion profiling",
"orexin/hypocretin system"
] |
[
"P",
"P",
"P",
"M",
"R"
] |
J_Med_Internet_Res-8-2-1550696
|
Health Information Literacy and Competencies of Information Age Students: Results From the Interactive Online Research Readiness Self-Assessment (RRSA)
|
Background In an era of easy access to information, university students who will soon enter health professions need to develop their information competencies. The Research Readiness Self-Assessment (RRSA) is based on the Information Literacy Competency Standards for Higher Education, and it measures proficiency in obtaining health information, evaluating the quality of health information, and understanding plagiarism.
Introduction
Background and Purpose of the Study
As society moves toward evidence-based medicine [1], health providers, health educators, and health care consumers must acquire not only basic health information literacy skills but also more advanced competencies [2]. These competencies include evaluation of the quality of health information resources, obtaining health information documents on narrow topics by conducting advanced searches, judging the trustworthiness of health information sources, and understanding the advantages and disadvantages of different media. The last point is of special concern because many individuals have come to rely on the Internet as a main source of health information. This research addresses the Healthy People 2010 Objective 11-2, currently worded as “to improve the health literacy of persons with inadequate or marginal literacy skills,” but which may be expanded to the entire US population instead of only to those with marginal or inadequate literacy skills [3]. In addition, it aims at providing needs assessment information that may aid in accomplishing Objective 11-3, which is related to increasing the proportion of health communication activities that include research and evaluation, and Objective 11-4, set to increase the proportion of health-related websites that disclose information that can be used to assess the quality of the sites.
Recent reports suggest that over 55% of Americans with Internet access seek health information online [4]. One of the most common complaints about online health information searches is the amount of time required to process the documents that are found [5], but this observation is likely to be related to the general nature of the searches conducted—few information consumers use advanced search features, precisely specify their keywords, or limit their searches in some other way. While Internet search engines help identify a very large number of health-related documents, their use calls for advanced competencies that not all information consumers may possess. For example, the vast majority of documents found on the Internet have not passed a rigorous peer-review process. The ability to conduct one’s own review is clearly an advanced skill. Arguably, health information consumers will be at a greater risk of making health decisions on the basis of noncredible information if they conduct a Google search as opposed to a search in a scholarly library database. This risk will be particularly high for individuals with poor health information competencies. Research comparing clinical evidence to Internet information reveals numerous examples of erroneous and potentially harmful information on such popular topics as cancer rates, smoking cessation methods, and fever management in children [6-8].
Internet users may tend to underestimate the effort and competence required for obtaining trustworthy health information. A decade ago, communication researchers who compared print and television media described this paradox:
[Individuals] have learned that print materials, so highly prized in school and elsewhere, are indeed more difficult to process, whereas TV can be processed for pleasure without much effort. However, this argument pertains only to the minimum effort needed for the satisfactory processing of materials; it says nothing about the amount of additional effort one could expend in processing televised material if one aimed at a deeper understanding of it[9].
Although the Internet provides access to a vast number of documents on health-related topics, it is hard to build evidence-based knowledge about a health issue if one cannot determine the credibility of websites and the trustworthiness of the online documents. The minimum effort required for identifying millions of websites on a particular health topic is in sharp contrast with the average effort required to sift through the gigabytes of information in order to sort out the most credible documents, or at least those that appear as such.
Higher education institutions in the United States provide access to an unprecedented quantity of digital information via library archives, licensed online databases, and the public-access Internet. To differentiate between publicly accessible Web documents and password-protected scholarly databases, which can be accessed by paid members via the Web, we refer to the former as the “the public-access Internet.”
Our study explores three basic questions: How proficient are university students at finding and evaluating health-related information? How well do they understand the difference between peer-reviewed scholarly resources and opinion pieces or sales pitches? How aware are they of their own level of health information competencies? The main goal of this project was to identify approaches to building Information Age competencies of young health consumers, specifically a cohort of 18- to 23-year-old students enrolled in higher education programs.
Literature Review: Health Information and the Internet
In accordance with the Healthy People 2010 health communication objective [3], public health professionals attempt to assist consumers seeking health information via the Internet, for instance, by reinforcing the need for quality standards and widespread criteria for evaluating health information [10-14]. Cline and Haynes [10] note that, while critics are fast to question the quality of online health information, limited empirical research on this topic does not allow any broad conclusions to be drawn. In a study published the same year, Eysenbach and colleagues [15] reported that Internet coverage of health information was often inconsistent, although the accuracy was generally good, and that search engines and simple search terms did not provide efficient access to health information. Crespo [16] reviewed several studies on online health information seekers and concluded that most users seemed to focus on finding information quickly rather than on evaluating the information found. Similarly, Eysenbach and Kohler [17] found that individuals explored only the first few links obtained from a search using a general search engine. Although some Internet users attempted to assess the credibility of sites by, for example, examining their source and professional designs, many people did not read the “about us” sections of websites, learn about the authors or owners of the sites, or review disclaimers and disclosure statements. Very few Internet users later remembered from which websites they retrieved information or who stood behind the sites [17].
Thus, abundance of health information does not always translate into informed choices. Hibbard and Peters [18] suggest that three factors should be considered in selecting information presentation strategies: (1) the complexity and amount of information; (2) the nature of the choice—degree to which there is a right or best option; and (3) the experience, motivation, and skills of users. The third point, deficient information skills, may prevent members of the public from recognizing that key information is missing, from understanding the difference between biased and unbiased information, from distinguishing evidence-based claims, and from interpreting the information intended for health professionals [10]. Researchers, having observed individuals who, on average, spent about one-half hour looking for health information, concluded that information consumers should have at least a tenth grade reading level to process Web materials. Many websites presented to the participants of this study contained material at a college level [15].
Online health care is having a growing cultural impact, affecting the practitioner-patient relationship and opening up the possibility of new roles for social workers and educators in the provision of health services [19]. The increasing use of the Internet draws scientists’ attention to modeling individual behavior, contributing to the development and refinement of individual health theories and models, such as the Theory of Planned Behavior, The Health Belief Model, and The Transtheoretical Model [20]. The theoretical framework for this study is largely based on the information processing theories and concepts discussed below.
Schneider and Shiffrin [21] distinguish two qualitatively different modes: (1) conscious, intentional processing of information that is capacity limited (controlled processing), and (2) quick and efficient automatic processing of information that has greater capacity, for example, when several tasks can be done at the same time. Automaticity requires less attentional resources than controlled processing, and it is developed through extensive practice under the condition of consistent stimuli and response requirements. When surfing the Internet, for example, health information consumers limit their exposure to inconsistent conditions—they tend to use the same search engines and the same searching methods, such as entering keywords into the nonadvanced search window. The assessment of health information competencies in this study incorporates tasks that call for automatic processing and tasks where stimuli and response requirements of the task are inconsistent with most health information consumers’ information search practices.
We also draw upon Anderson’s ACT theory [22,23], which explains skill acquisition. It incorporates research on automaticity and explains the development of cognitive skills important for processing digitized health information from a variety of electronic sources [24]. According to Anderson [22,23], skill development has three stages: (1) the declarative knowledge stage, when knowledge of facts is built, such as facts about reputable sources of health information and general procedures for obtaining information; (2) the knowledge compilation stage, which is characterized by proceduralization and composition; and (3) the procedural stage. To illustrate the second stage, consider a health information consumer who follows a set sequence of specific steps to search for a health-related terms (proceduralization) and reapplies this sequence until sufficient information on a health topic is found (composition). Once at the knowledge compilation stage, a consumer can perform an information search task at a higher speed and with fewer errors than at the declarative knowledge stage. High speed and low error rate are both important markers of skilled performance. However, a disadvantage of knowledge compilation is the rigidity of behavior, when individuals find it increasingly difficult to attend to intermediate feedback (e.g., step-related results) and engage in strategy modification (e.g., by adopting a search strategy that produces a greater number of trustworthy health information resources) [24]. Declarative and procedural knowledge are discussed in greater depth in the Methods section.
An Interdisciplinary Research Partnership
Our research originated from the collaboration of a psychologist, a health educator, and a librarian who set out to understand and improve health information competencies of the Information Age generation. The collaboration enhances our research in several ways. The psychologist contributes expertise in the area of psychometrics and test design, whereas the health educator contributes knowledge of health consumers’ behavior and intervention designs. The librarian contributes expertise in training and enhancing patrons’ health information–seeking skills [25], as well as knowledge about gateways to authoritative consumer health information, for example, Medline Plus [26,27]. Linnan and colleagues [28] believe that library/public health partnerships are capable of increasing information access, the quality of available health information, and the technological expertise of all community members. Neighborhood libraries often serve the online health information needs of consumers who may not have Internet access at home, such as the elderly, ethnic groups, and low-income and undereducated populations [29,30], whereas university libraries also serve as gateways to scholarly health materials that are not available on the public-access Internet. In addition to public-access health resources available online, this research focuses on scholarly health resources in academic libraries and their use by students who are training to become health professionals.
Methods
Participants
A sample of 400 college-age students was selected because this cohort is the first Information Age generation that has been exposed, for up to one-half of their lives, to the Internet. Students enrolled in three courses in the College of Health Sciences at a Midwestern university were invited to participate in the study. The first class was a high-enrollment introductory course on the determinants of health. Although only undergraduate students (n = 354) participated in this course, they represented all levels of undergraduates—freshman (59%), sophomores (22%), juniors (9%), and seniors (10%). The second class was an advanced course in health administration in which both undergraduate (n = 19) and graduate students (n = 3) were enrolled. The third class was a mid-level health education course (n = 25) for undergraduate students. All students enrolled in the advanced health administration course and the mid-level health education course were majoring in health professions. About one third of the introductory course students with declared majors were majoring in a health-related discipline, and 31% of students had not made up their minds about a major field of study.
Introductory course students completed the assessment for extra credit, while others did it to learn more about their own skills. The instructors emphasized that the purpose of the assessment was to help students become competent consumers of health-related information.
Measures
Health Information Competencies
Ivanitskaya and Casey developed the Research Readiness Self-Assessment (RRSA) to measure basic research skills based on the Information Literacy Competency Standards for Higher Education developed by the Association of College and Research Libraries [2,31]. The RRSA designers’ original intent was to measure information competencies, both general and discipline specific, of students attending colleges and universities. A health information version of the RRSA is discussed in this paper; it was created to specifically evaluate health information competencies. Competencies are knowledge/skills sets essential for accomplishing a goal, in this case, finding quality information on a specific health topic. The RRSA measures competencies linked to such college-age health information consumer behaviors as determining possible sources of health information, conducting health information searches, evaluating the quality of documents found, and using those documents appropriately. One of the relevant competencies is knowledge of plagiarism because it can be applied to properly recognize ideas contributed by others and to evaluate health-related documents. The RRSA designers aimed at measuring foundational competencies that are (1) transferable to other knowledge domains (e.g., social sciences in addition to health sciences); (2) applicable to a large number of health information consumers; (3) consistent with typical behaviors or experiences of health information consumers who seek information from electronic sources; and (4) that capture the nature and spirit of critical thinking, life-long learning, and advances in information technology. It is important to note that the RRSA instrument does not measure higher order skills that characterize experienced researchers, such as the design of clinical trials [31]. The word research in the assessment’s title matches the language commonly used by the lay population, as in “going to Google to research a health topic,” which is indicative of such behaviors as searching, judging, and making decisions.
The RRSA contains the following items: (1) multiple choice or true/false questions that measure declarative knowledge; (2) interactive, problem-based exercises that measure procedural knowledge; (3) demographic questions; and (4) a question that asks for a self-report about the level of the respondent’s research skills [31].
Declarative knowledge, defined as knowledge of facts or verbal knowledge, is a precursor to higher-order learning, which is needed, for example, to complete a sequence of steps to critically analyze a website or to employ elegant information search strategies [32]. Declarative knowledge questions in the RRSA measure knowledge of plagiarism, health information sources, and research-related terminology. For example, the following item is used to measure knowledge of research-related terminology:
A journal article abstract is…
an annotated list of references used in the articlea summary of the article’s contenta summary of other research on this topica note or paragraph about the authors of the articlea glossary of abstract concepts included in the researcher’s model
Compared to declarative knowledge, procedural knowledge is related to skills and problem solving. Essential for reproduction of learned behaviors, procedural knowledge is defined as knowledge of the process used to complete a task (e.g., how an information search process can be sequenced, organized, or controlled) [32]. In the RRSA, problem-based interactive exercises are used to measure procedural knowledge. Procedural knowledge questions include links to websites, library catalogs, and interactive search modules designed specifically for the RRSA. Students demonstrate their database navigation skills by setting up basic and advanced searches. For example, the following item is used to measure skill in conducting a search using Boolean operators (and, or, not):
You are interested in gathering information about work stress but are not interested in its medical side effects. Set up a document search in a separate window using the following keywords: stress
medical. Click here to begin your search [a hyperlink to an interactive online module similar to searches in health-related library databases, such as Medline, with text fields for entering key words and a choice of Boolean operators]. Report the number of documents you found: a) 255; b) 555; c) 700; d) 1164; e) 55164.
In addition, students evaluate the quality of research publications, make judgments about website trustworthiness, and detect plagiarism. For example, the following item is used to measure evaluation of the trustworthiness of websites:
You are looking for information on various nutritional supplements. You found three websites. Click on the links below to examine each site and to evaluate its content. Which of these websites is the most trustworthy? a) cognitogenic aids [a hyperlink]; b) dormitogenic aids [a hyperlink]; c) vescorogenic (gustatogenic) aids [a hyperlink].
Instrument Piloting and Validation
To pilot test an earlier version of the RRSA instrument and to gather initial evidence about its validity and reliability, we administered a 60-item assessment to undergraduates (n = 100) and doctoral students (n = 45), as well as professional librarians (n = 5) and health professionals (n = 3). The feedback from librarians and health professionals offered preliminary evidence in support of the instrument’s face validity and content validity. Specifically, the librarians confirmed that the items included in the RRSA assessment conformed to the Information Literacy Competency Standards and addressed knowledge and skills important to health information consumers. The wording of several items, both stems and response options, was revised based on librarians’ recommendations. In addition, the librarians completed the assessment themselves. Their scores were then compared to the scores of students at two academic levels, undergraduate and doctoral. The results indicated that individuals with greater training and experience in managing digital health information performed better than individuals with less experience. Undergraduate students’ overall scores were the lowest (about 66% correct responses), followed by doctoral students’ scores (73%) and librarians’ scores (95%). These results offer preliminary evidence of the assessment’s criterion-related validity. The pilot test indicated an acceptable internal consistency value (Cronbach alpha > .70), although it could be improved (approach .80) if four items were removed. Therefore, four RRSA items that reduced the overall internal consistency were deleted.
The revised assessment contains 56 items, including 16 multiple-choice questions and 40 true/false questions grouped under 7 stems (Multimedia Appendix 1). For example, knowledge of information sources is measured by a stem that states, “Which of these citations are to journal articles?” The participants then check all that apply from the list of 6 true/false items (3 references to journal articles, 1 book reference, and 1 book chapter reference). Items are scored as +1 if the answer is a correct positive or a correct negative and +0 if the answer is a false positive or a false negative. Further description of the development of the stimulus materials used in website evaluation appears in the Results section, under Proficiency in Evaluating Health Information.
The RRSA assessment was designed to be useable by more than one institution. Its content can be adapted to the needs of various educational programs. Specifically, instructions to participants, the text of individual questions, detailed feedback, links to additional resources, and disclaimers (e.g., about participants’ rights and how the information they provide will be used) can be revised, without help from programmers, using the password-protected online control panel. This has been done by three US universities and one Canadian university that adopted the RRSA for use in their academic programs. For example, all four institutions revised search questions to enable their students to search for documents in their own university’s library catalog. The original RRSA designers provide coaching and training in order to ensure that the changes made to the RRSA do not have a negative impact on its reliability and validity. Ongoing validation studies provide a quality control mechanism and allow the testing of new or revised questions suggested by the partner institutions. The administration of the RRSA to partner institutions is supported through grants, partner donations, and volunteer efforts by the RRSA design team members.
Other Measures
We asked the study participants to share information about their age, gender, and education. Self-reported level of research skills was measured with a single item, “How do you rate your research skills?” with six response options ranging from 1 (nonexistent) to 6 (excellent).
Procedures
The RRSA instrument was administered online. Each student was issued a unique pass to access RRSA questions. The students had the option of submitting an incomplete survey and then returning to it at a later time to finish the remaining questions. This feature promoted better information processing and relieved the students from the need to rush and finish the entire assessment on their first attempt. The average estimated RRSA completion time was 26 minutes. Upon answering all questions, the students received an individualized results page that summarized their performance in different areas by providing a score, a maximum possible score, and percent attained. In addition to the numerical RRSA results, the Web page displayed individually tailored feedback composed by an experienced librarian. The Web page was programmed to compare, within each performance category, each individual student’s performance to the performance of a norm group. In accordance with the student’s competency level, the feedback provided suggestions for skill improvement and an explanation of factors that may have contributed to low, average, or high performance in each area. Finally, students who completed the RRSA were given the option to request additional materials for remedial learning, such as an explanation of the difference between scholarly and nonscholarly resources. The links to these additional materials were delivered to students via email.
Data Analyses
Descriptive statistics were used to examine respondents’ performance in four areas—searching for health-related information, understanding plagiarism, evaluating health information, and self-reported skill level. To examine the relationship between self-reported skill level and actual performance, we computed composite scores. A composite overall score, which is indicative of the health information competency level, was created by adding points for 56 items, which were either true/false or multiple choice. Composite score calculations were preceded by an internal consistency reliability analysis that determined the appropriateness of combining responses from multiple items. We used a Spearman correlation to assess the relationship between the actual skill level (overall score) and self-reported skill level. A multiple regression analysis was used to examine the relationship between actual performance and perceived skill while holding the amount of education (number of credit hours earned) constant.
Results
Our research questions were the following: How proficient are university students at searching for and evaluating health-related information? How well do they understand the difference between peer-reviewed scholarly resources and opinion pieces or sales pitches? How aware are they of their own level of health information competencies? The results for each question are presented below, preceded by a sample description.
Respondent Characteristics
The participation rate was 77%. Nonrespondents (n = 92) differed from respondents (n = 308) in terms of their academic level (t
400 = 2.29, P = .02). Freshmen were slightly more likely not to participate in the RRSA than seniors; the participant group included 7% less freshmen and 10% more seniors than the nonparticipant group. Most respondents were female (77%) and between 18 and 23 years of age (95%). The vast majority of respondents (98%) did not have a bachelor’s degree, and the remaining students were working toward their master’s degrees. Because we administered the RRSA to students in health professions courses, over one third of respondents were majoring in health sciences. Common majors were athletic training and sports medicine, health administration, physical education, pre-physical therapy, and public health promotion. On average, the undergraduates who participated in the study had completed 40 or fewer semester credit hours of university coursework. A quarter of respondents reported earning over 71 credit hours.
Proficiency in Searching for Health Information
Table 1 summarizes performance in searching for health information. The data indicate that most students recognize common health journal titles and can perform a basic search in a library catalog, for example, by entering an exact book title into the title search. Few students, however, can perform an advanced search for a book when they know the book’s author (with a very common last name), general topic, and publication date. We call this search advanced because imprecise book specifications make it hard to find the book without performing a search that takes into account all or nearly all of the available information.
The data also show that two thirds of study participants are unable to understand or apply Boolean operators, such as and, or, and not. Boolean operators are used in most search engines, including those used for navigating the Internet (Google or Yahoo), library databases with scholarly journal articles, and library catalogs. Even though most students (89%) understand that a one-keyword search is likely to return too many documents, few are able to narrow a search by using multiple search categories simultaneously or by employing the Boolean operators. In addition, nearly half of the respondents have trouble discriminating between primary and secondary sources of information, as well as between references to journal articles or other published documents, such as books or book chapters.
Proficiency in Evaluating Health Information
One of the most important markers of a competent health information consumer—critical judgment of information—is assessed in two ways: (1) the first set of questions calls for a review of three full-text articles from journals, and (2) the second set of questions calls for a comparison of three health-related websites.
The three journal articles are on the topic of job satisfaction, a topic relevant to any profession, and come from a full-text library research database. They include a rigorous empirical study, a case study, and an opinion article. Only the empirical study has a bibliography and an explicit statement about the author’s affiliation. The opinion article, clearly the least authoritative source, makes no mention of the author’s affiliation. As shown in Table 1, most respondents can determine the article publication date; it appears at the top of a full-text article. Many respondents can also identify an opinion article. Fewer respondents know how to determine if an article includes a research review and are able to check for the author’s affiliation.
The three Web pages about nutritional supplements are realistic looking interactive screens that appear to be live websites. The content of these mock websites, developed specifically for the RRSA, includes graphics, hyperlinks, and text about nonexistent classes of nutritional supplements—cognitogenics, dormitogenics, and gustatogenics. Each website is dedicated to one class of supplement and explains its purpose (e.g., cognitogenics help people with learning disabilities), prevalence (e.g., “gustatogenic aids have been available in Germany and Canada for over five years”), and safety. Even though the descriptions of nutritional supplements were fictitious, all three websites accurately stated that the US Food and Drug Administration did not evaluate the safety or benefits of these nutritional supplements.
Table 1
Searching and evaluating health information: performance on select measures (n = 308)
Respondents With Correct Answers
n
%
Searching for Health Information
Knowledge of a scholarly source, Journal of American Medical Association (7)
293
95
Demonstration of a skill in locating a book in a university library catalogue based on its exact title (16)
286
93
Understanding that a one-keyword generic search may return too many documents—an overwhelmingly large number of resources on a variety of topics (4)
275
89
Use of a proper research strategy—thinking about a broad topic to identify a sub-area of interest (2)
268
87
Ability to detect a journal citation that is incomplete—lacks a year of publication (17)
241
78
Understanding of a term “article abstract”—a summary of the article’s content (8)
234
76
Knowledge that a journal is a source of scholarly (analytical) information on a narrowly specialized topic (6)
214
70
Understanding of a term “bibliography”—a list of references or citations (9)
213
69
Identification of a primary source of health information: medical record (14)
195
63
Identification of references to journal articles from a list of references that includes both book references and article references (11)
187
61
Knowledge of a peer-reviewed journal article as an authoritative source of specialized health information (12)
185
60
Identification of a primary source of health information: hospital annual report (14)
173
56
Demonstration of a skill in locating a book in a university library catalogue based on a non-unique authors’ name and a general topic (15)
111
36
Knowledge of Boolean operators (and, not, or) (3)
105
34
Demonstration of a skill in setting up and performing a search with Boolean operators (and, not, or) (13)
98
32
Evaluation of Information: Full-Text Journal Articles
Evaluation of journal articles: Identification of an article published prior to year 2000 (22)
248
80
Evaluation of journal articles: Identification of an article based on opinion rather than well-supported evidence (19)
242
79
Evaluation of journal articles: Identification of an article based on a review of existing research (20)
166
54
Evaluation of journal articles: Identification of an article written by an author whose affiliation is unknown (21)
148
48
Evaluation of Information: Websites on Nutritional Supplements
Evidence-based decision-making: Disagree that “all three websites make a good case for taking nutritional supplements” (25)
187
61
Evaluation of health-related websites: Identification of the most trustworthy website (23)
154
50
Evaluation of health-related websites: Ability to identify the purpose of a website—to sell services (24)
42*
46
Evidence-based decision-making: Agree that “none of the websites makes a good case for taking nutritional supplements” (25)
67
22
*This question was added later, and, therefore, it had a smaller number of respondents (n = 92).
Note: RRSA question numbers are shown in parentheses; see Multimedia Appendix 1 for exact question wording.
To facilitate comparison of the three websites, we built in standard features that provided clues about high or low credibility. The standard features are URLs (two websites were .org and one was .com), links to the authors’ biographies, dates of publication, references, disclaimers, and links to organizations with which the authors are affiliated. These features act as contextual clues that maximize or minimize the trustworthiness of the websites. A review of such features is part of many website evaluation recommendations (for example, in their 1999 publication, Kotecki and Chamness [11] draw evaluators’ attention to a website’s features rather than its text), yet it is unclear if health information consumers are able to compare these features across multiple websites.
These standard features, rather than the text content, are intended to differentiate the websites in terms of their credibility. Because all respondents are equally uninformed about the nutritional supplements described in the text, they must attend to other features when making quality-related judgments. This purposeful design was motivated by the desire to avoid the confounding influence of pre-existing knowledge about the subject matter described in the document that is being judged. A good measure of one’s ability to critically evaluate Web pages is being able to disentangle the judgment of a website’s features from the judgment of its content. Study participants may have had preconceived notions about the quality of nutritional supplements depending on their purpose (e.g., cognitogenics are for sleeping disorders and gustatogenics are for appetite suppression). To avoid a possible interaction between the untrustworthy features of a website and the believable description of the nutritional supplement, we asked a group of students (n = 52) to judge the trustworthiness of the supplements’ descriptions presented as Microsoft Word documents rather than as websites. Although the level of trustworthiness was about the same for all nutritional supplement descriptions, the least trusted nutritional supplements were placed on the website with the highest number of untrustworthy features.
When five subject matter experts independently reviewed the three websites and rated their trustworthiness using the Kotecki and Chamness [11] website evaluation tool, they reached 100% agreement regarding the most trustworthy site. In comparison, undergraduates’ performance was much poorer: only 50% of respondents were able to identify the most trustworthy website (see Table 1).
Understanding the Difference Between Scholarly Resources and Sales Pitches
Less than half of respondents determined the purpose of the least trustworthy website, which was to sell products and services. The visitors to this .com website are charged for reprints of the content, offered discounted products, and provided with multiple prompts (e.g., a running line) to book a consulting appointment with a private nutritionist who has few relevant qualifications. Customer testimonials posted on this site describe fantastic outcomes achieved within an unrealistically short time frame.
Less than a quarter of study participants reached the correct conclusion that none of the websites made a good case for taking the nutritional supplements, whereas 39% of respondents thought that all three websites made a good case for taking the supplements.
Understanding Plagiarism
Health care professionals are expected to share health information with others, for example, by summarizing information from a variety of sources and distributing it to patients and clients. Higher education programs prepare students to apply standard rules for acknowledging contributions by others and referencing idea sources. Because this skill set is expected to become an integral part of their professional ethics, we built the RRSA to include measures of students’ knowledge of plagiarism, their ability to recognize it, and their awareness of its penalties. Our results indicate that the vast majority of students (92%) know that their university may impose a severe penalty for plagiarism, up to and including expulsion. Table 2 and Table 3 display responses to sample questions that measure declarative knowledge of plagiarism. They show that many students are aware that common knowledge can be reproduced without references, whereas words written by others should be enclosed in quotation marks and accompanied by a complete reference. But when presented with more ambiguous examples of plagiarism, some study participants demonstrated misconceptions about what constitutes plagiarism. A surprisingly large number of respondents believed that it is appropriate to present another person’s ideas as their own without citing a specific source, especially if this person is a relative or if the original words have been slightly modified.
Table 2
Understanding plagiarism: when references are needed (n = 308)
Which of the following can be reproduced without proper reference? Check all that apply:
Respondents With Correct Positive or Negative Answers
n
%
Common knowledge*
294
96
Hospital board member’s point of view
264
86
My classmate’s ideas
232
75
Unpublished works
223
73
Spoken word
209
68
My dad’s political opinions
156
51
*Common knowledge can be reproduced without proper reference.
Note: Items are scored as +1 if the answer is a correct positive or a correct negative and +0 if the answer is a false positive or a false negative.
Table 3
Defining plagiarism (n = 308)
Which of the following are plagiarism examples? Check all that apply:
Respondents With Correct Positive or Negative Answers
n
%
Submitting a free research paper that was downloaded off the Internet.*
290
95
Reproducing a sentence that you found quoted in a book without referring to the original source.*
276
90
Enclosing the word-for-word sentence in quotation marks, accompanied by a citation.
271
88
Copying from the source verbatim without any quotation marks but adding a citation.*
215
70
Putting someone’s idea in my own words without citing a specific source.*
201
65
Using similar sentence structure to express another person’s ideas without referring to the original source.*
169
55
*These items are examples of plagiarism..
Note: Items are scored as +1 if the answer is a correct positive or a correct negative and +0 if the answer is a false positive or a false negative.
To measure procedural knowledge of plagiarism, we ask respondents to compare a sentence from a Health Affairs article by Lapetina and Armstrong [33] to two other sentences that may have been plagiarized (question 20). Over two thirds of respondents (82%, n = 253) detected plagiarism in a sentence that closely follows the original but provides no reference to the original source. The percent of respondents who correctly identified a sentence without plagiarism (89%, n = 275) was comparable to the percent of respondents who knew that they should enclose the word-for-word sentence in quotation marks and cite the source (88%, n = 271, as shown in Table 3).
Awareness of Personal Health Information Competencies
When asked “How do you rate your research skills overall?” most respondents (84%) believed that their skills were good, very good, or excellent. To compare self-reported and actual skill levels, we computed an overall health information competency score for each participant. An acceptable level of internal consistency reliability (Cronbach alpha = .78) for 56 right/wrong items indicates that it is appropriate to calculate the overall score as the sum of points of these 56 items. The overall scores ranged from 20 to 54 with a mean of 37 (SD = 6.35) and did not significantly depart from a normal distribution.
Actual performance was examined by self-reported skill level. The group differences were mostly in the expected direction (see Table 4), but there was a large amount of variation in the overall score within each self-reported skill level. This indicates that the overall health information competency score was high for some students and low for other students, despite the fact that their self-reported competency was the same.
Table 4
Means for health information competency overall score by self-reported skill level
How do you rate your research skills?
n
Mean Overall Score
SD
Nonexistent
0
-
0
Poor
3
36.33
4.04
Fair
47
34.89
5.52
Good
162
36.89
6.29
Very good
83
37.64
6.89
Excellent
13
36.77
6.10
Total
308
36.78
6.35
Health information competencies may vary as a function of education; therefore, we regressed undergraduates’ overall scores on the amount of credit hours earned toward the bachelor’s degree (Step 1) and self-reported skill level (Step 2). The level of education was operationalized as the number of credit hours earned (0-9, 10-24, 25-40, 41-70, and more than 71). The analysis was conducted for 302 undergraduate students (six graduate students were removed from this analysis). Age could not be used as a control variable because most students (95%) fell into the same category of 18 to 23 years of age. The variables entered on Steps 1 and 2 account for 8% of variance in the overall score (R2 = .08). The amount of education significantly predicted the overall score (β = .28, P < .001). When credit hours earned were held constant, self-reports of skill fail to explain a significant amount of variance in the overall score (β = .08, P = .23). Overall, the results suggest that although students’ self-ratings of research skills tend to increase with the increasing level of education, these self-reports may not be an accurate predictor of students’ actual health information competencies.
Discussion
Interpretation of Findings
The present study represents a systematic effort to measure health information competencies using a standardized and reliable measurement tool, the Research Readiness Self-Assessment (RRSA). The data were obtained from a diverse sample of 308 respondents (77% response rate). Nonrespondents (n = 92) differed from respondents (n = 308) in terms of their academic level: freshmen were slightly more likely not to participate in the RRSA than higher-level students. The most likely explanation for nonparticipation is a lack of interest in extra credit rather than the computer-assisted administration of the RRSA. It is possible, of course, that students with particularly poor computer skills found the online administration a barrier. However, a semester after we collected the data reported in this paper, there was a 100% participation rate by 180 undergraduates in two introductory courses where the instructors required RRSA completion. The two course instructors reported no student complaints about not being able to follow emailed instructions on how to complete the assessment.
The data indicate that many students lack important competencies that may limit their ability to make informed health choices. We observed deficiencies in the areas of conducting advanced searches, discriminating among different types of information sources, referencing other people’s ideas, and evaluating information from Web pages and journal articles. Our data suggest that undergraduate students are inaccurate judges of their own competencies and hold a very positive view of their ability to do research. This finding may reveal an important barrier to building health information competencies of college-age students.
We found that there is a large competency gap between the average and the best information consumer. An average undergraduate in our sample is able to solve only 68% of problems that are solved by the best performing study participant (an average score of 37 versus a maximum score of 54). Health information competencies are applied to transform health-related information into knowledge that is consistent with the most current medical practice. High competence variability is a proxy indicator of students’ varying ability to make evidence-based decisions. In the past, limited access to information may have prevented health information consumers from acquiring knowledge and making informed choices. The new generation of health information consumers has, for the most part, easy access to information; yet it may not be able to take full advantage of this convenient access.
Our study shows that individuals with limited health information competencies may fail to locate the best available information due to employing poor search strategies. Searches that do not take into account all of the important criteria often produce low-relevancy documents or documents from commercial websites that promote products or services. These sites often present one-sided evidence, which can be detrimental to making a good decision about one’s health. Overall, many students are rather unsophisticated information consumers who rely on basic searchers and the easiest ways of retrieving information.
We found that many individuals know little about information sources—primary versus secondary, articles versus books, commercial versus noncommercial websites, and opinion pieces versus empirical studies. Information consumers who do not understand these distinctions are likely to engage in information processing that is shallow and superficial. They may, for example, follow a search path that produces the highest number of documents, rather than a path that produces documents of the highest quality. When the number of documents criterion is applied, Google and Yahoo significantly outperform all scholarly databases available through libraries. For instance, a Google search for the keyword health produces, in less than a second, over 8 million results ordered by popularity (as of June 2005, 25% of these results had .com URLs and 16% had .org or .gov URLs), where a similar search in Medline Plus produces 665 results, organized by health topic. With heavy reliance on public-access Internet search engines, an Information Age generation student may have an inaccurate conception that the Internet is the only place where society stores its best knowledge.
Once the plethora of documents is obtained, they need to be critically evaluated. Although health consumers are warned to critically examine websites to determine the document’s purpose, author’s affiliation, date of publication, and other features [11-14], these website evaluation criteria are only useful to those who know how to apply them. Many students in our sample appear not to possess these skills, and this finding is consistent with other observational studies (e.g., [17]). Our website evaluation exercise reveals both poor judgment and readiness to follow the lead, even when the authors of the online documents do not explicitly ask for purchase of their products. Although we measured a behavioral intent, rather than an actual behavior, there is still a significant potential for harm, ranging from financial losses to negative health effects, if only a few individuals execute their intent to take nutritional supplements that can be best described as “fake” or “bogus.” As we designed the most trustworthy website for the RRSA, it was alarming to witness the ease of misrepresenting or even falsifying health information. In designing the trustworthy site, we tried to meet as many website evaluation criteria as possible, and it became very apparent that these criteria do not guarantee information accuracy. Even completely false information about nonexistent food supplements can be made to appear trustworthy, as though it comes from an authoritative source.
Indeed, there is no substitute for good judgment when it comes to navigating information. Because this good judgment is a product of both critical thinking and extensive knowledge of the subject matter being researched, we believe that higher education programs are uniquely positioned to develop health information competencies. However, initial work on developing Information Age competencies needs to be done at the K-12 level when children are beginning to be exposed to various sources of information, including the Internet.
In this study, we reviewed three broad categories of information competencies—obtaining information, evaluating information, and using information. Using information includes such behaviors as reaching evidence-based conclusions and sharing information with others, a behavior guided by one’s understanding of plagiarism. One study of plagiarism revealed that cyberplagiarism, or inappropriate use of phrases and ideas published on the Internet, is prevalent even among scholars [34]. Our findings suggest that, in college students, the plagiarism behavior may originate not only from motivation to cut corners (e.g., to cut-and-paste text without citations) but also from the lack of nuanced knowledge about plagiarism. The information revolution has rapidly intensified the exchange of ideas, but the distinction between plagiarism and proper acknowledgment of others’ ideas continues to be poorly understood. Many students, for example, think that they do not need to provide references for paraphrased sentences or for sentences whose structure they modified. Perhaps these students view plagiarism as a violation of ownership of exact words rather than a violation of ownership of ideas. Similarly, some respondents believe that it is appropriate not to give credit for original ideas that are expressed orally (rather than in writing) or by people whom they know well. If carried into one’s professional life, this misconception can make it difficult to follow ethical norms for recognizing others’ knowledge contributions. Such ethical norms are strong in health professions, and their violation may lead to negative consequences.
Perhaps the most interesting finding is the fact that participants are so unaware of their own skill deficiencies. It is possible that students make global judgments about their research skills based primarily on their ability to access information. That is, one’s ability to access information may be confused with one’s ability to generate knowledge from the information accessed. But obtaining information is merely the first step of knowledge acquisition. All of our study participants can access the Internet, as demonstrated by completing the RRSA online, but not all may be able to make good use of the information they access. Extending the argument by Solomon and Leigh [9] from television to Internet search engines, we conclude that the effort an individual expends to locate millions of documents in Google is a poor indicator of the true effort needed to process the obtained material “if one aimed at a deeper understanding of it” [9]. The Information Age generation of college students may benefit from this point.
Implications for Health Promotion Practice
The findings of our study have several implications for individuals who practice health promotion for health information consumers. Health educators, librarians, and other professionals who play an active part in promoting health information literacy need to assist health information consumers in becoming more aware of their skill limitations. These professionals should develop their own proficiency in managing modern media and be able to find, evaluate, interpret, and present health-related information to other information consumers. Research on health information competencies of practicing health professionals remains limited, and we do not yet have a complete picture of their preparedness for evidence-based practice. But in one survey study of 1097 registered nurses, it was found that many respondents “had no exposure to the research process in their educational programs, do not appreciate the importance of research to practice, and have great difficulty understanding research articles” [35]. In this study, most registered nurses did not search databases such as Medline or felt skilled to do so. This preliminary evidence suggests that health professionals need to build their health information competencies.
The RRSA instrument offers an operational definition of information literacy, which remains an ill-defined concept. Upon examination of 97 Medline articles on the topic of information literacy for health care professionals, Saranto and Hovenga [36] found that the concept of information literacy has not yet been established. It is sometimes used interchangeably with computer literacy and informatics awareness or with the ambiguous term computer experience. The RRSA assessment used in the present study adds to the literature on health literacy by defining basic knowledge and skills needed for managing electronic health information resources.
Among the limitations of the present study is the narrowly focused sample, which limits our ability to generalize the study’s findings to the broader population of health information consumers. The students from a Midwestern university may not be completely representative of the entire population of US Information Age students, due to, for example, the relatively homogeneous ethnic composition and possible overrepresentation of individuals raised in rural communities. In our future studies, we intend to broaden the pool of RRSA participants by including multiple educational institutions as well as urban and rural communities located in different geographic regions.
In contrast with many health information literacy studies, this research presents the results obtained via direct measure of skills and knowledge rather than via self-reports by health information consumers. While the reliability of the RRSA assessment reaches acceptable levels, it is necessary to further assess its unidimensionality, content validity, and criterion-related validity. A comprehensive validation study of the RRSA instrument is currently under way.
Conclusions
Today, health consumers are actively seeking information and using it to make health decisions. The ease of accessing information may influence their perceptions of their ability to make informed health decisions. Our study shows that to become savvy information consumers, young people may need assistance in understanding the various health media, building awareness of their own skill sets, and improving their ability to make evidence-based decisions. Individuals with less education and exposure to information-related activities are expected to have even lower health information competencies than our study participants [37]. Health educators must continue to partner with a variety of groups that play an important role in promoting health information literacy, such as librarians and educators.
The assessment tool used in the present study is a self-administered instrument that provides a reliable account of health information competencies related to managing electronic health information. Data acquired through this research can be used to suggest curriculum improvements and estimates of the higher end level of skill held by health information consumers. It can also be used to educate health information consumers about their levels of skill necessary for managing health information from electronic sources. RRSA findings suggest that health information competencies of undergraduate students, many of whom will soon enter a variety of health professions, are limited. Health literacy educators can utilize RRSA findings to design educational interventions that impact information consumers’ skills and prepare them for the challenges of living and working in the Information Age.
|
[
"health information",
"electronics",
"consumer health information",
"electronic health information",
"evaluation of electronic resources",
"telecommunications",
"patient education",
"educational status",
"computer network"
] |
[
"P",
"P",
"P",
"P",
"R",
"U",
"R",
"M",
"M"
] |
Pediatr_Radiol-3-1-1915603
|
Nonpalpable testes: is there a relationship between ultrasonographic and operative findings?
|
Background Ultrasonography (US) as a diagnostic tool in the work-up of boys with nonpalpable testes (NPT) is still controversial.
Introduction
Failure of testicular descent is a common childhood anomaly, being seen in 0.8–2% of full-term and 18–30% of premature boys. In undescended testes (UDT) the testis fails to migrate into the processus vaginalis, failing to reach the scrotum at 35 weeks of gestation. It is unilateral in 60–70% of cases. Approximately 20% of UDT are nonpalpable on physical examination [1]. There are four possible explanations for this phenomenon. First, the testis may be situated in the scrotum or inguinal canal, but is too small or there is too much subcutaneous adipose tissue. Second, the testis is located inside an open processus vaginalis in the inguinal canal and may be intermittently inguinal and abdominal in position. Third, it is in a true abdominal position; and fourth, the testis is lacking [2].
The literature reports that nearly half of viable nonpalpable testes (NPT) are in an abdominal position and 5% are in the inguinal canal. The remaining 45% are atrophic or absent, mostly as a result of in utero spermatic cord torsion and are located abdominally, inguinally or scrotally [3, 4].
Experienced paediatric surgeons can accurately diagnose UDT by repeated and meticulous palpation of the scrotal and inguinal region. Since palpation is quite subjective, the diagnosis mostly depends on the examiner’s experience and ability [2]. Currently there is controversy as to the step to take when clinical examination fails to identify a testis. Further exploration might be done by means of laparoscopy or by noninvasive methods such as ultrasonography (US), CT, MRI, venography or arteriography [3, 5, 6]. Of these, US is the imaging method of choice in children because it is noninvasive and does not use ionizing radiation. Furthermore, it is cheap and widely available [7].
When NPT is correctly diagnosed there is consensus on the need for early treatment, as this may decrease the risk of testicular cancer [8–12] and fertility problems [13–15]. We performed a study aimed at assessing the value of US in the diagnostic work-up of NPT, and to this end evaluated the relationship between preoperative US data and operative findings.
Materials and methods
During a 7-year period, 880 boys referred to the Department of Paediatric Surgery, Erasmus MC-Sophia Children’s Hospital, Rotterdam, underwent primary surgery for UDT. US was performed in 137 of these children with 156 NPT. Prior to US the patients were clinically examined by a general practitioner, a paediatrician, a urologist or, in most cases, by a paediatric surgeon. Clinical examination by a paediatric surgeon was with the boy in the supine position, the lower half of the body undressed. The examiner placed the index finger and thumb of the right hand on either side of the inguinal canal, thus preventing testes lying distally from the inguinal canal from withdrawing into the inguinal region during palpation of the scrotum. With the examiner standing on the right side of the patient, the inguinal region was then examined with the fingertips of a warm left hand. If there seemed to be no palpable testis, an attempt was made to empty the inguinal canal by carrying out an ironing movement with the fingertips stroking in the direction of the scrotum. This may reveal a palpable testis at the level of the exit of the inguinal canal, immediately shooting back deep into the inguinal canal. If no testis could be located at all, the perineum, the base of the penis and the thigh were closely examined to exclude an ectopic testis.
US was performed using an Ultramark 9 HDI unit (Advanced Technology Laboratories, Bothell, Wash.) with a high-frequency, broadband, 7–10-MHz linear array transducer or a Philips ATL HDI 5000 unit, also with a high-frequency, broadband, 5–12-MHz linear array transducer. US was performed by eight different paediatric radiologists. After US all boys were clinically re-examined by a paediatric surgeon and again on the operating table when under anaesthesia.
Subsequently, all boys underwent surgery and testicular position and other operative findings were assessed intraoperatively.
Depending on the US results for NPT different operative procedures were used. Laparoscopy was usually performed when US could not identify a testis. In the presence of a viable abdominal testis, a one- or two-stage orchiopexy was performed. Anorchia was diagnosed when a blind-ending vas deferens was identified. When the vas deferens and testicular vessels entered the internal inguinal ring the procedure was completed by an inguinal exploration. When a viable testis was found, orchiopexy was performed. When a nubbin (vanished testis) or an atrophic testis was found, this was excised for pathological examination. Inguinal exploration was usually performed when US did show an inguinal testis. When a viable testis was found, orchiopexy was performed. A viable testis was defined as having a volume according to age [16].
All US findings were compared with intraoperative findings.
Results
Eligible for this study were 137 boys with 156 NPT. Excluded from analysis were two with Müllerian inhibitory factor deficiency syndrome, implying abnormal testis position. The ages of the remaining 135 boys ranged from 4 weeks to 16.2 years. Of the remaining 152 NPT, 70 were left-sided, 48 right-sided and 17 bilateral.
Four patients (four NPT) had been referred to the radiologist by a general practitioner and 24 patients (30 NPT) by a paediatrician. Most patients, 107 with 118 NPT, were first diagnosed by a paediatric surgeon before being referred to the radiologist. All boys were seen by a paediatric surgeon after the US examination for re-examination and to discuss therapy determined by the US findings.
With knowledge of the US results, 28% of the NPT previously diagnosed as not palpable now appeared to be palpable. Three of the four NPT (75%) in patients first seen by a general practitioner were palpable when re-examined by a paediatric surgeon, as was the case for 18 of 30 NPT (60%) referred by a paediatrician. Of the NPT first examined by a paediatric surgeon, 18% were palpable on re-examination with knowledge of the US results.
US was able to locate 103 of 152 NPT (68%): 16 were found in the abdomen and 87 in an inguinal position; therefore 49 could not be found.
Comparing the US results with the surgical findings, there was a 100% positive predictive value (PPV) for the 16 abdominal testes seen on US; all were indeed found abdominally. Nevertheless, 14 viable abdominally located testes were missed by US. At surgery, all these testes were found in the lower abdomen—on the iliopsoas muscle, in the pelvis, or close to the internal inguinal ring. US located 87 inguinal testes, 84 of which were indeed found in the inguinal region (97% PPV); the other three were viable and found intraabdominally. Thirty-five of the NPT were defined as small, atrophic or vanished at operation.
Ultimately, US failed to locate 49 NPT. Nevertheless, 16 of these at surgery appeared to be viable, with 14 located intraabdominally and two in the inguinal region. Two atrophic testes were found in the abdomen. A further 22 atrophic testes were found inguinally or scrotally. Anorchia was present in nine boys.
In summary, 92 (85%) of the 108 NPTs defined as viable at operation were located by US. While nearly all viable inguinal testes (74/76) were correctly located by US, only half of the viable abdominal testes (16/33) were seen on US. The relevant data are summarized in Fig. 1.
Fig. 1Ultrasound findings in 152 nonpalpable testes in relation to findings at operative exploration
Discussion
Over the years, a wide range of investigations—from vascular techniques and imaging techniques to laparoscopy—have been used in detecting NPT [17]. Inguinal exploration used to be the standard surgical approach for nonpalpable testes. If the testis was not found at or below the internal inguinal ring, the procedure was extended to the abdomen. Laparoscopy has gained greater acceptance in diagnosing and treating NPT. Cortesi et al. [18] first reported its use in 1976 in an 18-year-old adolescent. While diagnostic laparoscopy is highly sensitive in detecting NPT [1, 2, 19, 20], it carries an approximately 1% risk of major or minor operative complications. Furthermore, the long-term incidence of peritoneal adhesions following laparoscopic procedures in children is approximately 10% [4, 21].
The literature is far from unanimous on the usefulness of US in UDT. Some authors recommend US for its feasibility and potential to settle the subsequent operative procedure [7, 21–27]. Others are sceptical because they found poor outcome for NPT, and because US does not exclude the necessity for laparoscopy [3, 28–30]. Most of the studies included only small groups of patients [24–26, 28, 30]. Not all studies investigated only NPT; some included all UDT [26, 27, 30]. Most studies compared US results with findings during the operative procedure [3, 7, 21, 22, 26–29], others compared CT and/or MRI with US [23–25, 30] in relation to the operative procedure. The most relevant conclusions of these studies are summarized in Table 1.
Table 1Review of the literature on the use of US in boys with NPT (n.a. not applicable, because all testes were ‘true’ NPT)ReferenceStudy designTestes (n)Located by USLocated “true” NPTAccuracy of US (%)aPredictive value of US (%)aRe-examinationRemarksAdvice in NPTPositiveNegativeTestes palpablePrevious examiner[22]Retrospective/medical charts5528/55 (51%)n.a.8410067n.a.–US[29]Prospective/cohort6237/62 (60%)n.a.761000n.a.Vanished not included in accuracy and predictive valuesLaparoscopy[3]Retrospective/medical charts6612/66 (18%)0/21 (0%)18b100b0b45/66 (68%)Referring physician82% US ordered by referring physician; 93% US performed elsewhereLaparoscopy[23]Retrospective/medical charts170111/170 (65%)n.a.72b100b19bn.a.–US[21]Prospective/cohort7448/74 (65%)16/29 (55%)65b100b0b45/74 (61%)Referring physicianOnly nubbins palpable on re-examinationUS[24]Prospective/cohort2113/21 (62%)n.a.769250n.a.US twice and blind to physical examinationUS → MRI[30]Retrospective/medical charts225/12 (42%)3/7 (43%)581002913/22 (59%)Referring physician–Laparoscopy[7]Prospective/cohort4545/45 (100%)6/8 (75%)9393–36/45 (80%)Referring physicianPhysical examination and US blind to each otherUS[26]Prospective/cohort3817/38 (45%)1/18 (6%)6188381/21 (5%)ExpertUS blind to physical examinationUS if obese20/41 (49%)Referring physician[28]Prospective/cohort142/14 (14%)n.a.211008n.aSurgery blind to USLaparoscopy[27]Prospective/cohort6961/69 (88%)4/12 (33%)9910088n.aUS[25]Prospective/cohort2315/23 (65%)n.a.919175n.a.US and CT blind to each otherUSaAccuracy and predictive values were (re)calculated: negative surgical exploration/absent testes were counted as true negative and testicular remnants/“nubbins” as false negative.bAbsent testes and atrophic testes grouped; number of absent testes not (exactly) mentioned in article (calculated as false negative).
An important message emerging from these articles is the very high PPV of US and the importance of thorough examination by a physician with experience in small genital examination. More than two-thirds of testes of referred patients were palpable on examination by experienced physicians. Many US investigations could have been avoided if patients had first been referred to an expert. The (paediatric) radiologist’s experience is also vital and US should be undertaken in the hospital where further treatment will be given.
US remains the modality of choice in our centre because we value its noninvasiveness, child-friendliness and cost-effectiveness. It also facilitates planning the correct surgical procedure (Fig. 2) [7, 29].
Fig. 2Flow chart representing the diagnostic and therapeutic approach in boys with NPT
In summary, we found 97% sensitivity of US for viable inguinal testes and 48% sensitivity for viable abdominal testes. When US located a testis it was also found at that site during surgery in 97% of NPT (PPV 97%). In our study, only three viable testes were not located correctly, being in the inguinal canal at US and within the abdomen at surgery. This can be explained by the fact that there was a patent internal inguinal ring with a mobile testis. More than two-thirds of NPT could be localized with US. Approximately one-third of NPT that were not found with US appeared to be viable at surgery. Of 34 testes considered nonpalpable by a general practitioner or paediatrician, 21 (62%) were palpable on re-examination by a paediatric surgeon after US. In contrast, of 118 testes considered nonpalpable by a paediatric surgeon, only 21 (18%) were palpable on re-examination.
Conclusion
We feel our results underpin the importance of referring boys with a presumed NPT to a physician with expertise in genital examination of small children. This will avoid unnecessary US investigations. Furthermore, we recommend US for all boys with NPT diagnosed by an experienced physician (Fig. 2). The findings would then determine the subsequent operation: diagnostic and/or therapeutic laparoscopy for all NPT with negative US, or intraabdominal testes located by US. In addition, we recommend inguinal exploration whenever US demonstrates an inguinal testis.
|
[
"nonpalpable",
"testes",
"children",
"ultrasound"
] |
[
"P",
"P",
"P",
"U"
] |
J_Mol_Biol-2-1-2279256
|
Predicting Protein Function from Structure—The Roles of Short-chain Dehydrogenase/Reductase Enzymes in Bordetella O-antigen Biosynthesis
|
The pathogenic bacteria Bordetella parapertussis and Bordetella bronchiseptica express a lipopolysaccharide O antigen containing a polymer of 2,3-diacetamido-2,3-dideoxy-l-galacturonic acid. The O-antigen cluster contains three neighbouring genes that encode proteins belonging to the short-chain dehydrogenase/reductase (SDR) family, wbmF, wbmG and wbmH, and we aimed to elucidate their individual functions. Mutation and complementation implicate each gene in O-antigen expression but, as their putative sugar nucleotide substrates are not currently available, biochemical characterisation of WbmF, WbmG and WbmH is impractical at the present time. SDR family members catalyse a wide range of chemical reactions including oxidation, reduction and epimerisation. Because they typically share low sequence conservation, however, catalytic function cannot be predicted from sequence analysis alone. In this context, structural characterisation of the native proteins, co-crystals and small-molecule soaks enables differentiation of the functions of WbmF, WbmG and WbmH. These proteins exhibit typical SDR architecture and coordinate NAD. In the substrate-binding domain, all three enzymes bind uridyl nucleotides. WbmG contains a typical SDR catalytic TYK triad, which is required for oxidoreductase function, but the active site is devoid of additional acid–base functionality. Similarly, WbmH possesses a TYK triad, but an otherwise feature-poor active site. Consequently, 3,5-epimerase function can probably be ruled out for these enzymes. The WbmF active site contains conserved 3,5-epimerase features, namely, a positionally conserved cysteine (Cys133) and basic side chain (His90 or Asn213), but lacks the serine/threonine component of the SDR triad and therefore may not act as an oxidoreductase. The data suggest a pathway for synthesis of the O-antigen precursor UDP-2,3-diacetamido-2,3-dideoxy-l-galacturonic acid and illustrate the usefulness of structural data in predicting protein function.
Introduction
The acceleration of genome sequencing has created a challenge for modern biology, namely, interpretation of the vast and growing amount of genetic information in the databases. Search algorithms such as the basic local alignment search tool (BLAST)1 can efficiently find related genes, but only in a small minority of cases is the precise function of a novel enzyme equivalent to that of the highest scoring hit. In straightforward cases the output is useful to generate hypotheses for the novel enzyme's function, based on common catalytic chemistry, substrates or both, and these postulates are then amenable to direct biochemical investigation. To make full use of genome data it is important that where such standard bioinformatic analysis cannot predict precise roles and biochemical function assays are not possible, tools are developed to differentiate the functions of genes. A good example of such a challenging case is where gene products are annotated as short-chain dehydrogenase/reductases (SDRs). Members of this family share low sequence identity (typically 15–30%) and catalyse a wide range of chemical reactions, including oxidation, reduction, epimerisation, dehydration and decarboxylation (reviewed in Ref. 2).
Bordetella
parapertussis and Bordetella bronchiseptica are pathogens of mammals. B. bronchiseptica is associated with respiratory tract infections in many animals including acute tracheobronchitis in dogs (kennel cough)3 and atrophic rhinitis in pigs,4 whereas B. parapertussis can cause whooping cough in humans.5 A separate lineage of B. parapertussis infects sheep.6
The major component of the outer leaflet of the outer membrane in Gram-negative bacteria is lipopolysaccharide (LPS). The complete structure of this macromolecule in B. bronchiseptica and B. parapertussis has been recently described.7 The lipid A domain of LPS, which consists of a diglucosamine backbone substituted with fatty acyl chains, forms the outer leaflet of the outer membranes of the bacteria. Lipid A is linked to a complex, branching oligosaccharide known as the core. Lipid A-core comprises a proportion of the LPS that is exposed on the cell surface and is known as band B. In B. bronchiseptica, lipid A-core can be substituted with a trisaccharide and this structure is known as band A. Synthesis of the band A trisaccharide requires functions encoded in the wlb gene cluster (previously named bpl).8
Both B. bronchiseptica and B. parapertussis also substitute their LPS with an O antigen. This O antigen contains 12 to 15 2,3-diacetamido-2,3-dideoxy-l-galactosaminuronic acid (l-GalNAc3NAcA) residues9 and is required for full virulence in animal and in vitro models of infection.10 O-antigen biosynthesis requires genes in the wbm cluster that is adjacent to the wlb genetic locus. The wbm locus contains three neighbouring SDR genes (wbmF, wbmG and wbmH), all of which have been annotated as nucleotide-sugar epimerases/dehydratases.
The structure of complete O antigen and the homology of wbm genes with genes of known function have led us to propose a pathway for biosynthesis of this polysaccharide and we are currently testing various steps in this scheme as part of an ongoing project to determine the functions of all 24 wbm genes. This report specifically concerns the roles of the SDR genes wbmF, wbmG and wbmH.
O-antigen residues are synthesised as sugar nucleotide precursors. The probable substrate for wbm locus-encoded biosynthesis of the nucleotide-activated l-GalNAc3NAcA is UDP-2,3-diacetamido-2,3-dideoxy-d-mannuronic acid (UDP-d-ManNAc3NAcA). This compound is related to the l-galacto configuration by inversion of the stereochemistry at the hexose 3 and 5 positions. In the bordetellae, UDP-d-ManNAc3NAcA is formed as a precursor for band A trisaccharide synthesis by 2-epimerisation of UDP-2,3-diacetamido-2,3-dideoxy-d-glucuronic acid, catalysed by WlbD.11 Because there are precedents for either a single12 or multiple13–16 SDR enzymes catalysing 3,5-epimerisation conversions of sugar nucleotides, we hypothesise that one or more of WbmF, WbmG and WbmH operate in this biosynthetic pathway to catalyse the conversion of UDP-d-ManNAc3NAcA to UDP-2,3-diacetamido-2,3-dideoxy-l-galacturonic acid (UDP-l-GalNAc3NAcA). Because of the low percentage conversion of the WlbD-catalysed reaction, UDP-d-ManNAc3NAcA is not currently available for biochemical studies and therefore this hypothesis cannot be directly tested at the present time.
In this report we demonstrate the involvement of wbmF, wbmG and wbmH in O-antigen expression by mutation of the genes in B. bronchiseptica. Because the putative substrate is unavailable, we elected to probe the catalytic functions of WbmF, WbmG and WbmH by characterising these proteins' three-dimensional structures. With soaking and co-crystallisation experiments, densities corresponding to cofactors and nucleotide portions of substrate analogues can be identified within the active sites. Biochemical and crystallographic studies have defined the structural basis for the catalytic chemistries of a range of other SDR enzymes (reviewed in Ref. 17). Our knowledge of these structure–function relationships enables us to interpret the structures of the wbm SDRs. We have analysed the potential for acid–base chemistry in the active sites to differentiate their potential roles in vivo. Based on the composition of O antigen, the genetic context of these wbm genes and information from the structural studies, we propose roles for these genes in O-antigen biosynthesis. The results demonstrate the usefulness and limitations of X-ray data in elucidating biochemical pathways where catalytic activity cannot be directly measured.
Results
Mutational analysis of wbmF, wbmG and wbmH
To test the hypothesis that wbmF, wbmG and wbmH are involved in O-chain biosynthesis, each gene was disrupted by insertion of a tetracycline-resistance cassette. Mutation of the chromosomal genes was confirmed by Southern hybridisation (data not shown). The effects of these mutations were assessed by silver-stained SDS-PAGE analysis and immunoblotting of mutant LPS (Fig. 1). Mutation of wbmG or wbmH results in apparent abrogation of O-antigen synthesis, as the LPS from these mutants lacks O band as detected by either silver stain or Western blot. O-antigen expression in the wbmF mutant, CNF0a, is dramatically reduced compared with the wild type, but this strain retains its ability to express a small amount of material with the electrophoretic mobility of O-band LPS and which binds the O-antigen-specific monoclonal antibody, D13B11.
Each of the mutations was complemented by expression of the cognate wild-type allele in trans (Fig. 1). Each coding sequence was cloned into the broad host range vector pBBR1MCS(kan) under the control of the flaA promoter to generate complementation vectors. Complementation of the wbmG and wbmH mutations restored O-antigen expression to near wild-type levels. O-antigen expression in the complemented wbmF mutant was increased relative to the wbmF mutant, but was still considerably less than in wild type. The reason for incomplete complementation of the wbmF mutation is unknown, but is not likely to be due to polar effects of the mutation because, due to the organisation of genes in the wbm locus, wbmF is at the 3′ end of an operon.18 Introduction of the empty complementation vector had no effect on O-antigen expression in any of the mutants.
The gene products WbmF, WbmG and WbmH are all in the extended SDR family
The SDR family is divided into the classical and extended subfamilies primarily on the basis of protein length and the sequence pattern of the glycine-rich cofactor-binding region close to the N terminus. According to both criteria, the products of wbmF, wbmG and wbmH, which are predicted to have 357, 310 and 313 amino acids, respectively, and all have the GxxGxxG motif near the N terminus, are extended SDRs2 (Fig. 2).
WbmG and WbmH are closely related to each other, sharing 38% amino acid identity. WbmF, WbmG and WbmH are all homologous to Pfam01370 (NAD-dependent epimerase/dehydratase family). The most closely related, characterised homologues of these Wbm SDRs in the protein databases are deoxythymidine diphosphate (dTDP)-glucose 4,6-dehydratases from Escherichia coli and Streptomyces venezuelae with 27% or 28% sequence identity (Table 1).19 This level of sequence conservation between members of the SDR family does not imply conservation of either catalytic chemistry or substrate; rather this functional information can be better determined by examination of the residues at key positions in the active sites and binding pockets. Several conserved motifs within the extended SDR subfamily facilitate alignment of sections of the Wbm SDR sequences with homologues (Fig. 2) but many of the critical functional amino acids are located outside of these motifs, and with this low level of conservation, it is not possible to make full sequence alignments with confidence that all of these residues are correctly identified.
Crystallisation of His6-WbmF, His6-WbmG and His6-WbmH
Thus far the data implicated each of these wbm SDRs in O-antigen synthesis, but did not indicate precise roles for each gene. Therefore, in order to elucidate their individual functions we characterised the gene products by X-ray crystallography. The overexpression, purification and crystallisation of these proteins are described elsewhere.20 The structures discussed in the present work are summarised in Table 2.
General architecture of His6-WbmF, His6-WbmG and His6-WbmH
These proteins all exhibit the typical SDR family architecture (Fig. 3). The structures each comprise two domains. The first of these is the N-terminal Rossmann fold domain in which a central β sheet is flanked by two layers of α helices. The cofactor-binding motif GxxGxxG is located at the C-terminal edge of this β sheet, which has seven parallel β strands running in the order 3, 2, 1, 4, 5, 6, 7. The second domain is largely made up of C-terminal sequence and contains all of the residues involved in binding of the nucleotide portion of the substrate. The catalytic sites in SDR enzymes are located at the interface of these two domains where, in sugar-nucleotide-modifying enzymes, the substrate hexose is brought into proximity with the cofactor nicotinamide ring. The main architectural differences between these three proteins lie in their C-terminal regions (Fig. 3). The C terminus of His6-WbmG consists of a loop that stretches out from the C-terminal domain to interact with the Rossmann domain. In His6-WbmH, this loop is not visible in the density, although this difference may not reflect any distinction between the real structures of the two proteins in solution. In WbmF, the last 30 residues form a large bent helix that covers two faces of the C-terminal domain. There is also an insertion of 14 amino acids in WbmF relative to the other structures, including all residues from Gly198 to Arg212. This extra loop extends over the nicotinamide end of the cofactor-binding pocket.
SDRs typically form dimers or higher oligomers. The crystal packing of these proteins suggests that WbmF, WbmG and WbmH have the conserved four-helix bundle usually found at the dimer interface. Analytical gel filtration chromatography indicated a hydrodynamic radius for each of these proteins consistent with either a monomer or dimer (data not shown). We conclude, therefore, that these proteins are present in solution predominantly in dimer form. In the WbmG and WbmH structures, the dimer interaction buries 1316 Å2 (5.7% of total protein surface) and 1562 Å2 (7.2%) of surface per dimer, respectively. In the WbmF structure the interface region is extended to include a short antiparallel β sheet composed of the residues Ile151 to Ser153 from each monomer as well as the loops that connect these strands to the bundle (Leu154 to Pro160); 2131 Å2 is buried in the WbmF dimer interface, which is equivalent to 8.6% of the total protein surface. In all three cases, the majority of the residues buried at these interfaces are hydrophobic or aromatic.
Cofactor binding
NAD can be modelled into the electron density in all three proteins. His6-WbmG and His6-WbmH copurify with this compound from the E. coli expression host, but in order to stabilise the protein in solution His6-WbmF required the addition of exogenous cofactor,20 and this is the most likely source of this molecule in these His6-WbmF structures. The preference in each case for NAD cofactor as opposed to NADP is indicated by the close interaction (2.7 Å hydrogen bonds) of an aspartate side chain with the 2′-hydroxyl of the adenylic ribose. This side chain would disfavour binding of NADP by both steric and electrostatic repulsion of the NADP 2′-phosphate. In some SDR structures [e.g., GDP-4-keto-6-deoxymannose 3,5-epimerase/reductase (GMER)21 and RmlD22] the bound cofactor has direct access to solvent, indicating a pathway via which spent NAD/NADP can be replaced from solution. In these structures, the bound cofactor is buried within the core of the protein; in the case of the His6-WbmF structure, the 14-amino-acid insertion supplies an additional occluding loop, further blocking release of cofactor in this conformation.
Binding of substrate
Soaking and co-crystallisation experiments with His6-WbmF and His6-WbmG revealed the binding mode for the nucleotide portion of the sugar nucleotide substrates of these enzymes (Fig. 3). The X-ray data from a His6-WbmG crystal soaked with UDP-glucose revealed extra density that could be modelled as far as the α-phosphate. The location of this molecule is consistent with the nucleotide-binding site for SDR homologue structures and with the hydrogen-bonding chemistry expected to bind uridine. No density was visible corresponding to the hexose portion of the soaked compound nor the second phosphate, suggesting that either this portion of the molecule is disordered within the crystal, or the compound was partially hydrolysed and the smaller uridine monophosphate (UMP) preferentially soaked into the substrate-binding pocket. In either case, the data indicate that the glucose-1-phosphate moiety of UDP-glucose is not strongly bound in a single conformation by WbmG. A similar soaking experiment but with GDP-mannose did not indicate any binding of this sugar nucleotide (data not shown), indicating a possible preference for the uracil base. The hydrogen bonding between the WbmG pocket and UMP suggests a basis for this preference (Fig. 3b). Although we were unable to obtain any complex of WbmH with a substrate analogue bound, the nucleotide-binding domain of WbmH is conserved in sequence and structure with WbmG, implying identical specificity and a common binding mode for the nucleoside portion of their substrates (Fig. 3c). UDP density to the α-phosphate is also visible in the WbmF/NAD/UDP ternary complex. The UMP density in this structure and the chemistry of its interaction with the protein suggests that this protein, like WbmG and WbmH, recognises a UDP-sugar substrate.
Both α and β nucleotide phosphates could be modelled in the data from a UDP soak into a WbmF-NAD+ co-crystal and co-crystallisation of His6-WbmG with UDP. In both structures the β-phosphate is coordinated by a conserved arginine (Fig. 4a and b), although the precise location of this phosphate is different in the two structures and both are inconsistent with the β-phosphate positions of bound substrate in the structure of a D128N, E129Q mutant of dTDP-glucose 4,6-dehydratase, DesIV, from S. venezuelae (Fig. 4c).23 In this DesIV structure [Protein Data Bank (PDB) ID 1R6D] the position of the substrate is consistent with the necessary overlap of molecular orbitals required for hydride transfer from the glucose C-4 to the cofactor. It is likely, therefore, that the β-phosphate positions in WbmF and WbmG structures do not represent the positions of equivalent phosphates when the native substrates for these enzymes are bound in the active site.
One common feature of substrate binding in SDR enzymes is pi stacking of the nucleobase with an amino acid side chain. In WbmF, this interaction involves Glu231 (Fig. 3a). Unusually, in WbmG the pi-stacking residue (Phe190) is the one that hydrogen bonds with the uracil N-3 (Fig. 3b). In WbmF, E. coli UDP-galactose 4-epimerase (GalE),24
S. venezuelae DesIV,23 and Pseudomonas aeruginosa WbpP,25 the hydrophobic contact is provided by the residue in the n + 2 position, whose backbone amide hydrogen bonds with the base's O2 group (Fig. 5). This unusual feature of the substrate binding in WbmG could not have been predicted from sequence alignments alone.
Potential catalytic chemistry
SDR enzymes have a conserved catalytic triad that is involved in their oxidoreductase activity and that is classically composed of a spatially conserved serine, tyrosine and lysine (SYK) triad, for example, Ser142, Tyr166 and Lys170 in WbpP. The first member of this triad is sometimes found as a threonine (TYK),23 and the tyrosine can be replaced by a methionine (SMK).26 The SDR triad is conserved in WbmG and WbmH as TYK but in WbmF Ala131 superimposes onto the Ser/Thr position (Fig. 6a).
In the context of O-antigen synthesis in Bordetella we are interested in the potential of the wbm SDR enzymes to catalyse secondary reactions of their substrates beyond the oxidoreductase reactions common to all SDRs. Of particular significance is the identification of a potential 3,5-epimerase. Despite the lack of a structure in which a full sugar nucleotide substrate analogue is bound, the cavity into which the substrate hexose binds is easily identifiable by comparison with homologue structures and because it is necessarily interposed between the UMP-binding site and the cofactor nicotinamide ring. The active sites of WbmG and WbmH are devoid of amino acid side chains capable of acid–base chemistry except for their respective TYK triads and, in WbmH, Ser176. SDR enzymes that catalyse 3,5-epimerisation of their substrates have a spatially conserved catalytic cysteine and a basic side chain. In GMER these residues are Cys109 and His179;21 in GDP-mannose 3,5-epimerase (GME) they are Cys145 and Lys217.12 In WbmG and WbmH these amino acids superimpose onto hydrophobic side chains, but in WbmF this catalytic cysteine is conserved (Cys133) and there are two candidates for the basic side chain, Asn213 and His90 (Fig. 6b). Cys133 is disordered in all of our WbmF crystals, indicating that an element of induced fit may be required to bring all of the active-site components into place. His90 has a very unusual backbone conformation, with a cis peptide bond to the following residue. This is strongly suggestive of a functional role for this residue, since such conformations are rarely observed except where they are required to place a key side chain in the active site.
Modelling of the putative substrate into the WbmF active site
The analysis of the potential catalytic chemistry of WbmF suggested that its role in O-antigen biosynthesis may be to catalyse the 3,5-epimerisation required in the overall conversion of UDP-d-ManNAc3NAcA to UDP-l-GalNAc3NAcA. The 4-keto derivative of UDP-d-ManNAc3NAcA is therefore a likely substrate or reaction intermediate for WbmF. This compound contains a bulkier sugar than any of the substrates of the characterised sugar-nucleotide-modifying SDRs. An important test of our proposed pathway is that the enzymes must be able to accommodate these unusually large substrates. We modelled 4-keto UDP-d-ManNAc3NAcA into the active site of WbmF (Fig. 7). This showed that it is feasible for the substrate to occupy the active site of WbmF in a manner that is consistent with the experimentally determined binding site for UMP, the proximity of the most likely catalytic site, and good geometry.
Discussion
The abrogation or reduction of O-antigen expression as a result of wbmF, wbmG or wbmH mutations implicates all three genes in the biosynthesis of this molecule, and is consistent with the hypothesis that these three adjacent SDR genes catalyse the UDP-d-ManNAc3NAcA to UDP-l-GalNAc3NAcA conversion, which is probably required for O-antigen synthesis.
Structural analysis of these three wbm SDR gene products also enables us to distinguish between them in terms of potential catalytic chemistry. Both WbmG and WbmH have the conserved SDR catalytic triad (TYK in these cases), which suggests that they should be competent to function as oxidoreductases. In both of these enzymes, the pocket that surrounds the active site is hydrophobic and featureless compared with other SDR enzymes; the lack of side chains capable of acid–base chemistry indicates that the functions of these enzymes are likely to be limited to oxidation and/or reduction and probably rules out specific 3,5-epimerase activity. The hydroxyl of Ser176 in WbmH may be capable of facilitating proton exchange; however, an amino acid pair capable of acid–base chemistry is required for 3,5-epimerase function.
Some SDRs that catalyse both oxidation and reduction bind cofactor irreversibly and recycle it within the active site.27 In other SDRs, this binding is reversible and the cofactor-binding site is situated at the bottom of a solvent-accessible groove (e.g., in E. coli GMER, PDB ID 1BWS).28 No such groove is visible in the WbmF, WbmG and WbmH structures, where the cofactor-binding sites are almost completely occluded. However, we do not take this observation as being diagnostic of irreversible cofactor binding in these proteins for the following reasons. Firstly, exogenous cofactor is required for WbmF to be stable in solution with respect to precipitation.20 Presumably either cofactor that copurifies with WbmF can be released, resulting in unfolding and aggregation of the protein, or apo-WbmF must bind cofactor supplied from the medium in order to fold correctly. Either case would suggest that NAD can be exchanged between WbmF and solvent. Secondly, human GalE reversibly binds cofactor,29 and yet does not have an open binding site like that of GMER. Finally, a conversion of configuration from d-manno to l-galacto can be catalysed by a single tight NAD+-binding enzyme such as GME,12 but we cannot envisage a pathway in which two or three such enzymes contribute to this transformation. Instead, we suggest that the protein loops that cover the NAD-binding site may be flexible and move to allow release and replacement of spent cofactor.
WbmF lacks the serine/threonine of the SDR catalytic triad. The conservation of the classical SDR catalytic triad is not absolute amongst members of the family. For example, in the very unusual SDR enzyme, biliverdin IXβ reductase (which catalyses the reduction of a range of substrates including some flavins and non-α-isomers of biliverdin), a histidine occupies the usual place of the catalytic lysine and tyrosine residues.30 The shapes and electron delocalisation exhibited by biliverdin IXβ reductase substrates imply that compared with the putative substrates of the Wbm SDR enzymes, they will have rather special requirements for binding geometry and reaction intermediate stabilisation during catalysis. In those SDRs that oxidise or reduce sugar nuclotides at the hexose C-4 position, however, the catalytic triad serine/threonine residue is important because it modifies the pKa of the C-4 hydroxyl by forming a short hydrogen bond.23 Mutation of the triad threonine (Thr134) to alanine in E. coli dTDP-Glc 4,6-dehydratase results in a 200-fold drop in kcat,31 whereas mutation of Ser124 to alanine in E. coli GalE reduces epimerase activity (kcat) almost 3000-fold.32 The incomplete WbmF catalytic triad raises doubts about the competence of this enzyme as a sugar nucleotide oxidoreductase. It is possible that the function of the missing side chain is provided by another active-site residue; this implies a different substrate-binding geometry and further studies will be required to investigate this possibility. Conversely, if a standard substrate-binding orientation is assumed, the active site of WbmF resembles that of 3,5-epimerases and our structures of this protein identify Cys133 and either Asn213 or His90 as potential mediators of proton exchange in the epimerisation reaction—one group abstracts a proton from the sugar while the other donates a proton to the opposite face.
By modelling the compound 4-keto UDP-d-ManNAc3NAcA into the WbmF structure we have demonstrated that there is space within this active site to accommodate this bulky sugar nucleotide. Other solutions produced in the modelling process demonstrated alternative plausible substrate positions (data not shown). For this reason, we do not suggest that the solution presented here (Fig. 7) represents the true active conformation. While we did not attempt similar modelling experiments with WbmG and WbmH, visual inspection of these structures indicates that they also have spacious active sites compared with other SDRs and can therefore accommodate diacetamido uronic acids in their hexose-binding pockets.
We therefore propose the following biosynthetic pathway (Fig. 8) for the 3,5-epimerisation of UDP-d-ManNAc3NAcA: One of the putative oxidoreductases, WbmG or WbmH, catalyses the initial oxidation at C-4, and the other oxidoreductase is responsible for the final reduction step. Introduction of the keto group lowers the pKa of the C-3 and C-5 hydrogen atoms, allowing WbmF to catalyse proton exchange effecting 3,5-epimerisation. This scheme is the most consistent with all of the available data. The function suggested for WbmF is very unusual in that we suggest that this enzyme does not catalyse oxidation or reduction of its substrate. If this is the case, WbmF would be the first such member of the SDR family, and this would imply that the cofactor in WbmF has a purely structural role, without participating in the catalytic cycle. It has been suggested that the transcriptional regulation protein NmrA from Aspergillus nidulans is a NAD-binding SDR family member that is not an enzyme at all. NmrA appears to function in control of nitrogen metabolism through physical interaction with the GATA family transcription factor AreA, possibly by controlling the rate of entry of AreA into the nucleus,33 and the structural equivalents of the SDR triad tyrosine and serine/threonine in NmrA are Met127 and Met113, respectively.34 Since the NmrA structure was reported, however, several SDR enzymes with SMK triads have been reported (e.g., WbpM from P. aeruginosa),26 so that the nonconservation of the SDR catalytic triad in NmrA should not rule out an enzyme function as absolutely as first appeared.34 An alternative possibility to the biosynthetic pathway we have outlined (Fig. 8) is that WbmF does in fact catalyse either oxidation or reduction as well as the 3,5-epimerisation, in which case the third SDR may participate in O-antigen expression by performing a function in the synthesis of the complicated sugar residue that adorns the nonreducing terminus of the O-chain.35 The nonredox role we propose for WbmF is consistent, however, with the lack of the conserved SDR triad in this enzyme, and if this protein operates purely as a 3,5-epimerase this may also suggest a mechanism by which the wbmF mutant is able to express O antigen, albeit in greatly reduced amounts. The product of the initial oxidation reaction is activated for proton exchange α- to the keto group and it may be that racemisation at these positions occurs at a biologically significant rate without an absolute requirement for enzyme catalysis.
Recent structural studies of the 3,5-epimerase in the dTDP-rhamnose pathway, RmlC,36 suggest that the rate-limiting step in its reaction involves a transition state in which the anomeric linkage is in the disfavoured equatorial position (which is also adopted by the reaction product), and that in order to avoid steric clashes and satisfy the requirement for axial proton abstraction, the reaction mechanism requires a twist-boat conformation such as that proposed in the GME reaction.12 Binding of substrate in a C-1 equatorial conformation may be part of the way that RmlC reduces kinetic barriers to this reaction. For the same reasons, the product of the proposed WbmF-catalysed 3,5-epimerisation will have the nucleotide attached to the l-GalNAcNAcA in the equatorial position in order to avoid 1,3,5-triaxial clashes. This transformation therefore involves considerable kinetic and thermodynamic barriers and will be slow without enzyme catalysis. However, if sufficient starting material is generated by the upstream oxidase, and the 3,5-epimerisation product is efficiently and selectively removed by the downstream reductase, the flux through this bottleneck may be sufficient to allow O-chain synthesis. Furthermore, the slower 3,5-epimerisation rate resulting from mutation of the wbmF gene may make relatively little impact on O-antigen expression levels if this conversion does not usually comprise the rate-limiting step for O-antigen synthesis. A similar effect may operate in the dTDP-l-noviose pathway in Streptomyces spheroides where observation of a naturally occurring l-rhamnoside analogue of novobiocin37 shows that 3,5-epimerisation of the dTDP-noviose precursor (dTDP-6-deoxy-d-xylo-4-hexulose) does occur despite the fact that in vitro, the 3,5-epimerisation catalysed by NovW has a 2000-fold lower kcat than RmlC38 (discussed in Ref. 39).
In conclusion, X-ray crystallography of WbmF, WbmG and WbmH enabled definition of these enzymes' NAD-cofactor preference and the substrate nucleotide specificity. We have been able to distinguish the likely function of WbmF from WbmG and WbmH, but cannot differentiate the roles of WbmG and WbmH. In order to achieve this, these proteins must be characterised in vitro. This will require chemical synthesis of UDP-d-ManNAc3NAcA and is beyond the scope of the present study. Subject to future experimental verification of predictions we have made, our data illustrate the usefulness of structural studies for investigating such challenging problems. While they also show why this methodology cannot universally supplant direct in vitro characterisation, this kind of structural analysis will prove essential for researchers to make full use of the gene databases.
Materials and Methods
Bacterial strains, plasmids and culture conditions
Bacterial strains used in this study are described in Table 3. Bordetella was grown on Bordet–Gengou agar (Difco) supplemented with 10% defibrinated horse blood (TCS Cellworks Ltd). E. coli was cultured in Luria–Bertani (LB) broth or on LB agar. All strains were incubated at 37 °C and ampicillin (100 μg ml− 1), kanamycin (50 μg ml− 1), tetracycline (10 μg ml− 1 for E. coli, 5 μg ml− 1 for B. bronchiseptica) or streptomycin (200 μg ml− 1) were added where required. Suicide plasmids were based on the host-restricted pEX100T backbone40 and broad host range shuttle vectors were based on a Kmr derivative of pBBR1MCS.41 For preparation of LPS, B. bronchiseptica was grown in medium supplemented with 50 mM MgSO4, as this maximises O-antigen expression in this strain.
DNA methods
Standard methods were used for DNA manipulations. Oligonucleotides were supplied by Sigma-Genosys. PCR was performed with template from boiled bacteria42 and Taq DNA polymerase (Promega) or KOD Hot Start DNA polymerase (Novagen).
Construction of wbmF, wbmG and wbmH mutants
The multiple cloning site (MCS) from pBluescript II SK+ (Stratagene) was excised using SacI and KpnI, blunt-ended and ligated into SmaI-cut pEX100T, to generate the MCS-containing suicide vector pEXMCS. The wbmF allelic exchange construct was prepared as follows. The wbmF gene in the plasmid Bb540g06 (a clone generated in the B. bronchiseptica RB50 genome sequencing project)43 was disrupted by insertion of a blunt-ended tetracycline-resistance (Tc) cassette into the internal NsiI site. The mutant allele was excised using Acc65I and XbaI, blunt ended, and then cloned into EcoRV-digested pEXMCS. For wbmG, the gene was amplified using primers 5′-ATATCTAGACATATGCGTATTCTGATCACCG-3′ (XbaI site underlined) and 5′-ATAAAGCTTTGATTACTGGCAACTCTTC-3′, and the PCR product was topoisomerase cloned into pCR2.1-TOPO using a TOPO-cloning kit (Invitrogen). wbmG was excised using XbaI and cloned into the XbaI site in pEXMCS. The wbmG gene was then disrupted by cloning a blunt-ended Tc cassette into the internal Acc65I site. The wbmH mutant allele was obtained by in vitro transposon-mediated mutagenesis of the wbm locus-containing cosmid, BbLPS118 (accession number AJ007747) using an EZ-Tn5™ <Tet-1> insertion kit (Epicentre). The <Tet-1> transposon, plus flanking wbmH DNA, was cut out by partial digestion with AluI and ligated into SmaI-cut pEX100T. Allelic exchange constructs were electroporated into E. coli SM10λpir and transferred to B. bronchiseptica by conjugation with E. coli SM10λpir as donor.44 Loss of the plasmid-encoded sacB gene in allelic exchange mutagenesis of B. bronchiseptica was selected for by growth on LB agar with reduced salt concentration and supplemented with 10% (w/v) sucrose.45
Complementation of mutations
The B. bronchiseptica
flaA promoter was amplified using primers 5′-GCTCTAGATAGGCGCATGCCATGGCC-3′ (XbaI site underlined) and 5′-AAGGATCCCATATGGAGGCTCCCAAGAGAGAA-3′ (BamHI and NdeI sites underlined), and cloned into the XbaI and BamHI sites in pBBR1MCS-kan to generate the vector pCompEmpty. wbmF was amplified using primers 5′-AAAAAAACATATGTTGCCAGTAATCATGAATGC-3′ (NdeI site underlined) and 5′-AAAAAGCTTAGATCTAGTCCGCCGTCTTATTTG-3′ (HindIII site underlined) and cloned into pCompEmpty using NdeI and HindIII to generate the wbmF complementation vector pCompF. wbmG was PCR amplified using primers 5′-AAAAAAACATATGCGTATTCTGATCACC-3′ and 5′-AAAAAGCTTAGATCTGCAACTCTTCAGGTCTTG-3′ and cloned into pCompEmpty using NdeI and HindIII, generating pCompG. wbmH was amplified using 5′-GAGAATTCCATATGAAGAAAGTATTTATTACGG-3′ and 5′-AAAAAGCTTAGATCTTTGTCGATGACCAGGATT-3′ and cloned into pCompEmpty using NdeI and HindIII, generating pCompH. Shuttle vectors were moved into B. bronchiseptica by conjugation with E. coli CC118 as donor46 and trans-acting transfer functions provided by E. coli S17-1 pNJ5000 as helper.44,47
SDS-PAGE, silver stain and immunoblot analysis of LPS
LPS was obtained from B. bronchiseptica using a modification of the method of Hitchcock and Brown48 as has been described.49 SDS-PAGE of LPS was performed using Novex precast 16% tricine gels (Invitrogen). LPS was oxidised in-gel with periodic acid50 and visualised with the Silver Stain Plus kit (BioRad). Western blotting was performed as previously reported51 using the monoclonal antibody D13B11 that specifically binds the O-antigen-containing LPS of B. bronchiseptica strain CN7635E.35
X-ray crystallography
Protein production, purification, crystallisation and data collection were performed as described.20 Data were processed using DENZO and SCALEPACK (version 1.97).52 Data were truncated and converted to structure factors using TRUNCATE53 from the CCP4 Suite.54 Molecular replacement was carried out using PHASER55 as described.20 For the initial WbmG structure, phases were determined experimentally from selenomethionine-labelled protein: peak, inflection point and high-energy remote data were used to determine initial phases using the PHENIX package.56 Model building was carried out using COOT,57 and refinement was performed with REFMAC version 5.0.58 Initial models were improved using ARP/wARP version 6.1 where the data quality permitted.59 The CCP4i interface60 was used where appropriate. TLS parameters were assigned (where used) using the TLSMD server.61 CNS62 was used to perform simulated annealing. Structures were validated using COOT, PROCHECK,63 WHATCHECK,64 and RAMPAGE.65
Table 2 provides a summary of the crystallographic statistics. Protein structures were analysed using Pymol,66 COOT,57 and CCP4MG.67 Surface area calculations were performed using the Richards' Rolling Probe method.68 More extensive details of the crystallographic methods are provided in the Supplementary Data.
Modelling of putative substrate into WbmF structure
Coordinates and parameters for the 4-keto derivative of UDP-2,3-dideoxy-2,3-diacetamido-d-mannuronic acid were prepared using the CCP4 monomer library sketcher with refinement in REFMAC 5.2.58 An initial structure was prepared by modelling the proposed substrate into the structure of WbmF co-crystallised with UDP in COOT.57 The observed density was used to place the atoms present in UMP, and to place sugar atoms in positions occupied by water or glycerol in the observed structure. Different orientations were used in the two molecules in the asymmetric unit to increase the sampling of conformational space. The structure of WbmF co-crystallised with UDP, with 4-keto UDP-d-ManNAc3NAcA placed in, was truncated to include only protein, NAD and 4-keto UDP-d-ManNAc3NAcA atoms. This was used as an input to MODELLER version 9.2 to determine an optimised structure for the complex with the sugar nucleotide.69 After modelling, the ligand coordinates were regularised in COOT to correct minor distortions. Ten modelling trials were performed with different seeds, and the best was selected by eye on the basis of geometry, agreement with the determined structures and lack of bad interactions.
Protein Data Bank accession codes
X-ray amplitudes, phases and the derived atomic coordinates have been deposited with the PDB under the accession codes 2PZJ, 2Q1T, 2Q1U, 2Q1S, 2PZK, 2PZL, 2PZM and 2Q1W.
|
[
"short-chain dehydrogenase/reductase",
"bordetella",
"o-antigen biosynthesis",
"lipopolysaccharide",
"x-ray crystallography",
"dtdp, deoxythymidine diphosphate",
"l-galnac3naca, 2,3-diacetamido-2,3-dideoxy-l-galacturonic acid",
"gme, gdp-mannose 3,5-epimerase",
"gmer, gdp-4-keto-6-deoxymannose 3,5-epimerase/reductase",
"lps, lipopolysaccharide",
"pdb, protein data bank",
"blast, basic local alignment search tool",
"sdr, short-chain dehydrogenase/reductase",
"udp-d-mannac3naca, udp-2,3-diacetamido-2,3-dideoxy-d-mannuronic acid",
"udp-l-galnac3naca, udp-2,3-diacetamido-2,3-dideoxy-l-galacturonic acid",
"ump, uridine monophosphate",
"mcs, multiple cloning site"
] |
[
"P",
"P",
"P",
"P",
"P",
"M",
"M",
"R",
"R",
"R",
"R",
"M",
"R",
"M",
"M",
"R",
"R"
] |
Eur_J_Epidemiol-3-1-1914270
|
Inter-scan reproducibility of coronary calcium measurement using Multi Detector-Row Computed Tomography (MDCT)
|
Purpose To assess inter-scan reproducibility of coronary calcium measurements obtained from Multi Detector-Row CT (MDCT) images and to evaluate whether this reproducibility is affected by different measurement protocols, slice thickness, cardiovascular risk factors and/or technical variables.
Introduction
A considerable proportion of the western society is at risk of suffering a cardiovascular event during life. Atherosclerosis is one of the main underlying processes. Non-invasive assessment of atherosclerosis is important since it allows studies into the etiology and consequences of early and advanced atherosclerosis in populations at large [1]. The last two decades, measurement of coronary artery calcification (CAC) using computer tomography (CT) has been used to assess coronary atherosclerosis non-invasively. The presence, and more importantly, the quantity of CAC, relates well with the overall severity of the atherosclerotic process [2]. Several studies have demonstrated a strong relation between coronary calcium burden and the incidence of myocardial infarction, a relation which was independent of age [3, 4].
Most of the evidence on determinants and consequences of coronary calcium is based on data obtained with electron beam CT (EBCT) [5–7]. The availability of EBCT scanners is modest, whereas the Multi Detector-Row CT (MDCT) scanners are more widely available and also allow for detection of coronary calcium. Current data suggest that EBCT and MDCT give comparable results [8, 9]. In contrast to EBCT, however, data on reproducibility of CAC measurements using MDCT images is not widely available [10, 11], but information is relevant. Furthermore, due to technical improvement, slice thickness of the images has become smaller which may affect the likelihood of detecting coronary calcium, and hence its reproducibility.
We set out to study inter-scan reproducibility of coronary calcium measurements from MDCT images and to evaluate whether reproducibility is affected by different measurement protocols, slice thickness, selected cardiovascular risk factors and technical variables.
Materials and methods
Participants were recruited from the PROSPECT study [12], cohort of 17,357 healthy breast-cancer screening participants, aged 49–70 years, living in Utrecht and surroundings, enrolled between 1993 and 1997. Between October 2002 and December 2004, a random selection of 1,996 women were invited by mail and 1,000 (50.1%) who were postmenopausal and did not use contraceptives or hormone replacement therapy answered positively. Of these 1,000 women, a random selection of 573 underwent a MDCT examination during a single visit and 76 of them were scanned twice. The Medical Ethical Committee of the University Medical Center Utrecht approved the study and written informed consent was obtained from all participants.
Current cardiovascular drug use (blood pressure lowering, lipid lowering and glucose lowering drugs) was assessed by asking women to bring all packages to the study centre. Smoking behavior, medical history and cardiovascular family history were assessed by a questionnaire. Height and weight were measured and body mass index was calculated as weight divided by height squared (kg/m2). Waist-to-hip ratio (WHR) was assessed. Systolic and diastolic blood pressures were measured at both arms with an automated and calibrated blood pressure device (DINAMAPTM XL, Critikon, Johnson & Johnson, Tampa, Florida, USA) with the subject in supine position. A venous blood sample was drawn after an overnight fast of at least eight hours. Plasma total cholesterol, plasma triglycerides, and plasma glucose were measured using standard enzymatic procedures. High-density lipoprotein (HDL) cholesterol was measured by the direct method (inhibition, enzymatic). Low-density lipoprotein (LDL) cholesterol was calculated using the Friedewald formula.
Coronary imaging and calcium measurements
The amount of calcium in the coronary arteries was assessed with a Multi Detector-Row CT (MDCT) scanner (Mx 8000 IDT 16, Philips Medical Systems, Best, The Netherlands). Subjects were positioned within the gantry of the MDCT scanner in supine position. During a single breath hold, images of the heart, from the level of the tracheal bifurcation to below the base of the heart, were acquired using prospective ECG triggering at 50–80% of the RR-interval, depending on the heart rate. Scan parameters were 16 × 1.5 mm collimation, 205 mm field of view (FOV), 0.42 s rotation time, 0.28 s scan time per table position, 120 kVp and 40–70 mAs (patient weight <70 kg: 40 mAs; 70–90 kg: 55 mAs; >90 kg: 70 mAs). Scan duration was approximately 10 s, depending on heart rate and patient size. We had the participant get up from the table and lay down again since in studies on change in CAC over one year it is not realistic to assume exactly the same position of the participant at both occasions. Therefore our patients sat up and consequently moved slightly between scans to mimic two separate scan runs.
From the acquired raw data, the whole volume was reconstructed with an intermediate reconstruction algorithm in non-overlapping data sets of 1.5 mm and 3 mm thick sections. Quantification of coronary calcium was performed on a separate workstation with software for calcium scoring (Heartbeat-CS, EBW, Philips Medical Systems, Best, The Netherlands). All regions with a density over 130 Hounsfield units were identified as potential calcifications.
After completing a training-program, one scan reader (AR) who was unaware of the scores of the first scan, manually selected the calcifications within one of the coronary arteries (left main, left anterior descending, left circumflex, right coronary artery, and PDA) and scored the second scan of the participants. To reduce the influence of noise, the minimum size of a calcified lesion was set at 0.5 mm2. The peak density in Hounsfield units and the area in mm2 of each selected region were calculated. The Agatston [13] calcium score was obtained by multiplying the area by a weighting factor that is dependent on the peak signal anywhere in the lesion. The scores of individual lesions were added to obtain the Agatston calcium score for the entire coronary tree. The total calcium volume was calculated by multiplying the area of the calcified lesion (measured in square millimeters) by section thickness (1.5 mm and 3.0 mm). The calcium volume for each coronary vessel was computed by summing the volumes of the lesions in that vessel for all sections. Finally, the total volume from all the vessels became the calcium volume for a subject. The mass method uses volumetric, density information and a calibration phantom of hydroxyapatite to calculate the actual mass of the calcified plaques [14].
In addition, information on breathing artifact (inconsistency of sternum bone in sagital section in mm), noise (standard deviation of enhancement in fixed cardiac area of 212 mm2) and mean heart rate (beats/min) during scan acquisition was collected.
Data analysis
The mean and standard deviations (SD) of coronary calcium were calculated for all scoring methods separately. Because of the skewed distribution of scores, medians were also computed. The Intra-class correlation coefficient was estimated for between scans data and for 1.5 and 3.0 mm slices thicknesses separately. The mean difference in score between scans was calculated as well as the absolute and relative differences.
To distinguish between random differences or systematic difference, information on mean and absolute differences is needed. One may assume a priori a non-differential misclassification in the calcium scores, but one has to show that with the results. When the chance of the 2nd result being higher or lower is equal, one would expect a mean difference of zero, with some standard deviation. The absolute difference will not be zero since all differences are ‘absolutised’, but it is expected that at least the mean difference is much less than the absolute difference. If however the chance of a higher or lower value in the 2nd scan is not equal, the mean difference will be plus or minus a certain value. In addition, the absolute difference will have a value close to that of the mean difference. Therefore we need both parameters.
To estimate a weighted kappa as measure of agreement of categorical variables, subjects were divided into four groups according to the mean Agatston score as proposed by Rumberger et al. [15]: A: 0–9 (absent-minimal), B: 10–99 (mild), C: 100–399 (moderate) and D: (400 (severe degree of calcification). This categorization is specifically for the calcium scoring method according to Agatston. Therefore we additionally categorized all scoring methods in their quartiles to calculate kappa as measure of agreement for all scoring methods.
The relation between risk factors, technical variables and measurement error was assessed using Spearman correlation coefficients. In a similar manner the relation between calcium level and measurement error was examined. Since logarithms of coronary calcium scores have generally been used in statistical analyses in other papers, we also studied the reproducibility of logarithmic transformed calcium score. Logarithmic analysis of coronary calcium scores was performed by calculating natural log of coronary calcium scores +0.001 (ln (CCS + 0.001)) because the logarithm of coronary calcium scores alone excludes all subjects with zero scores [16]. We defined relative difference as absolute difference divided by the mean calcium level multiplied in 100 and expressed in percent. Data analysis was performed with SPSS for windows, version 12.0. A statistically significant difference was assumed when the two-sided P-value was less than 0.05.
Results
Mean age was 67.3 ± 5.2 years. Fifty-five participants (72.4%) had a coronary calcification more than zero in Agatston (1.5 mm slice thickness). Table 1 shows the general characteristics of the 76 women who had two MDCT scans.
Table 1Characteristics of studied population (N = 76)MeanStd. deviationAge (year)67.35.2BMI (Kg/m2)26.33.9WHR0.840.06SBP (mmHg)133.918.9DBP (mmHg)71.79.1Total cholesterol (mmol/l)6.090.86LDL cholesterol (mmol/l)4.310.97HDL cholesterol (mmol/l)1.510.36Triglycerides (mmol/l)1.280.62Glucose (mmol/l) 4.050.69Heart rate (beat/minute)7211Current smoking (%)*11Former smoking (%)43Previous CVD (%)1Family history of CAD in either parents (%)10BMI = Body Mass Index; CAD = Coronary Artery Diseases; DBP = Diastolic Blood Pressure; LDL = Low Density Lipoprotein; HDL = High Density Lipoprotein; SBP = Systolic Blood Pressure; WHR = Waist to Hip Ratio* Percentages have been rounded
Table 2 presents information on calcium distributions by various scoring techniques and reproducibility results, by slice thickness. Overall, calcium scores were higher when based on the 1.5 mm slice thickness than based on the 3.0 mm slice thicknesses. The kappa agreement and Intra-class correlation coefficients between the two scans were high for all scoring methods, indicating that with respect to ranking of subjects all three methods are doing well. In addition, the mean differences in scores were relatively small compared to the absolute differences for all measurements, suggestion no systematic measurement errors. Finally, results for the scans with 1.5 mm slice thickness were similar to those for the 3.0 mm slice thickness (Table 2).
Table 2Characteristics of different coronary calcium scoring methods; effect of slice thickness on inter-scan reproducibilityMass 1st ScanMass 2nd Scan Volume 1st Scan Volume 2nd Scan Agatston 1st Scan Agatston 2nd Scan Slice thickness 1.5 mmMean32.2131.88154.52149.40170.33163.63Median6.156.0539.9736.5231.8532.00Agreement (k) Rumberger categories0.970.890.87Agreement (k) Quartiles0.840.810.88Mean difference0.35.16.7Absolute difference4.022.324.3Relative difference (%)12.414.614.5ICCC*0.990.990.98Slice thickness 3.0 mmMean25.5725.45131.45126.98140.06135.82Median4.003.6530.3021.9020.3018.00Agreement (k) Rumberger categories0.920.830.73Agreement (k) Quartiles0.840.840.84Mean difference0.14.44.2Absolute difference3.518.721.3Relative difference (%)13.714.715.4ICCC*0.990.980.98* Intra-class correlation coefficient
Table 3 presents the relation of cardiovascular risk factors with inter-scan mean difference. No consistent relations were found between risk factor levels and measurement error. Importantly, however was the observation that calcium level or the logarithm of the coronary calcium scores were not related to the mean difference between scans, whereas they were significantly related to the absolute and relative differences (Table 4, Figs. 1 and 2). These observations suggest that measurement error increases with increasing CAC levels, yet that this occurs in a random way.
Table 3Relationship between cardiovascular risk factors and inter-scan mean difference of coronary calcium scoring methods by MDCT (Slice thickness 1.5 mm)Inter-scan mean differenceCCS methodsMassVolumeAgatstonBiological variablesrP-valuerP-valuerP-valueBMI (Kg/m2)0.040.730.030.740.020.80Age (year)0.180.100.310.000.280.01Smoking(Categorical)−0.000.980.040.710.070.49WHR−0.030.730.080.480.130.24SBP (mmHg)0.100.370.160.140.240.03DBP (mmHg)0.160.140.050.610.110.34Cholesterol (mmol/l)−0.270.05−0.120.40−0.200.17LDL (mmol/l)−0.180.10−0.190.09−0.090.40HDL (mmol/l)−0.040.72−0.160.14−0.110.34Triglyceride (mmol/l)−0.020.850.130.240.110.34Glucose (mmol/l)0.160.24−0.000.980.000.98Mean heart rate−0.030.77−0.030.73−0.020.81Technical variablesMean breathing artifact0.010.88−0.030.78−0.020.87Mean SD of noise0.130.260.080.490.070.52Coronary calciumMean mass score0.000.98Mean volume score0.030.75Mean Agatston score0.020.86Mean log mass score0.000.99Mean log volume score0.030.76Mean log Agatston score0.020.85BMI = Body Mass Index; DBP = Diastolic Blood Pressure; LDL = Low Density Lipoprotein; HDL = High Density Lipoprotein; r = spearman correlation coefficient; SBP = Systolic Blood Pressure; WHR = Waist to Hip RatioTable 4Relationship between cardiovascular risk factors and inter-scan absolute and relative difference of coronary calcium scoring methods by MDCT (Slice thickness 1.5 mm)Inter-scan relative differenceCCS methodsMassVolumeAgatstonBiological variablesrP-valuerP-valuerP-valueBMI (Kg/m2)0.070.530.080.460.090.43Age (year)0.210.060.240.030.l170.12Smoking(Categorical)−0.030.73−0.070.51-0.140.20WHR0.070.550.050.660.050.66SBP (mmHg)0.060.570.040.680.110.32DBP (mmHg)0.320.0040.310.0050.330.003Cholesterol (mmol/l)0.130.370.100.500.001.00LDL (mmol/l)−0.140.21−0.170.12-0.180.11HDL (mmol/l)0.070.520.040.670.060.57Triglyceride (mmol/l)0.030.780.070.490.000.99Glucose (mmol/l)0.230.090.260.050.240.08Mean heart rate−0.010.910.010.930.000.97Technical variablesMean breathing artifact0.100.440.090.490.150.23Mean SD of noise0.190.090.190.090.180.11Coronary calciumMean mass score0.290.009Mean volume score0.330.003Mean Agatston score0.380.001Mean log mass score0.290.010Mean log volume score0.330.003Mean log Agatston score0.370.001Inter-scan absolute differenceTechnical variablesMean breathing artifact0.120.320.120.330.150.22Mean SD of noise0.200.080.190.090.150.17Coronary calciumMean mass score0.86<0.001Mean volume score0.84<0.001Mean Agatston score0.89<0.001Mean log mass score0.86<0.001Mean log volume score0.83<0.001Mean log Agatston score0.89<0.001BMI = Body Mass Index; DBP = Diastolic Blood Pressure; LDL = Low Density Lipoprotein; HDL = High Density Lipoprotein; r = spearman correlation coefficient; SBP = Systolic Blood Pressure; WHR = Waist to Hip RatioFig. 1Relation between mean calcium score and inter-scan difference in mean calcium scores (Bland-Altman plots)Fig. 2Relation between mean calcium score and inter-scan absolute difference
Discussion
With respect to ranking of subjects, the inter-scan reproducibility of coronary calcium measurements by MDCT using Agatston, volume and mass scoring algorithms is excellent. The inter-scan reproducibility showed no major differences between scoring methods. The slice thickness did not affect reproducibility, nor did heart rate and technical parameters. Measurement error was related to increased coronary artery calcification, although our findings suggest that the error in the measurements is a random phenomenon.
Our findings, i.e., no major differences between scoring methods are in contrast with several reports on reproducibility based on EBCT scanning. Direct comparison of the findings of these studies with those of other is difficult since the parameters used to indicate reproducibility differ between studies. Furthermore, potentially the prevalence of CAC and its extent may affect reproducibility, as our findings suggest that measurement error increases with increasing CAC levels. Also the sizes of the studies differ which have undeniable effects on reproducibility results. However, our results are similar to those of by Rumberger and Kaufman [17], who compared these three methods and did not find any one method preferable to another in terms of reproducibility of results from consecutive scans in a patient.
Although the correlation between inter-scan measurements is excellent [18, 19], it still occurs that subjects with small deposits of calcium in scan one may have larger deposits of calcium in the 2nd scan, which leads to proportionally larger error in reproducibility. This has triggered other studies [20] on reproducibility to suggest that “the variability is partially a function of the absolute calcium score and inversely related to it”, implicating that low coronary calcium scores may not be reproducible. However, our results could not confirm this.
Besides different algorithms for calcium scoring, slice thickness has been reported to affect the reproducibility of scoring protocols. In our study, the reproducibility of the coronary calcium measurements by MDCT was similar for 1.5 mm as for 3.0 mm slice thickness, and equal for Agatston, volume and mass measurements confirming the results by Rumberger and Kaufman [17].
The implications of our main findings depend on the research question that is asked in studies using CAC measurements. When the interest is using CAC measurements for prognostic studies our results for kappa and ICCC show that ranking of subjects is adequate based on one CT scan. So the need for duplicate CAC scan is absent. The fact that measurement error increases with increasing CAC values, is in prognostic studies not of major importance since the categorization of individuals seems adequate. When the interest is in etiologic studies using CAC as outcome parameter, our findings show that risk factor relations will be validly estimated since none of the risk factors relates to measurement error. When the interest is in using CAC as risk factor for future events (assessment of relative risks), it is most likely that in analyses with CAC as continues variable the magnitude of association of high CAC levels with events reflects an underestimation of the true magnitude. The direction of the relation will not change since based on our results measurement error is random, leading to random misclassification of the exposure variable. When the interest is in diagnostic value of CAC measurements, which is usually done in categories of CAC, again the relations will be valid given our high kappa coefficients. Although our study was performed in healthy postmenopausal women, we expect that the finding will also be applicable for men.
Our findings are important in the light of the wider availability of MDCT in countries compared to EBCT. One reason for that is lower equipment cost. Other advantages of MDCT over EBCT have been suggested to be less quantum noise, thinner section thickness, and simultaneous acquisition of four sections (with 16-slice or with 64-slice ), which is reported to reduce misregistration artifact.
In conclusion, our findings demonstrate that coronary calcium measurements by MDCT are highly reproducible and are not affected by scoring protocols, slice thicknesses and technical factors.
|
[
"reproducibility",
"multi detector-row ct (mdct)",
"atherosclerosis",
"coronary artery calcification",
"epidemiology"
] |
[
"P",
"P",
"P",
"P",
"U"
] |
Biochim_Biophys_Acta-1-5-1906864
|
Oxidation of chylomicron remnant-like particles inhibits their uptake by THP-1 macrophages by apolipoprotein E-dependent processes
|
The influence of the oxidative state of chylomicron remnants (CMR) on the mechanisms of their uptake and induction of lipid accumulation by macrophages derived from the human monocyte cell line, THP-1, during foam cell formation was investigated using chylomicron-remnant-like particles (CRLPs) at 3 different levels of oxidation. The oxidative state of CRLPs was varied by exposure to CuSO4 (oxCRLPs) or incorporation of the antioxidant, probucol (pCRLPs) into the particles. oxCRLPs caused significantly less accumulation of triacylglycerol in the macrophages than CRLPs, and their rate of uptake was lower, while pCRLPs caused more lipid accumulation and were taken up faster. Uptake of all 3 types of particles was inhibited to a similar extent when entry via the low density lipoprotein (LDL) receptor related protein (80–90%), LDL receptor (− 30–40%), CD36 (− 40%) and phagocytosis (− 35–40%) was blocked using lactoferrin, excess LDL, anti-CD36 and cytochalasin D, respectively, but blocking scavenger receptors-A or -B1 using poly inosinic acid or excess HDL had no effect. These findings show that oxidation of CRLPs lowers their rate of uptake and induction of lipid accumulation in macrophages. However, oxidation does not change the main pathways of internalisation of CRLPs into THP-1 macrophages, which occur mainly via the LRP with some contribution from the LDLr, while CD36 and phagocytosis have only a minor role, regardless of the oxidative state of the particles. Thus, the effects of CMR oxidation on foam cell formation contrast sharply with those of LDL oxidation and this may be important in the role of dietary oxidized lipids and antioxidants in modulating atherosclerosis.
1
Introduction
Chylomicron remnants (CMR) carry lipids of dietary origin from the gut to the liver for processing [1] and there is now a large and growing body of evidence indicating that these lipoproteins are strongly atherogenic. They have been shown to be taken up into the artery wall as efficiently as LDL [2–5]; remnant-like particles containing apolipoprotein E (apoE) have been isolated from human aortic intima and atherosclerotic plaque [6,7]; and delayed clearance of CMR from the circulation correlates with lesion development [8,9]. Moreover, we and others have shown that CMR induce extensive lipid accumulation causing foam cell formation in human monocyte-derived macrophages (HMDM) [10] and in human and murine monocyte/macrophages cell lines [11,12].
Low density lipoprotein (LDL) plays a major role in atherogenesis and in foam cell generation, but oxidation of the lipoprotein particles, a process which can occur within the artery wall, is necessary before extensive lipid accumulation is induced [13]. In striking contrast, CMR do not require prior oxidation to cause macrophages to form foam cells [10–12]. However, our studies have demonstrated that incorporation of lipophilic antioxidants into the particles enhances, rather than inhibits, lipid uptake and accumulation in the cells [14,15], suggesting that the oxidative state of CMRs may play a role in their induction of foam cell formation, but in the opposite way to that of LDL. Oxidized CMR could occur either in the artery wall by the action of the cell-associated lipoxygenase and myeloperoxidase which are believed to oxidize LDL, or in the circulation, because dietary oxidised lipids, which are produced when fat is cooked at high temperatures, have been shown to be transported in these lipoproteins [16,17]. Clearly, therefore, it is important for the understanding of the atherogenicity of CMR to establish how their oxidation influences their uptake and induction of foam cell formation and the pathways by which CMR are internalised by the cells.
It has been demonstrated that CMR are taken up by the liver by apolipoprotein E (apoE)-dependent pathways mediated by the LDL receptor (LDLr) and the LDL receptor-like protein (LRP) [1]. The exact mechanisms by which CMR are taken up by macrophages, however, are not yet definitively established, and nothing is known about the effects of oxidation of the particles on the routes by which they are internalised. The LDL receptor (LDLr) appears to play a part [1], but as it is down-regulated by the influx of cholesterol into cells, native LDL does not induce foam cell formation [16], and our studies [18] have suggested that the delivery of cholesterol to macrophages by CMR has a similar effect. Thus, other mechanisms are also likely to be involved, and evidence from experiments using antibodies to the LDLr and animals lacking the LDLr supports this view [10]. Candidates include the LRP [19], the apoB48 triacylglycerol-rich lipoprotein receptor (apoB48r) [20], an as yet unidentified 43 kDa protein described by Elsegood et al. [21] and scavenger receptors such as scavenger receptor A (SR-A) and CD36 [1]. Phagocytosis has also been suggested as a possible mechanism [22]. Since the route of uptake of LDL by macrophages is profoundly changed after oxidation from the regulated LDLr to the unregulated scavenger receptors [16], it is important to establish whether the oxidative state also alters the mechanisms of uptake of CMR by the cells.
The aim of this study is to investigate the effects of the oxidative state of CMR on their uptake by macrophages and on the accumulation of lipid within the cells, and to determine how oxidation affects the pathways involved in the internalisation of the particles. Chylomicron remnant-like particles (CRLPs) at three different levels of oxidation (CRLPs, oxidized CRLPs (oxCRLPs) and CRLPs containing the antioxidant probucol (pCRLPs)) and macrophages derived from the human monocyte cell line THP-1 were used as the experimental model, and the mechanisms of uptake were evaluated using specific inhibitors of the processes believed to be involved. The findings clearly demonstrate that oxidation of CRLPs reduces the rate of their uptake by THP-1 macrophages and decreases lipid accumulation in the cells, and further show that this is due to differential interaction with apoE dependent receptors.
2
Materials and methods
RPMI 1640 medium, fetal bovine serum (FBS), l-alanyl-l-glutamine (glutamax) penicillin/streptomycin and β-mercaptoethanol were obtained from Gibco (Paisley, UK). FBS was heat inactivated (56 °C, 30 min) before use. Trypan blue, fatty acid-free bovine serum albumin (BSA), phospholipids, cholesterol, cholesteryl oleate, phorbol 12-myristate 13-acetate (PMA), Oil red O, probucol, poly inosinic acid (poly I), lactoferrin and cytochalasin D were supplied by Sigma (Poole, UK). 1′1′-dioctadecyl-3,3-3′,3′-tetramethylindo-carbocyanide perchlorate (DiI) was from Cambridge Bioscience (Cambridge, UK). The blocking antibody for CD36 was obtained from Immunodiagnostic Systems Ltd (Tyne and Wear, UK) and cholesterol oxidase from Merck Biosciences Ltd (Nottingham, UK). DiI-labelled acetylated LDL (acLDL) was purchased from Molecular Probes (Paisley, UK).
2.1
Preparation of lipoproteins
LDL and high density lipoprotein (HDL) were isolated from human plasma (National Blood Service, London UK) by ultracentrifugation. Plasma was layered under 0.9% NaCl (d = 1.006 g/ml), centrifuged for 5 h at 100,000×g (4 °C), and the top fraction discarded. The density of the bottom layer was raised to 1.063 g/ml with KBr, layered under KBr (d = 1.063 g/ml) and centrifuged for 16 h at 175,000×g (4 °C). LDL was collected from the top fraction by tube slicing. For the preparation of HDL, the bottom fraction was then adjusted to d 1.21 g/ml with KBr, layered under KBr (d 1.21 g/ml) and centrifuged at 175,000×g/ml at 4 °C for 18 h and the top fraction containing HDL was collected by tube slicing. LDL and HDL were dialysed against saline for 48–72 h prior to used.
CRLPs were prepared by sonication (power setting 22–24 μm, 20 min, at 56 °C) of a lipid mixture containing 70% trilinolein, 2% cholesterol, 3% cholesteryl ester and 25% phospholipids in 0.9% NaCl (w:v) in Tricine Buffer (20 mM, pH 7.4) followed by stepwise density gradient (2.5 ml d 1.065 g/ml, 2.5 ml d 1.020 g/ml, 3 ml d 1.006 g/ml) ultracentrifugation as described by Diard et al. [23] at 17,000×g for 20 min at 20 °C. The upper layer of grossly emulsified lipids was then removed and replaced with an equal volume of NaCl solution (d = 1.020 g/ml) and tubes were centrifuged at 70,000×g for 1 h (20 °C). For apoE binding, lipid particles collected from the top layer were incubated with the dialysed d 1.063–1.21 g/ml fraction of human plasma (National Blood Transfusion Service, North London Centre, UK) prepared as described above at 37 °C with shaking for 4 h (1 volume of particles: 2 volumes plasma). The CRLPs containing apoE were then isolated by ultracentrifugation at d = 1.006×g/ml (120,000×g for 12 h at 4 °C), harvested from the top layer, purified by a second centrifugation at the same density (202,000×g for 4 h at 4 °C) and stored at 4 °C under argon until required. All preparations were used within 1 week. For DiI-labelled CRLPs and pCRLPs, probucol (1 mg) and/or DiI were added to the lipid mixture prior to sonication. CRLPs were oxidized by incubation with CuSO4 (20 μM) with shaking for 5 h at 37 °C and the CuSO4 was then removed by dialysis (0.9% NaCl, 24 h, 4 °C). The oxidation process had no effects on the fluorescent properties of the DiI label.
2.2
Culture of THP-1 cells
THP-1 monocytes were maintained in RPMI 1650 culture medium containing 10% (v/v) FBS, glutamax (2 mM), penicillin/streptomycin (100 U/ml/100 μg/ml) and β-mercaptoethanol (20 μM) (culture medium). The cells were differentiated into macrophages by incubation with PMA (200 ng/ml) for 72 h at 37 °C in 5% CO2/95% air. The medium containing PMA and any non-adherent cells were then removed and the macrophages were washed with PBS (3 × 1 ml) and incubated with CRLPs or DiI-labelled CRLPs (30 μg cholesterol/ml) in the presence or absence of specific inhibitors as indicated in the text. The inhibitors were added 1 h prior to the addition of the CRLPs. Cell viability as assessed by Trypan blue exclusion was > 95% and was not affected by any of the CRLP types or conditions used. After the incubation, the macrophages were washed (culture medium 3 × 1 ml) and lipid accumulation was assessed by staining with Oil red O, or harvested for lipid analysis [24]. For studies with DiI-labelled CRLPs, the fluorescence associated with the cells was assessed by viewing with a Zeiss LMS 510 laser -scanning confocal microscope and quantified by absorbance volume analysis, or by fluorescence-activated cell analysis (FACS) using a BD FACS Calibur flow cytometer (BD Biosciences, Oxford, UK). For FACS analysis, cells were harvested in PBS containing EDTA (5 mM) and lidocaine -HCl (8 mg/ml), incubated for 20 min at 37 °C and centrifuged at 10,000×g (2 min). The cell pellet was then resuspended and fixed in PBS containing 4% formalin.
2.3
Analytical methods
For mRNA analysis, total RNA was extracted from THP-1 macrophages before or after incubation with CRLPs (30 μg cholesterol/ml) using a kit from Sigma (Poole, UK) and first strand synthesis was carried out using a kit supplied by Promega (Southampton, UK) according to the manufacturer's instructions. mRNA levels for the LDLr, LRP and the housekeeping gene glyceraldehyde phosphate dehydrogenase (GAPDH) were determined by real time polymerase chain reaction using a SYBR green quantitative PCR kit (Sigma, Poole, UK) with an Opticon light cycler under the following conditions; denaturation at 94 °C for 2 min followed by 40 cycles of 94 °C for 15 s, 58 °C for 1 min, with a final extension at 72 °C for 1 min. The primers used and the product sizes are shown in Table 1. The Ct values were determined by automated threshold analysis using Opticon Monitor 2 software. Data were normalized with the values obtained for GAPDH and the fold change in mRNA expression in CRLP-treated as compared to untreated macrophages was determined by the method described by Pfaffl [25].
Lipids were extracted from cell samples by the addition of chloroform: methanol (2:1 v:v, 20 volumes) followed by 0.88% KCl (v:v) (40% total volume). The mixture was shaken and the chloroform layer containing the lipids was collected and dried under nitrogen. The triacylglycerol (TG) content of CRLPs and lipid extracts was determined using a fully enzymatic assay kit (Alpha Laboratories, Eastleigh, Hants, UK) in which glycerol released from triacylglycerol by a lipase is measured using glycerol phosphate oxidase and peroxidase. The total cholesterol (cholesterol + cholesteryl ester) content of CRLPs and lipid extracts was determined using cholesterol oxidase and cholesteryl esterase (kit supplied by Sigma, Poole, UK), and the cholesterol content of lipid extracts was determined enzymatically using cholesterol oxidase [26]. Cholesteryl ester concentrations were calculated by subtracting the values for cholesterol from those for total cholesterol. The extent of oxidation of CRLPs was determined by measuring the level of thiobarbituric acid-reacting substances (TBARS) in the preparations [27].
The apolipoprotein content of CRLPs was assessed by SDS-PAGE using 10% gels [28]. Proteins were visualised by staining with Coomassie blue and the bands corresponding to apoE were quantified by optical density volume analysis.
Significance limits were calculated using ANOVA repeated measures or one way ANOVA followed by Student's paired t test for experiments with multiple time points or a single time point, respectively, except where indicated otherwise.
3
Results
3.1
Characteristics of CRLPs
The lipid composition of CRLPs, oxCRLPs and pCRLPs is shown in Table 2. The concentration of TG and TC in CRLPs and pCRLPs was similar, but that of oxCRLPs was a little lower, reflecting the dilution of the preparations during the oxidation procedure. The mean TG:TC ratio, however, was between 5 and 6 for all 3 CRLP types. Previous work in our laboratory has shown that the phospholipid content of CRLPs, oxCRLPs and pCRLPs prepared by the methods used here are not significantly different [15,29]. Analysis of the TBARS content of the particles showed that there were significant differences in their oxidative states, with values being higher in oxCRLPs as compared to CRLPs, while those for pCRLPs were lower (one way ANOVA, Bonferroni's test post hoc) (Table 2).
The apoE content of CRLPs, oxCRLPs and pCRLPs was evaluated by SDS PAGE electrophoresis. The results showed that the particles contained apoE, but no apoE was detected in the top fraction of the d 1.063–1.21 g/ml fraction of human plasma incubated and centrifuged in the absence of lipid particles (Fig. 1A), indicating that the apoprotein was bound to the CRLPs during the incubation. No apoCs or other apolipoproteins were detectable. There were no significant differences in the apoE content of the CRLPs, oxCRLPs or pCRLPs as assessed by optical density volume analysis (Fig. 1B).
3.2
Effect of the oxidative state of CRLPs on the induction of lipid accumulation in macrophages
Incubation of THP-1 macrophages with CRLPs or oxCRLPs for 5,24 or 48 h caused a marked increase in the total lipid found in the cells (Fig. 2A) at all time points, but lipid accumulation was greater with CRLPs as compared to oxCRLPs (P < 0.001). This effect was mainly due to greater accumulation of TG in the presence of CRLPs (P < 0.001) (Fig. 2B), as the trend for an increase in TC in CRLP- as compared to oxCRLP-treated macrophages did not reach significance (Fig. 2C).
3.3
Effect of the oxidative state of CRLPs on their uptake by macrophages
The uptake of DiI-labelled CRLPs, oxCRLPs and pCRLPs by THP-1 macrophages was assessed by confocal microscopy and FACS analysis (Fig. 3). Examination of cells incubated with DiI-labelled particles for periods up to 24 h with the confocal microscope showed that the fluorescence associated with the cells increased with time in all cases, but that there was clearly more in pCRLP- and less in oxCRLP-treated as compared to CRLP-treated cells (Fig. 3A). Quantification of the cell-associated fluorescence and analysis by ANOVA repeated measures showed that the rate of uptake of CRLPs over 24 h was significantly higher than that of oxCRLPs (P < 0.01) and significantly lower than that of pCRLPs (P < 0.01) (Fig. 3B), and a similar result was obtained using FACS (P < 0.05, CRLPs vs. oxCRLPs or pCRLPs (Fig. 3C).
3.4
Role of apoE in the uptake of CRLPs of different oxidative states
To investigate whether apoE is necessary for uptake of CRLPs, THP-1 macrophages were incubated with DiI-labelled CRLPs or CRLPs without apoE for 2 h and the fluorescence associated with the cells was determined by FACS analysis. In the absence of apoE, the uptake of CRLPs was reduced by about 90% (fluorescence values: CRLPs, 98.6 ± 7.1; CRLPs without apoE, 8.4 ± 0.9, n = 3).
The potential role of apoE dependent receptors in the uptake of CRLPs by the cells was studied using excess LDL and the LRP ligand, lactoferrin, to block entry via the LDLr and LRP, respectively. The effects of lactoferrin (2 mg/ml) on lipid accumulation in macrophages exposed to CRLPs or oxCRLPs as assessed by Oil red O staining are shown in Fig. 4. Lactoferrin had no effect on the lipid content of cells incubated in the absence of CRLPs, but reduced that in macrophages treated with either CRLPs or oxCRLPs (P < 0.05, both cases).
When the effects of excess LDL (200 μg cholesterol/ml) and lactoferrin (2 mg/ml), on the fluorescence associated with the macrophages after incubation for 1, 4, 16 and 24 h with CRLPs, oxCRLPs or pCRLPs were assessed by confocal microscopy (Fig. 5A–C), no significant inhibition was observed in the presence of excess LDL, but lactoferrin caused a decrease of > 90% in experiments with CRLPs (Fig. 5A) and oxCRLPs (Fig. 5B) and > 80% with pCRLPs (Fig. 5C) (P < 0.0001, all cases). Addition of LDL and lactoferrin together completely abolished detectable uptake of all three types of particles (Fig. 5A–C). FACS analysis of macrophages treated with CRLPs, oxCRLPs or pCRLPs in the presence or absence of lactoferrin (2 mg/ml), LDL (300 μg cholesterol/ml) or LDL + lactoferrin for 2 h also showed that lactoferrin caused a marked decrease of about 75% in the uptake of all three types of CRLPs (P < 0.01), and in this case significant inhibition (− 30–40%) was observed in the presence of LDL (P < 0.05) (Fig. 5D). Moreover, in a further separate experiment in which the excess of LDL added was increased by lowering the concentration of CRLPs to 10 μg cholesterol/ml and raising that of LDL to 500 μg cholesterol/ml, the uptake of both CRLPs and oxCRLPs was also significantly inhibited (fluorescence values (n = 3): CRLPs, 94.8 ± 6.2; CRLPs + LDL, 58.7 ± 5.8 (P < 0.05); oxCRLPs 56.7 ± 1.5; oxCRLPs + LDL, 29.2 ± 7.3 (P < 0.05)). As observed in the experiments using confocal microscopy, adding both LDL and lactoferrin to the incubations caused a further decrease to similar minimal levels in the uptake of all three types of CRLPs.
Assessment of the fluorescence associated with THP-1 macrophages after incubation with DiI-labelled acetylated LDL (acLDL) (10 μg cholesterol/ml) for 2 h showed that acLDL uptake, in contrast to that of CRLPs, was not significantly inhibited by excess LDL (500 μg cholesterol/ml) or lactoferrin (2 mg/ml) (Fig. 6).
3.5
Role of scavenger receptors and phagocytosis in the uptake of CRLPs of different oxidative states
The effects of blocking the class A scavenger receptor SR-A, the class B receptors CD36 and SR-B1 or phagocytosis on the uptake of CRLPs by THP-1 macrophages was investigated using poly-I, a known ligand for SR-A, a blocking antibody to CD36, excess HDL, which binds to SR-B1, and cytochalasin D, which blocks the polymerization of actin microfilaments [30]. Evaluation of the fluorescence associated with the cells after incubation with DiI-labelled CRLPs, oxCRLPs or pCRLPs for 1, 4,16 or 24 h by confocal microscopy (Fig. 7A–C) show no significant change in the presence of poly-I (5 μg/ml) or excess HDL (300 μg/ml). Significant decreases, however, were observed in the presence of anti-CD36 (1 μg/ml) (− 38% after 24 h) (P < 0.05) and cytochalasin D (10 μg/ml) (− 27% after 24 h) (P < 0.01) in experiments with CRLPs (Fig. 7A), although no change in the uptake of oxCRLPs or pCRLPs was detected with these treatments (Fig. 7B, C).
FACS determinations after incubation of macrophages with CRLPs, oxCRLPs or pCRLPs for 2 h confirmed that poly-I and HDL did not inhibit the uptake of the particles (Fig. 7D). In contrast, however, the uptake of DiI-labelled acLDL was decreased by about 90% in the presence of poly-I (50 μg/ml) (Fig. 6). As found in the experiments using confocal microscopy, anti-CD36 and cytochalasin D significantly inhibited the uptake of CRLPs, but in this case effects of a similar magnitude were also seen with oxCRLPs and pCRLPs (After 24 h:anti-CD36; CRLPs, − 38%, oxCRLPs − 40%, pCRLPs, − 39%: cytochalasin D; CRLPs − 39%, oxCRLPs − 35%, pCRLPs − 38%), although the changes with oxCRLPs did not reach significance (Fig. 7D).
3.6
Expression of mRNA for the LDLr and LRP in THP-1 macrophages
In THP-1 macrophages exposed to CRLPs (30 μg cholesterol/ml) for 24 h, LRP mRNA levels were increased in CRLP- treated THP-1 cells (fold change, 2.7 ± 1.01, n = 3), while LDLr mRNA abundance was decreased (0.27 ± 0.11(range), n = 2).
4
Discussion
Since homogeneous CMR cannot be obtained easily from human blood without contamination with other lipoproteins of a similar density such as chylomicrons and very low density lipoprotein (VLDL), model CRLPs were used in this study. These particles are similar in size, density and lipid composition to physiological remnants [23], and also contained human apoE, thus they differ from physiological CMR only in lacking apoB48. Importantly, extensive previous studies in both humans and experimental animals have demonstrated that chylomicron- and chylomicron remnant-like particles without apoB48 are cleared from the blood and metabolised in a similar way to the corresponding physiological lipoproteins [31–34]. In addition, CRLPs lacking apoB48, but containing apoE from the appropriate species, have been found to have effects which mimic those of physiological remnants in rat hepatocytes and pig endothelial cells [23,35], and our earlier work has shown that CRLPs cause extensive lipid accumulation in THP-1 macrophages and human monocyte derived macrophages (HMDM) which is comparable to that found in experiments with physiological CMR from rats and the murine macrophage cell line J774 [24,36]. For the current investigation, the use of CRLPs facilitated the manipulation of the oxidative state of the particles by inclusion of the lipophilic antioxidant probucol or by exposure to oxidising conditions similar to those used extensively in studies with oxLDL [16]. In previous work, we have shown that incorporation of probucol into CRLPs increases their uptake by THP-1 macrophages [15], and that this is due to protection of the particles from oxidation, since lycopene, a chemically unrelated antioxidant, has a similar effect [14]. For the present study, therefore, we used pCRLPs in addition to CRLPs and oxCRLPs, so that 3 different oxidative states were tested. Assessment of TBARS in the particles used showed that the oxidative states of CRLPs, oxCRLPs or pCRLPs were significantly different, and there were no significant differences in their apoE content, thus they provided a convenient and suitable model for the study.
Earlier work in our laboratory and others has demonstrated that CMR cause the extensive lipid accumulation associated with foam cell formation in HMDM, THP-1 macrophages and murine macrophage cell lines [24,36,37]. In experiments with THP-1 macrophages and HMDM, we found that CRLPs induced a greater increase in the intracellular total lipid content than oxLDL (at the equivalent cholesterol level), and that, as might be expected, this was mainly due to greater accumulation of TG, while TC levels were raised to a comparable extent with both types of lipoprotein [24]. In the current study we compared the effects of oxidized and non-oxidized particles and, although both caused rises in cellular lipid content compared to untreated cells, the increase was clearly smaller with oxCRLPs as compared to CRLPs, mainly because of decreased accumulation of TG (Fig. 2). These results suggest that the uptake of CRLPs by the cells may be inhibited by oxidation, and this was confirmed in confocal microscopy and FACS studies (Fig. 3), which showed that the rate of uptake of CRLPs was decreased by oxidation and increased by incorporation of probucol into the particles to protect them from oxidation (Table 2). These results clearly demonstrate that oxidation of CMR inhibits their uptake by macrophages and attenuates foam cell formation. This is an important new finding, since this effect is strikingly different from that found with LDL, where oxidation is required to induce macrophages to form foam cells.
CMR are known to be cleared from the circulation by the liver mainly by apoE-dependent pathways involving the LDLr and the LRP, and the present study shows that apoE also plays an important role in the entry of CRLPs into THP-1 macrophages, since a markedly reduced rate of uptake was observed when apoE was not present in the particles. These finding are in agreement with earlier work which has suggested that both the LDLr and the LRP are involved in the uptake of CMR by macrophages [10,19,38]. Our results indicate that both receptors are able to mediate the uptake of CRLPs regardless of their oxidative state, although lactoferrin, a ligand for the LRP, caused a marked reduction in lipid accumulation and uptake after exposure of THP-1 macrophages to all three types of CRLPs (Figs. 4, 5), while the maximum effect of excess LDL on uptake observed was more modest in all cases (Fig. 5). This lesser role of the LDLr may be due to its down-regulation on the influx of lipoprotein into cells. We have shown previously that the expression of mRNA for the LDLr in THP-1 macrophages is decreased by CRLPs while the expression of LRP mRNA is increased [18], and the present study, which shows a decrease of about 73% in LDLr mRNA and a 2.7fold rise in LRP mRNA after exposure of the macrophages to CRLPs, is in agreement with these findings.
The differences in rates of uptake of oxCRLPs, CRLPs and pCRLPs were retained in the presence of inhibitors of both the LDLr and the LRP, although when both the LDLr and the LRP were blocked, uptake of oxCRLPs, CRLPs and pCRLPs was reduced to a similar low level (Fig. 5), suggesting that the difference in their rate of uptake is apoE-dependent. Thus, an important novel finding of this study is that oxidation of CRLPs, unlike LDL, does not change their major routes of uptake by macrophages, with apoE-dependent receptors being of major importance and the LRP playing the predominant role, regardless of the oxidative state of the particles. ApoE does not bind to the LDLr family in its lipid-free state, as interaction with lipid is necessary to induce a conformational change which promotes high affinity for the receptors [39]. In addition, apoE has been shown to adopt different conformations when complexed to different lipids. Thus, changes in the lipid composition of lipoproteins such as an increased content of oxidized lipids or the presence of lipophilic compounds such as probucol may alter the conformation of the protein on their surface, and not all apoE molecules on a particular remnant particle may be able to act as ligands [39,40]. The different rates of uptake of CRLPs oxCRLPs and pCRLPs by macrophages demonstrated here, therefore, could be explained by effects on interaction with the LDLr and the LRP caused by differences both in the conformation of apoE and in the number of apoE molecules able to bind to the receptors, even though the total amount of apoE associated with the particles is not changed. Although we cannot rule out the possibility that the faster internalization of pCRLPs is due a specific effect of probucol on the surface structure of the particles, the finding that CRLPs are taken up more slowly after oxidation and more rapidly when the antioxidant is present suggests the changes are more likely to be related to the amounts of oxidized lipid in the particles.
The expression of scavenger receptors such as SR-A and the class B receptors CD36 and SR-B1 is a characteristic feature of macrophages, and SR-A and CD36 are know to play a part in the induction of foam cell formation by oxidized or chemically modified LDL [41]. It is possible, therefore, that this type of receptor may also be involved the uptake of oxCMR by macrophages. In the present study, poly I, a ligand for SR-A, reduced the uptake of acLDL by THP-1 cells by about 90% as expected, but had no significant effect on the uptake of CRLPs, irrespective of their oxidative state. In addition, excess HDL, which binds to SR-B1, did not prevent the entry of CRLPs, oxCRLPs or pCRLPs into the cells. In contrast, in the presence of anti-CD36, the uptake of all three CRLP types was inhibited by about 35–40% so that the differences in their rates of uptake were retained (Fig. 7). Thus, SR-A and SR-B1 do not appear to play a significant part in the uptake of oxCRLPs by macrophages, although CD36 may have a role which is unaffected by the oxidative state of the particles.
Mamo et al. [22] have reported previously that the phagocytosis may be important in the uptake of CMR by rabbit alveolar macrophages, although in a later electron microscopy study with HMDM they found no evidence for entry via this route [42]. In our experiments with THP-1 macrophages, however, cytochalasin D inhibited the uptake of all CRLPs types tested by 35–40%, suggesting that some phagocytosis of CRLPs of all oxidative states does occur in these cells (Fig. 7).
It has been suggested that the apoB48r may be involved in the uptake of chylomicron remnants by macrophages [1,20], and Kawakami et al. [43] have reported that it is responsible for the induction of macrophage foam cell formation by remnant lipoproteins from hyperlipidemic patients. Since apoB48 is an integral protein incorporated during the assembly of chylomicrons in intestinal cells, it is not possible to bind it to model CRLPs in a physiological way. The particles used in the present work, therefore, do not contain apoB48 and we were unable to study involvement of this receptor in the uptake of CMR by macrophages. However, antibodies to apoB48 have been found not to inhibit the uptake of chylomicron remnants by rat macrophages [44], and Elsegood et al. [21] who were unable to detect binding of chylomicron remnants to a protein with a molecular weight corresponding to the apoB48 receptor in THP-1 macrophages have suggested that it may be specific for VLDL remnants rather than chylomicron remnants.
The results of this study demonstrate that oxidative modification of CMR as compared to LDL has profoundly different effects on the uptake of the particles and the subsequent induction of lipid accumulation in macrophages. Instead of markedly enhancing foam cell formation, oxidation of CMR slows their uptake by macrophages and reduces the amount of lipid subsequently accumulated in the cells. This difference may be due to the different receptor mechanisms involved, since oxidation of LDL shifts the main route of uptake from the regulated LDLr to the unregulated scavenger receptors, while our experiments suggest that CMR are taken up mainly by the LRP with some contribution from the LDLr, with CD36 and phagocytosis playing only minor roles, irrespective of their oxidative state. These findings provide important new information about the way in which oxidation of CMR influences their induction of foam cell formation and the mechanisms involved, and has important implications for the role of dietary factors such as oxidized lipids and antioxidants which are transported in CMR in the promotion of atherosclerosis.
|
[
"chylomicron remnants",
"macrophages",
"foam cells",
"atherosclerosis",
"oxidized lipoproteins"
] |
[
"P",
"P",
"P",
"P",
"R"
] |
J_Behav_Med-3-1-2080858
|
Repression: Finding Our Way in the Maze of Concepts
|
Repression is associated in the literature with terms such as non-expression, emotional control, rationality, anti-emotionality, defensiveness and restraint. Whether these terms are synonymous with repression, indicate a variation, or are essentially different from repression is uncertain. To clarify this obscured view on repression, this paper indicates the similarities and differences between these concepts. Repression is the general term that is used to describe the tendency to inhibit the experience and the expression of negative feelings or unpleasant cognitions in order to prevent one’s positive self-image from being threatened (‘repressive coping style’). The terms self-deception versus other-deception, and socially related versus personally related repression refer to what is considered to be different aspects of repression. Defensiveness is a broader concept that includes both anxious defensiveness and repression; the essential difference is whether negative emotions are reported or not. Concepts that are sometimes associated with repression, but which are conceptually different, are also discussed in this paper: The act of suppression, ‘repressed memories,’ habitual suppression, concealment, type C coping pattern, type D personality, denial, alexithymia and blunting. Consequences for research: (1) When summarizing findings reported in the literature, it is essential to determine which concepts the findings represent. This is rarely made explicit, and failure to do so may lead to drawing the wrong conclusions (2) It is advisable to use scales based on different aspects of repression (3) Whether empirical findings substantiate the similarities and differences between concepts described in this paper will need to be shown.
Introduction
People differ in their tendency to openly show, or to hide their negative emotions. This is an important topic in behavioral medicine, since studies have shown that repression is a potential health risk factor for disorders as diverse as chronic pain (Beutler et al. 1986) and cancer (Jensen 1987; Weihs et al. 2000). Another reason why repression may be considered a relevant topic for research in this field is that the tendency to avoid expressing negative emotions (also labeled ‘repressive coping style’) is known to distort the assessment of a patient’s distress. As a result, this tendency to repress negative emotions may lead to making false conclusions. For instance, if patients report levels of distress similar to healthy individuals but show more repressive tendencies, they may in fact be more distressed. This repressive tendency may even influence the reporting of somatic symptoms and quality of life (Koller et al. 1999).
The possible influence of repression on disease development, health behavior and symptom reporting has been investigated in many studies. Summarizing the findings proves problematic, however, as authors use different labels for ‘repression-like’ concepts, such as repression, suppression, non-expression of negative emotions, emotional control, emotional inhibition, rationality, anti-emotionality, type C response style, defensiveness, restraint, concealment, type D personality, denial, alexithymia and blunting. It is unclear whether this array of terms actually refers to the same concept or altogether different concepts. For instance, is forgetting details of traumatic events (e.g., sexual abuse or war experiences) comparable to not wanting to show one’s emotions because of one’s preference to rationalize? Is the tendency to minimize one’s problems and to emphasize the positive aspects of experiences comparable to non-expression of negative emotions because one is afraid of personal confrontation? And yet in all these instances, the term repression is used.
The meanings of the various terms used in this field are defined below and an attempt is made to analyze their relationships on a conceptual level. This treatise does not discuss theories of repression and related constructs, but concerns the meaning of concepts. In our view, confusion exists more on the level of concepts than on the level of theories, and these theories are not very helpful in clarifying the conceptual confusion that exists. For instance, there is no theory that brings to notice that repression and concealment—whereas the literal meaning of these words may suggest overlap—refer to fundamentally different concepts. Nor is there any theory that alerts against the special use in some Behavioral Medicine texts of the term ‘denial’ in the sense of minimizing the seriousness of a disease and not as denial of negative emotional states in general (Brown et al. 2000; Butow et al. 1999, 2000; Greer et al. 1979). It is important to add here, that several terms in this field have been introduced not on the basis of theory, but during the process of developing a measurement method.
We make a distinction between those concepts that in our view are related to repression, and other concepts that are sometimes defined as similar to or related to repression, but are clearly different. Figure 1 illustrates the conceptual network.
Fig. 1A conceptual diagram indicating which concepts fall under the headings of repression and anxious defensiveness, respectively, and which concepts are sometimes associated with, but theoretically different from defensiveness (voluntary suppression, repressed memories, denial and alexithymia)
Repression
Repression is the general term that is used to describe the tendency to inhibit the experience and the expression of negative feelings or unpleasant cognitions in order to prevent one’s positive self-image from being threatened. A typical example of a person with repressive tendencies would be a sociable and cheerful man who rarely complains about any misfortune including disease, and whose self-image is one of a positive-minded person who is in control of his life. When he encounters someone who discusses an emotional problem, he is inclined to quickly change the subject in an attempt to avoid entering a world of anxiety, sadness or worry, which would imply that he has lost control.
Other authors have presented comparable definitions, such as “Individuals who avoid focusing on ego-threatening material are termed repressors” (Ashley and Holtgraves 2003), or “repression can be defined as the avoidance of threatening information” (Baumeister and Cairns 1992). The motive of “preventing one’s positive self-image from being threatened” is added in order to exclude some forms of non-expression, namely those due to shyness, social phobia and introversion. Shy or social-phobic people fear social situations, which inhibits their emotional expression, whereas repressive people do not fear or avoid social situations. Another difference is that shyness, social phobia and introversion refer to non-expression of both negative and positive feelings, whereas repression refers only to non-expression of negative feelings.
We use the term ‘non-expression of negative emotions’ here as a synonym for repression. This term suggests perhaps a focus on expressive behavior only, but it actually refers to the inhibition of both emotional experiences and behaviors. An explicit restriction is made to negative emotions, rather than emotions in general. This asymmetry, implied in the definition of repression, is mirrored by empirical findings. A repressive tendency appears to be related to downplaying negative aspects rather than to overstating positive aspects of a person (Myers and Brewin 1996), and to a memory deficit for real life events associated with negative emotions but not for events associated with positive emotions (Davis 1987).
There is no reason to label repression as either positive or negative on a phenomenological level. Repressive people will generally not bother other people with their problems and may even facilitate social situations by their positive attitude (Furnham et al. 2003). Conversely, this coping style may impoverish intimate social interactions and may in the long run have a negative impact on the person’s own functioning, due to a lack of insight into their own psychological functioning, a decrease in the variety of their coping repertoire and overlooking signals that lead to seeking medical help in time. In the long term repression may have negative somatic consequences, including an increased risk of various disorders.
Below is first discussed a proposal by Weinberger to make a distinction between repression and anxious defensiveness (Weinberger et al. 1979; Weinberger 1990; Weinberger and Schwartz 1990). This distinction is useful for placing repression-like concepts under one of these two headings.
Defensiveness
Whereas most elements in this discussion can be approached on a merely conceptual level, the discussion of the concept of Defensiveness needs to be introduced by mentioning empirical findings. Weinberger (1990) makes a distinction between two types of defensiveness: He defines repression as scoring high on defensiveness but low on anxiety, and anxious defensiveness as scoring high on both defensiveness and distress self-report scales.1 In fact, Weinberger formed six groups, based on a tripartition of restraint scores and a dipartition of distress scores, but the remaining four groups are not relevant to this discussion. In an earlier publication he describes a more familiar division of defensiveness types based on a dipartition of social desirability scores and anxiety scores, which is comparable to his more recent division (Weinberger et al. 1979). The two groups appeared to be different on a number of personality variables. Compared to the other groups, the anxious defensive group scored low with respect to assertiveness, ability to express themselves in close relationships, sensitivity to their own needs and feelings, self-esteem and self-control. They also scored high on avoidant personality (shyness), dependency (emotional reliance on others and approval dependence), obsessive worrying, and (minor) physical illnesses. The repressive group, on the other hand, was characterized by high scores for intimacy, self-esteem, self-control (tendency to use self-management techniques), defensiveness and alexithymia2, while low on avoidant personality. These differences in a broad spectrum of personality traits indicate that the division into the two defensiveness groups is more than the product of an arithmetic procedure; it refers to a constellation of essential individual differences (Weinberger and Schwartz 1990).
Defensiveness, therefore, covers a broader category than repression. Defensiveness concerns different strategies to protect oneself against being hurt psychologically. One strategy is to behave—more or less anxiously—in a socially acceptable way, to be nice in order not to get hurt, and to avoid social confrontations. Another strategy is to inhibit thoughts about negative aspects of oneself, and to consider oneself as the social person one would rather be. The first condition, anxious defensiveness, includes the awareness of negative emotions, whereas the second condition, repression, denies these emotions.
Weinberger’s division into two forms of defensiveness is highly useful. If high levels of anxiety or other forms of distress are implied in the definition of a repression-like concept, one should place this concept under the heading of anxious defensiveness, rather than under the heading repression. On an empirical level, a low level of distress reporting in repression would be expected (negative relationship between repression and distress), and a relatively high level in anxious defensiveness (positive relation).
(Un)consciousness
Repression is a tendency that a person may be (partly) aware of, which in psychodynamic theories is referred to as suppression, or unaware of, referred to as repression in these theories. However, (un)consciousness is often not explicitly included in the definition of repression as a distinctive characteristic. To give some citations: “Repression can be defined as the avoidance of threatening information” (Baumeister and Cairns 1992, p. 853). “Repressors are individuals who habitually and efficiently control their emotions” (Boden and Dale 2001, p. 122). Some authors—on the contrary—stress the role of the unconscious: “Repressive-defensiveness is characterized by a non-conscious avoidance of threatening information” (King et al. 1992, p. 87). Repressive-defensiveness is defined here according to Weinberger’s operational definition of repression (Weinberger et al. 1979; see above). Quite confusing is that other authors, using a different term namely repressive coping but referring to the same operational definition of Weinberger, claim the opposite: “Repressive coping is thought to modulate the conscious experience of negative affect following the appraisal of threat” (Newton and Contrada 1992, p. 160). Most contemporary authors describe repression in terms of active cognitive processes, such as selective inattention and motivated forgetting rather than in terms of an unconscious defense mechanisms (Baumeister and Cairns 1992; Newton and Contrada 1992; Mendolia et al. 1996). The most explicit view on this comes from Erdelyi (1993, 2001), who states that empirical findings do not support any distinction between conscious and unconscious forms of emotional inhibition. He also uses an historical argument. Whereas many authors referred to Freud when discussing the difference between repression and suppression, this is—in Erdeleyi’s view—historically unwarranted, given that Freud used these terms interchangeably. The notion that repression must be unconscious is not Sigmund Freud’s, but his daughter’s, Anna Freud (1946).
If any distinction is made between suppression and repression, the focus is usually placed on processes, i.e. on time-limited cognitive acts. The distinction between conscious and unconscious forms of inhibition does seem possible and it is perhaps useful when applying it to acts, but this distinction is more difficult to apply to traits. The discussion below is about repression as a trait (‘repressiveness’). One may on occasion be aware of one’s tendency to inhibit the experience and expression of negative feelings, but most of the time be vaguely aware, and more often totally unaware of them.
Empirical findings indicate that repressors are often unaware, or at least not fully aware of their emotional avoidance style. As indicated below, repressors genuinely perceive themselves as being low in anxiety and are primarily self-deceivers. Another indication for the (mainly) unconscious character of repression is the repressor’s decreased ability to recall personal experiences associated with a negative affect. When memories are recalled by repressors, thus having become conscious, they are not processed more slowly than by the non-repressors (Davis 1987).
Self-deception and Other-deception
Expressing negative emotions may be deliberately avoided as part of the tendency to make a favorable impression on other people. This tendency is called impression management or, originally, other-deception. It is distinguishable from self-deception, in which case the person actually believes his or her positive self-reports (Paulhus 1984; Rogers and Kristjanson 2002).
Impression management is modestly and negatively related to reports of negative emotions and somatic symptoms, while a self-deceptive response style reduces symptom reporting above the effects of deliberate impression management (Linden et al. 1986). Given that underreporting negative emotions is the hallmark of repression, these findings seem to indicate that self-deception is closer to repression than impression management. Another characteristic of repression is a deficiency in the memory for emotional events (Furnham et al. 2003; Davis 1987). This deficiency appeared to be related to self-deception, but not to impression management (Ashley and Holtgraves 2003).
Two studies found that repressors scored high on both other-deception and self-deception questionnaires (Derakshan and Eysenck 1999; Furnham et al. 2002). The study of Derakshan and Eysenck (1999) also showed that repressors are more self-deceivers than other-deceivers. This was demonstrated with the so-called ‘bogus pipeline’ method, where participants are connected via electrodes to a piece of apparatus resembling a lie detector, which could allegedly detect whether they are telling the truth. Compared to a control condition, people are generally more willing to report truthfully about their emotional states when subjected to the bogus line condition, even if this report is seen as socially undesirable or embarrassing for that person. Repressors generally did not show any difference in anxiety scores between both conditions. This finding suggests that repressors genuinely perceive themselves as being low in anxiety. They are mainly self-deceivers, though the questionnaire data indicated that they also showed some tendency to present themselves deliberately in a socially desirable light. In another experiment, Baumeister and Cairns (1992) showed that repressors who privately received threatening feedback spent the least amount of time reading it, whereas repressors who received the same feedback publicly spent considerably more time reading it. Non-repressors were unaffected by the favorability of the evaluation (threatening or not) or the private versus public nature of the situation. These findings suggest that repressors abandon self-deceptive strategies in favor of impression management strategies in an attempt to invalidate the socially undesirable information about their personality.
Doubts have arisen about the validity of self-deception (SD) and other-deception (OD) scales, including the widely used Balanced Inventory of Desirable Responding (BIDR) of Paulhus (1984). Studies have shown that the scores to SD and OD scales can be influenced by instructions (Stober et al. 2002; Pauls and Crost 2004), such as a fake good instruction (e.g. “Present yourself as much as possible in a favorable light”), a fake being competent instruction (e.g. “Present yourself as much as possible as being competent and self-confident”), or a fake social harmony instruction (e.g. “Present yourself as much as possible in an agreeable and conscientious light”). The SD and OD scores changed differently, dependent on the type of instruction. For instance, SD scores were most sensitive to the competence instruction and OD scores were most sensitive to the social harmony instruction (Pauls and Crost 2004). SD and OD were measured in this study with the widely used Balanced Inventory of Desirable Responding. The authors of this study argue that this questionnaire does not so much measure SD and OD, as is generally believed, but more so overconfidence versus need for social harmony. This may imply that the question whether repression is predominantly associated with self-deception or other-deception cannot be answered with so-called SD and OD scales, but only with experimental designs such as applied in the studies of Derakshan and Eysenck (1999) and Baumeister and Cairns (1992).
What is the precise relationship between these two concepts and repression? The character of self-deception, i.e., believing one’s positive self-report, is completely compatible with the definition of repression. There is, however, no complete overlap between repression and impression management. An extreme form of impression management would be consciously trying to project a positive image towards others while being fully aware of one’s negative feelings and negative cognitions. This condition of other-deception without self-deception is not compatible with the description of repression. Because of this conceivable exception, other-deception is depicted in Fig. 1 as only partly overlapping with repression.
Different Aspects of Repression
Certain explanations of repression seem to emphasize the social aspect, whereas other explanations do not describe the repressive tendency as specifically socially related. For instance, Weinberger and Schwarz, who used the term ‘self-restraint’, state that it concerns ‘domains related to socialization and self-control and refers to repression of egoistic desires in the interest of long-term goals and relations with others’. The term ‘self-restraint’ also encompasses ‘tendencies to inhibit aggressive behavior, to exercise impulse control, to act responsibly, and to be considerate of others’ (Weinberger and Schwartz 1990, p. 382). To label such tendencies, the term ‘socially related repression’ would therefore seem appropriate. These tendencies may be part of the broader tendency to behave in a socially acceptable way, which should not be conceived as simply the need to follow external norms, but reflect a self-concept that depends on the approval of other people.
We suggest using the label ‘personally related repression’ for a second aspect of repression, which is not primarily socially related and may be defined as the general tendency to inhibit one’s expression of anxiety, depression, worry and anger, and not to let oneself be influenced by these negative feelings. This heading subsumes several concepts. Watson and Greer (1983) use the term emotional control and describe it as ‘the extent to which individuals report controlling their reactions when a particular emotion is experienced’, especially anger, anxiety and depressed mood. This description does not specifically suggest social determination. Rationality is another example and is described by Spielberger (1988) as ‘the extent to which an individual uses reason and logic as a general approach to coping with the environment’.
Empirical support for our proposed division was found in a study that applied a secondary factor analysis to investigate the interrelationships between various subscales of two repression questionnaires and three distress questionnaires (Giese-Davis and Spiegel 2001). The scales applied in this study were the three subscales of the Courtauld Emotional Control Scale (CECS) used to measure emotional control, the five subscales of the Weinberger Attitude Inventory (WAI), used to measure self restraint, and three distress questionnaires (POMS, CES-D and IES; four subscales in total). There is no reference to social situations in the items included in the CECS (Watson and Greer 1983). An example item is “When I feel afraid or worried, I smother my feelings”. Most of the items of in the WAI subscales (D. A. Weinberger, unpublished ) refer to social situations, such as the item “I think about other people’s feelings before I do something they might not like”. One would expect personally related repression scales (CECS) to load on one dimension and scales assessing socially related repression (WAI) on a second dimension (and the distress scales on a third dimension). The factor analysis yielded a four-factor solution, including (1) subscales of the CECS (2) WAI restraint scales (3) WAI defensiveness scales and (4) the distress scales. We cannot explain why the WAI repression scales loaded on two different factors. However, the finding that the CECS scales clustered in one factor and the WAI repression scales in two other factors, could be seen as support for our division into personally related and socially related repression.
Concepts Different from Repression
Concepts that are in our view different from repression are: the act of emotional suppression, repressed memories, habitual suppression, concealment, type C coping pattern, type D personality, denial, alexithymia and blunting. The first concept refers to an ‘act,’ whereas repression is discussed in this paper as a tendency or coping style. Four of these concepts—habitual suppression, concealment, type C coping style and type D personality should, in our view, be interpreted as types of anxious defensiveness rather than forms of repression. Our motive for discussing these ‘non-repression’ concepts is that the literature may suggest that these concepts are related to repression.
The Act of Voluntary Suppression of Emotionally Charged Material
In an experimental context people can be asked to refrain from showing emotional reactions, or to try not to think about an emotional condition. Suppression of emotional behavior or thoughts can also occur spontaneously in everyday life. Given that such acts can be performed incidentally, both by people high or low in repression, the act should be distinguished from the habitual response style. It cannot simply be assumed that the consequences of the act of emotional suppression are similar to the psychological and somatic concomitants of being an habitual repressor. For instance, in the long term emotional disclosure leads to a decrease in reported psychological and somatic symptoms (Smyth 1998). On the other hand, compared to non-repressors, habitual repressors also report less distress (Ward et al. 1988; Weinberger and Schwartz 1990; Weinberger, unpublished ; Swan et al. 1992; Tomaka et al. 1992; Bleiker et al. 1993). Therefore, the act of emotional expression as well as the response style of non-expression are both associated with low reported distress.
There is an area of research that studies the effects of inhibiting emotional behavior (Gross and Levenson 1993) and another area of research that is interested in the effect of thought suppression (Abramowitz et al. 2001). Participants in such studies are asked to refrain from emotional behavior, such as facial expressions, or not to think of a certain image. Gross uses the term ‘emotional suppression’ in the sense of an act and describes it as the conscious inhibition of behavioral signs of emotion, while being emotionally aroused (Gross and Levenson 1993).
Repressed Memories
Repressing memories of traumatic events concerns a complex of cognitions and emotions that is mainly limited to a certain theme or event, such as sexual abuse in childhood. This is different from repression, which concerns the tendency not to express negative emotions in general. Repression of memories is initiated by traumatic events, whereas repression is a habitual style applied in a variety of situations. Although repressing memories of traumatic events could lead to an habitual style of repression, or magnify an existing tendency to repression that does not undo the conceptual difference.
Strengthening of existing repressive tendencies in response to a traumatic condition was demonstrated in a study among women who were awaiting the outcome of diagnostic tests for breast cancer, which may be conceived as a traumatic event. An increase in the number of repressors was found after the diagnosis of breast cancer was made known to the patients, whereas no increase was found in women who appeared to be free of cancer (Kreitler et al. 1993).
In a series of studies, McNally presented evidence that those who believe they were sexually abused as children, but have no memory for these events (“repressed memories”) show a particular style of information processing, which is different from those who have never forgotten their childhood sexual abuse or have never been sexually abused (McNally 2001). Individuals with repressed memories exhibited symptoms of psychological distress, elevated levels of dissociation and absorption, superior forgetting abilities for trauma-related material and memory distortions. Most of these characteristics were also found in individuals who report having recalled long-forgotten episodes of childhood sexual abuse (“recovered memories”). These characteristics may reflect a propensity for repressing traumatic memories, a propensity for forming false memories of trauma, or a consequence of abuse itself (assuming it occurred). Anyhow, this personality profile relates to information processing distortions of trauma-related material, not to emotionally loaded material in general as found in repression. Moreover, the high distress scores of individuals with repressed memories are incompatible with the concept of repression.
Habitual Suppression
In later publications on emotion regulation, Gross shifted his attention from suppression as an act to the habitual use of suppression, which was described as a form of response modulation that involves inhibiting ongoing emotion-expressive behavior (John and Gross 2004). Inhibition of emotional experiences is not assumed, as is the case in the definition of repression. Gross even presumes that using suppression in everyday life might actually be associated with greater negative emotion experience. Acknowledging these characteristics, habitual suppression must come under the heading of anxious defensiveness.
Empirical findings indicate that habitual suppression is different from, and in some respects, the opposite of repression. Whereas habitual suppressors experience less positive and more negative emotions than other people (John and Gross 2004), repressors do not differ from other people with respect to the experience of positive emotions and they experience less negative emotions than non-repressors (Furnham et al. 2003). Moreover, habitual suppressors have a lower self-esteem and a less optimistic outlook (John and Gross 2004), whereas the opposite was found for repressors (Myers and Reynolds 2000).
Self-concealment
Larson and Chastain (1990) introduced the concept self-concealment as the trait version of the act of inhibition, studied by Pennebaker et al. 1990). The authors define self-concealment as a “predisposition to actively conceal from others personal information that one perceives as distressing or negative,” and they say that “self-concealed personal information is consciously accessible to the individual” (Larson and Chastain 1990, p. 440).
How is this concept related to repression? There are three gradual differences with repression: (1) Self-concealment concerns specific distressing secrets, whereas repression concerns negative feelings in general, although it should be said that there is a rather thin line between these two elements; (2) Self-concealment is explicitly a tendency towards voluntary and conscious inhibition, whereas repression is conceptualized as incorporating both unconscious and conscious coping strategies; (3) Self-concealment implies the awareness of distressing thought contents, whereas repression implies the inhibition to become fully aware of such thought contents. Especially this last aspect implies that self-concealment could be better placed under the heading anxious defensiveness, rather than repression.
Empirical findings support this supposed conceptual difference. While repression is often negatively related to distress reporting (Weinberger and Schwartz 1990; Weinberger, unpublished ; Swan et al. 1992; Tomaka et al. 1992; Bleiker et al. 1993; Gick et al. 1997; Vetere and Myers 2002), a positive association has been found between self-concealment and depression, anxiety and physical symptoms (Larson and Chastain 1990) and rumination (King et al. 1992). A negative relationship between self-concealment and repression (King et al. 1992; Ritz and Dahme 1996) has also been reported. In fact, (Ritz and Dahme 1996) found the lowest scores on the Self-Concealment scale (SCS; Larson and Chastain 1990) for repressors and the highest SCS scores among the (truly or defensive) high-anxious persons.
Type C Coping Style
The Type C concept was first introduced in 1980 in an abstract presented by Morris and Greer (1980), who considered this coping style as characteristic of cancer patients. The style was described as being emotionally contained, especially in stressful situations. Temoshok independently developed a similar concept, which included several elements (Kneier and Temoshok 1984), and described this coping style as ‘abrogating one’s own needs in favor of those of others, suppressing negative emotions, and being cooperative, unassertive, appeasing, and accepting. The Type C individual is considered nice, friendly and helpful to others, and rarely gets into arguments or fights... The Type C individual may be seen as chronically hopeless and helpless, even though this is not consciously recognized in the sense that the person basically believes that it is useless to express one’s needs... The Type C individual does not even try to express needs and feelings; these are hidden under a mask of normalcy and self-sufficiency’ (Temoshok 1987, pp. 558–560).
One of the repressive types, as distinguished by Weinberger and Schwartz (1990), in our view shows a remarkable resemblance to the Type C coping pattern. The characteristics mentioned for anxious defensiveness resemble the above-mentioned characteristics for individuals who use Type C coping. Both descriptions mention unassertiveness, low sensitivity to one’s own needs and feelings, abrogating one’s own needs in favor of others, emotional reliance on others, being cooperative, appeasing and accepting, and high levels of distress (obsessive worrying/ helplessness and hopelessness). Because of this similarity and the inclusion of helplessness and hopelessness in the description of this concept, the Type C response pattern seems closer to anxious defensiveness than repression.
Type D Personality
The term ‘Type D personality’ was introduced by Denollet (1997) to describe those people who are distressed, but who also inhibit the expression of emotions. Denollet developed this concept, while working in the field of cardiovascular disorders. It consists of a combination of two factors that seemed predictive for the development of coronary heart disease and hypertension. The first factor was high distress levels (anger, depression, anxiety and vital exhaustion) and the second factor was the inhibition of emotional expression. This second factor was specified as reflecting social inhibition (and introversion).
It is important to be aware that according to the description of the Type D person, the negative emotions of anger, anxiety, and depression are experienced consciously. Whereas repression and distress are often negatively related, a high level is implicated in the Type D personality. Type D persons are categorized as scoring high on distress and high on social inhibition (Denollet 2005). Therefore, the Type D personality style is explicitly the anxious defensive type.
Denial
Denial is conceived here as denying or minimizing the seriousness of a medical condition, not as denial of emotions or painful events as is commonly the case, and not as an unconscious defense against painful and overwhelming aspects of external reality, as described in psychoanalytic theory. The way denial is described in this paper—denial of diagnosis or denial of impact—is one of the several definitions quoted in the literature (Vos and Haes 2007; Moyer and Levine 1998). Greer et al. described denial in breast cancer as “apparent active rejection of any evidence about their diagnosis which might have been offered, including the evidence of breast removal, such as “it wasn’t serious, they just took off my breast as a precaution” (Greer et al. 1979, p. 786). Minimizing the impact of cancer is a milder and more realistic form of denial, and was measured in studies by Butow. (Butow et al. 1999, 2000; Brown et al. 2000). Our definition of denial indicates a clear conceptual difference between denial and repression. Repression does not specifically refer to the emotional consequences of a disease, but rather to negative emotions in general. A person might repress these emotions, while not denying the seriousness of the disease. Denial or minimizing can either be an act (an event-driven coping response) or it can reflect a habitual style of minimizing the seriousness of unpleasant events.
It is interesting that cancer studies showed opposite consequences of the two phenomena. In studies using a prospective, longitudinal design to investigate the role of psychological factors on the course of cancer, two studies found that repression predicted an unfavorable course (Jensen 1987; Weihs et al. 2000). Obversely, in four other studies denial or minimizing was found to predict a favorable course of cancer (Greer et al. 1979; Dean and Surtees 1989; Greer et al. 1990; Butow et al. 1999, 2000).
Alexithymia
The concept of alexithymia is derived from clinical observations of a cluster of specific cognitive characteristics among patients suffering from psychosomatic diseases, substance use disorders, and post-traumatic stress disorders (Nemiah et al. 1976; Bagby et al. 1997). It evolved into a theoretical construct, with the following salient features: (1) difficulty identifying feelings, (2) difficulty describing feelings, and (3) externally oriented thinking (Bagby et al. 1997).
Due to the difficulty of identifying feelings, one might assume that emotions are not expressed either. However, alexithymic persons should not be considered to be emotionally flat. Nemiah et al. reported a proneness to sudden outbursts of crying and anger in these persons, though they were unable to connect these behaviors with thoughts and fantasies (Nemiah et al. 1976). Corresponding to these observations, Sifneos reported that it was common for his patients to mention anxiety or to complain of depression (Sifneos 1967), although they used a limited vocabulary to describe their emotions. The emotions of alexithymic individuals appear to be rather diffuse, poorly differentiated and not well-represented. Taylor et al. concluded that alexithymia should be regarded not as a defense against distressing affects or fantasies, but rather as the reflection of an individual difference in the ability to process and regulate emotions cognitively (Taylor et al. 1997). They suggested that this construct is different from ‘other emotion-related constructs such as inhibition and the repressive-defensive coping style.’
The difference between alexithymia and repression is empirically supported. First, several studies have shown that the Toronto Alexithymia Scale (TAS)—the most widely used and well-validated questionnaire for assessing alexithymic traits—appeared to be unrelated or negatively related with various measures of repression (King et al. 1992; Newton and Contrada 1994; Myers 1995; Linden et al. 1996; Taylor et al. 1997). Second, the TAS is unrelated (Linden et al. 1996), or positively related with self-reported distress (Taylor et al. 1997; Verissimo et al. 1998), whereas measures of repression are negatively related with distress.
Based on psychometric comparisons, alexithymia shows some correspondence to the sensitizing style of high-anxious persons, rather than the avoidant style of repressors. Repressive individuals often report that they are not upset despite objective evidence to the contrary, whereas alexithymic individuals acknowledge that they are upset, but have difficulty in specifying the nature of their distress.
Blunting
Miller (1987) distinguishes two types of individual differences in dealing with threatening stimuli. Monitors are those people who tend to seek information when coping with a threat, such as going to the dentist, being taken hostage or flying (information seekers), and blunters tend to avoid information when faced with a threat (distractors). There is some similarity between blunting and repression, as both repressors and high blunters avoid distressing information. Repressors avoid mainly personally relevant, emotionally loaded information. High blunters, on the contrary, avoid material about external conditions that people generally regard threatening. There is no implication in the definition that high blunters avoid expressing negative emotions because that would threaten their self-image, which is explicitly included in the description of repression.
Discussion
The way in which most people use the term ‘repression’ in an everyday context indicates that they generally understand what it actually refers to. Whether science was right to introduce the current assortment of subtle differences thus exposing the gross simplicity of society’s everyday use of the term, or whether science has ultimately entangled the term in a maze of unclear terminology is debatable. It is undebatable, however, that there is a lack of consensus about what repression is. The impression is that the many terms used in this field—repression, suppression, non-expression of negative emotions, emotional control, emotional inhibition, rationality, anti-emotionality, defensiveness and restraint—denote something similar to repression, but there is no certainty whether they can be considered synonymous, indicate a subtle variation of repression, or indicate an associated, but essentially different concept. One reason for this confusion is that definitions are rarely given, and that hardly ever is indicated how a new term relates to regularly used terms.
Repression has been defined as the tendency to inhibit—consciously or unconsciously—the experience and expression of negative emotions or unpleasant cognitions in order to prevent one’s positive self-image from being threatened. The term is used to describe an act, such as avoiding a specific memory, or a tendency or coping style (‘repressive coping style’). This paper deals with repression as a tendency or coping style. Terms whose definitions appear to agree with the definition of repression, which can therefore be considered synonyms of repression, are: non-expression of negative emotions, emotional inhibition, emotional control, anti-emotionality, rationality and self-restraint. Although these terms may all be subsumed under the heading repression, their definitions also suggest some differences concerning the motives for repression.
We tentatively made a difference between socially related and personally related repression (see Fig. 1). An example of socially related repression is ‘self-restraint’, which encompasses ‘tendencies to inhibit aggressive behavior, to exercise impulse control, to act responsibly, and to be considerate of others’ (Weinberger and Schwartz 1990, p. 382). An example of personally related repression is rationality, which is described as “the extent to which an individual uses reason and logic as a general approach to coping with the environment” (Spielberger 1988). In the first category, the tendency to inhibit the experience and expression of negative emotions in order to prevent one’s positive self-image from being threatened is (more) socially related, which reflects a self-concept that depends (more) on the approval of other people more so than personally related repression. Future research will show whether this distinction is useful or not. A study has indeed provided some empirical evidence for its validity, showing that restraint scales and emotional control scales loaded on different dimensions in a secondary factor analysis (Giese-Davis and Spiegel 2001).
By definition, repression implies (some degree of) self-deception, whereas repression may or may not include other-deception. Self-deception implies honestly believing one’s positive self-report. The overlap with repression is evident, given that “the inhibition of the experience of negative emotions or unpleasant cognitions in order to prevent one’s positive self-image from being threatened” implies self-deception. Other-deception is described as deliberately avoiding expression of negative emotions as part of the tendency to make a favorable impression on other people. Other-deception without self-deception, therefore, seems to be incoherent with the definition of repression.
Defensiveness is a broader concept than repression. Defensiveness concerns different strategies to protect oneself against being psychologically hurt, which include repression and anxious defensiveness (Weinberger and Schwartz 1990). This distinction was made on the basis of operational criteria. Although high levels of defensiveness characterize both forms, repressors report relatively low distress levels, whereas anxious defensive persons report relatively high distress levels. The division into the two defensiveness groups appears to be more than the product of an arithmetic procedure; it refers to a constellation of essential individual differences (Weinberger 1990). One essential difference concerns the two defensiveness groups’ association with distress.
We have also indicated which concepts, although sometimes associated with repression, are basically different from repression: voluntary suppression, repressed memories, habitual suppression, self-concealment, type C coping pattern, type D personality, denial, alexithymia and blunting. The first concept concerns an ‘act,’ whereas we have discussed repression as a tendency or coping style. Four of these concepts were placed under the heading anxious defensiveness, because their definitions imply experiencing high levels of negative emotions: Habitual suppression, self-concealment, type C coping style and type D personality.
Our conceptual analysis has motivated us to propose the following recommendations for future research: (1) In studies on the character and the consequences of repression one should ideally include measures of personally related and socially related repression, and—as a contrast—a measure of anxious defensiveness. (2) An acute distinction should be made when summarizing literature findings between repression and concepts that are related to, but essentially different from repression (3) Future research will need to show whether relationships between questionnaires substantiate the similarities and differences between the concepts described in this paper.
Our objective with this treatise on defensiveness-related concepts is to provide more clarity in this field. The next step in finding our way in the current maze of repression points to a review on defensiveness-related questionnaires.
|
[
"repression",
"emotional control",
"defensiveness",
"restraint",
"self-deception",
"type c",
"denial",
"alexithymia"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
J_Biol_Inorg_Chem-3-1-1915633
|
Crystal structure of the ferritin from the hyperthermophilic archaeal anaerobe Pyrococcus furiosus
|
The crystal structure of the ferritin from the archaeon, hyperthermophile and anaerobe Pyrococcus furiosus (PfFtn) is presented. While many ferritin structures from bacteria to mammals have been reported, until now only one was available from archaea, the ferritin from Archaeoglobus fulgidus (AfFtn). The PfFtn 24-mer exhibits the 432 point-group symmetry that is characteristic of most ferritins, which suggests that the 23 symmetry found in the previously reported AfFtn is not a common feature of archaeal ferritins. Consequently, the four large pores that were found in AfFtn are not present in PfFtn. The structure has been solved by molecular replacement and refined at 2.75-Å resolution to R = 0.195 and Rfree = 0.247. The ferroxidase center of the aerobically crystallized ferritin contains one iron at site A and shows sites B and C only upon iron or zinc soaking. Electron paramagnetic resonance studies suggest this iron depletion of the native ferroxidase center to be a result of a complexation of iron by the crystallization salt. The extreme thermostability of PfFtn is compared with that of eight structurally similar ferritins and is proposed to originate mostly from the observed high number of intrasubunit hydrogen bonds. A preservation of the monomer fold, rather than the 24-mer assembly, appears to be the most important factor that protects the ferritin from inactivation by heat.
Introduction
Ferritin is a protein involved in metal homeostasis by reversible storage of iron and presumably also in protection against oxidative stress by scavenging reactive oxygen species. It is a small protein of approximately 20 kDa and its main structural motif is a bundle of four parallel α-helices and a fifth short α-helix tilted towards the bundle axis and masking the one end of the cylindrical bundle. The functional ferritin molecule is an approximately 500 kDa hollow spherical assembly of 24 subunits, with outer and inner diameters of 12 and 8 nm, respectively. The 24-meric ferritin agglomerate may be a homopolymer or, in higher eukaryotes, a heteropolymer consisting of homologous subunits designated as H, M and/or L. Ferritin subunits, except the L-type subunit, incorporate a ferroxidase center (FC) that consists of a μ-oxo bridged dinuclear iron group and is responsible for iron oxidation. Additionally, a third iron site is observed in the subunits of some bacterial ferritins [1, 2]. The binuclear iron cluster of the FC and the third iron site are referred to as A, B and C iron sites, respectively. Some ferritins, also called bacterioferritins (BFRs), enclose 12 heme groups that are located between homodimeric subunit pairs, and may be involved in electron transfer [3]. The homologous protein DNA protection during starvation (DPS) and also called a “small ferritin,” is a spherical 12-subunit assembly with smaller dimensions (outer and inner diameters of 9 and 4.5 nm, respectively) and iron storage capacity (500 iron atoms vs. 3,000 in ferritin) than ferritins [4]. While iron is believed to be the main cationic substrate of ferritins, zinc accumulation in ferritin has also been reported [5].
Ferritins occur in a wide variety of organisms, from prokaryotes to mammals. In humans, their function is related to iron deficiencies or iron-overload disorders such as thalassemia, sickle cell anemia, hemochromatosis and pulmonary hemosiderosis [6–9]. Understanding the mechanism of the action of ferritin will be of great value in the treatment of iron-related diseases. Besides its medical relevance, this knowledge is also of fundamental scientific interest, especially in case of anaerobes. To date, an insight into the iron-uptake reaction has only been achieved for aerobic organisms. In these organisms, Fe(II) ions are oxidized, possibly together with phosphate oxoanions, and stored inside a ferritin shell as a soluble Fe(III) mineral, believed to be a ferrihydrite mineral with approximate composition Fe2O3·0.5H2O. Ferric ions from the core can be reduced and released from the ferritin molecule when needed by a cell. Nevertheless, the chemical nature of the physiological reductants for iron release and in the case of anaerobes that of the physiological oxidants for iron uptake have not yet been established.
Crystal structures of ferritins from bacteria, fungi, plants, insects and vertebrates are available, whereas those from archaea are limited to a ferritin from the sulfate-reducing anaerobe and thermophile Archaeoglobus fulgidus (AfFtn) (Protein Data Bank, PDB, codes 1sq3 and 1s3q) [2] and a DPS protein from Halobacterium salinarum (PDB codes 1tjo, 1tk6, 1tko and 1tkp) [10].
In AfFtn, a novel 23 point-group symmetry of the 24-mer assembly was found, which is unusual for ferritins and was only known previously in 12-meric assemblies, such as the DPS proteins [10, 11] and also in the ferritin from the bacterium Listeria innocua [12]. As a result, four outsized triangular pores about 45-Å wide are present in the protein shell of AfFtn [2].
Pyrococcus furiosus is a strict anaerobe and hyperthermophilic archaeon, living optimally at 100 °C. It was isolated from the hot marine springs off the beach of Porto di Levante in Vulcano, Italy. P. furiosus is a fermentative heterotroph that grows on starch, maltose, peptone and yeast extract and produces CO2, acetate, alanine, H2 and H2S. The expression of the native ferritin from P. furiosus (PfFtn) has been confirmed by N-terminal sequencing of the purified protein and the recombinant ferritin has been overproduced in Escherichia coli in order to afford extensive biochemical studies [13].
Herein we present the 3D structure of PfFtn, refined at 2.75-Å resolution. In contrast to the previously reported first structure of an archaeal AfFtn, the functional PfFtn 24-mer has the typical 432 point-group symmetry observed in other ferritins and BFRs. There are no four large pores such as were found in AfFtn. Instead, the characteristic threefold and fourfold channels are present in PfFtn.
The extreme thermostability of PfFtn, which in vitro can withstand 1 day of incubation at 100 °C, has been analyzed at the structural level, in comparison with structurally related ferritins. Although P. furiosus grows at the highest temperature, 100 °C, among the organisms used in thermostability analysis, our results suggest that the ferritin from Thermotoga maritima (TmFtn), which has an optimal growth temperature of 80 °C, may be even more thermostable than PfFtn. The hyperthermal stability of PfFtn and other ferritins appears to result mainly from the preservation of the monomer fold rather than the 24-mer assembly, owing to a high number of intramolecular hydrogen bonds between main-chain atoms, and between main-chain and side-chain atoms.
The FC stability was analyzed by electron paramagnetic resonance (EPR). Without further experiments, the lack of FC iron sites B and C in nonsoaked PfFtn crystals obtained does not allow any conclusions supporting or contradicting the FC as a stable catalytic cofactor.
Materials and methods
Protein production and crystallization
PfFtn was obtained and purified as previously described [13]. Ferritin iron loading prior to crystallization was performed by adding a freshly and anaerobically prepared aqueous solution of 20 mM iron sulfate, containing 0.1% HCl, to a solution of apoferritin in 50 mM N-(2-hydroxyethyl)piperazine-N′-ethanesulfonic acid (Hepes), pH 7, under aerobic conditions. Crystals of PfFtn were produced using the hanging drop vapor diffusion method at room temperature [14]. The crystals used for data collection were grown against 0.5 ml of 2 M ammonium sulfate reservoir solution, with drops containing 2 μl of 5 mg ml−1 protein solution in 50 mM Hepes, pH 7, and an equal amount of the reservoir solution. The crystals of the as-isolated ferritin, containing approximately 17 Fe per 24-mer, grew in 3 weeks and those loaded with approximately 1,000 Fe atoms per 24-mer prior to crystallization grew in 4 months. The crystals exhibited an irregular pyramidal shape, with base side and height between 150 and 250 μm. The iron content of as-isolated ferritin was determined with ferrene as previously described [14] by measuring the absorption of the iron–ferrene complex at 593 nm, ε = 35,500 M−1 cm−1.
Iron soaking of ferritin crystals was performed by transferring them to a 10–20-μl drop of a solution with composition 20 mM FeSO4, 2 mM Na2S2O4, 25% glycerol and 2 M ammonium sulfate, for 15 min. Zinc soaking with ZnCl2 was carried out in a similar fashion, with the exception that dithionite (Na2S2O4) was omitted from the solution.
X-ray data collection
Cryoprotecting conditions consisted of briefly dipping the crystals into a modified crystallization solution containing 20% glycerol in the case of the as-crystallized and Fe-loaded crystals, and 25% glycerol in the case of the Fe-soaked and Zn-soaked crystals, prior to flash-freezing at −173 °C in a nitrogen-gas stream, using an Oxford Cryosystems low-temperature device.
Diffraction data from an as-crystallized, an Fe-soaked and a Zn-soaked crystal of as-isolated PfFtn, as well as from a loaded PfFtn crystal, with approximately 1,000 Fe per 24-mer, were collected at the ESRF beamline BM-14, from flash-frozen single crystals at −173 °C. Two datasets were collected from the Zn-soaked crystal: the first at the Zn K-edge (9.6760 keV, 1.2810 Å, Zn peak) as determined from a fluorescence scan, and the second (9.500 keV, 1.305 Å, Zn low-E rm) on the low-energy side of the absorption edge. The diffraction images were integrated with the program MOSFLM [15]. The processed data were scaled, merged and converted to structure factors using the CCP4 program suite [16]. Data-processing statistics are summarized in Table 1. Table 1X-ray diffraction data collection and processing statisticsBeamlineESRF BM-14ESRF BM-14ESRF BM-14ESRF BM-14ESRF BM-14DetectorMAR 225 CCDMAR 225 CCDMAR 225 CCDMAR 225 CCDMAR 225 CCDCrystalAs-isolatedFe-loadedFeSO4 soakZnSO4 soak, Zn peakZnSO4 soak, Zn low-E rmWavelength (Å)1.77121.73401.45851.28101.305Space group C2221C2221C2221C2221C2221Unit cell (Å) a255.00254.32254.30255.37255.69 b341.42343.16342.88342.06342.36 c265.52266.26266.22265.99266.21Resolution range (Å)58.2–2.75 (2.90–2.75)63.2–2.95 (3.11–2.95)42.0–2.80 (2.95–2.80)42.9–2.80 (2.95–2.80)42.9–2.80 (2.95–2.80)Observations1,685,078 (151,184)903,956 (85,605)1,011,149 (102,079)1,082,567 (111,068)1,086,719 (111,952)Unique reflections295,335 (42,009)237,583 (32,389)280,407 (38,583)279,537 (37,607)280,474 (37,777)Completeness (overall) (%)99.7 (97.8)98.1 (92.2)99.1 (94.1)98.7 (91.6)98.7 (91.9)Redundancy5.7 (3.6)3.8 (2.6)3.6 (2.6)3.9 (3.0)3.9 (3.0)Rmerge0.085 (0.300)0.113 (0.532)0.081 (0.354)0.075 (0.288)0.077 (0.344)I/σ(I)16.5 (3.1)10.8 (1.6)8.2 (2.0)8.8 (2.5)8.6 (1.9)Anomalous completeness (%)98.0 (88.3)89.2 (71.1)94.4 (64.8)97.7 (85.4)97.6 (85.5)Anomalous redundancy2.9 (1.9)2.0 (1.5)1.8 (1.6)1.9 (1.5)2.0 (1.5)Ranom0.043 (0.204)0.078 (0.414)0.058 (0.294)0.063 (0.252)0.059 (0.288)Estimated Boverall (Å2)53.362.760.554.358.5Values in parentheses refer to the last resolution shell
Structure determination of as-isolated and Fe-loaded PfFtn
Previously [14], the 3D structure of PfFtn was solved by the molecular replacement (MR) method, using the program PHASER [17] and the homologous TmFtn (PDB entry 1vlg) as the search model. The sequence of this 165 amino acid ferritin shares a 51% amino acid sequence identity (88 amino acid residues) with that of PfFtn (174 amino acids), and the only insertions or deletions occur at the termini of the subunit chains. However, despite the success of the MR calculations we believed that, because of the low resolution of the diffraction data, relatively low amino acid sequence identity with the search model, as well as the complexity of the structure, experimental phase information would be needed in order to obtain the best possible structural model from the data. To that end, the first BM-14 dataset was collected at a long wavelength (7.0000 keV, 1.7712 Å) from a crystal of as-isolated PfFtn with the intention of using S-SAD in phasing, whereas a second dataset was collected just above the Fe K-absorption edge (7.1500 keV, 1.7340 Å) from an Fe-loaded PfFtn crystal, with the aim of using Fe-SAD in phasing and also for better location of the incorporated Fe atoms. However, neither dataset could provide the desired independent phase information, owing to complications that arose from radiation damage, which prevented us from collecting the highly redundant dataset required for application of the S-SAD method to the first dataset, as well as the lack of a sufficiently high ordered Fe content, which limited the usefulness of the anomalous dispersion data obtained from the second dataset.
Therefore, the 3D structure of PfFtn was solved by repeating the MR calculations previously described [14] with the new “as-isolated” 2.75-Å dataset, using the standard PHASER protocol as implemented through the CCP4 graphical user interface, with reflection data up to 3.25-Å resolution. Previously, we had established [14] that the asymmetric unit of the PfFtn crystal structure contained 36 monomers, arranged as 1.5 24-mers. The search model used was half of a TmFtn 24-mer, constructed from the PDB 1vlg coordinates without the Fe sites, and three such models were searched for in the crystal structure of PfFtn. The MR calculations with PHASER gave a single solution. Prior to model rebuilding, electron-density map improvement was carried out both by using Arp/wArp [18] in atom update and refinement mode as well as by using DM [19] with 36-fold noncrystallographic symmetry (NCS) averaging. In this way, the initial PHASER figure of merit of 0.502 for the full resolution range of the 2.75-Å dataset was improved to 0.697 (Arp/wArp) and 0.849 (DM). Rebuilding of the protein chain to take into account the differences in sequence and chain length between TmFtn and PfFtn was carried out with the program TURBO [20] using one monomer and the 36-fold averaged electron-density map obtained from the DM calculations. The coordinates for each of the remaining 35 monomers were obtained from those of the first by applying the appropriate NCS transformations, followed by model correction with TURBO using the electron-density maps obtained from both the DM and the Arp/wArp calculations. In these maps, peaks corresponding to Fe occupation of the A sites only in the FC were observed, and therefore only 36 iron sites (one per monomer), were included in the model.
The structure of the Fe-loaded PfFtn was also determined by MR with PHASER using a similar procedure; however, the |Fo| density map, as well as an anomalous electron density map, calculated with MR phases, showed that only the A site in the FC was occupied by Fe atoms. Since the “as-isolated” dataset had higher resolution and better quality, further work on the “Fe-loaded” dataset was abandoned.
Structure refinement of as-isolated PfFtn
Refinement was carried out with the program REFMAC [21] using medium main-chain and weak side-chain NCS restraints between the 18 dimers in the asymmetric unit. At a later stage, a translation–libration–screw (TLS) rigid-body motion refinement [22] was made prior to restrained refinement of atomic positions and thermal motion parameters. One rigid body was defined for each of the 36 independent monomers. In the final refinement stages, a total of 832 solvent molecules were located with Arp/wArp [23]. On the basis of their B factors, observed electron density and interactions with neighboring protein residues, 791 were assigned to water molecules, while the remaining 41 were modeled as sulfate ions. Individual restrained B factors were refined for all non-hydrogen atoms and, for consistency with the structure refinements of the Fe-soaked and Zn-soaked crystals (see below), only tight NCS restraints (0.1 Å positional and 1.0 Å2 thermal parameter) were applied to all the protein residues involved in the FC metal coordination, instead of defining the coordination geometry for each individual protein–metal bond as in the earlier stages of refinement.
The final values of R and Rfree were 0.195 and 0.247, respectively. Neither TLS rigid-body refinement nor full NCS restraints (i.e., extending over the whole monomer and relating all the chains in the asymmetric unit) were used in the final refinement, as they were seen to lead to higher values of both R and Rfree. Rfree [24] was calculated from a randomly chosen subset of the data, containing approximately 5% of the total number of independent reflections. Throughout the refinement, the model was periodically checked and corrected with TURBO against 2|Fo| − |Fc| and |Fo| − |Fc| electron-density maps. The refinement statistics are included in Table 2. The maximum-likelihood estimate of the overall coordinate error, obtained with REFMAC, was 0.20 Å. For all 36 independent monomers in the PfFtn crystal structure, only the first 168 residues were included in the model. Although some electron density was visible beyond Gly168, it was not possible to build any more C-terminal residues into it, and it is likely that the C-terminal tail is disordered in all the independent PfFtn monomers. On the basis of a B-value comparison with the protein residues directly coordinating the iron atoms, it was realized that at least 17 out of the 36 metal sites were not fully occupied, and were assigned occupation factors of 0.8 (0.7 in one case). In addition, 788 side-chain atoms could not be seen in either 2|Fo| − |Fc| or |Fo| − |Fc| electron-density maps and were therefore given zero occupancy during the refinement. Table 2Refinement and structure analysis statistics for the ferritin from Pyrococcus furiosus (PfFtn)CrystalAs-isolated, as-crystallizedFe-soakedZn-soakedResolution (Å)56.8–2.7542.0–2.8042.8–2.80No. of reflections in work set/test set280,375/14,914266,173/14,187265,356/14,134Rwork/Rfree0.195/0.2470.197/0.2490.201/0.251Number of atoms Protein—total number of atoms49,82449,82449,824 Protein—atoms with zero occupancy788718720 Fe/Zn atoms36108108 Water molecules (sulfate ions)791 (41)413 (26)162 (44)Mean B factors (Å2) Protein main chain (side chain)32.6 (34.6)38.0 (39.9)30.9 (32.7) Fe/Zn atoms57.553.236.9 Water molecules (sulfate ions)26.5 (67.1)29.0 (76.8)21.9 (74.0)Root mean square deviations from ideal values Bond lengths (Å)0.0140.0140.015 Bond angles (°)1.441.461.48 Mean positional error (Å)0.200.220.22PROCHECK Ramachandran analysis Total no. analyzed residues5,5445,5445,544 Residues in core regions, no. (%)5,255 (94.79)5,241 (94.53)5,239 (94.50) Residues in allowed regions, no. (%)277 (5.00)285 (5.14)289 (5.21) Residues in generous regions, no. (%)11 (0.20)10 (0.18)9 (0.16) Residues in disallowed regions, no. (%)1 (0.02)8 (0.14)7 (0.13)PROCHECK G factors Mean (range) dihedral0.00 (−0.13 to 0.07)−0.04 (−0.23 to 0.03)−0.04 (−0.21 to 0.09) Mean (range) covalent0.38 (0.32–0.45)0.38 (0.29–0.46)0.36 (0.26–0.44) Mean (range) overall0.15 (0.07–0.19)0.12 (0.00–0.18)0.12 (0.01–0.17)
The structure was analyzed with PROCHECK [25] and its stereochemical quality parameters were within their respective confidence intervals. A Ramachandran [26] ϕ,φ plot showed that of the 5,544 non-glycine and non-proline residues, 11 (0.20%) were found to lie in the generously allowed regions and only one (0.02%) outlier was observed. All these residues with more unusual ϕ,φ conformations were located in a loop region containing Lys145, for which the electron density was generally poor. The final atomic coordinates were deposited with the PDB [27], to be released upon publication, with the accession code 2jd6.
Structure determination and refinement of the Fe-soaked and Zn-soaked PfFtn crystals
In view of the very small changes in cell parameters observed in comparison with those of the as-isolated crystal, the structures of the Fe-soaked and Zn-soaked crystals of PfFtn were determined directly by using the final refined coordinates of the as-isolated structure, omitting the iron sites and the solvent atoms. Ten cycles of rigid-body refinement were first carried out, followed by a TLS rigid-body motion refinement [22] prior to restrained refinement of atomic positions and thermal motion parameters. At this stage, the 2|Fo| − |Fc| or |Fo| − |Fc| electron-density maps were inspected with COOT [28] for inclusion of the FC A, B and C metal sites in the models. Because of the complexity of the metal coordination in the FC, the low resolution of the diffraction data and the realization that some of the FC metal sites were not fully occupied (see “Discussion” for details), tight NCS restraints (0.1 Å positional and 1.0 Å2 thermal parameter) were applied to all the protein residues involved in the FC metal coordination, instead of defining the cordination geometry for each individual protein–metal bond.
In the final refinement of the Fe-soaked crystal structure, 423 solvent molecules and 26 sulfate ions were located with Arp/wArp [23] and included in the model, and individual restrained B factors were refined for all non-hydrogen atoms. On the basis of a B-value comparison with the protein residues directly coordinating the iron atoms, it was realized that not all metal sites were fully occupied: all the B sites and one C site were assigned an occupation of 0.5. The final values of R and Rfree were 0.197 and 0.249. The refinement statistics are listed in Table 2. The structure was analyzed with PROCHECK [25] and its stereochemical quality parameters were within their respective confidence intervals. A Ramachandran [26] ϕ,φ plot showed that of the 5,544 non-glycine and non-proline residues, ten (0.18%) were found to lie in the generously allowed regions and only eight (0.14%) outliers were observed. With three exceptions, all these residues with more unusual ϕ,φ conformations were located in a loop region containing Lys145, for which the electron density was generally poor.
In the final refinement of the Zn-soaked crystal structure, 162 solvent molecules and 44 sulfate ions were located with Arp/wArp [23] and included in the model, and individual restrained B factors were refined for all non-hydrogen atoms. A B-value comparison with the protein residues directly coordinating the metal atoms, combined with the observation of a mixed iron/zinc population at the A sites (see below for details) led us to postulate a mixed (50:50 Zn/Fe) occupancy for the A sites, which was modeled as a Zn atom with 0.9 occupancy. Also, the B sites were modeled as partially occupied by zinc (50%) and the C sites as partially occupied by iron (50%). The final values of R and Rfree were 0.201 and 0.251. The refinement statistics are included in Table 2. The structure was analyzed with PROCHECK [25] and its stereochemical quality parameters were within their respective confidence intervals. A Ramachandran [26] ϕ,φ plot showed that of the 5,544 non-glycine and non-proline residues, nine (0.16%) were found to lie in the generously allowed regions and only seven (0.13%) outliers were observed. With two exceptions, all these residues with more unusual ϕ,φ conformations were located in a loop region containing Lys145, for which the electron density was generally poor.
As for the as-isolated structure, neither TLS rigid-body refinement nor full NCS restraints were used in the final refinement of the Fe-soaked or Zn-soaked structures, as they were seen to lead to higher values of both R and Rfree. Rfree [24] was calculated from a randomly chosen subset of the data, containing approximately 5% of the total number of independent reflections, and with the same h,k,l indices as for the “as-isolated” data. These test-set reflections were not chosen in thin resolution shells. However, while it is true that the 24-mers have very high NCS, in all three refined crystal structures the Rfree/Rwork ratios are typical for the data resolution and refinement parameters used, [29] and therefore there is no indication that the more simplistic test-set choice may have introduced a bias in the Rfree values. Throughout the refinement, the models were periodically checked and corrected with COOT [28] against 2|Fo| − |Fc| and |Fo| − |Fc| electron-density maps. Since the asymmetric unit of both structures contains 36 monomers, and each monomer contains one FC with three metal sites, a total of 108 metal atoms are present in the model. In addition, 718 side-chain atoms could not be seen in either 2|Fo| − |Fc| or |Fo| − |Fc| electron-density maps for the Fe-soaked structure (720 for the Zn-soaked structure) and were given zero occupancy during the refinement. The final atomic coordinates were deposited with the PDB [27], to be released upon publication, with the accession codes 2jd7 (Fe-soaked structure) and 2jd8 (Zn-soaked structure).
Results and discussion
Structure of the PfFtn monomer
The PfFtn monomer exhibits the typical four-helical bundle fold which is encountered in ferritins, BFRs, DPS proteins and rubrerythrins. According to the DSSP algorithm [30] the four helices comprise residues 4–34 (helix A), 37–64 (helix B), 83–111 (helix C) and 114–144 (helix D). A short, C-terminal α-helix (helix E, residues 148–160), bordering one of the cylindrical bundle openings, followed by a tail without any secondary structure also occurs in the PfFtn monomer fold (Fig. 1a, b). For convenience of terminology, the two extremities of the cylindrical α-helical bundle will be further referred to as the open side and the E-helix sides of the monomer. Fig. 1a Topology diagram of the ferritin from Pyrococcus furiosus (PfFtn). b Ribbon diagram of PfFtn showing the location of the ferroxidase center sites A, B and C (orange spheres) near the center of the four-helical bundle. c Superposition of the PfFtn Cα chain with those of its eight closest homologous structures. In b and c, the molecule colors change from blue to red in the N-terminal to C-terminal direction. b, c Prepared with PyMOL [50]
The PfFtn monomer is built of hydrophobic amino acids at its open and E-helix sides and polar and hydrophilic residues in the middle of the subunit, similarly to the subunits from other ferritins. The open side of the monomer contains three leucine residues (Leu10, Leu104 and Leu120), two phenylalanine residues (Phe57 and Phe123), Met6, Tyr60 and Ile101, whereas the E-helix side has three phenylalanine residues (Phe87, Phe31 and Phe39), Trp42 and Met27 (denoted by asterisks in Fig. 2). The residues defining the open and E-helix sides of the monomer are conserved among PfFtn, the ferritin from E. coli (EcFtnA), AfFtn and human H chain ferritin (HuHF) in terms of their polarities, apart from Phe123, which in other ferritins is a tyrosine. Fig. 2Amino acid sequence alignment of PfFtn, the ferritin from Archaeoglobus fulgidus (AfFtn), the ferritin from Escherichia coli (EcFtnA) and human H chain ferritin (HuHF). Asterisks show the residues around the open and E-helix sides of the monomer; A, B and C denote the residues coordinating the iron sites of the ferroxidase center (FC); red bars highlight the main residues located around the threefold channels and gray bars those around the fourfold channels. The alignment was prepared with ClustalW [51]
The central part of the subunit comprises the negatively charged and polar amino acids involved in metal binding. These are Glu17, Glu49, Glu50, Glu94, Glu126, Glu129 and Glu130 and His53, which are shown in Fig. 2 with A, B or C labels to denote the iron site they coordinate (see “The FC”). In addition, Tyr24 and Gln127, which in EcFtnA [1] indirectly contribute to the A and B sites via hydrogen bonds to Glu94 and Glu17, are also oriented towards the FC in PfFtn.
It has been noted [2] that Gly67, Ser84 and Gly147 (in AfFtn numbering) are conserved among AfFtn, EcFtnA and HuHF, and define the BC loop and the end of D helix, respectively. In the case of PfFtn, only Ser84 is conserved, while the two glycine residues are changed into Asn65 and Lys145.
A surface hydrophobic pocket at the PfFtn dimer interface
A large surface pocket, largely lined by hydrophobic residues, was located at the PfFtn dimer interface. Interestingly, one of these is Ile51, corresponding to Met 57 in the BFR from Desulfovibrio desulfuricans ATCC 27774 [31], which binds the heme group. Instead of a heme, the pocket appears to be occupied by a disordered string of water molecules (Fig. 3) which form a continuous region of electron density. This feature was observed for all the PfFtn structures investigated. A similar situation was reported by Yu et al. [32] for the hydrophobic cavity of human interleukin 1β. Fig. 3Stereoview of a dimer interface in the as-isolated PfFtn structure, showing a difference electron density feature, drawn at the 2.5 map root mean square (RMS) level, which results from a disordered chain of water molecules trapped within a hydrophobic pocket centered around Ile51. This feature is observed in all dimers of all PfFtn structures investigated. The bulk of the PfFtn dimer is represented by its Cα trace (blue-gray). The side chains (including Cα atoms) of the Ile51 residues in both monomers are represented in ball-and-stick mode and are colored gold. The view is down the noncrystallographic twofold symmetry axis of the dimer, looking towards the inside of the PfFtn 24-mer. The figure was prepared with DINO [52]
Subunit structure comparison
In order to compare the PfFtn structure with the structures of other known ferritins, a protein structure comparison against the PDB using the secondary structure matching (SSM) tool at the European Bioinformatics Institute (http://www.ebi.ac.uk/msd-srv/ssm) [33] was performed. This resulted in 377 hits, which, after elimination of duplicate protein chains, were seen to correspond to 63 different protein structures. A multiple structural alignment was then carried out with a subset of eight SSM hits, listed in Table 3 and shown as a superposition of Cα chains in Fig. 1c. The members of this subset were chosen such that the percentage of their aligned residues was equal to or greater than 25%, all five secondary structure elements (α-helices) were aligned, the Q score, of their pairwise SSM alignment with PfFtn was greater than 0.5, and considering only one crystal form of each different protein structure from each organism. This subset includes representatives from archaea, bacteria and eukarya. The structural alignment showed PfFtn to be highly similar to TmFtn, followed by AfFtn and EcFtnA (Table 3). The deviations occur mostly in loop regions: AB loop from Glu32 to Leu34, end of the BC loop from Lys78 to Pro83, CD loop and helix D up to Phe123 and DE loop from Lys145 to Leu151 (in PfFtn numbering). Table 3Proteins with known 3D structure most closely homologous to PfFtnPDB IDProtein nameSource organismChain lengthIdentity (%)Identity SS-aligned (%)Q scoreaRMSD (Å)1vlgFerritin (TM1128) Thermotoga maritima1765555.30.920.491s3qFerritinArchaeoglobus fulgidus1735048.20.940.531krqFerritinCampylobacter jejuni1673939.20.811.021eumFerritin ECFTNAEscherichia coli1653838.40.860.901r03Mitochondrial ferritinHomo sapiens 1824030.60.681.552fhaH ferritin K86Q mutantHomo sapiens 1833729.40.621.981mfrM ferritinRana catesbeiana1763629.60.681.671rcgL ferritinRana catesbeiana1733225.00.612.03PDB Protein Data Bank, SS secondary structure, RMSD root mean square deviationaParameter that takes the alignment length and the RMSD between the superimposed atoms into account, and varies from 0 in the case of poor matches to 1 in the case of identical proteins
Structure of the PfFtn 24-mer and comparison with AfFtn
A highly conserved quaternary structure appears among all known ferritins—a spherical tetraeicosameric assembly with 432 point-group symmetry. An interesting exception to this rule is the recently reported structure of the AfFtn 24-mer [2], which exhibited 23 point-group symmetry, normally an attribute of the dodecameric DPS proteins. Even though PfFtn and AfFtn are highly similar archaeal ferritins (50 and 70% amino acid sequence identity and similarity, respectively), the PfFtn 24-mer has the canonical 432 point-group symmetry instead of the 23 point-group symmetry found in the AfFtn 24-mer. Although there may be other archaeal tetraeicosameric ferritins with 23 point-group symmetry like AfFtn, our structure clearly suggests that it is not a feature of all archaeal ferritins.
When comparing the two quaternary structures (Fig. 4), it can be seen that the 432 and 23 arrangements in PfFtn and AfFtn result from a different organization of the four fundamental units, hexamers formed by three dimers around a 90° vertical turn from a threefold NCS rotation axis. The 23 arrangement leads to the appearance of four outsized triangular pores about 45-Å wide in the protein shell of AfFtn [2], which are absent in the 432 arrangement of PfFtn and all other known structures of tetraeicosameric bacterial ferritins. Fig. 4Ribbon diagrams of the PfFtn and AfFtn 24-mers. I View down a threefold axis, showing one “fundamental hexamer.” II View rotated by approximately 90° about a vertical axis, showing the different hexamer arrangement, which leads to the appearance of large pores in AfFtn. The figure was prepared with PyMOL [50]
Threefold and fourfold channels
The eight threefold channels are formed by helices C and D from three adjacent dimers, in such a way that, with respect to the hollow tetraeicosameric protein shell, the C-terminal ends of helices C define the outer entrance to the channel and the N-termini of helices D define the inner entrance (Fig. 5). The PfFtn threefold channel is lined by a mix of hydrophobic and hydrophilic amino acids Ala106 and Glu109 at the outer side of the channel, Tyr114 and Arg117 in the middle and Ala118 and Glu121 at the inner entrance of the channel (Fig. 5). In general, and similarly to EcFtnA and AfFtn, this channel is less hydrophilic than in vertebrate H and L chain ferritins, suggesting a rather hydrophobic threefold channel in microbial ferritins in comparison with vertebrate ferritins. Nevertheless, when compared with that in EcFtnA and AfFtn, the threefold channel in PfFtn is more similar to that of HuHF, in the sense that is it more negatively charged at its inner entrance and more positively charged in its central region (Fig. 5). Fig. 5Profile view of the threefold and fourfold channels in PfFtn, AfFtn, EcFtnA and HuHF. The exterior of the shell lies on the left side and the inner cavity on the right side of each cartoon as shown schematically in a. The red residues correspond to acidic Glu and Asp residues and the blue ones to positive Lys, Arg and His residues, and highlight the arrangement of positive and negative amino acids along the channels. The figure was prepared with PyMOL [50]
On the basis of the crystallographic, kinetic, mutagenesis and electrostatic calculations studies on HuHF [9, 34–36], the threefold channels were proposed to be involved in the entry of iron ions into the ferritin molecule. The inhibition of iron uptake in HuHF upon Asp131 and Glu134 substitutions [34] and the coordination of Ca2+ by Asp131 and Glu134 [37] have been observed in this protein. The inner narrow entrance of the threefold channel in PfFtn resembles that of HuHF, suggesting that a similar mechanism is possible in PfFtn. Nevertheless, the mutations of His118 and Cys130 also influenced the iron-binding capacity of HuHF and inhibited the binding of different metals by the channel to different extents. The different amino acid arrangements in the threefold channel of the various ferritins suggests the possibility of different iron incorporation mechanisms.
The six fourfold channels, formed around the fourfold symmetry axes of the PfFtn 24-mer, are lined, from the outside to the inside of the protein shell, by Gln149, Met153, Lys156 and Glu157 (Fig. 5). Owing to a nozzlelike shape formed by Gln149, the PfFtn fourfold channel is longer (approximately 16 Å) than its counterparts in EcFtnA (approximately 14 Å) and HuHF (approximately 12 Å). In PfFtn, Lys145 in the DE loop, which forms an outer gate to the fourfold channel, occurs in place of EcFtnA Gly145 or HuHF Gly159, making this gate more polar and charged in PfFtn. In addition, the DE loop and its amino acids are somewhat oriented outwards from the fourfold channel, resulting in a hourglass shape of this channel, which is less strongly developed in EcFtnA and even less so in HuHF, where the pore has an almost cylindrical shape. The outer entrance of the PfFtn fourfold channel is polar and hydrophilic, the region inside the channel is nonpolar and the inner entrance is polar again. The fourfold channels in EcFtnA have a similar arrangement of polar residues at both channel entrances and are nonpolar in the middle, while in HuHF they are mostly apolar with Leu165 and Leu169 and polar His173 on the inner side of the channel (Fig. 5). In electrostatic terms, the PfFtn fourfold channel is similar to that of both EcFtnA and HuHF; the outer entrance of it is uncharged as in HuHF and the inner entrance has a compensated charge from the sequence of positive Lys156 and negative Glu157 as in EcFtnA.
The FC
The active site responsible for the oxidation of Fe(II) to Fe(III) in ferritins, the FC, is located in the central part of each subunit (Fig. 1b). It is generally believed that in most ferritins the initial stages of the core formation take place via the FC. Subsequently, iron incorporation and removal is thought to occur predominantly through one or the other of the channels formed around the threefold and fourfold symmetry axes of the 24-mer. In this discussion, the nomenclature for the FCs in ferritins and BFRs presented in a recent review [38] will be followed. The FC in PfFtn is of the FtnA (EcFtnA) type, which is strictly conserved in all the closest prokaryote structural homologues of PfFtn besides EcFtnA (Table 4). The EcFtnA FC contains three metal binding sites (A, B and C), in contrast with the HuHF and E. coli BFR FC, which only contain two (A and B). The FC of the as-crystallized PfFtn showed iron bound to site A only. Iron binding to sites B and C was observed only after soaking the PfFtn crystals in Fe(II). Table 4The ferroxidase center in PfFtn and its homologous crystallized ferritinsPDB IDProtein nameSource organismResolution (Å)FC typeMetal site occupation–FerritinPyrococcus furiosus2.75FtnAA (Fe)A, B, C (Fe,Zn)1s3qFerritinArchaeoglobus fulgidus2.10FtnAA, B (Zn)aA, B, C (Fe)1vlgFerritin (TM1128) Thermotoga maritima2.00FtnAA (Fe)1krqFerritinCampylobacter jejuni2.70FtnAA, B (Ow?)b1eumFerritin ECFTNAEscherichia coli2.05FtnAA, B (Zn)cA, B, C (Fe)1r03Mitochondrial ferritinHomo sapiens sapiens1.70HuHF–2fhaH ferritin K86Q mutantHomo sapiens sapiens1.90HuHF–1mfrM ferritinRana catesbeiana2.80HuHF–1rcgL ferritinRana catesbeiana2.20––FC ferroxidase centeraZn in 1s3q and Fe in 1sq3 [2]b1krq contains four water molecules, two of which are located at the FC, in positions corresponding to A and B sites. They might be low-occupancy metal sites but there is no indication of whether it is a native or a soaked crystalc1eum corresponds to the native structure [1]; these sites were located in the Zn2+ and Fe3+ derivatives, respectively, for which no coordinates were deposited in the PDB
In PfFtn, iron site A is coordinated by Glu17, Gly50 and His53, iron site B by Glu50, Glu94 and Glu130, and iron site C by Glu49, Glu126, Glu129 and Glu130. Site B is 3.0 Å distant from site A. Site C is located 7.5 and 6.3 Å away from sites A and B, respectively (Table 5, Fig. 6). Glu17 and Glu94 appear to be monodentate ligands; therefore, sites A and B are coordinated by side chains from three amino acids, while site C is coordinated by side chains from four amino acids. The final 2|Fo| − |Fc| electron density around the FC of a selected PfFtn monomer in the as-isolated, Fe-soaked and Zn-soaked crystal structures are represented in Fig. 7a–c, respectively. Table 5Coordination geometry statistics of the ferroxidase center in PfFtnBondsaDistance (Å)As-isolatedbFe-soakedbZn-soakedbMA...Glu17 Oε12.24 (0.09) [2.00–2.46]2.21 (0.08) [2.00–2.47]2.18 (0.07) [2.01–2.32]MA...Glu50 Oε12.26 (0.11) [2.09–2.47]2.24 (0.09) [2.09–2.55]2.22 (0.07) [2.02–2.38]MA...His53 Nδ12.12 (0.09) [1.97–2.41]2.28 (0.09) [2.12 – 2.47]2.21 (0.08) [2.08 – 2.40]MA...MB3.02 (0.19) [2.62–3.63]2.93 (0.18) [2.54–3.49]MA ... MC7.41 (0.10) [7.18–7.63]7.51 (0.11) [7.30–7.77]MB...Glu50 Oε22.21 (0.12) [1.93–2.45]2.16 (0.12) [1.92–2.50]MB...Glu94 Oε22.34 (0.14) [2.12 – 2.70]2.40 (0.10) [2.18 – 2.65]MB...Glu130 Oε12.87 (0.16) [2.47–3.14]2.81 (0.20) [2.41–3.21]MB...MC6.29 (0.20) [5.81–6.77]6.25 (0.15) [5.93–6.55]MC...Glu49 Oε12.11 (0.08) [1.92 – 2.36] 2.29 (0.11) [2.13 – 2.56]MC...Glu126 Oε12.06 (0.09) [1.87–2.26]2.31 (0.16) [1.84–2.53]MC...Glu129 Oε12.31 (0.10) [2.12–2.58]2.02 (0.10) [1.81–2.26]MC...Glu130 Oε21.89 (0.08) [1.72–2.04]1.89 (0.09) [1.66–2.07]aM stands for metal (Fe in as-isolated and in Fe-soaked; Zn or Fe in Zn-soaked)bFor each column, the numbers listed are the mean value, calculated from the 36 independent distances in the asymmetric unit, followed by the corresponding standard deviation in parentheses, and finally the minimum and maximum values in the population, enclosed in bracketsFig. 6View of the FC down the PfFtn subunit from the E-helix side. The three metal sites are annotated as A, B and C. Metal occupation of sites B and C is observed only upon crystal soaking in Zn or Fe. The dashed lines correspond to the coordination geometries described in Table 5. The figure was prepared with PyMOL [50]Fig. 7Details of the FC of a selected monomer in the structures of as-isolated, Fe-loaded, Fe-soaked and Zn-soaked PfFtn crystals. a Final 2|Fo| − |Fc| electron-density map for the as-isolated PfFtn, contoured at 1.2 map RMS. b Final 2|Fo| − |Fc| electron-density map for the Fe-soaked PfFtn, contoured at 1.2 map RMS. c Final 2|Fo| − |Fc| electron-density map for the Zn-soaked PfFtn, contoured at 1.2 map RMS. d Anomalous Fourier map for the Fe-loaded PfFtn, contoured at 3.0 map RMS. e Anomalous Fourier map for the Fe-soaked PfFtn, contoured at 3.0 map RMS. f Anomalous Fourier map for the Zn-soaked PfFtn using the peak data, contoured at 3.0 map RMS. g Labeled view of the residues and metal sites in the Fe-soaked FC. h Anomalous Fourier map for the Zn-soaked PfFtn using the low-energy remote data, contoured at 3.0 map RMS. i Dispersive Fourier map for the Zn-soaked PfFtn, contoured at 2.8 map RMS. The anomalous electron-density maps were calculated using as amplitudes the anomalous difference coefficients obtained from each dataset, and the phases (rotated by 90°) obtained from the respective final structure refinement, except for the “Fe-loaded” data, for which the phases were taken from the “as-isolated” refinement. The Zn-dispersive Fourier map was calculated by first scaling together the peak and low-energy remote datasets with CCP4 SCALEIT [16] and then using as coefficients the difference F(low-energy remote) − F(peak) and the phases from the final “Zn-soak” refinement. a–c show that the occupation of site B in the as-isolated crystal is negligible; d shows a small residual occupancy of site B in the Fe-loaded structure and no evidence of site C occupancy; e shows a site B occupancy clearly lower than those of sites A and C in the Fe-soaked structure; f, h and i show that in the Zn-soaked structure, part of the originally present Fe in site A has been displaced to site C by the Zn ions, that site B is occupied by Zn only, and that site C is very likely occupied by Fe only. All panels were drawn with the same orientation. In all panels except g the bulk of PfFtn monomer is represented as a gray tube onto which the side chains (including Cα atoms) of the FC residues have been overlaid in ball-and-stick representation (carbon atoms blue-gray, oxygen atoms red, nitrogen atoms blue, iron atoms pink and zinc atoms cyan). In g, only the residues near the FC are represented in ball-and-stick mode, and are labeled for easier identification of the residues mentioned in the text. The figure was prepared with DINO [52]
For the Fe-loaded crystal structure, the anomalous difference Fourier map, calculated with the phases obtained from the final “as-isolated” refinement (Fig. 7d), shows a very small occupancy of site B. This map is very similar to that obtained from the MR phases (not shown). In the Fe-soaked crystal, the anomalous difference Fourier map calculated at the end of the refinement showed that sites A and C have a clearly higher Fe occupancy than site B (Fig. 7e).
In the Zn-soaked crystal, the anomalous difference Fourier maps obtained from the two datasets collected (just above and 176 eV below the Zn K-absorption edge) clearly show some iron occupation of sites A and C (Fig. 7h), whereas zinc populates sites A and B (Fig. 7f, i). The dispersive Fourier map calculated using the two datasets (Fig. 7i) shows Zn peaks at sites A and B but since no diffraction data were recorded at the inflexion point of the Zn K-absorption edge, this map is noisier and no conclusions can be drawn about any Zn occupation of site C. Since the as-isolated structure only showed Fe occupation of FC site A, we conclude that the Fe occupation of site C in the Zn-soaked structure results from iron initially present in site A, displaced to site C by the Zn soak. This is an interesting new result, since the previous Zn-soaked and Fe-soaked structures reported for AfFtn [2] and EcFtnA [1] were obtained from apoferritin crystals.
FC stability
Interestingly, while mutagenesis studies have shown both sites A and B to be important for iron uptake and oxidation in ferritins [38], in the crystallized proteins the occupancy of the both sites was found only in BFR from D. desulfuricans ATCC 27774 [31]. The PfFtn FC followed the common trend, and as mentioned already the occupation of FC sites B and C was only observed in Fe-soaked or Zn-soaked crystals. This contradicted our EPR studies that showed a fully developed dinuclear iron FC EPR signal in the sample prior to crystallization [39]. In order to check the integrity of the PfFtn FC, we loaded three samples of apo PfFtn with iron and incubated them for 1 day and 2 months at 4 °C in air and, mimicking the crystallization conditions, for 2 months in a 2 M ammonium sulfate solution at room temperature. The FC EPR signal from each sample was measured after titration to 130 mV, which is the potential at which the mixed-valence PfFtn EPR signal is maximum [39]. The amplitude of the FC signal from the sample incubated for 2 months in air was noticeably but not drastically smaller than that of the sample incubated for 1 day (Fig. 8). However, the FC EPR signal amplitude of the sample incubated for 2 months in a 2 M ammonium sulfate solution was fivefold smaller than that from the sample incubated for 2 months in air (Fig. 8). The coordination distance from Glu130 to site B (Table 5) is quite long (more than 2.8 Å); therefore, site B may be rather labile, which could account for its lower occupation in the Fe-soaked and Zn-soaked crystals, as well as for the easy removal of its iron atom by the precipitating agent ammonium sulfate during crystallization of the as-isolated ferritin. Fig. 8The Fe(III)–O–Fe(II) FC electron paramagnetic resonance (EPR) signal of 6 μM PfFtn 24-mer, titrated to 130 mV. The decrease in the amplitude of the signal depending on incubation time and the presence of crystallization solution in the sample illustrates the stability of the center. Prior to EPR measurement, the three samples tested were incubated, respectively, for 1 day after loading with iron (blue line), for 2 months (red line) and for 2 months in the presence of a crystallization solution, 2 M ammonium sulfate (black line)
There is an ongoing discussion in the literature on whether the FC iron sites A and B are transient in the process of iron uptake–oxidation–translocation into the protein core, or whether they form a stable cofactor [38]. In the current experiment, the decrease in the amplitude of the FC EPR signal in the sample incubated in air for 2 months appears to be too small to be caused by iron translocation into the protein cavity. Instead, as shown by the sample incubated for 2 months in ammonium sulfate, the most likely cause for the lack of the B iron site in the as-crystallized PfFtn is iron complexation and removal by the crystallization agent.
Thermostability
The thermal stability of a protein can often be described as an equilibrium between the protein’s folded and unfolded state, and frequently correlates with the growth temperature of the organism. Analysis of thermostability factors is often made via comparison of the genome/protein sequences from the hyperthermophiles, thermophiles and mesophiles, or via mutagenesis. In the literature, a number of factors have been proposed to be involved in thermostability. These include salt bridges, hydrogen bonds, van der Waals interactions, distribution of amino acids, higher β-strand content, more charged amino acids, fewer uncharged polar amino acids, more hydrophobic β-branched amino acids and others [40–46]. However, no single consistent set of determining factors has emerged from the different studies that have been carried out to date; therefore, it appears that protein thermal stability is the cumulative result of various factors.
Ferritins are large and highly symmetrical agglomerates of 24 single subunits. This suggests that these proteins possess forces that keep the big structure together. Horse spleen ferritin has been shown to resist inactivation by up to 10 M urea [47] and temperatures up to 93 ºC [48].
PfFtn has been shown to be an extremely thermostable protein. No melting temperature could be found with differential scanning calorimetry up to 120 °C and the protein could withstand incubation at 100 °C for 1 day or autoclaving at 120 °C for half an hour without loss of its iron-uptake activity [13]. There are no data in the literature regarding the thermostability of other thermophilic and hyperthermophilic ferritins.
In our analysis of PfFtn thermostability (summarized in Table 6), and given that no thermostability data are available in the literature on ferritins from microorganisms other than P. furiosus, we assumed that it is the most thermostable ferritin among those shown in Table 6, and we used the optimal growth temperature of the organism as a measure of relative thermostability for the other ferritins listed in Table 6. However, it should be noted that the denaturation temperature of a protein may be higher than the optimal growth temperature of the organism where it is produced. Table 6Hydrogen bonds and salt bridges in the PfFtn 24-mer and its closest structural homologues, calculated with HBPLUSPDB IDSource organism aTotal H-bonds bMMMSSSSalt bridges cIntramolecularIntermolecularIntramolecularIntermolecularIntramolecularIntermolecularIntramolecularIntermolecular–Pyrococcus furiosus (100 °C)4,6303,3772443972544174172671vlgThermotoga maritima (80 °C)4,7673,237305401176332101891201s3qArchaeoglobus fulgidus (80 °C)4,4572,78824412804531761141341krqCampylobacter jejuni (42 °C)4,6323,360246487238414472721eumEscherichia coli (37 °C)4,7203,320248363241296604MM hydrogen bonds between main-chain atoms only, MS hydrogen bonds between main-chain and side-chain atoms, SS hydrogen bonds between side-chain atoms onlyaOptimal growth temperature in parenthesesbCalculated using default program parameterscA subset of SS hydrogen bonds, considering interactions between side chains of negatively charged Asp/Glu residues with side chains of positively charged Arg/His/Lys residues
The hydrogen bonds were calculated with HBPLUS [49] using default parameters, and were divided into three classes: between main-chain atoms only (MM); between main-chain and side-chain atoms (MS); and between side-chain atoms only (SS). The salt bridges were determined as a subset of SS hydrogen bonds, considering interactions between side chains of negatively charged Asp/Glu residues with side chains of positively charged Arg/His/Lys residues. Furthermore, a distinction between intramolecular (within the same monomer) and intermolecular (between neighboring monomers) hydrogen bonds was made. The MM bonds are mostly intramolecular and are responsible for the stability of the secondary structure elements (α-helices). The intramolecular MS and SS hydrogen bonds contribute to stabilize the 3D fold of the monomeric subunits. The intermolecular MS and SS hydrogen bonds are those that contribute most to the stability of the 24-mer.
A surprising result of this calculation was that the total number of hydrogen bonds in the ferritin 24-mers was quite similar. If we exclude the lowest number (4,457 for AfFtn) which can possibly be attributed to its unusual quaternary structure, PfFtn actually has the lowest total number of hydrogen bonds, two fewer than the ferritin from Campylobacter jejuni (CjFtn), 90 fewer than EcFtnA and 137 fewer than TmFtn. There is no clear single trend in the data listed in Table 6, which only emphasizes the complex nature of the structural factors governing thermostability, even within a group of closely related structures. However, a careful inspection shows some interesting features, which may have a bearing on the structural basis for the different thermostabilities of the different ferritins listed in Table 6. The most striking single feature in Table 6 is the marked increase in intramolecular salt bridges between the ferritins of mesophilic (EcFtnA and CjFtn) and hyperthermophilic (AfFtn, TmFtn and PfFtn) organisms. Indeed, a similar but less clear trend is observed for the SS hydrogen bonds overall. These structural features could explain the higher resistance to thermal inactivation in the latter. However, this trend by itself does not validate our initial assumption that PfFtn is the most thermostable ferritin in the group. A secondary trend can be perceived for the main-chain intramolecular hydrogen bonds in the hyperthermophilic organisms (AfFtn, TmFtn and PfFtn), which is greatest in PfFtn, and may offset the effect of a lower number of intramolecular salt bridges in comparison with TmFtn. Therefore, the most important factor that protects PfFtn and other ferritins from thermal denaturation appears to be the preservation of the monomer fold, rather than the 24-mer assembly.
Conclusions
In this study, we reported the crystal structure of PfFtn. High sequence similarity between PfFtn, TmFtn and AfFtn suggests a separate group of thermostable ferritins. PfFtn is the second archaeal ferritin structure to be reported, following that of AfFtn. In spite of the high sequence identity (50%) between these two proteins, their quaternary structures are significantly different. PfFtn exhibited the 432 symmetry commonly found in other known ferritins, whereas AfFtn showed the 23 symmetry similar to dodecameric DPS proteins. The 23 symmetry of the AfFtn 24-mer resulted in four triangular pores with an approximate size of 45 Å, making large openings in the AfFtn shell. Therefore, the 23 symmetry of the AfFtn 24-mer remains an exception among ferritins, and further research is required to explain such an unusual conformation.
Three iron sites were observed in the FC of PfFtn monomer. Site A was present in the as-crystallized protein and sites B and C were observed only upon crystal soaking with either Fe or Zn ions. Combining these results with our measurements of the FC EPR signal prior to crystallization, where samples were incubated in air for 1 day and 2 months, and for 2 months in 2 M ammonium sulfate (the crystallization solution), we conclude that the lack of iron atoms at the B site is very likely due to their complexation by ammonium sulfate, and that the absence of these sites in the as-crystallized protein cannot be taken per se as evidence of a transient nature of the FC, in PfFtn and other ferritins.
The possible iron entry and exit routes of PfFtn, the threefold and fourfold channels, are similar to those from bacterial ferritins in the sense that the threefold channel is less hydrophilic and the fourfold channel is more polar than the channels of mammalian H and L chain ferritins.
The hyperthermostability of PfFtn was analyzed at the structural level by comparing the number of salt bridges and hydrogen bonds in PfFtn with those in AfFtn, TmFtn, CjFtn and EcFtnA. Although there are no data on the thermal stability of other thermostable ferritins, our results suggest that TmFtn, the ferritin from T. maritima, a bacterium which grows optimally at 80 °C, is probably more thermostable than PfFtn. Owing to a high number of intramolecular hydrogen bonds between main-chain atoms, and between main-chain and side-chain atoms, the main factor contributing to the high thermostability of PfFtn as well as the other structurally similar ferritins from hyperthermophilic microorganisms appears to be the preservation of the monomer fold rather than the 24-mer structure.
|
[
"ferritin",
"hyperthermophile",
"archaeon",
"ferroxidase center",
"iron"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Bioprocess_Biosyst_Eng-2-2-1705536
|
Production of a recombinant polyester-cleaving hydrolase from Thermobifida fusca in Escherichia coli
|
The hydrolase (Thermobifida fusca hydrolase; TfH) from T. fusca was produced in Escherichia coli as fusion protein using the OmpA leader sequence and a His6 tag. Productivity could be raised more than 100-fold. Both batch and fed-batch cultivations yield comparable cell specific productivities whereas volumetric productivities differ largely. In the fed-batch cultivations final rTfH concentrations of 0.5 g L−1 could be achieved. In batch cultivations the generated rTfH is translocated to the periplasm wherefrom it is completely released into the extracellular medium. In fed-batch runs most of the produced rTfH remains as soluble protein in the cytoplasm and only a fraction of about 35% is translocated to the periplasm. Migration of periplasmic proteins in the medium is obviously coupled with growth rate and this final transport step possibly plays an important role in product localization and efficacy of the Sec translocation process.
Introduction
Since more than 15 years, the interest in biodegradable plastics (polymeric materials which can be cleaved by enzymes and assimilated by microorganisms, respectively) is steadily raising and nowadays a number of commercial products are on the market. Polymers containing ester groups, especially polyesters, play a central role as they are easily subject to microbial attack, the degradation rate being largely dependent on their physical structure and chemical composition. First attempts to produce biodegradable plastics were based on natural polymers like poly(β-hydroxybutyrate) (PHB). However, their material properties proved to be insufficient and could not compete with those of synthetic polyesters. Tokiwa and Suzuki [1] already showed that also synthetic aliphatic polyesters are accessible to a biological degradation. However, aliphatic polyesters exhibit similar disadvantages concerning their material properties (e.g. minor tensile strength, low melting points) and therefore are hardly applicable for every day life products. On the other hand, polyesters such as poly(ethylene-terephthalate) (PET) and poly(butylene-terephthalate) (PBT) exhibit excellent material properties, but were classified as non-biodegradable. Therefore, aliphatic–aromatic copolyesters, which combine both acceptable use properties and sufficient biodegradability [2] were developed. The material properties desired for specific applications can be easily adjusted by the ratio of the aliphatic to the aromatic diacid [3, 4]. Commercial products are available now under the trade name Ecoflex® (BASF AG, Germany).
Recently, a thermophilic actinomycete was isolated and identified as Thermobifida fusca (former name: Thermomonospora fusca) exhibiting a remarkable degradation capability for aliphatic–aromatic copolyesters [5, 6]. The extracellular hydrolase (TfH) responsible for the degradation of these copolyesters was purified from the culture broth and classified as a serine hydrolase with the highly conserved G-H-S-M-G motif [7]. The mature hydrolase consists of 261 amino acids, corresponding to a molecular weight of 28.2 kDa. Maximum activity was found at 65–70°C. The extracellular hydrolase gene was located on an operon containing two nearly identical genes (bta1 and bta2) with an amino acid identity of 92% [6].
The enzyme is only expressed by T. fusca if a polyester, such as poly(1,4-butandiole adipate-co-1,4-butandiole terephthalate) (BTA) or poly (ε-caprolactone) (PCL) is present in the medium. Gouda et al. [8] observed, that the amount of enzyme excreted in the medium correlates with the amount of polyester added to the culture. This induction behaviour as well as adsorption and inhibition phenomena [8] complicate a large-scale production of this enzyme (TfH) and therefore, a simpler to handle recombinant expression system for producing TfH, independent of the presence of the polymer was highly desirable, particularly, as recent investigations showed that TfH is capable of degrading even PET [9].
In this paper, we report on the heterologous expression of the polyester-degrading hydrolase from T. fusca in E. coli TG1. To this end, the TfH-encoding gene (bta1) was cloned together with a C-terminal His6-affinity tag and the ompA-secretion signal under the control of a thermo-inducible promoter. This strategy should enable translocation of the gene product by the Sec-transport system from the cytoplasm into the periplasmic space, where the properly folded TfH should accumulate after cleavage of the OmpA sequence. Several cultivation strategies and purification routes were considered. The aim of the investigations was to provide sufficient amounts of TfH for larger-scale polyester degradation studies (low purification grade) and a highly purified TfH for mechanistic studies and structural analysis.
Materials and methods
Cloning of the recombinant His6-tagged TfH
Escherichia coli TG1 (DSM 6056) was used for the production of the recombinant hydrolase from Thermobifida fusca (rTfH). TG1 is characterized by the following genetic markers: supE hsdΔ5 thiΔ(lac-proAB) F´[traD36 proAB+ lacIq lacZΔM15] [10].
The vector pCytexP1-OmpA-bta1 [6] was digested with NdeI and EcoRI to obtain a DNA fragment consisting of 876 bp. This DNA fragment was used as a template to clone a His6 tag at the C-terminal end using the primer set bta1-rev (5′-TAC TTA GGA TCC TTA TCA GTG GTG GTG GTG GTG GTG CTC GAG GAA CGG GCA GGT GGA GCG-3′) and bta1-for (5′-GTC ATG GGC CAC TCC ATG GGC-3′). The DNA fragment was amplified using VENT DNA polymerase (New England Biolabs, USA), the cycle programme was run at 95°C for 2 min, 30 cycles of 30 s at 95°C, 30 s at 50°C and 2 min at 55°C, followed by an extension reaction for 5 min at 55°C. The reaction mixture was designed according to the description of the manufacturer. The resulting PCR-fragment rbta1–His6 was purified using the JETQUICK PCR Fragment purification kit (GENOMED GmbH, Germany). rbta1–His6 was ligated to a DNA fragment obtained from the digestion of the origin vector with NcoI and BamHI. The resulting plasmid pCYTEXP1-OmpA-bta1-His6 with the secretion signal of OmpA, the hydrolase from T. fusca (TfH) and the affinity tag was used for the expression of rTfH.
Transformation of pCYTEXP1-OmpA-bta1-His6 into E. coli TG1 was done by electroporation. Expression of the recombinant TfH (rTfH) was controlled by the strong tandem promoter λpRpL, which is expressed by the thermolabile repressor cI857.
Media
Louria-Bertani medium (LB medium): 10 g L-1 tryptone, 10 g L−1 NaCl 5 g L−1, yeast extract and 0.8 g L−1 glucose was used for cultivations of E. coli TG1 in the complex medium [11]. The medium was autoclaved at 121°C for 20 min. Before cultivation, ampicillin (100 μg mL−1) was added.
A defined medium as described by Korz et al. [12] was used for batch and fed-batch cultivation of E. coli TG1. The medium contained (per litre): (NH4)2PO4 4 g, KH2PO4 13.3 g, citric acid 1.7 g, glucose 25 g, MgSO4 1.2 g, trace elements (Fe(III)citrate·H2O 100 mg, CoCl2·6H2O 2.5 mg, MnCl2·4H2O 15 mg, CuCl2·2H2O 1.5 mg, H3BO3 3 mg, Na2MoO4·2H2O 2.5 mg, Zn(CH3COO)2·2H2O 13 mg, EDTA 8.4 mg) and 100 μL antifoam (Sigma Antifoam 204, Sigma-Aldrich, USA). MgSO4 and glucose solution were autoclaved separately from the phosphate buffer and trace elements at 121°C for 20 min. pH was adjusted to 6.8 with 5 M NaOH, before. Ampicillin 100 mg L−1 and thiamine hydrochloride 4.5 mg L−1 were sterilized by filtration (Sterivex, cut off of 0.2 μm, Millipore, USA) and added before starting the cultivation.
The feed solution for fed-batch cultivations contained the following components (in 1 L demineralized water): glucose 300 g, MgSO4 7.6 g, trace elements (Fe(III)citrate·H2O 16.2 mg, CoCl2·6H2O 1.5 mg, MnCl2·4H2O 8.9 mg, CuCl2·2H2O 0.9 mg, H3BO3 1.9 mg, Na2MoO4·2H2O 1.5 mg, Zn(CH3COO)2·2H2O 6.1 mg, EDTA 4.9 mg), thiamine hydrochloride 4.5 mg and ampicillin 100 mg. The feed solution was sterilized by filtration.
Cultivation
A bioreactor (Biostat B2 , B. Braun, Melsungen, Germany) with 2 L working volume connected to an exhaust gas analysis unit (S710, Sick Maihak, Germany) was used for the cultivations. Bioreactor control was done by a computer, operating the MFCS software (B. Braun). Feed and base reservoirs were stored on scales connected to the computer. Starting conditions of the reactor were set to 30°C, 500 rpm, pH 6.8 and 1 L min−1 airflow. pH was controlled by adding either 25% NH3 (in case of the complex medium: 5 M NaOH) or 1 M H3PO4. The minimum level of dissolved oxygen (pO2) was set to 20% and was adjusted automatically by the stirrer speed and aeration rate. If the stirrer speed and aeration rate run at maximum, pure O2 had to be enriched to the inlet airflow. Antifoam was added manually to the cultivation under sterile conditions, if necessary.
Pre-cultures were carried out in the respective medium of the main culture by inoculation with a glycerol stock of E. coli TG1 containing the plasmid pCYTEXP1-OmpA-bta1-His6 at 30°C and 180 rpm for 12 h. Batch and fed-batch cultivation in the bioreactor was started by adding 20 mL of preculture.
In batch cultivation, induction of rTfH production was started at an OD of 0.5–0.6 by raising the cultivation temperature from 30 to 39°C or 42°C, respectively, within about 15 min.
In the fed-batch cultivation, feed was started after the initial glucose was consumed as indicated by a sharp increase in pO2. An exponential feed profile was used. For a desired specific growth rate μset (h−1), the volumetric feed rate Fglc (L h−1) was calculated from (Eq. 1) [13]: where (XV)0 is the biomass (g) at feed start, tf the time (h) after feed start, YX/S the biomass yield coefficient (0.4 g g−1) and S the glucose concentration in the feed (g L−1). Substrate consumption by maintenance is not considered in Eq. 1. Therefore, the actual growth rate may be less than the set value (μset = 0.12 h−1). Feed rate and volume variations were corrected automatically by the MFCS software. The expression of rTfH was initiated at a biomass level of about 40 g L−1 by shifting the temperature to 39°C, which was found optimal in the preliminary runs. After induction, the feed-rate profiles were kept constant at the value of the induction time or it was continously increased as given by Eq. 1 for μset = 0.12 h−1.
Analytical methods
Cell-mass concentration was followed by OD at 600 nm (Ultrospec 3100 pro, Amersham Biosciences, Sweden). For the determination of the dry cell weight (DCW), 1 mL of the culture was added in triplicate to pre-weighed tubes and centrifuged for 5 min at 13,000g and 4°C. After a subsequent washing step, the pellet was dried until constant mass at 45°C under vacuum.
Extracellular glucose and metabolites such as acetate, formate, succinate, propionate and isobutyrate were determined in the supernatant. To this end, the supernatant was heated up to 80°C for 5 min and centrifuged (13,000g, 10 min, 4°C) to eliminate the precipitated proteins. The supernatant was additionally filtered (pore size 0.2 μm) before applied into the HPLC (Shimadzu, Japan) using an Aminex HPX-87H column (Biorad, USA). 5 mM H2SO4 with a flow rate of 0.6 mL min−1 was used as mobile phase at a temperature of 60°C.
Electrophoresis
SDS-PAGE was performed with a Mini Protean® 3 apparatus (Bio-Rad, USA) according to the manufacturer’s recommendations. The samples were run on 12% polyacrylamide gel. Molecular mass marker was purchased from Bio-Rad. Proteins were stained with Comassie-Blue [11] or silver [14].
Proteins were transferred onto a 0.45-μm PVDF membrane (Immobilon™-P Transfer Membrane, USA) using a Semi-Dry Transfer Unit (Semiphor™ Hoefer TE® 70 Series, Amersham Bioscience, Sweden). rTfH was detected by Comassie-Blue staining.
Hydrolase activity
Hydrolase activity was measured by monitoring the cleavage of p-nitrophenylpalmitate (pNPP) as described by Schmidt-Dannert et al. [15]. The released p-nitrophenol was determined spectrophotometrically (Ultrospec 3100 pro, Amersham Biosciences, Sweden). pNPP solution was freshly prepared by mixing 1 mL pNPP dissolved in 2-propanol at a concentration of 3 g L−1 with 9 mL 20 mM phosphate buffer (pH 7.1, sodium taurocholate 2.3 g L−1, gum arabic 1.1 g L−1). Immediately after addition of the enzyme sample the absorption was measured at 400 nm and 25°C against a sample without enzyme for 2 min. One unit was defined as the amount of enzyme that caused the release of 1 μmol p-nitrophenol per minute under test conditions. The extinction coefficient of p-nitrophenol is 9.62 cm2 μmol−1. To compare the photometrically determined pNPP activity units with a previously used activity test [8] the cleavage of the polyester bonds of PCL (poly(ε-caprolactone)) films in a pHstat titration system was additionally measured as described by Gouda et al. [8]. It was found that one pNPP enzyme unit corresponds to three PCL units.
Localization of rTfH
In this paper, it is discriminated among the various rTfH fractions located in the extracellular medium, the periplasmic space and the cytoplasm as soluble or insoluble protein. The extracellular fraction (EF) was obtained from the culture supernatant after centrifugation (13,000g, 5 min, 4°C). A variation of the protocol of Neu et al. [16] was used to obtain the periplasmic fraction (PF). Biomass (about 37 mg) corresponding to ten OD was resuspended in 1 mL ice-cold 30 mM Tris–HCl (pH 8.0) containing 20% sucrose and 1 mM EDTA and after incubation for 10 min at room temperature centrifuged at 13,000g and 4°C for 10 min. The supernatant was separated from the cell pellet and stored at an ice bath. The cell pellet was again resuspended in 1 mL ice-cold 5 mM MgSO4 solution and was slightly rotated on ice for 10 min. After centrifugation the supernatant was pooled with the previously obtained supernatant giving the total periplasmic proteins (PF). Each fraction (EF and PF) was stored at −20°C until further use.
Soluble proteins harboured in the cytoplasm were obtained by resuspending the cells in 275 μL PBS buffer [11] [pH 7.2, 10% (v/v) BugBuster, 0.1% (v/v) benzonase (both Novagen, USA), 275 U lysozyme], slightly rotated at room temperature for 20 min and centrifuged at 13,000g and 4°C for 25 min. The supernatant formed the soluble cytoplasmic fraction (sCF). To gain the insoluble cytoplasmic protein fraction (iCF) cells were again resuspended in PBS buffer (10% (v/v) BugBuster, 27.5 U lysozyme) and after incubation for 5 min diluted sixfold with PBS buffer (1% (v/v) BugBuster). After centrifugation at 5,000g and 4°C for 15 min cells were washed twice with 800 μL PBS (1% (v/v) BugBuster). Final centrifugation was done at 13,000g and 4°C for 25 min. The pellet was then resuspended in 275 μL of 1% (w/v) SDS.
Purification of rTfH
To gain a rTfH solution of low purification grade 75 mg biomass was resuspended in 2 mL ice-cold PBS buffer, pH 7.2 [11]. Cells were disrupted by ultrasonification (Sonopuls HD 2070, Sonotrode UW 2070, Bandelin electronics, Germany) in an ice bath using three intervals each of 30 s pulse and 2 min break. Subsequently, disrupted cell were denaturated by heat at 55°C for 10 min, and centrifuged (13,000g, 25 min, 4°C). The supernatant was stored at −20°C.
Highly purified rTfH was recovered from the extracellular and the PF (see Localization of rTfH section). Chelating Sepharose™ Fast Flow Gel (Amersham Bioscience) was activated as described by the manufacturer and added to the solution harbouring rTfH (1 mL chromatographic material was used for 400 U). The suspension was stirred for 1 h and filled then into a XK 26/60-column (Amersham-Bioscience) and after the chromatographic material has settled, the column was connected to the FPLC system (LCC-500 Plus, Amersham Bioscience, Sweden). Column was washed with 3 column volumes (CV) binding buffer containing 20 mM Tris (pH 7.5), 500 mM NaCl and 10 mM imidazole at a flow rate of 4 mL min−1. Elution of the bound rTfH was done by a linear increase of the imidazole concentration to 500 mM within 3 CV. Fractions with activity were pooled and concentrated by ultrafiltration (Amicon Stirred Cell 1050 or 1010, Amicon, USA) using a regenerated cellulose membrane (YM Ultrafiltration Membranes, cut off: 10 kDa, Amicon, USA). 6±0.5 mL of the concentrated enzyme was applied manually to the gel filtration column (HiLoad 26/60 Superdex 75 prep grade, Amersham Bioscience) equilibrated with 0.9% NaCl. The flow rate was set to 2 mL min−1 to elute the proteins. Fractions with activity were pooled and stored at −20°C until further use. Whenever necessary, rTfH solution was concentrated by ultrafiltration.
Protein determination
Protein was determined by Roti®-Nanoquant (Roth, Germany) as described by the manufacturer.
Results
Batch cultivations
E. coli TG1 harbouring the plasmid pCYTEXP1-OmpA-bta1-His6 was firstly cultivated batchwise with complex medium in 1 L scale monitoring OD, DCW, glucose and metabolites in the supernatant. At an OD of about 0.6, cultivation temperature was raised to either 39 or 42°C to induce rTfH production. After starting induction, generation of rTfH was pursued by the spectrophotometric assay in the extracellular and PFs. Results are shown in Fig. 1.
Fig. 1Expression of rTfH in E. coli TG1 during batch cultivation with LB medium. Cells grew at 30°C to an OD of 0.6 before starting induction by a temperature shift to 39 or 42°C, a growth and CO2 generation, b rTfH activity in different locations, extracellular fraction (EF; open square), periplasmic fraction (PF; open circle), sum of all fractions (continuous line), c glucose and acetate in broth
Glucose present in the complex medium is quickly consumed at about 1.4–2 h after induction (Fig. 1c). Consumption of glucose is accompanied by formation of acetate, its extracellular concentration peaks at 39°C at 2 h (3 h at 42°C, respectively), which is remetabolized after total glucose consumption. The consumption pattern is reflected by biomass development and CO2 generation (Fig. 1a) which show a two- and three-phase behaviour, respectively. Firstly growth rate is highest as long as glucose is available (induction at 39°C). The following consumption of the intermediate acetate leads to reduced growth rate. After acetate consumption the growth rate further declines and finally growth ceases, however, CO2 evolution is yet observed in the stationary phase. From the biomass and CO2 evolution, one has to conclude that in all cultivation phases considerable amounts of constituents of the complex fractions of the medium are also consumed. For instance, the 1.5 g L−1 biomass (2 h after induction at 39°C) cannot be generated alone by glucose consumption. Unfortunately, the complex constituents are not known. Therefore, a carbon balance is not possible.
The production of the hydrolase rTfH is shown in Fig. 1b. Both induction profiles, shifting the temperature to 39 or 42°C, revealed similar rTfH production kinetics. Shortly after induction rTfH is only found in the periplasm. However, in the course of the cultivation, rTfH did not accumulate in the PF as expected due to the leader signal OmpA. Surprisingly rTFH was entirely found in the extracellular medium at the end of the cultivation. At an induction of 42°C the promoter λpRpL in front of the bta1 gene is stronger than at 39°C, but expression of rTfH is nearly finished about 6 h after induction resulting in a lower volumetric activity (only 60%). About 4 h after induction, a remarkable increase of extracellular rTfH activity was observed which obviously correlates with a high metabolic activity as indicated by the second strong increase in CO2 evolution. The rise in rTfH activity cannot solely originate from translocation but suggest additional secretion from cytoplasm. Apparently expression of rTfH is faster than its transport through the cytoplasmic membrane in the early phase shortly after induction. In shake-flask cultivation, the rTfH production was monitored in the cytoplasm as soluble and insoluble cytoplasmic fractions (iCFs; data not shown). Maximum hydrolase activity was achieved after about 7.5 h summing up all compartments. Subsequently, the total hydrolase activity was slightly decreased, and finally rTfH was only detected in the supernatant.
To study the influence of medium composition on recombinant TfH production and to follow the transport of the enzyme into the medium, batch cultivations with the defined medium containing 23 g L−1 glucose were carried out (Fig. 2). rTfH was again induced at low OD. The course of the cultivation with the defined medium is remarkably different from that using the complex medium. Raising the temperature to either 39°C or 42°C caused firstly a low growth rate. The consumption of glucose was coupled with acetate formation; 8 h after induction at 39°C, growth rate rised to about 0.5 h−1 (0.6 h−1 at an induction at 42°C) and both glucose and acetate were consumed while a low amount of succinate was generated which was remetabolized after exhaustion of other C-substrates. As expected, growth was directly coupled with CO2 evolution. The respiration coefficient RQ was around 1, the yield coefficient (biomass on glucose) was determined as 0.4 g g−1 for the batch cultivation at 39°C. Over the entire induction period (about 23 h), carbon balances of the cultivation with the defined medium were close to 100 ± 3%.
Fig. 2Expression of rTfH in E. coli TG1 during batch cultivation with the defined medium. Cells grew at 30°C up to an OD of 0.6 before starting induction by a temperature shift to 39°C or 42°C, a growth and CO2 generation, b rTfH activity in different locations, EF (open square), PF (open circle), sum of all fractions (continuous line), c glucose and metabolites in broth
In contrast to the cultivation in LB medium, succinate was detected in the extracellular medium under both the induction profiles. Succinate is a typical product of the anaerobic metabolism of E. coli. It is produced during the mixed acid fermentation and the fumarate reduction. Both pathways are strictly regulated by Fnr the global regulator of the anaerobic metabolism [17]. With the applied experimental approach, anaerobiosis can be excluded. The formation of succinate was coupled with the CO2 evolution and hence with a high metabolic activity. This may lead to an imbalance in the citric cycle or reduction equivalents. Unfortunately, other metabolites of the citric cycle were not determined in the supernatant supporting this assumption. Similar observations were made by Korz in batch cultivations using the same defined medium. He correlated the succinate formation with the amount of CO2 raising in the exhaust gas [18].
While secretion of rTfH in the periplasm started immediately after induction, the rate of rTfH release in the cultivation medium correlates with the growth rate, i.e., growth and product transport from the cytosol via the periplasm into the extracellular space are coupled. At the end of the cultivation, almost the entire product was located outside the cells. The final results of the batch cultivations at controlled pH with the complex and the defined medium are summarized in Table 1. For induction in both media, by shifting the temperature to 39°C gives better results than the increase to 42°C. While the volumetric activities achieved for the complex and the defined media at the same induction temperature are comparable, the specific activities differ significantly due to the higher biomass obtained with the defined medium. As the induction time is considerably longer for the cultivation in the defined medium, the specific productivities differ even more. These results clearly indicate that production of the rTfH in batch cultivation becomes more attractive when the complex medium is employed.
Table 1Summary of results on rTfH production in batch cultivationsComplex mediaDefined mediaInduction temperature(°C)39423942Biomass (gDCW L−1)2.22.19.89.6Induction time (h)9.962323.7Volumetric activity (103 U L−1)8.084.437.52.84Specific activity (103 U gDCW−1)3.672.110.770.30Volumetric productivity [103 U (L h)−1]0.820.740.330.12Specific productivity [103 U (gDCW h)−1]0.3730.3520.0340.013
During the induction phase, the protein content in the culture supernatant was analysed by SDS-PAGE (Fig. 3). About 4 h after the temperature upshift, the first protein bands can be recognized, particularly, for the cultivation with LB medium, where rTfH and other protein bands appear simultaneously. In cultivations with the defined medium, the pattern looks differently. In the first phase, up to almost 4 h after induction, a single protein of about 57.4 kDa is found. By N-terminal sequencing, this band could be identified as the periplasmic located dipeptide binding protein (DBP), which is involved in the transport of dipeptides [19]. In the cultivation with complex medium, leakage of the outer membrane increased significantly 6 h after induction when C-starvation began, growth was slow but CO2 generation was still high. With the defined medium protein release in the extracellular medium was drastically rising at about 12 h after induction when all glucose was consumed. The findings shown in Fig. 3 are in accordance with the results reported by Rinas and Hoffmann [20], who studied in detail the leakage of host-cell proteins of recombinant and non-recombinant E. coli. These authors concluded that irrespective of the cell line, proteins of periplasmic origin are predominantly released in the medium, specifically, under stress conditions like temperature increase, nutrient starvation and high cell densities.
Fig. 3SDS-PAGE under denaturating conditions: EFs of different cultivations are shown after increasing the temperature to 39°C to induce rTfH production, a batch cultivation with complex medium, b batch cultivation with the defined medium, dashed line marks the beginning of glucose starvation
The SDS-PAGE shown in Fig. 3 indicates that the further protein bands observed for each media differ only in their intensity but not in their composition. This suggests that protein release in the medium cannot be attributed to cell lysis but is mainly due to an increasing leakage of proteins located and transported to the periplasm from the cytoplasm by the Sec-transport system.
This conclusion is also confirmed by analysis of the N-terminal sequence of rTfH, which clearly revealed that the OmpA leader sequence has been cleaved. However, cleavage was not uniform. Only 15% of the secreted rTfH found in the extracellular medium was cleaved at the expected cleavage site (Fig. 4). The remaining rTfH still carried parts of the OmpA leader peptide. The incomplete processing of signal sequences has been reported before [21]. The cleavage motif of a signal petidase is the A-X-A box [21] and fission occurs after the X. The N-terminal analysis of the secreted rTfH revealed that the cleavage motif is kept. These observations may allow the conclusion that the Sec-transportation pathway is overloaded, which might explain the improper cleavage of the signal peptidase.
Fig. 4Amino acid sequence of the signal peptide OmpA (underlined) and the beginning of rTfH; arrows mark the cleavage site and their relative frequency
Fed-batch cultivation
To enlarge volumetric rTfH activities, fed-batch cultivations were carried out. The basic assumption is that the volumetric productivity Qp [U (L h)−1] is given by equation with the specific productivity qp [U (gDCW h)−1], which is anticipated to be approximately constant when increasing the biomass concentration. As observed in the controlled batch cultivation it was also thought that the target enzyme may be released into the cultivations medium from the periplasmic space under high cell density conditions, thus facilitating down-stream processing [20]. High cell density cultivation (HCDC) for E. coli can be effectively performed by using an appropriate defined medium. After the batch phase with unlimited growth and consumption of glucose, the cultivation is continued by feeding glucose at a feed rate resulting in a reduced growth rate (μset < μmax; Eq. 1). Thus, oxygen starvation can be prevented and the culture grows under glucose limitation in the feed phase. The oxygen partial pressure was kept constant at 20% as described in the Cultivation section. Two typical HCDC runs are shown in Fig. 5.
Fig. 5Expression of rTfH in E. coli TG1 during HCDC. Cells grew at 30°C before induction by shifting the temperature to 39°C (dashed line). a After 12.8 h, the batch phase ended and exponential feed was started and set to a growth rate 0.12 h−1. At the time point of 25.5 h, feeding was set to a constant rate (12.8 gglc h−1). b Fed batch with an exponential feeding profile with μset=0.12 h−1 over the entire feed phase including induction
Starting with an initial glucose concentrations of 24 g L−1, the biomass concentration is about 12 gDCW L−1 at the end of the batch phase (Fig. 5a, b). Due to the metabolic overflow, acetate was generated in the unlimited growth phase (μmax = 0.47 h−1), which was remetabolized at the end of the batch phase (Fig. 6). Then the glucose feed was started. The feed rate was adjusted to give a growth rate of about 0.12 h−1 according to Eq. 1, by assuming a yield coefficient YX/S of 0.4 g g−1. The actually measured growth rates were 0.09 (Fig. 5a, b). In Run A, after a feed phase of 12.8 h, a biomass concentration of 38.5 gDCW L−1 was obtained and at this point, the expression of rTfH was induced by shifting the temperature from 30 to 39°C within 15 min. The glucose feed rate was not further increased (12.8 gglc h−1) but kept constant to the value when the induction was started. Therefore, the growth rate steadily decreased and after about 14 h, the induction growth ceased. In the first phase after induction (14 h), an average growth rate of roughly 0.03 h−1 can be estimated from the data. However, over the entire period of 22.4 h with the constant feed rate, CO2-generation rate was also constant. Therefore, the fed glucose is used only for maintenance in the second part of the induction phase when there was zero growth . In the fed-batch phase of Run B, the same exponentially increasing glucose-feed rate as for Run A was applied, i.e., Fglc was calculated from Eq. 1 for μset = 0.12 h−1 and YX/S = 0.4 g g−1. After starting the induction at 42 g L−1, DCW exponential glucose feeding was continued for the same μset (0.12 h−1). In spite of the exponential glucose feeding, growth decreased and ceased about 6 h after induction. An average growth of 0.07 h−1 can be estimated from the experimental biomass data at the first 6 h of the induction phase. CO2 generation increased exponentially even when the growth was negligible in the second half of the induction phase. After 11.4 h induction, a final biomass concentration of 67.8 gDCW L−1 was obtained.
Fig. 6Glucose and metabolites formed during HCDC (Run A). cf. Fig. 5
In the course of both runs, the respiratory quotient was always about 1 (± 0.02). This indicates that consumed glucose was exclusively used for ATP production. Indeed, an intermediate formation of fermentative by-products such as acetate, succinate and formate was only observed at low level at the beginning and end of the induction period (Fig. 6). Both during the fed-batch and the induction phase, the glucose concentration was always zero. Carbon balances were checked by the installed computer program with consideration of volume changes by feeding and sampling and found to be close to 100 ± 2%.
The kinetics of the appearance of the heterologous gene product (rTfH) in the periplasm and cultivation broth is shown in Fig. 7 for Run A with a long induction phase (> 20 h). The activity in the periplasm steadily increased up to a value above 70 U mL−1 while in the first 10 h, no rTfH was released in the supernatant. The final activity in the medium was rather low (12 U mL−1). A similar release kinetics as shown in Fig. 7 was also observed for Run B, which differs largely from the behaviour found for batch cultures (Figs. 1, 2). In contrast to the batch cultivations the periplasmic activity does not pass a maximum and, particularly, little product (say less than 15%) is found in the medium.
Fig. 7Production kinetic of rTfH in the fed-batch cultivation (Run A). At a biomass of 38.5 gDCW L−1 rTfH expression was induced by raising the temperature to 39°C
However, despite the low leakage of rTfH into the medium a SDS-PAGE (Fig. 8) made from the supernatant after the enzyme induction showed that the cells were already leaky at the beginning of the induction, which is proven by the appearance of the periplasmic DBP. During the induction phase, an increase both of DBP and rTFH is observed.
Fig. 8SDS-PAGE: EFs of different cultivations are shown after increasing the temperature to 39°C to induce rTfH production, HCDC (Run A)
The hydrolase activity (pNPP test system) was also measured in the cytoplasm before the induction and at the end of the HCDC. While no activity was found prior to induction, a surprisingly high intracellular activity was detected at the end of the cultivation process. Indeed, about two-thirds of the entire rTfH activity was located in the cytoplasm, apparently in soluble and active form. Any significant inclusion-body formation was not observed. The operational parameters and the results received in the two HCDC runs shown in Fig. 5 are summarized in Table 2. In spite of the different feed rates and the total induction times, the rTfH activities found in the different compartments are in reasonable agreement. The constant feed rate after induction in Run A yields higher specific activities obviously due to lower biomass formation. However, as a result of the long induction time, the productivity is lower for Run A than for Run B. In comparison to the batch runs (Table 1) the accumulation of product in the periplasm and cytoplasm is most noticeable. Thus the total volumetric activity is about 30-fold larger than that achieved in the batch runs. These higher overall product concentrations are advantageously for rTfH recovery and purification.
Table 2Applied operation parameters and achieved rTfH activities in fed-batch runs (starting glucose concentration 24 g L−1, pH 7, T = 30°C, Tind = 39°C)Cultivation Run ARun BDCW at induction (g L−1)38.542Feed rateConstant (12.8 gglc h−1)Exponential (μ = 0.12 h−1)Estimated mean growth rate after induction (h−1)0.030.07Total induction time (h)22.411.4Final DCW (g L−1)53.267.8SupernatantVolumetric activity (103 U L−1)12.045.16Specific activity (103 U gDCW−1)0.2270.076PeriplasmVolumetric activity (103 U L−1)74.2251.69Specific activity (103 U gDCW−1)1.3950.762Cytoplasm (soluble)Volumetric activity (103 U L−1)141.24156.74Specific activity (103 U gDCW−1)2.6552.312TotalVolumetric activity (103 U L−1)227.5213.59Specific activity (103 U gDCW−1)4.283.15Productivity [U (103 gDCW h)−1]a0.190.28aTotal induction time
Purification
Two different strategies were applied to purify rTfH with regard to different uses. To gain rTfH with a high purification grade, the His6-affinity tag cloned at the C-terminal end of the bta1 gene allowed a simple and effective chromatographic procedure. rTfH was purified from the culture broth or the periplasm, which was obtained by an osmotic shock treatment (Localization of rTfH section). In the following, the purification of rTfH is described, which originated from the periplasmic space. About 37 g of wet biomass (corresponding to ca. 6.5 gDCW) was subjected to osmotic shock treatment (1 L scale). After separating the cells by centrifugation, 25 mL Ni2+ activated sepharose was directly added to the supernatant harbouring the periplasmic protein. After 1 h incubation, only low amounts of rTfH activity (0.1%) could be detected in the supernatant. The suspension was then filled into a glass column. Particle settling could be improved by connecting a peristaltic pump to the column outlet. After connecting the column to a FPLC system, rTfH was washed with bindung buffer; rTfH was eluted by applying a linear gradient of imidazole (Fig. 9a). rTfH elution was observed at an imidazole concentration of about 200 mM. In the pooled fractions, a rTfH activity of 6.53 U mL−1 could be detected. The His6-affinity step resulted in a 29-fold purification and a loss of about 30% rTfH activity.
Fig. 9Purification of rTfH from the PF. E. coli TG1 cells derived from a fed-batch cultivation (Run A); 37 g biomass was used for osmotic shock treatment, a Ni2+-affinity chromatography with Chelating Sepharose Fast Flow, b gelfiltration with Superdex 75 prep grade; pooled fraction are marked by a line
To further eliminate impurities and to remove high concentrations of NaCl and imidazole from the protein sample, gelfiltration was done using a Hi Load 26/60 Superdex 75 prep grade column with a 0.9% NaCl buffer (Fig. 9b). Before applying the sample to the column, pooled fractions with rTfH from the Ni2+-affinity chromatography were concentrated by ultrafiltration (concentration factor 7; cut-off 10 kDa). Gelfiltration assured purification to homogeneity as shown by SDS-PAGE (Fig. 10) although only a minor increase of the specific activity was observed (Table 3). Gel filtration resulted in a further loss of rTfH activity of about 10%. Totally, rTfH could be purified 30-fold with a 57% recovery in activity by the applied purification route. Table 3 gives an overview of the procedure to obtain the highly purified rTfH, which will be used for mechanistic studies (adsorption, kinetic) and structural analysis.
Fig. 10SDS-PAGE: rTfH at different purification steps (silver staining), 37 g biomass was used for the osmotic shock treatment, cells derived from HCDC (Run A) standard: molecular mass protein standard, 1 supernatant from the periplasm, 2 Ni2+-affinity chromatography, 3 pooled fraction after Superdex 75 gel chromatographyTable 3Summary of two routes used to recover and purify rTfH from HCDC; the highly purified product was obtained from the periplasmic fraction (Run A), the low grade product was recovered from the cytoplasm and periplasm by ultrasonification (Run B)Purification stepTotal units (103 U L−1)Total protein (mg L−1)Specific activity (U mg−1)Purification-factorYield (%)High grade rTfH Supernatant with periplasmic protein8.57587151100 Ni2+-affinity chromatography6.53154362971 Gel filtration4.89114453057Low grade rTfH Crude extract26.522,57010.321100 Heat treatment22.511,47015.31.585 Ultrasonification + heat treatment28.891,41020.52109aaCrude extract (ultrasonification plus centrifugation)
Contrary to the purification profile of the TfH from the culture broth of T. fusca cultivations, the overall yield could be increased from 14 to 57% [7]. Prior to performing the first purification step, the supernatant containing TfH had to be concentrated by ammonium sulphate precipitation. By this method, a final recovery of only 37% could be achieved because of the limited amounts of the precipitated proteins that could be redissolved [7]. By inserting the His6-affinity tag at the C-terminal end of rTfH, precipitation could be avoided and the recombinant enzyme could be easily obtained from E. coli-culture broths independently of its composition.
For industrial processes, purification with Ni2+-affinity chromatography is too costly of course. Therefore, an alternative purification route was developed, which aimed at a rTfH preparation of a low purification grade but sufficient activity and stability for larger-scale investigations on polyester degradation. Activity test of the soluble protein in the cytoplasm obtained during HCDC showed that most of the expressed rTfH was not secreted into the periplasm but remained in the cytoplasm. Hence, the cells were disrupted by ultrasonification to get rTfH both from the cytoplasm and the periplasm. In preliminary experiments, different ultrasonication conditions were tested. Three-fold pulses of 30 s each with 2 min break were found as the best treatment for the test conditions of 75 mg biomass resuspended in 2 mL PBS buffer. Here, biomass from the HCDC Run B was used. After the separation of cell debris by centrifugation, an activity of 26.52 U mL−1 was detected corresponding to a specific activity of 4.1 U gDCW−1. Chemical disruption of the cells by a detergent formulation (BugBuster, see Localization of rTfH section) resulted only in a specific activity of 3.07 U gDCW−1 implying that cell disruption is more effective than the chemical treatment. Protein concentration of the crude extract was 2.57 mg mL−1. Taking the specific activity of the high-grade rTfH into account (445 U mgProtein−1), 1 mg protein harboured 2.3% heterologous produced rTfH. Because rTfH is very stable against higher temperatures [7], heat treatment was carried out at 55°C for 10 min. Two routes were tested, namely either heat treatment of the disrupted cells immediately after the ultrasonication or heat treatment after the separation of the cell debris. Activity tests revealed that a heat treatment directly after the cell disruption yielded the best results. Thus the purification factor could be increased from 1.5 to 2 and a higher yield was observed. Table 3 summarized the results in detail.
Discussion
The recombinant strain E. coli TG1 harbouring the plasmid pCYTEXP1-OmpA-bta1-His6 proved to be an appropriate system for expression, secretion and purification of the hydrolase from T. fusca. In comparison to cultivations with T. fusca and induction of TfH by adding polyesters a 1.5–2-fold higher volumetric activity could be obtained in batch cultivations of the recombinant E. coli strain. The volumetric activity could be further improved in fed-batch cultivations with glucose limitation. Considering only the extracellular and PFs, the increase is 20-fold. If one also accounts for the activity found in the cytoplasm, the total volumetric activity is over 50-fold higher than that obtained from T. fusca. The overall volumetric productivity [U (L h)−1] increased by a factor more than 100. Hence, the production of rTfH in recombinant E. coli is an attractive and promising approach for providing hydrolytic activity for polyester cleavage in degradation and recycling processes. This view is supported by the simple recovery route to give a crude and stable rTfH preparation, which did not exhibit activity losses after storage at −20°C for several weeks.
A further improvement of rTfH production in E. coli is the insertion of the His6 tag at the C-terminal end of the enzyme. Purification of the wild-type TfH proved to be quite ineffective. A three-step purification strategy (precipitation, ion-exchange chromatography and hydrophobic interaction chromatography) resulted in only 14% yield in activity [7]. With the His6 tag, 57% activity yield of rTfH could be obtained in a two-step purification route starting from the supernatant and PFs. A 30-fold purification was achieved yielding a specific activity of 445 U mg−1. From this value, the overall concentration of the produced rTfH can be estimated as about 0.5 g L−1.
In the batch runs with LB and the defined medium, better results were found by induction to only 39°C. Therefore, fed-batch cultivations were carried out with this induction temperature. While both the media yielded comparable volumetric activities, the specific activities differ significantly (about fivefold in LB). LB medium obviously supplies E. coli with all nutrients to easily synthesize the recombinant protein at high specific productivity and to secrete it across the cell envelope. Increased growth rates and facilitated protein production in LB medium are frequently reported in the literature [22, 23]. In batch cultivations with both media, rTfH expression as well as its transport across the cell membrane in the periplasm and the subsequent release into the cultivation media are seemingly coupled with growth and CO2 generation (Figs. 1, 2).
In the fed-batch cultivations with the defined medium, roughly tenfold higher activities could be achieved in the supernatant and periplasm than in batch operation. Surprisingly, the specific activities in the glucose limited fed-batch process are even twofold higher. If one accounts for the soluble cytosolic fraction the activity yields further improve remarkably. The total specific activity of the HCDCs is in the same range as those obtained for the batch process with LB medium. An influence of the His6 tag and the full or truncated OmpA leader sequence on TfH activity can hardly be estimated from the results of this study. However, the high activities of rTfH obtained suggest that small sequences at the C- and N- terminals might have little effect on TfH activity. The fraction of soluble active protein located in the cytoplasm is remarkably high but its amount can be estimated to be less than 2% of the total protein content of the cells (assuming 50% protein of DCW). Therefore, it is understood that inclusion body formation could not be detected.
Among the different mechanisms to transport protein from the cytoplasm across the cell membrane to the periplasm, the Sec pathway is the most prominent for prokaryotic organisms [24]. The Sec-translocation process is thought to be generally responsible for the transport of newly synthesized but unfolded proteins out of the cytosol to the periplasm. This process is accompanied by the cleavage of the leader sequence (OmpA) from the pre-protein and proper folding in the periplasm. If we assume that the Sec pathway is also active for the transport of the heterologous rTfH out of the cytoplasm, it is obvious that due to its accumulation in the cytoplasm during the fed-batch runs, the translocation rate is less than the expression rate. Hence, transport into the periplasm apparently presents a limiting step. It might be possible that the entire translocation process is subjected to an equilibrium phenomenon in as far as high protein concentrations in the periplasm reduces the transport rate. This is evident if the rTfH masses (or units) in periplasm and cytoplasm are not referred to cultivation volume as given in Table 2 but to their respective actual volumes. If it is assumed that the volume of periplasm is about 5% of the cell volume then the rTfH concentration in the periplasm is more than tenfold higher than in the cytoplasm. This might explain that no cytoplasmic rTfH activity was found in the batch runs, where the periplasmic protein was entirely released in the extracellular space. In contrast, relatively little protein was transported across the outer cell wall into the medium in the fed-batch runs with high cell densities and higher rTfH activities were found in both the cytoplasm and periplasm.
As to the Sec pathway, the rTfH in the cytoplasm should be unfolded and hence is not expected to be in the active form. However, when gaining the cytoplasmic fraction (see Localization of rTfH section) immediate activity was observed. This suggests that either folding takes place spontaneously in the PBS buffer or the cytosol already harbours the active folded rTfH and its translocation follows another mechanism. Unfortunately, our study does not allow to distinguish between the two possibilities.
Though it is often assumed that E. coli laboratory strains do not secrete proteins across the cell envelope in the extracellular medium at normal growth conditions [25], leakage of the outer cell membrane is a common phenomenon observed for a variety of cultivation conditions as reported in recent papers [20, 25, 26] and review articles [21, 27]. Permeability of the outer cell wall may depend on medium composition, growth rates, cell concentration, changes in cultivation conditions, etc. For instance, Mg2+ and Ca2+ increase leakage of the outer membrane [28]. Rise of temperature causes changes in membrane permeability and fluidity due to alterations in the lipid composition of the cell wall. This aspect was particularly studied in continuous cultures of E. coli [26] and an increased permeability was found at higher growth rates. Rinas and Hoffmann [20] studied in detail the leakage of host-cell proteins during HCDCs of recombinant and non-recombinant E. coli. These authors reported that the proteins found extracellularly are of periplasmic origin except the heat-shock protein DnaK. Various and interrelated stress factors in the temperature-induced HCDC are suggested to be responsible for the migration of periplasmic proteins in the extracellular medium.
In all cultivations performed in this study, rTfH could be detected in the medium without applying any additional measures to permeabilize the outer cell membrane. Cell lysis as a possible mechanism of rTfH release can be excluded as the cleavage of the OmpA sequence in the extracellularly found hydrolase indicates transport across the cytoplasmic membrane by the Sec-translocation mechanism. Any significant cell lysis would also release cytoplasmic rTfH still containing the complete signal sequence.
In batch cultivations (Figs. 1, 2) with LB and the defined medium, induction of rTfH expression was started at low cell densities. Immediately after the temperature upshift, rTfH appeared in the periplasm but its periplasmic activity passes a maximal value just in the first part of the exponential growth phase. In accordance with the findings of Shokri et al. [26], the migration of the periplasmic rTfH in the medium is obviously growth coupled. This effect may be attributed to changes in cell-wall composition which facilitate transport into the extracellular medium. On the other hand, during cell growth and cell-wall biosynthesis, transient openings of the outer membrane and the murein layer are unavoidable [29], thus the protein of periplasmic origin can migrate into the extracellular medium. Furthermore, vesiculation and fragmentation of the outer membrane may occur as a natural phenomenon during the exponential growth phase [30, 31] which is thought to be caused by an imbalance of the outer-membrane biogenesis and peptidoglucan synthesis. Therefore, several growth-coupled phenomena may contribute to the migration of periplasmic proteins in the medium.
At the end of the batch cultivations, the produced rTfH has entirely accumulated in the medium. As can be discerned from the SDS-PAGE shown in Fig. 3 the transport of rTfH across the cell envelope is accompanied simultaneously by release of other periplasmic proteins [20]. The kinetics of rTfH accumulation in the extracellular medium (Figs. 1, 2) agrees qualitatively with the results shown in Fig. 3, particularly, no significant migration is observed at low growth rates after the induction when the culture adapts to the higher temperature. Rinas and Hoffmann [20] found DBP as the most prominent periplasmic protein transported to the medium. This observation and even the release kinetics reported by these authors for high cell-density conditions are in striking agreement with our findings for batch conditions (relatively low cell densities) as observed in the defined medium (Fig. 3b).
In contrast to the batch runs, rTfH accumulates in the cytoplasm and periplasm during the glucose-limited fed-batch cultivations after induction at higher cell densities. The relative fractions of rTfH released extracellularly is low. For an induction period of 11 h, only 5 and 10% of the PF are released for Run A and B, respectively. As in Run B, the growth rate was higher as a result of exponential glucose feeding; this finding underpins the view that the permeability of the outer membrane is related to growth rate. As shown in Fig. 8 the outer membrane is already leaky before the temperature shift from 30 to 39°C as among other periplasmic proteins the DBP is predominantly found in the extracellular medium as reported previously [20]. Of course, in the induction phase a further increase of DBP is observed (Fig. 8). Though the transport of rTfH in the periplasm as well as its migration into the extracellular medium continued (Fig. 7), the cultivation was terminated to prevent cell lysis after growth ceased. However, in the non-growth phase, glucose does not accumulate in the medium but is consumed as is also exhibited by the monotonous increase of CO2 generation. This points to the increased maintenance and ATP demand for translocation of the heterologous gene product (rTfH) across the inner cell membrane.
These mechanisms, which govern the local distribution of rTfH are not yet understood. Though cell-specific productivities are in a comparable range, product distributions differ largely. Our results suggest that the outer membrane permeability is related with the growth rate and hence all rTfH can migrate into the extracellular medium at high growth rates (batch runs). Therefore, final product release across the cell envelope could play an important role in the overall translocation process as product accumulation in the periplasm seemingly reduces the translocation rate from the cytoplasm by the Sec pathway. This hypothesis requires further investigations, which are presently in preparation. Therefrom, new strategies for cultivation and downstream processing with improved product qualities are expected.
|
[
"purification",
"batch culture",
"sec pathway",
"recombinant protein expression",
"fed-batch culture"
] |
[
"P",
"P",
"P",
"R",
"R"
] |
Psychopharmacologia-3-1-2082659
|
Effects of dopaminergic modulation on electrophysiological brain response to affective stimuli
|
Introduction Several theoretical accounts of the role of dopamine suggest that dopamine has an influence on the processing of affective stimuli. There is some indirect evidence for this from studies showing an association between the treatment with dopaminergic agents and self-reported affect.
Introduction
Dopaminergic neurotransmission has been suggested to be involved in several functions of the brain that are associated with reward, such as the experience of rewarding feelings (Wise and Bozarth 1985), reward-related reversal learning (Mitul et al. 2001), motivational aspects of reward (Robinson and Berridge 1993), and reward processing (Day and Carelli 2007). In addition, dopamine neurons mediate reward signals in appetitive learning of visual stimuli (Schultz 1997). Recently, Robbins and Everitt (2007) stated that “the role of enhanced DA activity is to increase responsivity to cues paired with reinforcement”. In addition to these functions, there are suggestions that dopamine is involved in the regulation of affect (Saeedi et al. 2006). Most studies addressing the role of dopamine in affect-regulation have been carried out among animals. Few studies address the role of dopamine in affect among humans. Studies among Parkinson’s disease patients, whose mesencephalic dopamine activity is decreased, show that dopamine is able to modulate emotional responding (Wieser et al. 2006). In addition, it has been demonstrated that dopamine is also involved in the perceptual processing of fearful stimuli in Parkinson disease patients (Tessitore et al. 2002). No studies are known which address whether dopamine is able to modulate the processing of affective stimuli in normal subjects. However, there are some indirect suggestions pointing towards a role of dopamine in affective processing. Takahashi et al. (2005) showed among healthy volunteers watching unpleasant pictures that the dopamine D2 receptor antagonist sultopride decreased responses in the limbic areas (amygdala, hippocampus, anterior cingulate), which are associated with affective processing. In contrast, they found that sultopride produced greater activations in the cortical areas (frontal, temporal, and parietal), and no effect of sultopride on self-reported affect was found, suggesting a complex role of dopamine in affective processing.
With regard to psychopathology, what several theoretical accounts of psychopathological behavior have in common is that they state that dopamine may have a function in signaling relevant or salient information (Franken 2003; Kapur 2003; Robinson and Berridge 1993) and suggest that dopamine is involved in the processing of information having motivational relevance. In concordance with this notion, Franken et al. (2004) showed that the attention-capturing properties of heroin-related stimuli in heroin-dependent patients were decreased as result of treatment with the dopamine antagonist haloperidol. Indirect evidence that dopamine is involved in the processing of affective information comes from a study of Taylor et al. (2005) among schizophrenics. They found that structures with extensive dopaminergic innervations, such as the ventral striatum, show blunted responses to emotional salience in schizophrenics.
In addition, several theories suggest that dopamine is involved in the experience of self-reported positive affect (Ashby et al. 1999; Burgdorf and Panksepp 2006). In concordance with these notions, several studies examining the effects of dopamine-antagonists, in particular D2 antagonists such as haloperidol, report a blunting of the self-reported affective response (e.g., Mizrahi et al. 2007). In contrast, dopamine agonists result in self-reported enhanced mood (Saeedi et al. 2006; Willner et al. 2005).
In the present study, we use event-related potentials (ERPs) to investigate the processing of affective stimuli, in particular reward-associated stimuli, in humans. ERPs measured during the perception of pictures are known to vary with their affective contents (Cacioppo et al. 1994; Cuthbert et al. 2000; Keil et al. 2001; Schupp et al. 2000). Affective pictures, both pleasant and unpleasant, elicit larger late positive potentials (LPP; starting 300–400 ms after picture onset until picture offset) than neutral pictures. LPPs capture controlled processes and are generated by a diffuse network including anterior areas such as the orbitofrontal cortex (Rule et al. 2002; Vuilleumier 2005) and higher order visual areas such as parietal and occipito-temporal regions (Keil et al. 2002). In addition to these late effects, the viewing of affective pictures results in reduced ERP amplitudes in the early time window (150–300) as compared to neutral stimuli (Codispoti et al. 2007; Schupp et al. 2006). This early ERP effect is sometimes referred to as the early posterior negativity (EPN; Schupp et al. 2003b). The EPN, which is associated with the perceptual encoding of emotional material, may reflect the call for resources in the capacity-limited second stage of processing (Codispoti et al. 2007; Schupp et al. 2006) and is generated in the visual occipito-temporal regions (Schupp et al. 2006).
In passive viewing contexts such as in the present study, the EPN and LPP modulation by affective stimuli are similar in respect to the fact that they are both pronounced for high arousing pleasant and unpleasant stimuli (Schupp et al. 2006). However, it seems that the LPP, in contrast to the EPN, is enlarged when allocating top-down attentional resources to emotion, suggesting that the EPN is relatively insensitive to voluntarily attentional processes. Previous studies show that similar designs that study the electrophysiological activation to affective pictures are able to measure the modulation of affective processing by psychopharmacological agents (Franken et al. 2007; Kemp et al. 2004).
This study will investigate whether a D2 agonist or a D2 antagonist can modulate the electrophysiological response to affective pictures. Decrease in dopaminergic activity will be induced by a single oral dose of 2 mg haloperidol. This dose does not produce serious side effects (Franken et al. 2004; Kahkonen et al. 2001), and 2 mg was shown to be able to induce modulation of the processing of neutral and rewarding information (Franken et al. 2004; Kahkonen et al. 2001). Increase in dopamine will be induced by a single oral dose of 2.5 mg bromocriptine. Generally spoken, there are several indications that bromocriptine can modulate the response towards rewarding stimuli (Kirsch et al. 2006).
As dopamine is associated with an enhanced signalling of reward-related stimuli, it was expected that bromocriptine enhanced the processing of pleasant stimuli and haloperidol decreased the processing of pleasant stimuli. Based on the literature, the hypotheses for unpleasant stimuli could be less clearly formulated. Because some studies show that dopamine only signals stimuli associated with reward while others do find that dopamine also signals punishment-related stimuli, we expected that bromocriptine decreased the processing of unpleasant stimuli and haloperidol enhanced the processing of unpleasant stimuli. Neutral stimuli served as control conditions.
Materials and methods
Subjects and procedure
A randomized, double-blind, placebo-controlled crossover design was employed. Participants were 32 healthy right-handed male undergraduate psychology students and were recruited by advertisement at the university campus. All participants were non-smokers. Eleven participants could not be included in the analysis because of excessive electroencephalographic (EEG) artifacts in one of the three sessions (n = 3), drop out because of side effects (nausea) in the bromocriptine treatment (n = 6), or drop out because of side effects (low blood pressure, n = 1 and nausea, n = 1) in the placebo treatment. In total, 21 participants completed all three the measurement sessions, and all analyses were conducted using this sample. The mean age of the subjects was 19.8 years (SD = 1.2, range = 18–22).
Two weeks before the experiments, participants who were willing to participate were informed about the procedure. One week before the start of the experiments, subjects underwent a clinical psychiatric and physical examination by an experienced psychiatrist (L.P.). Eligible participants were instructed to take the medication at 8:00 a.m., 10:00 a.m., or 12:00 a.m. 4 h before each session in the laboratory. Further, they were instructed to abstain from coffee, grapefruit, alcohol, and drugs at the days of the experiment. In addition, they were instructed to abstain from alcohol and drugs the day before the experiment. The time between each session was always 1 week. All participants received a financial compensation of 50 euro. This study was approved by the Medical Ethics Committee of the Erasmus Medical Centre, Rotterdam.
Treatment
All subjects received a single oral dose of placebo (lactose), bromocriptine (2.5 mg), and haloperidol (2 mg) in a counterbalanced order. The medication was provided by the pharmacy of the Erasmus Medical Centre in indistinguishable capsules. It has been demonstrated that a 2-mg dose of haloperidol does not produce serious side effects in healthy subjects (Franken et al. 2004; Kahkonen et al. 2001). Further, 2.5 mg of bromocriptine has been shown to modulate cognitive processes and in absence of side effects. Based on the literature, it was not expected that the 2.5 mg of bromocriptine would result in serious side effects (Jarvik et al. 2000). However, after several sessions, we noticed nausea occurring in some sessions. Because we excluded all participants with nausea (with reference to the possible interaction between nausea and the viewing of affective pictures), this resulted in relatively high dropout rates. After inclusion of 15 participants, all participants in all sessions were provided with domperidone (10 mg) to prevent nausea. Participants were instructed to take the domperidone together with the research medication1.
Stimuli and task
Before the first session, participants were instructed, and they signed informed consent. Subsequently, participants filled out the Positive and Negative Affect Scales (PANAS) questionnaire and two single-item self-rating scales that were used to assess the subjective effects of the medication (see “Questionnaires” section). After this, the participants were seated on a comfortable chair in a light and sound-attenuated room. Stimuli were presented on a 21″-monitor 1.5 m away from the eyes. After attaching the electrodes (approximately 15 min), they were instructed to pay close attention to the pictures that would be presented. To be sure that they paid attention to the pictures, they were told that questions about the pictures would be asked after the experiment.
All pictures were taken from the International Affective Picture System (IAPS; Lang et al. 1999). To reduce test–retest recognition effects, three pictures sets which were closely matched on valence and arousal ratings were used in a counterbalanced order (see Table 1). A one-way analysis of variance (ANOVA) confirmed that there were no differences in arousal and valence ratings between the three sets. Further, a repeated measures ANOVA confirmed that there were significant valence rating differences between the three affective categories, F(2,118) = 1324.2; p < 0.001. As expected, valence ratings for pleasant stimuli were higher than for neutral stimuli, and for neutral stimuli higher than unpleasant stimuli (higher scores indexes more pleasantness). In addition, analysis of the arousal ratings showed also significant differences between the three affective categories, F(2,118) = 304.5; p < 0.001. Both pleasant and unpleasant pictures had higher arousal ratings than neutral pictures. In addition, pleasant stimuli had slightly higher arousal ratings as compared to the unpleasant stimuli. The neutral pictures consisted mainly of buildings, household objects, and neutral persons. The pleasant pictures consisted mainly of erotic and romantic scenes, while the unpleasant pictures consisted mainly of mutilated bodies and threat scenes.
Table 1Mean (SD) valence and arousal ratings of the three sets of IAPS pictures according to the normative ratings of the IAPS database (Lang et al. 1999) Set 1Set 2Set 3Differences between sets (F values, ns)Neutral valence4.9 (0.3)4.9 (0.4)4.9 (0.2)0.01Arousal2.7 (0.4)2.8 (0.5)2.7 (0.6)0.16Pleasant valence7.2 (0.6)7.2 (0.6)7.2 (0.6)0.06Arousal6.3 (0.9)6.5 (0.7)6.2 (1.1)0.58Unpleasant valence2.5 (0.6)2.6 (0.5)2.6 (0.7)0.05Arousal5.8 (1.0)5.6 (0.9)5.7 (0.9)0.11ns Not significant
Questionnaires
The PANAS (Watson et al. 1988) were administered as a measure of positive and negative affect. The PANAS is a 20-item bidimensional mood inventory. Positive affect reflects the extent to which a person feels enthusiastic, active, and alert (Watson et al. 1988). Negative affect is a general dimension of subjective distress and unpleasurable engagement that subsumes a variety of aversive mood states including anger, contempt, disgust, guilt, fear, and nervousness (Watson et al. 1988). Psychometric properties of the PANAS scales are good (Boon and Peeters 1999; Watson et al. 1988).
Two single-item self-rating scales were used to assess the subjective effects of the medication. A five-point Likert scale (1 = no effect, 5 = a very strong effect) was used to assess the magnitude of the effect. A visual analogue scale (0 = very unpleasant, 100 = very pleasant) was used to assess the pleasantness of the effect.
Electroencephalographic recording
The electrophysiological signals were recorded through the Active-Two amplifier system (Biosemi, Amsterdam, The Netherlands) from 64 scalp electrodes (10–20 system) using Ag/AgCl electrodes mounted in an elastic cap. Furthermore, six additional electrodes were attached to left and right mastoids, two outer canthi of both eyes (HEOG), infraorbital, and supraorbital regions of the eye (VEOG). Two additional scalp electrodes were used as reference and ground electrodes, respectively. Online signals were recorded from DC to 134 Hz. All signals were digitized with a sample rate of 512 Hz and 24-bit A/D conversion. Data were off-line re-referenced to an average reference. Off-line, EEG and EOG activity was filtered with a bandpass of 0.05–30 Hz (phase shift-free Butterworth filters; 24 dB/octave slope). After ocular correction (Gratton et al. 1983), epochs including an EEG signal exceeding ±75 μV were not included in the average. The mean 200-ms pre-stimulus period served as baseline. After baseline correction, average ERP waves were calculated for artifact-free trials at each scalp site across the three stimulus conditions and across the three treatment conditions. Brainvision Analyzer (Brain Products, München, Germany) software was used for all EEG analyses.
Data analysis
Definitions of the peaks of interest were based on previous studies using the same IAPS paradigm (Franken et al. 2007; Schupp et al. 2000, 2003a). Inspection of voltage maps of the difference waves (emotional–neutral) revealed that the EPN was maximal present at the following posterior sites: P5/6, P7/8, P9/10, PO3/4, PO7/8, and O1/2 (see Fig. 1). This site is comparable to that used by previous studies (Franken et al. 2007; Schupp et al. 2003b). To reduce the number of covarying variables and to reduce the possible influence of single spurious electrodes, these individual sites were collapsed by averaging the electrodes into one region per hemisphere (Dien and Santuzzi 2005; Schupp et al. 2003b).2 The EPN in this area was defined as the mean value of ERP activities in 150- to 300-ms time windows. Note that the EPN can have positive values: The EPN is a negative going potential showing reduced positivity for emotional stimuli.
Fig. 1Scalp topography of the EPN component (mean value in the 150- to 300-ms time window). Voltages represent difference scores: Pleasant–neutral (left panel) and unpleasant–neutral (right panel). Note that the anterior activity is positive and posterior activity is negative
Inspection of voltage maps of the difference waves (emotional–neutral) revealed that the LPP was maximal at central and centro-parietal sites C5/6, C3/4, C1/2, CP5/6, CP3/4, CP1/2, P5/6, P3/4, P1/2 (see Fig. 2). Again, these individual sites were collapsed by averaging these electrodes into one region per hemisphere. The LPP was defined as mean value of ERP response within 400–700 ms. The LPP is a positive deflection in the EEG, and emotional stimuli increase the amplitude of this component. In addition, because in previous studies valence effects on the LPP have shown hemispheric differences (e.g., Keil et al. 2001), we exploratively included hemisphere as factor in all ANOVAs.
Fig. 2Scalp topography of the LPP component (mean value in the 400- to 700-ms time window). Voltages represent difference scores: Pleasant–neutral (left panel) and unpleasant–neutral (right panel). Note that the anterior and posterior activities are negative and that the central activity is positive
Statistical analysis
To test the influence of dopaminergic manipulation on the EPN and LPP components, a 3 (medication: placebo, bromocriptine, haloperidol) × 3 (stimulus category: neutral, pleasant, unpleasant) × 2 (hemisphere: left, right) repeated measures ANOVA was employed with all variables as within-subjects factors. For the manipulation check, a three-way repeated measures ANOVA was used for each self-reported measure (pleasantness and magnitude of medication effect, positive and negative affect). In all instances, Greenhouse–Geisser adjusted p values were used. In case of a significant ANOVA result, partial Eta squared (η2) as measure of effect-size is reported. Significant ANOVA effects were followed up by pairwise comparisons with Bonferroni adjustment.
Results
Self-reported effects
There was no significant effect of treatment on the magnitude of the subjective effects of the medication, F(2,40) = 0.8; p = 0.45. In addition, no differences in experienced pleasantness of the medication between the conditions was observed F(2,40) = 0.0; p = 0.96. Further, no effects of the medication condition on negative, F(2,40) = 0.8; p = 0.45, nor positive, F(2,40) = 0.4; p = 0.66, affect was observed. Summarized, no self-reported effects of medication on magnitude and pleasantness of effects and affect were observed.
ERP data
EPN component
Table 2 displays all mean values of the EPN and LPP components, including standard deviations. Figs. 3, 4, 5, and 6 show the difference ERP waves (emotional minus neutral) for affective stimuli at the left and right hemisphere sites. As expected, a significant main effect for stimulus category was observed, F(2, 40) = 81.0, p < 0.001, η2 = 0.80. Pairwise follow-up analysis show that both pleasant (p < 0.001) and unpleasant stimuli (p < 0.001) yielded smaller ERP values as compared to neutral stimuli, suggesting an enhanced early perceptual encoding of affective stimuli. However, no significant medication × stimulus category interaction effect was observed F(4, 80) = 0.17, ns, suggesting that dopaminergic medication had no influence on this early stage of picture processing. In addition, no other significant main or interaction effects were observed.
Table 2Mean (SD) values of the EPN and LPP components in left and right hemisphere towards neutral, pleasant, and unpleasant pictures for the placebo, bromocriptine, and haloperidol treatment conditions (n = 21) PlaceboBromocriptineHaloperidolEPN neutral left9.4 (3.6)8.4 (3.7)9.4 (4.4)Right8.6 (3.1)8.5 (4.2)8.9 (3.6)EPN pleasant left7.0 (4.3)6.0 (4.0)6.9 (4.2)Right6.7 (3.8)6.3 (4.5)6.8 (3.8)EPN unpleasant left8.6 (4.0)8.0 (3.8)8.6 (4.4)Right8.5 (4.0)8.2 (4.4)8.6 (4.3)LPP neutral left3.5 (1.6)3.2 (1.9)3.4 (1.8)Right3.8 (1.4)3.1 (1.7)3.3 (1.7)LPP pleasant left4.9 (1.9)4.6 (1.8)4.8 (2.1)Right4.6 (2.0)4.2 (1.8)4.2 (1.7)LPP unpleasant left5.4 (1.7)4.7 (1.7)5.5 (2.0)Right4.8 (1.9)4.2 (1.8)4.5 (1.9)Fig. 3Differences waves (pleasant–neutral) for pleasant stimuli at the left and right posterior cluster (P5/6, P7/8, P9/10, PO3/4, PO7/8, and O1/2) for placebo (black line), bromocriptine (gray line), and haloperidol (dotted line) treatment conditions separatelyFig. 4Differences waves (unpleasant–neutral) for unpleasant stimuli at the left and right posterior cluster (P5/6, P7/8, P9/10, PO3/4, PO7/8, and O1/2) for placebo (black line), bromocriptine (gray line), and haloperidol (dotted line) treatment conditions separatelyFig. 5Differences waves (pleasant–neutral) for pleasant stimuli at the left and right central/centro-parietal clusters (C5/6, C3/4, C1/2, CP5/6, CP3/4, CP1/2, P5/6, P3/4, P1/2) for placebo (black line), bromocriptine (gray line), and haloperidol (dotted line) treatment conditions separatelyFig. 6Differences waves (unpleasant–neutral) for unpleasant stimuli at the left and right central clusters (C5/6, C3/4, C1/2, CP5/6, CP3/4, CP1/2, P5/6, P3/4, P1/2) for placebo (black line), bromocriptine (gray line), and haloperidol (dotted line) treatment conditions separately
LPP component
Again, a significant main effect for stimulus category was observed, F(2, 40) = 48.1, p < 0.001, η2 = 0.71. Pairwise follow-up analysis show that both pleasant (p < 0.001) and unpleasant stimuli (p < 0.001) yielded a larger LPP component as compared to neutral stimuli, suggesting an enhanced processing for affective stimuli in the late elaborative phase. In addition, a small but significant main effect of medication was found, F(2, 40) = 5.2, p < 0.05, η2 = 0.21. Follow-up analysis indicated that bromocriptine resulted in overall reduced LPP waves as compared to the placebo treatment (p < 0.05), suggesting that bromocriptine results in an overall reduced processing of pictorial stimuli in this stage. Further, a small but significant stimulus-category × hemisphere interaction effect was observed, F(2, 40) = 3.7, p < 0.05, η2 = 0.16. Because none of the post hoc tests resulted in a significant pairwise comparison, and this interaction had not our major interest, we will not discuss this finding in further detail. Most important for our hypothesis, no significant stimulus category × medication interaction effect was found, suggesting that the employed dopamine antagonist and agonist had no selective influence on affective processing.
Discussion
To our knowledge, this is the first study examining the influence of dopamine on ERP correlates of affective stimulus processing. The present study demonstrates that an acute low dose of bromocriptine resulted in globally reduced LPP waves as compared to the placebo treatment, suggesting that bromocriptine results in decreased cortical processing of pictorial information. Although there are no previous studies using an acute low dose of bromocriptine examining the processing of emotional pictures (as in the present study), these results are in line with a study of Oranje et al. (2004). Although in that study a different paradigm was employed and different components were studied, these authors found a globally reduced ERP response as result of bromocriptine treatment. In contrast to bromocriptine, an acute low dose haloperidol did not result in a non-specific modification of cortical processing of pictorial information. Previous studies addressing information processing using haloperidol and ERP measures yielded mixed results. For example, Kahkonen et al. (2001) did observe a global reduction of ERP waves using 2 mg (oral) of haloperidol. In contrast, other studies did not observe a general effect of haloperidol on ERP amplitudes (e.g. Ford et al. 1994). Clearly, the effects of haloperidol on ERPs are dependent on dose, ERP component, characteristics of the participants, and experimental paradigms that are used.
Most importantly, we could not confirm the hypothesis that D2 receptor modulation results in a modulation of cortical processing of affective stimuli. Neither an acute low dose of bromocriptine nor an acute low dose of haloperidol did differentially modulate the brain’s electrophysiological response towards affective stimuli. This finding suggests that dopamine D2 receptors are not involved in the cortical processing of affective information. Because the studied ERP components measured in the present study are associated with attentive motivational processing (Schupp et al. 2006, 2007), it can be concluded that motivational attentive processing is not affected by haloperidol nor bromocriptine. Given that the ERPs in the present study reflect attentive processing, the results are in line with Oranje et al. (2006) who did not find any effects of l-dopa nor bromocriptine on ERP indices of selective attention in healthy volunteers.
In addition to the main finding that modulation of the D2 receptor did not result in early (EPN) or late (LPP) effect on ERPs indexing early perceptual encoding of emotional material and subsequent elaborative processing, we did not observe an effect of haloperidol and bromocriptine on the subjective self-reported affect. Neither medication had an effect on positive or negative affect. However, it must be noted that the present study was not specifically designed to measure the subjective effects of the medication. In addition, although the present results show that dopamine is not associated with the perception of reward-associated stimuli per se, it still might be associated with establishing stimulus–reward associations or functions dealing with reward-related motor responses. In addition, it might be that the reward processing is only modulated by dopaminergic agents in persons with a genetically heightened reward sensitivity, for example, carriers of the A1 allele of the dopamine D2 TaqIA gene (Kirsch et al. 2006).
The present study suggests that low doses of bromocriptine in Parkinson’s disease patients and haloperidol in schizophrenia patients may have limited influence on affective processing and self-reported affect in patients having these disorders. However, it must be noted that we did only study the effects of an acute dose of these medications; it is not known whether these effects differ from those associated with long term use of medication.
The present study has several limitations. First, there was a substantial dropout in the bromocriptine treatment. It may be that this is a specific subgroup of persons with heightened sensitivity for dopamine agonists. Because this group is not included in the analysis, this might have resulted in lowered generalizability of the results. Second, because of risks of unwanted side effects, we employed rather low doses of bromocriptine and haloperidol, 2 mg and 2.5 mg, respectively. Third, a possible limitation of this study is that some participants received domperidone and others did not, which may have influenced the results. However, it must be noted that domperidone does not cross the blood–brain barrier but exerts its effect peripherally, and as such, does not have a major influence on the central dopamine system (Barone 1999; Champion et al. 1986). In addition, additional statistical analyses show that domperidone did not influence any of the outcome variables (electrophysiological nor self-reported) used in the present study.
Summarizing, the present results indicate that neither an acute low dose of bromocriptine nor an acute low dose of haloperidol has selective effect on electrophysiological indices of affective processing and self-reported affect. In contrast, bromocriptine decreased overall processing of all stimulus categories regardless of their affective content. The results indicate that dopaminergic D2 receptors do not seem to play a crucial role in the selective processing of affective visual stimuli.
|
[
"dopamine",
"affective processing",
"haloperidol",
"erp",
"bromocriptine"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
J_Biomol_NMR-3-1-1892527
|
All-atom molecular dynamics simulations using orientational constraints from anisotropic NMR samples
|
Orientational constraints obtained from solid state NMR experiments on anisotropic samples are used here in molecular dynamics (MD) simulations for determining the structure and dynamics of several different membrane-bound molecules. The new MD technique is based on the inclusion of orientation dependent pseudo-forces in the COSMOS-NMR force field. These forces drive molecular rotations and re-orientations in the simulation, such that the motional time-averages of the tensorial NMR properties approach the experimentally measured parameters. The orientational-constraint-driven MD simulations are universally applicable to all NMR interaction tensors, such as chemical shifts, dipolar couplings and quadrupolar interactions. The strategy does not depend on the initial choice of coordinates, and is in principle suitable for any flexible molecule. To test the method on three systems of increasing complexity, we used as constraints some deuterium quadrupolar couplings from the literature on pyrene, cholesterol and an antimicrobial peptide embedded in oriented lipid bilayers. The MD simulations were able to reproduce the NMR parameters within experimental error. The alignment of the three membrane-bound molecules and some aspects of their conformation were thus derived from the NMR data, in good agreement with previous analyses. Furthermore, the new approach yielded for the first time the distribution of segmental orientations with respect to the membrane and the order parameter tensors of all three systems.
Introduction
Using anisotropic media, such as partially oriented bicelles or macroscopically oriented membranes, NMR investigations can reveal a wealth of information about molecular properties, namely conformation, orientation and dynamics. In many solid state NMR studies of membrane-active peptides and transmembrane proteins, the samples are conveniently prepared with macroscopically oriented bilayers to obtain structural information. The NMR data analysis relies on a uniform alignment of all molecules with respect to the static magnetic field, as it makes use of the orientation dependence of the chemical shift, quadrupolar coupling or dipolar coupling interactions. In contrast to single crystal studies, where the molecules are immobilized in a unique conformation, in the case of lipid membranes and liquid crystalline systems one has to consider a wide distribution of molecular orientations and anisotropic motions. Here, we apply for the first time a new MD strategy to deduce such structural and dynamics information on three representative compounds in biomembranes with increasing complexity: (i) pyrene, (ii) cholesterol, and (iii) the antimicrobial peptide PGLa.
A convenient nucleus for solid state NMR investigations of oriented membranes is deuterium for several reasons (for reviews see e.g., Vold 1994; Davis 1983; Ulrich and Grage 1998). First, the influence of 2H substitutions on the molecular structure and dynamics is negligible. More importantly, in most biologically relevant cases the quadrupolar interaction tensor of a carbon-bound deuterium is to a good approximation aligned parallel to the 2H–C bond and directly reflects the local molecular orientation and dynamics of the labelled segment. Using 2H-NMR, the alignment and dynamic behaviour of lipids, small organic guest molecules and membrane-active peptides has been studied in numerous examples. Most previous strategies to extract the molecular orientation from NMR data, however, have been restricted to molecules with a rigid conformation, e.g. peptides assuming an ideal α-helix with fixed backbone torsion angles. In these studies, several individual labels are usually placed into strategic positions on the rigid molecular part. The measured quadrupolar splittings are then compared in a least squares RMSD analysis with the predicted values upon systematically varying the molecular alignment. The best-fit molecular orientation is determined from a 3D error plot as the one yielding minimal differences between the observed and calculated NMR parameters. However, this RMSD (root mean square deviation) analysis can only give preliminary results for flexible molecules such as peptides or proteins with a high degree of internal mobility.
A proper way to account for motional averaging of the NMR parameters caused by molecular mobility is to run all-atom MD simulations. When applied to biomembranes, however, such simulations pose two closely connected problems: (i) large size of the system and (ii) long simulation times. In particular, water surrounding the lipid bilayer has to be included into the MD simulations, and time-spans have to be covered approaching the NMR time scale. Moreover, the volume of interest has to be surrounded by multiple copies of the central cell to avoid boundary effects. Nevertheless, such simulations have been demonstrated to be feasible for moderately sized membrane segments hosting medium size molecules (see e.g. Tieleman et al. 2001). In the present contribution, we propose an alternative strategy, in which the oriented medium is not explicitly considered, but instead the ordering membrane environment is replaced in the MD simulations by pseudo-forces derived from the measured NMR parameters. This way sufficiently long MD runs become possible and allow to calculate the motionally averaged parameters observed in the NMR experiments. As a result, the full information about molecular orientation, order, segmental motions and even aspects of the molecular conformation could be revealed.
There is one remarkable advantage in the calculations involving orientational constraints. Since an experimentally observed NMR value usually represents a time average over the motions of the molecules, knowledge of the geometry and strength of these molecular motions is a prerequisite for properly interpreting the NMR data in terms of structural parameters. The order parameter, which is typically used to describe the effect of molecular motion on the NMR spectra, however, is often not known a priori. In the present approach, the molecular motions are intrinsically included and the order parameters are obtained as part of the simulation result.
Theory
NMR interaction tensors and coordinate transformations
The orientation dependence of the observed frequencies or splittings of an NMR property can be described by a 2nd rank tensor P, whose components are specified in the context of a coordinate system. On the one hand, a description in a molecule-fixed frame of reference is required to follow changes in orientation, but on the other hand the NMR observables are extracted from the interaction tensor in the laboratory frame related to the static magnetic field. We will define the molecule-fixed frames in the following way from the coordinates of the three bonded atoms A–B–C (see Fig. 1):
Fig. 1Local coordinate system used to define a tensorial NMR property on atom A. The bond from A to atom B is taken as the local z-direction, the plane spanned by A, B and C defines the x-direction in-plane, and the y-direction is perpendicular to that plane
Atom A is selected as the position of the nucleus of interest (for instance 2H), whose property PA is to be calculated. In case there are several possible ways of selecting atoms B and C, they will be defined such that the valence of bond B–C is a maximum (the valence can be estimated from the bond length, see O’Keefe and Brese 1991). This selection ensures that the y-axis will be perpendicular to a π-system if it exists, and the x-axis will be positioned within the π-plane.
Once we have selected a coordinate system that is fixed to the molecule, we can calculate the columns of the transformation matrix D from the unit vectors along of this frame, given in coordinates of the laboratory frame.
Matrix D can be used to transform the tensorial properties from the molecular to the laboratory frame. The tensor components of P in the molecular frame are thus transformed to yield P′ in the laboratory frame according to
where D is the transformation matrix as given in Eq. 2 (the Einstein sum convention is used for Greek subscripts and α, β = {x, y, z}, the second index indicating the unit vector).
For most deuterium quadrupolar couplings, as well as vicinal dipolar couplings like 15N–1H and 13C–1H the tensor PA will be to a good approximation diagonal in this coordinate system, hence we can easily assign the principal values of PA to the three axes. In cases where the chemical shift tensor is to be analysed, the same molecular coordinate system can be used as in those calculations based on the bond polarization theory (BPT, see Sternberg 1988). Therefore, all these types of NMR interactions can be treated within the same formalism.
In static NMR experiments the tensor components are observed in a coordinate system of reference whose z-axis is aligned with B0. In this frame the measured frequency or splitting is given by the zz-component of the tensor P. However, in the approach presented here we can incorporate further tensor components in the MD simulations, which are accessible for example by measuring the sample at further tilt angles.
Constraints and calculation of molecular properties
To include experimental constraints in the MD simulations, pseudo-energy terms are added to the molecular energy provided by the force field. These pseudo-energies are defined as functions of the difference between the experimental and a calculated tensor property, Pi:
where k is a force constant which is chosen to adjust the size and unit of the energy. The first sum runs over all observed tensor components and does not need to include all possible elements. In a 1D NMR measurement of a single label in an oriented or monocrystalline sample, only the zz-value of the tensor of this site can be experimentally determined. However, any additional tensor component introduced into the calculations will improve the characterization of the orientation of the system. Because of the nature of the tensor transform (Eq. 3) each principal axis of the tensor can be inverted without changing the observed NMR frequency. Therefore, eight discrete tensor orientations are equivalent, even if all elements of the observed tensor are used as constraints. If, however, only the three principal values of the tensor are known, then in addition to this eightfold degeneracy a continuous set of orientations (instead of discrete possibilities) will be compatible with the experimental data from a single label. Therefore, a larger number of orientational constraints from several labels are necessary to determine the orientation of a rigid molecular segment in an unambiguous way. Obviously, if a molecule is intrinsically flexible, an even higher number of constraints will be required.
The experimental NMR values represent an ensemble average over a large number of molecules as well as a time average. The relevant time scale for averaging is given by the inverse of the frequency range that results from the anisotropy of the relevant interaction. When motions are faster than this time scale, a time average of the corresponding NMR parameters is observed. Motions slower than this time scale will result in a distribution of the respective NMR property. In all-atom MD simulations of larger molecules we can only perform calculations for a single molecular system and for time spans on the order of a Larmor period. Assuming that the NMR interactions are already completely averaged within this time, averaged properties can be simulated. This is likely to be the case for motions within the molecule and for the rotational diffusion of small molecules in oriented media such as the lipid bilayer considered here, which are on the time scale of 30 ps to 100 ns (see Brüschweiler 1994). We thus can assume that the time averages over the shorter period accessible for MD simulations do reflect the longer time averages observed by the NMR measurement. For the average of Ptheo we obtain using Eq. 3:
If the NMR property does not change with time, we have to perform the time averaging only over the products of the transformation matrix elements . For most deuterium quadrupolar couplings and vicinal dipolar couplings, any conformational influence on the NMR interaction can be neglected. In the case of chemical shift tensors, however, the change of the tensors in the course of the MD simulation has to be considered.
Since MD simulations are started with random velocities, we have to run many time steps before a thermal equilibrium is reached, and before the kinetic energy is redistributed properly between the different degrees of freedom of our system. During this equilibration phase of the MD the pseudo-energies will be large and unrealistic, since the proper mean values for the interaction tensors are not yet reached. Therefore the time averaging is performed here using an exponential memory function (Torda and van Gunsteren 1991):
The memory time constant is denoted with τ, and N is the norm of the integral. This memory function serves to scale down the contributions from the beginning of the MD calculation exponentially. By proper selection of τ it can be ensured that we average mainly over the late, relevant time span of interest. Since in conventional MD simulations the equations of motion are integrated in finite time steps Δt, the integral in Eq. 6 is converted into a discrete sum S. During the MD simulation, the sum Sn+1 of the time step n + 1 is calculated from the sum of the previous time step Sn in a recursive manner:
In the course of the MD simulation, S will contain the exponentially weighted time average of the property P, and in the case of a tensor the time averaging will be performed with all of its components.
Calculation of pseudo-forces
To include the NMR constraints into the equations of motion, pseudo-forces have to be calculated from the respective pseudo-energies (Eq. 4). They are obtained as the derivatives of the energies with respect to the coordinates of the atoms. In the case of orientational pseudo-forces, we have to derive the transformation matrices D with respect to the coordinates of the atoms that were used in their definition (see Eq. 1). The pseudo-forces in the Cartesian directions x, y and z (denoted with the Greek index γ) acting on the atoms A, B, and C (denoted with j, see Eq. 1) are then given by
The calculation of the orientational pseudo-forces is thus reduced to determining the derivatives of the elements of the transformation matrices D with respect to the Cartesian coordinates of atoms A, B, and C.
In the following expressions we will use a short notation for the derivatives, omitting all indices:
Using the definitions of Eqs. 1 and 9, these derivatives can be calculated in the following way:
The derivatives of the unit vectors can be used to construct the derivatives of the elements of the transformation matrices .
Because of the time dependence of the transformation matrices, the derivatives were calculated continuously during the MD simulation at each time step. The time average is only calculated for the NMR property.
As evident from Eq. 8, the pseudo-forces depend linearly on the difference between theory and experiment. To avoid too high pseudo-forces in the beginning of the MD simulation, when the deviation of the simulated and experimental values is still large, the contributions of the pseudo-forces are scaled down. Another possibility to avoid too large pseudo-forces is achieved by multiplying the pseudo-forces by weighting factors , which lead to nearly constant forces if exceeds a threshold ΔP (see Witter et al. 2001):
For this factor behaves similar to the derivative of the original energy expression (4). It is possible to choose different force constants k (Eq. 8) and width parameters ΔP for each type of NMR property of interest.
Order parameter calculation
Besides determining the structure of a molecule, it is equally important to describe its dynamic behaviour in an oriented medium such as a lipid bilayer. MD simulations provide a unique opportunity to obtain information on the molecular dynamics, which is difficult to determine experimentally. The traditional way to account for motional averages in NMR investigation is to introduce order parameters. To this aim, the time average of the transformation matrices D (Eq. 5) can be used to characterize the influence of restricted molecular motions on the anisotropic properties of a molecule. In a similar way, Saupe (1964) defined the so called order tensor to describe the average orientation of a molecule:
where 〈 〉 denotes the mean value and cosΘi are the z components of the unit vectors of a molecule-fixed frame within the laboratory system of reference. To describe the dynamic properties of flexible molecules, however, the Saupe approach is not sufficient, since molecular groups like peptide side chains may exhibit a larger mobility than the rest of the molecule. For this purpose Limmer (1989) introduced a segmental or site-specific order tensor W, which is the tensorial product of the Saupe order tensor and a conformational order tensor.
By calculating the NMR properties in the course of the MD simulation it is possible to obtain the segmental order tensor WA of the vicinity of atom A. The principal axis system of the tensor property is used to represent the instantaneous segmental orientation, which is then related to the average tensor P determined by Eq. 5. The transformation matrices DPAS, which link the average P with the principal axis system of P at each time step of the simulation, can then be used to determine W (〈 〉t denotes the time average):
This average is calculated in every step of the MD simulation for all sites A for which the tensorial NMR properties are calculated. In general the complete order tensor has 81 elements, which reduce to nine components if we are interested only in the zz component of a property (Eq. 13). Since the order tensor W (and S as well) is symmetric and traceless, we have to consider only five independent elements. The transformation matrices DPAS in Eq. 13 can be calculated from the eigenvectors corresponding to the principal values of PA. After a full-length MD simulation, the order tensors WA will contain the information describing the time averaged orientation of all sites A.
To calculate the molecular Saupe order tensor, a coordinate system has to be fixed to the molecule. In a perfectly rigid molecule any set of three atoms could be selected at will. However, as most molecules are intrinsically flexible, it does not make sense to define the coordinate system that way, as the Saupe order tensor would then depend on this selection. We therefore decided to use the principal axis system of the tensor of inertia to represent the orientation of the entire molecule (see Burnell and De Lange 1980). This way we can calculate a molecular order tensor that is independent of any particular choice of coordinates:
The axes of inertia are denoted with a, b and c, corresponding to the moments of inertia in increasing order. For rod-like molecules the long axis will thus be the a-axis of least inertia.
In many cases it is convenient to diagonalize the order tensor to introduce a principal frame of order (Low 2002). Within this frame there are only two independent parameters: the order parameter S and a biaxality parameter ξ:
By ordering the principal components as , one gets . If the biaxiality ξ is zero, the orientational distribution of the molecule is defined by a single axis, which is called director. In the biaxial case with a non-zero ξ, there exists a second distinct direction orthogonal to the director, and the molecular motions are characterized by two axes. A perfectly axially symmetric molecule that is oriented parallel to the membrane normal will lead to the following order tensor:
When the molecular Saupe tensor is used to interpret the NMR observables, the conformational tensors CA have to be known as well. These tensors mediate between the orientation and order of a site A and the molecular coordinate system. If the transformation matrices leading from the molecular coordinate system to the individual tensor systems of the sites A are known, CA can be calculated the same way as the order tensor (see Limmer 1989). Only in the case of rigid molecules, simple expressions for CA can be derived from the molecular geometry. Since in our MD simulations the segmental order tensors WA and the molecular Saupe order tensor are calculated directly, the conformational order tensors become obsolete.
Molecular dynamics simulation
The integration of the equations of motions is based on Verlet’s algorithm (Verlet 1967), and time steps of 0.5 fs were employed to sample all high frequency hydrogen atom vibrations. In constrained MD simulations it is generally necessary to control the temperature during the simulation time. This is accomplished by coupling the molecular system to a heat bath which dissipates the heat generated by the pseudo-forces. All prevailing differences between the constraints and their calculated values are sources of heat. To obtain an NTV assemble (with conserved particle number N, temperature T, and volume V), we introduced a proper thermostating procedure (see Evans and Morriss 1990). The coupling to the thermostat is controlled by a coupling time constant η which should be much larger than the time step. This time span η allows an adjustment of the range of thermal fluctuations in the simulated molecular system.
To prevent too large pseudo-forces at the start of the MD simulation the pseudo-forces were gradually increased towards their final values during the MD simulation. To this aim we introduced time dependent scaling factors , which approach the value 1.0 in an exponential fashion. The time constant was set in most cases to 200 ps, leading to a relatively smooth course of the temperature.
When applying the NMR orientational constraints during an MD run, the resulting pseudo-forces will “heat up” the system and enhance its rotational degrees of freedoms. Because the averaging procedure depends on the molecular re-orientations caused by the NMR constraints, some net rotational motion will prevail up to the end of the simulation. In standard MD simulations any overall molecular rotations and translations are subtracted from the velocities, since these external degrees of freedom are not of interest. In the present orientationally constrained calculations, however, only the net translations of the systems are removed.
Parametrization
For NMR spectroscopic considerations it is convenient to introduce the quadrupolar coupling tensor CQ in frequency units:
where Q is the quadrupolar moment of the deuterium nucleus, and V the electric field gradient tensor at the nuclear position. The trace of CQ is zero because of , and in the case of carbon-bonded deuterons the quadrupolar coupling tensor can be considered to a good approximation as diagonal in the coordinate system aligned with the C–2H bond. If the electron distribution is axially symmetric there is only one linearly independent tensor component, otherwise there are two. In the general high temperature case we observe two quadrupolar resonances with frequencies symmetric to zero:
If the z-axis of the local electric field gradient points into the direction of the magnetic field, the corresponding quadrupolar splitting Δ ν is
In the case of a C–2H bond the deuterium quadrupolar coupling depends only weakly on the molecular surrounding, hence representative values from static solid state NMR measurements of characteristic substances can be used as quadrupolar constants (Table 1). In this selection we have to keep in mind that the quadrupolar coupling constants depend strongly on the hybridization of the carbon, and to some extent also on the polarization of the C–2H bonds by partial charges caused by electronegative groups.
The coordinate system for the local tensors is defined in Fig. 1. The zz-component is aligned with the C–2H bond direction, and the yy-component is perpendicular to any π system. For an aromatic C–2H (sp2) bond an asymmetry parameter of with ) was assumed, and the tensor axes were assigned according to calculations of Bailey (1998). The value for the –CD3 group is obtained from the C(sp3)–2H value scaled by 1/3, assuming rapid rotation around the C–CD3 axis.
Program implementation
The routines to apply the proposed orientational constraints are included into the COSMOS-NMR force field, which has been used in a number of previous applications, but which had so far been restricted to distance and chemical shift constraints (see e.g. Sternberg et al. 2003; Witter et al. 2002), and for the force field (Möllhoff and Sternberg 2001; Sternberg et al. 2001). These two types of NMR constraints can be combined with the new orientational constraints as will be demonstrated in the example of PGLa below. The authors provide the backend version of the full COSMOS program, containing all computational procedures without the graphics and modelling interface (GUI) (see http://www.cosmos-software.de). The COSMOS-backend (C++) was compiled for several operation systems including Windows, Unix and Linux.
Applications
The use of NMR orientational constraints is particularly well suited to gain insight into the alignment and dynamics of molecules embedded in biomembranes. In the three examples presented here, we will first demonstrate our new MD approach on pyrene as a simple model compound dissolved in lipid bilayers, then we will apply it to cholesterol as an intrinsic membrane lipid, and finally to the antimicrobial peptide PGLa, which forms an amphiphilic α-helix in membranes.
Pyrene
Pyrene was used as a first example, not only because of its relatively simple and rigid structure, but also because a full set of data is available from a recent 2H-NMR analysis in oriented lipid bilayers (Hoff et al. 2005). The same publication included an all-atom MD simulation of four pyrene molecules in a POPC (1-palmitoyl-2-oleoyl-phosphatedylcholine) membrane. Those previous results will thus allow a direct comparison of our new strategy with established simulation techniques. The present orientational constraint-driven simulation utilizes the experimental 2H-NMR data of deuterium-labelled pyrene (see Fig. 2) embedded at 4 mol% in POPC bilayers. Spectra had been acquired at two discrete sample orientations, namely with the membrane normal aligned parallel and perpendicular to the B0 field. In both orientations two pairs of narrow lines were observed (corresponding to the sites 2H1 and 2H2 + 2H3, respectively), whose splittings changed by a factor of 1/2. This situation can only originate from axially symmetric averaged quadrupolar tensors that are aligned with the membrane normal. This means that the pyrene molecules reorient quickly (on the time scale of the NMR interaction) about the membrane normal. Since the principal axis system of each averaged quadrupolar deuterium coupling tensor coincides with the membrane-fixed coordinate system, the absolute values of all 6 or 5 linearly independent components of the averaged coupling tensor are known and can be included in the present MD simulation. In this case, we could even make use of the signs of the principal quadrupolar components, as they had been deduced from the previous MD simulations mentioned above (Hoff et al. 2005).Fig. 2Fully deuterated pyrene molecule with three kinds of labelled sites (2H1, 2H2 and 2H3), illustrating the principal axes of the local deuterium quadrupolar tensors and of the principal axes of inertia a, b and c
Using the 2H-NMR data from the literature, we performed two MD simulations on pyrene constrained by the complete tensors. In the first run, only the contributions to the pseudo-energy arising from the three diagonal elements (first sum in Eq. 4) were considered, and in a second run also the three off diagonal values (in this case 0.0) were included. In the first run, 11.4 ns were simulated, with the MD parameters set as given in Table 2. To study the time behaviour of the pseudo-forces they were not scaled up exponentially as in the other two case studies below. The width parameter was set equal to the experimental error as ΔP = 1 kHz, and the memory time was τ = 200 ps. The time development of the simulated tensor components for the first nanosecond can be seen in Fig. 3. The trajectories are shown for the 2H2 site, which possesses an experimental splitting of 40.5 kHz. After an initial time period of about 250 ps, the final average coupling tensor components are reached and their values fluctuate only weakly. The residual fluctuations depend on four parameters: (i) the thermal fluctuations controlled by the coupling to the heat bath, (ii) the pseudo-energy, (iii) the width parameter ΔP given in Eq. 11, and (iv) the memory time τ of Eq. 6. In the simulated 2H-NMR spectra (Fig. 4) these prevailing small fluctuations translate into line broadening, and the resulting line shapes are indeed very similar to the experimentally observed spectrum (Hoff et al. 2005).Table 1Deuterium quadrupolar coupling tensorsGroupCzzQ/kHzCyyQ/kHzCxxQ/kHzSourceC(sp2)–2H193−102.29−90.71Benzene, naphthalene (Rinné and Depireux 1974)C(sp3)–2H175−87.5−87.5Acenaphthene bridge (Rinné and Depireux 1974)–CD358.33−29.17−29.17C(sp3)–2Haa The values of the quadrupolar coupling constant are scaled by 1/3 to take into account the free methyl rotationTable 2General parameters for the MD simulations with orientational constraintsParameterValueTarget temperature293 KMD time step0.5 fsCoupling time η to the heat bath0.05 psPseudo-force width ΔP (Eq. 11)1 kHzMemory decay time τ for the average (Eqs. 6, 7)200 psTime constant ρ for the exponential rise of pseudo-forces200 psTotal MD duration1 nsFig. 3Trajectory of the time development of the averaging for the deuterium tensor components for 2H2 of pyrene. Under the influence of orientational pseudo-forces all components approach their final values after a period of about 250 ps. The component at about 20 kHz is experimentally observable in an oriented sample with the membrane normal parallel to B0. The component at −10 kHz would correspond to an experiment with a perpendicular sample alignmentFig. 42H-NMR spectrum from the MD simulation of deuterated pyrene, using orientational constraints from an NMR measurement in oriented POPC membranes (Hoff et al. 2005)
In this first simulation we had constrained only the three principal components of the quadrupolar tensor. It turned out that in this case small off-diagonal values did not approach zero but remain in the range of −1.5 to 2 kHz up to the end of the simulation (see Fig. 3). At the beginning of the simulation a high pseudo-energy of 31,000 kJ/mol was encountered, which dropped to 0.8 kJ/mol by the end of the simulation. Because of the large pseudo-energies at the start, we realized that it is preferable to first run a preliminary MD simulation (1–4 ns) with much smaller pseudo-forces, and then to step up the pseudo- forces after some initial averaging has been performed. In the next two cases studies the pseudo-forces were switched on exponentially to avoid long preliminary equilibration periods.
In a second MD simulation of pyrene over 7.8 ns, we also constrained the three off-diagonal tensor components Cxy, Cxz, and Cyz to zero, which are averaged by rotations of the molecule about the membrane normal. The corresponding values calculated from the MD run yielded off-diagonal elements with absolute values smaller than 0.23 kHz. The pseudo-energy increased by a factor of two, as it now contains contributions from both the diagonal and off-diagonal tensor components. In an attempt to compensate for the higher pseudo-energy, we set the memory time to a higher value of τ = 1,000 ps (see Eq. 6). Altogether we obtained smaller fluctuations of the coupling tensor components and much narrower lines in simulated spectra, which can be most likely attributed to the damping effect of the memory function outweighing the higher pseudo-energies.
Table 3 summarizes the calculated quadrupolar splittings corresponding to the results of the second simulation, in which all tensor components were constrained. The values were averaged for any sites on pyrene with mutually parallel C–2H bond directions, so that the 2H1 splitting is a mean value of two and the others are mean values of four sites each. The MD simulation is seen to reflect the experimental 2H-NMR data very well in several aspects. First of all, the calculated quadrupolar splittings reproduce the experimental values very well. The marginally lower values found for the simulated splittings might be explained by a slightly elevated temperature encountered during the simulation. Namely, the mean temperature of this simulation was 302 K and the final temperature 299 K. Since the target temperature of the simulation was 290 K, the NMR constraints produced a constant temperature increase of about 10 K. This difference would explain the slightly smaller simulated splittings compared to the experiment. In full agreement with the experimental data, no significant difference was found for the sites 2H2 and 2H3. In the simulations, the difference between their splittings was within the scatter of the individual sites, and this difference could neither be resolved experimentally. Furthermore, also the averaging of the pyrene molecule in the membrane is reflected correctly in the simulation. The values for the tensor components perpendicular to the B0 field direction are negative and half of the values of Table 3, indicating that the simulation does indeed lead to an averaged axially symmetric tensor aligned parallel to the membrane normal.Table 3Calculated quadrupolar splittings in a 7.8 ns constrained MD simulation of pyrene, compared with the experimental values obtained from a 2H-NMR measurement of the deuterated pyrene in oriented POPC membranesSiteSign of tensor componentaConstrained MD splitting Δν (kHz)NMR experimentalb splitting Δ ν (kHz)2H1+91.4 (91.39, 91.4)c93.02H2+39.9 (39.8–40.0)c40.52H3+39.7 (39.6–39.8)c40.5a The sign is given for the orientation of the membrane normal in B0 field directionb Hoff et al. (2005)c Range of simulated values for two or four sites, respectively
The experimental 2H-NMR data only reflect dynamic averages of e.g. the C–2H bond orientation, but can access only little information on the dynamic processes leading to the observed splittings. Here, the constraint-driven MD simulation provides a way to obtain details on molecular motions that are compatible with the experimental data. In the example of pyrene, the symmetry of the molecule and its motions in the membrane allowed us to derive information on the dynamic behaviour of the molecule from its observed splittings. Since in our case the coupling tensors are diagonal within the bond coordinate system and the segmental order tensor is also diagonal, simple expressions for the quadrupolar splittings are obtained:
and for the case of an axially symmetric coupling tensor CQ we get
The latter expression is especially useful, since the quadrupolar splitting of a site A is directly related to a single segmental order parameter. Assuming that the order tensor is diagonal in the laboratory frame, Eq. 21 applies to dipolar splittings as well, since the static dipolar coupling tensor is always axially symmetric and traceless, and the same holds for a rapidly rotating CD3-group. However, since the coupling tensor CQ for the C(sp2)–2H bonds is not axially symmetric (see Table 1), the three principal elements of the order tensors WA contribute to the splitting as in Eq. 20. It is therefore not possible to extract the order parameter tensor components from a single experimental splitting. From the course of our MD simulation, on the other hand, all components of the orders tensors WA (Table 4) are available, and the terms of Eq. 20 can be readily analysed.Table 4Simulated order parameters for pyrene in POPCSiteWxxWyyWzz| Wxy,Wxz,Wyz|2H1+0.021−0.342+0.321≤0.032H2+0.181−0.310+0.129≤0.072H3+0.165−0.294+0.129≤0.03SbbSccSaa|Sab, Sac, Sbc|Moleculea+0.048−0.275+0.228≤0.006a The values for the molecular Saupe order tensor are ordered in the columns such that the axes of inertia fit to the corresponding coordinate axes of the 2H1 site (see Fig. 2)
For a perfectly rigid pyrene molecule the molecular Saupe order tensor S and the segmental order tensors WA are linked by fixed conformation tensors. With our choice of the molecular coordinate system, S and WA are identical for the site 2H1, and for the other two sites the tensors are related by simple geometric expressions. Comparing the molecular order tensor S (last row in Table 4) and the segmental order tensor W (e.g. of the site 2H1, which is expressed in the same coordinate system, first row in Table 4), it is noticed that most of the molecular S tensor components are slightly smaller than the corresponding values of the segmental order tensor W. This effect is caused by bond vibrations and molecular twists occurring during the MD simulation. However, as these contributions to the dynamics (manifested in the difference between segmental and molecular order) are small, the mobility of the pyrene in the membrane is well described by the molecular order tensor S. Further qualitative conclusions on the behaviour of the pyrene molecule in the membrane can thus be drawn from the order tensor. Because Saa constitutes the largest (signed value) component, it is obvious that the corresponding axis a (in the system of inertia), and with it the long molecular axis of pyrene, shows some preference for an orientation in the direction of the membrane normal. Since Scc and Sbb are different, the molecular motions are not axially symmetric, but display a preference for a second axis, expressed in a biaxiality parameter ξ = 0.11 (see Eq. 16). This pronounced biaxiality is a consequence of the restricted motions of the flat pyrene molecule within the lipid matrix. A further indication for such a restricted motion is the fact that the smallest component of the order tensor S is found along the c-axis direction perpendicular to the aromatic ring system. This direction has thus the smallest average component along the membrane normal and is most conserved, as expected for the planar shape of the molecule. In their all-atom MD simulation including 128 lipid molecules and water, Hoff et al. (2005) obtained essentially the same results as in this work, with the exception that they found a much higher Saupe order component Szz = 0.42 (in this work denoted with Scc).
Cholesterol
Cholesterol is a constituent of many biomembranes, and its orientation and motion within the lipid bilayer is of great interest to explain its influence on membrane properties. The behaviour of cholesterol in membranes has been studied extensively, especially by 2H-NMR in view of its sensitivity towards molecular alignment and dynamics (Smith and Oldfield 1984). We have based our MD analysis on a recent study by Marsan et al. (1999), who had measured and interpreted 2H-NMR spectra of partially deuterated cholesterol in oriented DMPC (dimyristoylphosphatidylcholine) bilayers. The labelled sites in the sterol ring system are indicated in Fig. 5.Fig. 5Cholesterol molecule with 2H-labelled sites indicated, for which experimental orientational constraints were available. From the MD simulation quadrupolar splittings were calculated for all deuterons in the ring system
The authors examined six different models of cholesterol structures proposed in the literature using an RMSD analysis to assign all observed lines in the 2H-NMR spectrum. In this analysis the quadrupolar splittings were calculated from the respective rigid molecular models, treating the molecular orientation and the Saupe tensor as free variables. It turned out that the results depended crucially on the selection of the molecular coordinates, and only the data from a neutron diffraction analysis (McMullan et al. 1992) of a cholesterol derivative (20-CH3-methylpregnene-3,20-diol) produced convincing results. This strong dependence on the exact coordinates is regarded as one of the main drawbacks of RMSD analyses of rigid molecular models.
In the analysis presented here, we performed MD simulations for both methylpregnenediol and cholesterol, having modelled the latter with COSMOS (see Fig. 5). As constraints for the MD simulations we used the 2H-NMR splittings of the eight experimentally observed deuterium labels of Marsan et al. (1999), together with their best assignment (see Table 5). In addition, from the MD simulations we predicted the putative quadrupolar splittings of all other protonated sites of the sterol ring system. This new data from our MD simulations can thus be used in the future to assign further 2H-NMR spectra without the need for any special labelling schemes or particular molecular coordinates. As outlined above, we again used the full quadrupolar tensors as constraints. Two short preliminary MD runs were performed to determine the signs of the coupling constants, since this information is required for the full tensorial constraints. All quadrupolar splittings of cholesterol which had been measured in oriented samples with the membrane normal parallel to B0 yielded negative coupling constants, (giving correspondingly positive values for the perpendicular orientation, see Table 5). As in the example of pyrene, MD simulations over 1,000 ps were sufficient to reproduce all features of molecular motion. To avoid too large pseudo-forces at the start of the simulation, the pseudo-forces were exponentially increased to their final values during the MD run. The parameters used in the simulations are summarized in Table 2.Table 5Calculated deuterium quadrupolar splittings from a constrained MD simulation compared with the experimental 2H-NMR data (Marsan et al. 1999)SiteConstrained MD splitting of methyl-pregnenediol (kHz)Constrained MD splitting of cholesterol (kHz)Experimental NMR splitting (kHz)H2a−101.5−101.4101.68H2e−66.7−67.767.86H3_1−107.1−106.7107.30H4e−63.2−62.362.68H4a−94.3−93.094.98H6−7.29−6.696.44H7a−92.9−95.496.12H7e−91.6−91.391.48C18–CD3−33.0−33.8C19–CD3−37.5−34.7H1a−104.2−99.2H1e−69.4−46.1H8_1−107.0−100.3H9_1−112.6−103.0H11a−106.4−102.7H11e−50.7−74.7H12a−46.4−13.1H12e−103.1−101.1H14_1−103.4−100.3H15_1−102.1−81.8H15_2−93.5−88.9H16_1−17.6−40.9H16_2−23.1−27.6H17_1−85.7−92.7
The final results for the quadrupolar splittings obtained from the constrained MD simulations of both sterols are summarized in Table 5. The corresponding simulated 2H-NMR spectrum of cholesterol is shown in Fig. 6, indicating the peak positions of all ring deuterons. This presentation can be helpful in the future for assigning the signals of fully deuterated cholesterol derivatives. Overall, the theoretical and experimental values agree very well within 1 kHz (except for 2H6). The only noteworthy deviation is seen for 2H7a, where the calculated splitting of methylpregnenediol differs by 3.2 kHz from the experiment. Overall, the simulated cholesterol quadrupolar couplings agree slightly better with the experimental values than the respective methylpregnenediol values. This finding is also reflected in the lower mean temperature reached in the simulation of cholesterol (323 K) than for methylpregnenediol (336 K). The higher mean temperature of the latter is probably also responsible for the slight differences between the calculated splittings of the two molecules in some of the other sites, too (see Table 5).Fig. 6Calculated 2H-NMR spectrum from the MD simulation of deuterated cholesterol in an oriented membrane, with the membrane normal parallel to B0. Pseudo-forces were applied in the MD simulation to a sub-set of deuterons for which experimental data was available (see Fig. 5). The peak positions of all other deuterons connected to the ring skeleton are also predicted and indicated (see Table 5). The intensity of the CD3-signals has to be multiplied by a factor of three
The constrained MD simulation showed that both cholesterol derivatives essentially assume an upright orientation in the membrane and perform rotations mainly around their long molecular a-axis of inertia. This behaviour is reflected in the molecular Saupe order tensor (see Table 6), whose values are not far from the ideal axially symmetric case of Eq. 16 with a low biaxiality of ξ = 0.01. Previous calculations of the order tensor from the NMR data (Marsan et al. 1999) had given similar results as the present MD simulation. We find that the preferred orientation of the long a-axis of the molecule displays a small mean tilt angle Θ a of 13° with respect to the membrane normal. A similar result was obtained by Marsan et al. (1999), who reported a tilt of the long molecular axis (as defined in the paper of Dufourc et al. 1984) with respect to the membrane normal of 12° (provided that in their RMSD data analysis the restriction to a rotationally symmetric order tensor was dropped).Table 6Molecular Saupe order tensor derived from the constrained MD simulation, compared with the values derived from the previous RMSD analysis (Marsan et al. 1999)Tensor componentMD simulation of methylpregnenediolMD simulation of cholesterolStatic RMSD analysisSaa0.870.880.94Sbb−0.44−0.44−0.48Scc−0.43−0.44−0.46Sab0.0−0.11Sac0.0−0.12Sbc0.00.0
We note that the axially symmetric averaging of cholesterol does not imply that its motion resembles a smooth rotation about a fixed axis. In the MD simulation we observed frequent jumps of the long axis by 20 to 30°, accompanied by a change of the conformation of the aliphatic tail of the molecule. This behaviour can be examined more closely by inspecting the trajectory of the angle Θ a between the a-axis of inertia and the membrane normal (z-axis, see Fig. 7). Under the influence of the full pseudo-forces after 150 ps, the long axis of cholesterol frequently jumps off the most preferred orientation of Θ a = 4°. As a consequence, the mean orientation and the most frequent orientation do not coincide. Methylpregnenediol behaves clearly different in the MD simulation, since it has only a short aliphatic tail that does not give rise to conformational changes.Fig. 7Trajectory of the time development of the angle Θa between the a-axis of inertia and the z-direction (membrane normal). The axis frequently jumps off its most preferred orientation due to conformational changes in the aliphatic tail of the molecule, hence the mean tilt (13°) and the most preferred tilt (4°) do not coincide
PGLa
In our third case study, the new MD approach was applied to a membrane bound peptide. PGLa is a 21-residue cationic peptide (GMASKAGAIAGKIAKVALKAL-NH2) from the magainin family of antibiotics present in frog skin, which folds into an amphiphilic α-helix when bound to lipid bilayers. The mechanism of antimicrobial activity has been addressed in numerous studies and is attributed to the perturbation of bacterial membranes. Solid state NMR has yielded much insight into its structure and dynamic behaviour in model membranes (Bechinger 1999). For our MD simulations we used the constraints from the 2H-NMR investigations of Strandberg et al. (2005), who had labelled four native alanine residues (positions 6, 8, 10 and 14) and two isoleucines (positions 9 and 13) one by one with 2H3-alanine. At a peptide-to-lipid ratio of 1:200 in DMPC, it was found from the quadrupolar splittings of these six CD3-groups that the α-helical PGLa is aligned flat in the plane of the membrane, and the peptide undergoes fast rotational diffusion about the membrane normal at 35°C. From the same data we now derived the full quadrupolar coupling tensors for the six 2H-labelled sites, including the sign of the tensor elements, and used these as constraints for the MD simulations. For the quadrupolar coupling tensors we used the parameters of the CD3-group (Table 1), both for the native Ala substitutions as well as for the nominal constraint along the Cα–Cβ segment of isoleucine.
It is known from previous NMR studies (Strandberg et al. 2005) that membrane-bound PGLa forms an α-helix in the range of the labelled stretch, and further evidence for an α-helical conformation between residues 6 and 21 is provided by (Bechinger 1999). We therefore started the MD simulations with an idealized α-helix, and for all backbone hydrogen bonds we introduced 18 additional distance constraints of 1.86 Å to keep the molecular model helical. At the end of the simulation, the RMSD of the backbone hydrogen bond lengths from their ideal values was only 0.2 Å, thus confirming that that the molecule stayed indeed mostly helical during the MD simulation. It also had to be taken into account that the four lysine side chains and N-terminus are positively charged. Since charged NH3+ groups have a strong tendency to form hydrogen bonds, we added a water molecule near each hydrogen atom of a charged group. This way any undesired hydrogen bonds of lysines to the backbone could be prevented, and indeed at the end of the simulation all 15 water molecules were still in contact with the peptide.
Compared to the two small-molecule case studies above, the NMR data on PGLa provide only few constraints on the large and rather flexible molecule, as the pseudo-forces act only on six amino acids out of 21. Therefore, unrealistically large-pseudo forces at the start of the MD simulation have to be avoided. As a first measure, the pseudo-forces were increased exponentially to their final values during the MD simulation with a time constant of ρ = 100 ps, like in the cholesterol simulation. From the NMR studies it was known that PGLa rotates rapidly (on the NMR time scale) about the membrane normal. This motion opens up a second possibility to scale the pseudo-forces properly. We calculated rotational averages of the quadrupolar tensors in every step of the MD. Only three tensor values rotated by 120° around the director axis (in this case the membrane normal oriented parallel to the B0 field directions) are necessary to obtain a mean tensor within the limits of a fast rotation. All off-diagonal values of the simulated tensors thus become zero and only the principal values of the tensors are left as constraints. The mean temperature at the end of the simulation was only 11 K above the target temperature of 293 K, indicating that exceedingly high pseudo-forces have been successfully avoided indeed, which would otherwise have lead to additional heat production.
As in the other two case studies above, the constraint-driven MD simulations were able to reproduce the experimental 2H-NMR data correctly, as evident from the comparison in Table 7. All calculated frequencies are well within the experimental error margins (as estimated from the experimental line widths). For some additional sites the quadrupolar frequencies were predicted from the MD simulation, and it would be of interest whether future measurements can confirm these values.Table 7Calculated deuterium quadrupolar splittings from a constrained MD simulation of PGLa, compared with the experimental 2H-NMR data (Strandberg et al. 2005) in DMPCSiteMD (kHz)NMR (kHz)Ala3−22.2Ala6+15.7+15.6Ala8+17.1+17.2Ala10−15.0−15.0Ala14−26.8−26.6Ala17+17.1Ala20−25.5Ile9−4.9−5.2Ile13+26.226.4
The major aim of the present NMR data analysis was the determination of the time- and ensemble averaged alignment of the helical PGLa backbone in the membrane environment. As illustrated in Fig. 8, the principal axes of inertia offer a suitable molecule-fixed frame of reference to describe the peptide orientation. In the PGLa the a-axis of least inertia points along the helix, and the c-axis of highest inertia is located between the hydrophilic lysine-rich face and the hydrophobic face of the molecule formed by alanines, valines, leucines and isoleucines. Intuitively, we would expect the helix to be in contact with biomembranes just like it is displayed in this view, namely with the intermediate b-axis collinear with the membrane normal.Fig. 8View along the long axis of PGLa modelled as an idealized α-helix with only the lysine, alanine and isoleucine side chains displayed. The latter two types were labelled by 2H3-alanine, and their 2H-NMR splittings were included into the MD simulation. The principal a-axis of inertia, points along the helix, and the b-axis of inertia points radially through Cα of lysine 12
The outcome of the MD simulation is illustrated most appropriately in Fig. 9, which visualizes the alignment and dynamic behaviour of PGLa in DMPC. Here, the peptide orientations during the MD simulation are presented as a scatter plot of the inertia axes a, b and c on the surface of a unit sphere. A narrow range of orientations is seen for the a-axis representing the peptide helix. In contrast, the broad scatter for b and c indicates that a high degree of motional freedom exists for motions around the helix axis. The time-averaged tilt angle of the a-axis with respect to the membrane normal in found to be 99°, with a mean deviation of about ±4°. This value is in very good agreement with the tilt angle τ ≈ 98° of PGLa that had been deduced from an RMSD plot of the 2H-NMR by Strandberg et al. (2005) assuming a rigid idealized α-helix. We note, however, that such helix axis definition is not exactly comparable to the a-axis of inertia used in our MD calculations. The authors further described the azimuthal alignment of the amphiphilic peptide by defining an angle ρ between an axis pointing from the centre of the helix radially through the Cα-atom of lysine 12 and the membrane plane. In our case, the b-axis of inertia also happens to be directed to Cα of lysine 12 (see Fig. 8), but the centre of inertia does not exactly coincide with the helix axis. As can be seen from Fig. 10, the angle Θ b between the b-axis and the membrane normal oscillates between 10° and 70°, with a mean value of 33°. This average value corresponds to an angle ρ of approximately 123°, which compares well with ρ ≈ 115° extracted from the previous analysis (Strandberg et al. 2005). In this comparison we have to keep in mind that in our simulations the PGLa molecule was found to be far from rigid and possesses a high degree of mobility around the helix long axis. The lysine side chains are found to spread out much more than in the initial model, and the PGLa helix itself exhibits a small bend.Fig. 9Visualization of the alignment and motional behaviour of the α-helical peptide PGLa (only lysine side chains shown) in a DMPC membrane, as determined by a 1 ns MD simulation using experimental 2H-NMR constraints (Strandberg et al. 2005) The instantaneous orientation of the three axes of inertia a, b and c is displayed as a scatter plot on a sphere, representing snapshots of these axes orientations every picosecond. The poles of the sphere are oriented parallel to the membrane normal. The helix long axis displays only a small scatter with a mean tilt angle of 99°, while the two other axes oscillate about this direction producing a large scatterFig. 10Molecular dynamics trajectory of the angle Θb between the intermediate b-axis of inertia and the membrane normal
Similarly important as the alignment of the peptide are its dynamic properties, which also enter into the observed 2H-NMR data. Table 8 lists the components of the molecular Saupe tensor, which is again used to describe the average alignment a molecule fixed frame (i.e. the principal axes of inertia) with respect to the laboratory frame. However, in contrast to the two previous case studies, this Saupe tensor does not represents the fast reorientation of the molecule around the membrane normal. No pseudo-forces were driving this motion in the simulation, since a rotational mean value was introduced. Here, the order tensor represents only the wobble of the long a-axis and the pronounced oscillations about this molecular axis (see Fig. 10). As expected from the distribution of polar and hydrophobic residues on the different faces of the helix, it is seen that the Saupe tensor is not diagonal Table 8). The biaxiality of ξ = 0.09 (S = 0.76, see Eq. 15) is clear evidence that the molecule does not behave like a rod-shaped object, and rotation about its long axis is restricted. In fact, oscillations of about 60° around the long axis, but no complete rotation events were observed in the MD simulation (see Fig. 10). These oscillations start after about 100 ps when the pseudo-forces approach their full strength.Table 8Molecular Saupe order tensor from a constrained MD simulation of PGLaTensor componentMDSaa+0.49Sbb−0.28Scc−0.46Sab−0.46Sac+0.06Sbc−0.08
At a first glance the rather regular oscillations of PGLa in Fig. 10, or the rotations of cholesterol (Fig. 7), may not seem to reflect reality. However, we have to keep in mind that our MD simulations are performed with single molecules in vacuum, and the temperature is controlled by a continuously acting NTV “thermostat” adjustment. In nature, heat is transferred by stochastic interactions with other molecules, thereby introducing a stochastic behaviour of molecular rotations and re-orientations. This stochastic aspect is not present in our MD simulations, but nonetheless the amplitudes, velocities and directions of the motions adopt realistic values as the NMR constraints have to be satisfied. The molecule in the simulation cannot stop in one preferred orientation, because then no averaging would be performed anymore.
Conclusions
Solid state NMR is a valuable technique to gain insight into the behaviour of peptides and proteins in oriented media, provided the data can be interpreted in terms of molecular structure and dynamics. In this contribution we developed a new strategy in which all-atom MD simulations and NMR data obtained from oriented samples are combined to obtain such structural and motional information. To this aim, a molecular mechanics force field (in this case COSMOS-NMR) was extended to include pseudo-forces, which drive the molecular dynamics to meet the NMR constraints. They “heat up” molecular rotations or re-orientations, leading to proper averaging of the calculated tensor values such that the calculated tensor values agree with the corresponding experimental observations. The orientational constraints can be further combined with intramolecular constraints such as distances or chemical shifts. This way, similar results can be obtained as in full membrane MD simulations, but without the computational burden of having to perform a detailed simulation of the lipids and surrounding water molecules. Because they are performed in vacuum, the constrained MD simulations can be completed in relatively short simulation times (≤1 ns), still reaching a complete averaging of the NMR observables.
In three case studies, ranging from a small rigid compound to a 21-residue membrane-active peptide, the MD simulations with orientational NMR constraints succeeded to produce a detailed picture of the molecular motions and orientations in oriented membranes. It could be demonstrated that this new method is not limited to rigid molecules and does not depend on the choice of the initial coordinates. Deuterium quadrupolar splittings from 2H-labelled pyrene, cholesterol and PGLa in oriented lipid bilayers have been used as constraints in the present examples, but the general formalism presented will be applicable to all kinds of tensorial NMR properties.
|
[
"molecular dynamics simulations",
"pyrene",
"cholesterol",
"order parameters",
"2h-nmr",
"oriented samples",
"orientational nmr constraints",
"pgla peptide",
"force field calculations"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"R",
"R"
] |
Eur_J_Pediatr-3-1-1914268
|
Long-term cognitive and behavioral consequences of neonatal encephalopathy following perinatal asphyxia: a review
|
Neonatal encephalopathy (NE) following perinatal asphyxia (PA) is considered an important cause of later neurodevelopmental impairment in infants born at term. This review discusses long-term consequences for general cognitive functioning, educational achievement, neuropsychological functioning and behavior. In all areas reviewed, the outcome of children with mild NE is consistently positive and the outcome of children with severe NE consistently negative. However, children with moderate NE form a more heterogeneous group with respect to outcome. On average, intelligence scores are below those of children with mild NE and age-matched peers, but within the normal range. With respect to educational achievement, difficulties have been found in the domains reading, spelling and arithmetic/mathematics. So far, studies of neuropsychological functioning have yielded ambiguous results in children with moderate NE. A few studies suggest elevated rates of hyperactivity in children with moderate NE and autism in children with moderate and severe NE. Conclusion: Behavioral monitoring is required for all children with NE. In addition, systematic, detailed neuropsychological examination is needed especially for children with moderate NE.
Introduction
It has been widely accepted that severe NE following PA can lead to serious motor disabilities, mental retardation and seizure disorders [59]. Traditionally, research in this area has focused on developmental consequences at a very young age. In addition, the effects of NE have often been treated as an all-or-nothing phenomenon; patients have been believed to show a completely normal development or become severely disabled. So far, few studies have looked at milder consequences of NE at school age.
From perinatal asphyxia to neonatal encephalopathy
PA can be defined as impaired respiratory gas exchange accompanied by the development of acidosis [30]. The pathway of PA can have its onset in the antepartum, intrapartum or postpartum period. Estimates of the incidence of PA vary from 1 to 8 per 1,000 live births [4]. This wide range can be largely attributed to problems in selecting indicators to identify children with PA. As a consequence, researchers have used many different sets of inclusion criteria when studying the effects of PA. For this reason, it is difficult to compare results regarding developmental outcome of infants with PA across studies.
In the past, single-marker definitions were commonly used to diagnose infants with PA (e.g., low Apgar score). However, these single markers have relatively low sensitivity and specificity for accurate diagnosis [54]. Moreover, they lack predictive power [3, 20, 35–37, 40, 51]. Therefore, a shift has been made towards multiple-marker definitions based on combinations of indicators [16, 35]. Among the perinatal indicators are fetal distress (abnormal heart rate or meconium-stained amniotic fluid), delay in onset of spontaneous respiration, low Apgar score (<6 or 7 at 5 min), need for resuscitation and/or ventilation, and metabolic acidosis (cord pH <7.0 or 7.1 and base deficit >12 mmol/l). Postnatal indicators of PA include NE, multi-organ failure and abnormal findings on brain imaging. Of the above, the concept of NE is generally accepted as an essential criterion for the diagnosis of PA [4, 30]. NE has been defined as a clinical syndrome of disturbed neurological function in the earliest days of life in term infants, manifested by difficulty with initiating and maintaining respiration, depression of tone and reflexes, subnormal level of consciousness and often by seizures [30].
Recent studies of the protective effects of head cooling and systemic hypothermia after NE following PA used a combination of clinical and physiological criteria of PA (including a pH of 7.0 or less or a base deficit of 16 mmol per liter or more) and a neurological examination to determine the severity of NE [22, 52]. These inclusion criteria are closely related to a set of criteria postulated by The American College of Obstetricians and Gynecologists (ACOG). There has been extensive debate about the relationships between PA, NE and serious possible consequences such as CP [25]. Contrary to what was once believed, only a small part of all cases of CP are caused by intrapartum events [10]. Likewise, many cases of NE are not associated with a history of PA [6, 38]. However, NE has to be present in the pathway from PA to subsequent CP [25]. The ACOG and the International Cerebral Palsy Task Force proposed a set of criteria to define an acute intrapartum hypoxic event that could subsequently lead to CP [4, 30]. The presence of moderate or severe NE and metabolic acidosis (pH <7 and base deficit ≥12 mmol/l) are considered to be essential markers for hypoxic ischemic injury. In addition, these criteria underline the importance of excluding other identifiable causes of NE before assuming asphyxia as the underlying etiology.
Both in clinical practice and for scientific purposes, patients suffering from NE are often assigned an encephalopathy score. Several classification systems can be identified. Leviton and Nelson provide an overview of qualitative and quantitative grading schemes [27]. In addition, scoring systems were developed that are applicable in low-income settings. Moreover, a new grading system has gained interest in the literature [19, 33]. The classification system of Sarnat and Sarnat developed in 1976 [50] is still widely used and the basis for most modern classification systems. In this review, the Sarnat system is used to grade NE because in most studies that were included this system it had been applied. According to the Sarnat system children are assigned a score of 1, 2 or 3 (1= mild NE; 2= moderate NE; 3= severe NE). Stage 1 lasts less than 24 h and is characterized by hyperalertness, uninhibited Moro and stretch reflexes, sympathetic effects and a normal EEG. Stage 2 is marked by obtundation, hypotonia, strong distal flexion and multifocal seizures. Infants with the most severe stage of NE (stage 3) are stuporous and flaccid, and brain stem and autonomic functions are suppressed. In addition, the EEG shows abnormal patterns of brain activity.
Grades of neonatal encephalopathy in relation to developmental outcome
Encephalopathy scores have often been used to predict neurodevelopmental outcome [43, 44, 46]. Although strong relationships exist between the categories of NE and developmental outcome, diversity of outcome is reported within individual grades [41]. This diversity may be partly explained by individual researchers using different neurological signs for category inclusion [27]. The Sarnat grading system has helped to clarify the NE categories and improved their usefulness in studies of prognosis [45].
Most outcome studies have focused on neurological functioning and severe deficits in young children (<4 years). In general, very few children with mild encephalopathy show neurological impairments or have developed severe mental or motor retardation at preschool age. In contrast, children with severe encephalopathy nearly always die or develop severe impairments such as CP, mental retardation, epilepsy and in some cases sensorineural hearing loss or cortical visual impairment [4, 45, 47]. While dyskinetic CP or quadriplegia is more commonly seen following “acute near total type of asphyxia” with MRI abnormalities present in the deep grey nuclei and in the perirolandic cortex, milder forms of CP, such as spastic diplegia, are more common among those with “subacute partial asphyxia” with MRI abnormalities in the watershed areas. Hemiplegia can be seen following “perinatal arterial stroke.” Although this condition will present with NE and seizures, it is less commonly associated with PA [13, 34]. Children who have suffered moderate encephalopathy seem to form a more heterogeneous group; research shows varying rates of infant death and morbidity [16].
Need for long-term follow-up and focus on cognitive and behavioral outcomes
Because the majority of studies have focused on detecting major developmental abnormalities at a very young age, still little is known about the milder difficulties children may experience in the long run. Recently, however, research has started to focus more on the subtle problems of children without major neurological deficits [8, 23, 32, 35, 43, 44, 46]. Knowledge of the outcome for this group of children is important because these subtle problems can affect later scholastic achievement and behavioral adjustment.
Cognitive and behavioral difficulties can be expected because of the patterns of brain injury that have been associated with NE. The hippocampus and striatum are among the brain structures that can be affected [7, 21, 31, 42, 56]. These structures have been associated with specific cognitive functions such as memory and attention and hypothesized to play a role in the pathogenesis of ADHD, autism and schizophrenia [14–16, 28, 57].
To gain insight into the long-term effects of NE on development, follow-up is required throughout the school-age period. Specific cognitive functions continue to develop throughout childhood. As it is impossible to examine a function that has not developed yet, short-term follow-up cannot exclude minor cognitive and behavioral difficulties at school age. Moreover, subtle functional deficits usually do not become apparent until a child faces increasing demands to master complex abilities in school.
The aim of this review was to discuss the results of previously published studies on the possible effects of NE on general cognitive functioning, educational achievement, neuropsychological functioning and behavior from infancy through adolescence. A literature search was performed using Web of Science. Studies were screened that had been published before October 2006. Search terms were ’asphyxia neonatorum,’ ‘neonatal encephalopathy,’ ‘newborn encephalopathy,’ ‘hypoxic-ischemic encephalopathy’ and ‘perinatal asphyxia’ combined with the terms ‘cognition/cognitive’ and ‘behavior(-al)’. Reference lists of retrieved articles were screened for additional relevant publications. Studies were selected that reported the developmental outcome of NE or HIE following PA of children until the age of 18 years. The presence of NE was required for selection. When no additional indicators of PA were used, the asphyxial etiology of the NE had to be assumed by the authors. Studies were excluded when NE was partially caused by other conditions than PA, for example [17, 34]. Also, studies were excluded in which outcome was presented in preset categories (e.g., normal versus abnormal), in which several outcome measures were combined, for example [11, 18, 26]. In these cases it was not possible to judge whether a child had been assigned to the category ‘abnormal’ because of cognitive or behavioral problems or because of other problems, e.g., motor difficulties. A third reason for excluding studies from this review was the selection of children from a larger group of eligible patients. For example, some researchers selected only those children with abnormal findings on MRI in order to examine the correlation with outcome [61]. In this review, outcome is reported according to grade of NE (mild, moderate or severe). Most authors of the included studies used the classification system of Sarnat when determining the severity of the NE. However, in three cases the authors significantly deviated from this system [32, 47, 53]. In order to increase comparability across studies, the patients in these studies were reclassified according to Sarnat criteria. Articles were reviewed that reported general intellectual ability, levels of specific cognitive functions and/or behavioral problems. Details of the samples, inclusion criteria, outcome measures and results are described in Tables 1, 2, 3, 4, 5, 6, 7 and 8.
Table 1Griffiths Mental Development Scale* at 1 year CPNo CPCarli et al. (2004) [12]Moderate NE, N=13Moderate NE, N=27• N=10: <−2 SD• N=4: between −2 SD and −1 SD• N=1: between −2 SD and −1 SD• N=23: >−1 SD• N=1: >−1 SD• N=1: untestableGray et al. (1993) [24]Moderate NE, N=10Moderate NE, N=9• N=9: <55Mean =118 (N=1 not assessed)• N=1: 87Thompson et al. (1997) [55]No grades of NE recorded, N=16No grades of NE recorded, N=24Mean =47, range: 7–101Mean =16, range 94–128Barnett et al. (2004) [9]No grades of NE recorded, N=59Mean =100.34, SD=15.05, range: 55–130Rutherford et al. (1996) [49]Moderate NE, N=8Mild NE, N=4All below normal (cut-off point not reported)All normalSevere NE, N=1Below normal*Griffiths mean of standardization sample =100.18, SD=12.76Table 2IQ at 2–6 years (no CP) Age (years)Mild NEModerate NESevere NEBarnett et al. (2004)b [9]2N=45,f Griffiths: mean =103.53, SD=12.33 (range: 81–140)Robertson & Finer (1985)a [43] 3.5N=66, Stanford-Binet: mean IQ =101.5, SD=14.0N=94 (21.3% handicapped) Stanford-Binet2: mean IQ =92.3, SD=23.2N=7, Stanford-Binet2: mean IQ =37.1, SD=26.7Shankaran (1991)cb [53]5N=14,f McCarthy: 64%: >−1 SD, 86%: >−2 SDBarnett et al. (2002)bd [8] 5.5–6.5N=20, WPPSI-R: mean IQ =109.7, SD=14.6 (N=1: 76, rest: >90)N=12, WPPSI-R: (N=1: untestable) N=11: mean IQ =106.2, SD=11.8 (N=1: 84, rest: >94)Barnett et al. (2004)b [9]5–6N=53,f WPPSI-R 85%: >−1 SD, mean IQ =101.98, SD=16.06, range: 69–139Robertson & Finer (1988)e [44]5.5 N=56, Stanford-Binet: mean IQ =106, SD=12N=71, Stanford-Binet: mean IQ =99, SD=18)Griffiths Mental Developmental Scales: mean of standardization sample =100.18, SD=12.76; Stanford-Binet Intelligence Scales: mean of standardization sample =100, SD=16; McCarthy Scales of Children’s Abilities: mean of standardization sample =100, SD=16; Wechsler Preschool and Primary Scale of Intelligence-revised: mean of standardization sample =100, SD=15. aAll groups are significantly different from the other groups; p=0.001. bNo significance tests performed. cPatients were reclassified by author (MH) according to the classification system of Sarnat [50]. dIn the original article, results are presented for individual cases. Group means and standard deviations have been calculated by one of the authors (MH). eMild NE not significantly below comparison group; moderate NE significantly below mild NE and comparison group. fCombined group of mild, moderate and severe NETable 3IQ at 7–9 years (no CP) Age (years)Mild NEModerate NESevere NEMarlow (2005)b [32]7N=65, BAS-II:Group 1 (N=34): GCS=112.3, SD=11.2Group 2d (N=31): GCS=102.7, SD=13.2Robertson et al. (1989)a [46]8N=56, WISC-R: mean IQ =106, SD=13N=66, WISC-R: mean IQ =102, SD=175, WISC-R: mean IQ =36, SD=7Robertson (1997)c [47]9N=64, WISC-R: mean IQ =100, SD=14WISC-R: Wechsler Intelligence Scale for Children-revised: mean of standardization sample =100, SD=15; BAS-II: British Ability Scales, GCS: General Cognitive Score: mean of standardization sample =100, SD=15. aIQ of mild NE is not significantly lower than IQ of comparison group. IQ of moderate and severe NE is significantly lower than IQ of mild NE and comparison group (p < 0.001). bGSC of group 1 is not significantly lower than GSC of comparison group. GCS of group 2 is significantly lower than GCS of comparison group (p < 0.01). cMean IQ is significantly lower than IQ of comparison group. dMarlow refers to this group as ‘severe NE.’ This group is reclassified by one fo the authors (MH) according to the classification system of Sarnat [50]Table 4Scholastic abilities at 7–13 years (no CP) Mild NEModerate NESevere NEReadingRobertson & Finer (1989) [46]13% >1 grade level below expected level for age35% >1 grade level below expected level for agea100% >1 grade level below expected level for agebRobertson (1997) [47]41% >1 grade level below expected level for agedMarlow et al. (2005) [32]Group 1: < control groupcGroup 2f: < control groupdMoster et al. (2002) [35]NE < control groupdNE < control groupdNE < control groupdWritingMarlow et al. (2005) [32]Group 1: = control groupGroup 2f: < control groupeMoster et al. (2002) [35] NE = control groupNE = control groupNE = control groupSpellingRobertson & Finer (1989) [46]2% >1 grade level below expected level for age18% >1 grade level below expected level for agea100% >1 grade level below expected level for agebMarlow et al. (2005) [32]Group 1: < control groupcGroup 2f: < control groupdMoster et al. (2002) [35]NE = control groupNE = control groupNE = control groupMathRobertson & Finer (1989) [46]16% >1 grade level below expected level for age20% >1 grade level below expected level for agea100% >1 grade level below expected level for agebRobertson (1997) [47]39% >1 grade level below expected level for agedMarlow et al. (2005) [32]Group 1: = control groupGroup 2f: < control groupdMoster et al. (2002) [35]NE < control groupcNE < control groupcNE < control groupcRobertson & Finer (1989) [46]: age 8 years; Marlow et al. (2005) [32]: age 7 years; Moster et al. (2002) [35]: age 8–13 yearsaSignificant difference with mild NE and comparison group, p < 0.01; bstatistical analyses are not given because of small group and low scores; csignificant difference with comparison group, p < 0.05; dsignificant difference with comparison group, p < 0.001; esignificant difference with comparison group, p < 0.01; fMarlow refers to this group as ‘severe NE’. This group is reclassified by one of the authors (MH) according to the classification system of Sarnat [50]Table 5Neuropsychological functions [43, 44, 46]Tests3.5 years5.5 years8 yearsReceptive vocabularyMild NE: in the mean rangeMild NE = control groupMild NE = control groupModerate NE < mild NEaModerate NE< mild NE+ control groupbModerate NE< mild NE+ control groupaSevere NE < mild + moderate NEaSevere NE: 100% < −1 SD cSevere NE < mild NE + control group cVisual- motor integrationMild NE: in the mean rangeMild NE = control groupMild NE = control groupModerate NE < mild NEaModerate NE < mild NE + control groupaModerate NE < mild NE + control groupa Severe NE < mild NE + control group cSevere NE < mild + moderate NEaSevere NE: 100% < −1 SD cReceptive vocabulary: Peabody Picture Vocabulary Test, visual-motor integration: Developmental Test of Visual-Motor Integration. ap < 0.001; bp < 0.01; cstatistical analyses are not given because of small group and low scoresTable 6Neuropsychological functions (NEPSY): moderate NE compared to control group [32]Age: 7 yearsModerate NE group 1Moderate NE group 2*Attention and executiveGroup 1 = control group (p=0.08)Group 2 < control group (p < 0.01)LanguageGroup 1 < control group (p=0.01)Group 2 < control group (p < 0.01)SensorimotorGroup 1 < control group (p=0.04)Group 2 = control group (p=0.10)VisuospatialGroup 1 = control group (p=0.14)Group 2 < control group (p=0.02)Memory and learningGroup 1 = control group (p=0.61)Group 2 < control group (p < 0.01)*Marlow refers to this group as ‘severe NE.’ This group is reclassified by one of the authors (MH) according to the classification system of Sarnat [50]Table 7Verbal memory (Rey’s Auditory Verbal Learning Test)* [31]Age: 16 yearsModerate NETotal words learned trial 1–5NE: 52.5, SD=8.7Control group: 57, SD=5.8Delayed recall (p=0.034)NE: 11.6, SD=2.2Control group: 13.3, SD=1.6Recognition (p=0.011)NE: 14.2, SD=1.0Control group: 15, SD=0.0*Raw scores (number correct)Table 8Verbal memory (Children’s auditory verbal learning test)* [47]Age: 9 yearsModerate NELevel of learningaNE: 98, SD=16Control group: 104, SD=15Immediate recallaNE: 96, SD=18Control group: 105, SD=15Delayed recallaNE: 95, SD=15Control group: 103, SD=16Recognition accuracyNE: 22% delayedbControl group: 27% delayedTotal intrusionsNE: 19% delayedbControl group: 20% delayedImmediate attentionNE: 97, SD=15Control group: 100, SD=15*Mean of standardization sample =100, SD=15, ap < 0.01; bDelayed is below the 17th percentile of the standardization sample
General cognitive development
Infancy (0–24 months)
Eight studies were identified that measured developmental levels between the ages of 12 and 24 months in children with NE [2, 9, 12, 24, 39, 49, 50, 55]. Two of these were performed in low-income, non-western countries [2, 55]. Standardized testing with the Griffiths Mental Development Scales showed that almost all children with moderate NE who had been diagnosed to have CP at the age of 12 months were severely developmentally delayed (Table 1). In these studies, the rates of CP among children with moderate NE varied between 23% [24] and 82% [49]. Children with moderate NE without CP at this age performed in the average range or above, although the range of scores varies widely across studies. In addition, other studies in which the Denver Developmental Screening Test or the Bayley Scales of Infant Development were used reported normal development in children with mild NE, none of whom had CP.
Two to six years, without CP
In all of the selected studies [8, 9, 43, 44, 53], children with mild NE showed average general intellectual abilities that equaled those of a healthy comparison group (Table 2). Although children with moderate NE also performed in the average range, their abilities were significantly below those of children with mild NE and a reference group. The general intellectual abilities of children with severe NE were assessed in only one study. These children showed very low levels of functioning: a mean IQ of 37.1 was found [43]. In two studies, no grades of NE were recorded [9, 53], which hampers interpretation of these data. Rates of average or above-average functioning were 85% and 64%, respectively.
Seven to nine years, without CP
At school age, children with mild NE showed age-appropriate levels of general intellectual ability and performed as well as healthy comprison children on standardized testing [32, 46, 47] (Table 3). Although most children with moderate NE obtained significantly lower scores on intelligence measures than children with mild NE and comparison groups, their ability levels were in the average range.
Educational achievement
Relative strengths and weaknesses in cognitive functioning can become visible when children have to acquire specific scholastic abilities in school. A test of school readiness showed that 5.5-year-old children with mild and moderate NE were not significantly behind compared to comparison children [24]. However, children with moderate NE achieved lower scores in all domains, and 42% of them scored more than 1 SD below the average of the standardization group. Children with severe NE showed low school readiness scores at the age of 5.5 years. Table 4 gives an overview of the results of four studies, which assessed reading, writing, spelling and mathematics/arithmetic in school-aged children [32, 35, 46, 47]. Survivors of mild NE compared to a comparison group were not significantly more often more than one grade behind in reading, spelling and arithmetic. In these domains, all children with severe NE were more than one grade behind the expected level for their age. The group with moderate NE appeared to have distinct difficulties in reading, spelling and arithmetic/mathematics. Writing has been found to be less developed in only one subgroup of one of the studies.
Neuropsychological functions
Studies of specific neuropsychological functions yielded results that are similar to the findings on general intellectual ability and educational achievement (Tables 5, 6, 7 and 8). Children who suffered from mild NE tended to show normal receptive vocabulary and visual-motor integration at 3.5, 5.5 and 8 years of age [43, 44, 46]. As is the case in other cognitive domains, the severe group was impaired, and their ability levels were below those of children with mild or moderate NE and comparison groups. The moderate group showed an ambiguous pattern of performance. Deficits were found in receptive vocabulary, language and visual-motor integration. In contrast, attention and executive functions, visuospatial ability, and memory and learning were normal in a subgroup of 7-year-old children [29]. Although memory and learning did not seem to be impaired in this subgroup, these children had difficulties in the sub-domains narrative memory and sentence repetition, compared to intact memory for faces, names, everyday events and intact orientation. In addition, auditory verbal learning and delayed recall was impaired in a group of 9- and a group of 16-year-old children with moderate NE [31, 47]. Tests of other auditory skills in the former showed intact auditory attention and concentration, language and binaural separation, in contrast to impaired binaural integration, and sequencing and labeling [47].
Behavioral problems
So far, few researchers have addressed the issue of behavioral problems in children with a history of NE. Four studies were identified that met the selection criteria. In general, hyperactivity was more often present in children with moderate NE, but not in children with mild NE [32, 35, 44]. In addition, one study that used parent’s observations of their child’s behavior found more problems related to tractability, aggression, passivity and anxiety in a mixed group of children with NE compared to a control group [35]. No differences were found, however, related to attention, habituation and impulsivity. A recent study reported an unexpectedly high proportion of children with moderate and severe NE that developed an autism spectrum disorder [5]. However, the inclusion criteria differed considerably from those of Sarnat. Both groups can be classified as moderate NE when Sarnat’s criteria are applied. In addition, the authors explicitly did not exclude etiologies of NE other than PA.
Discussion and conclusion
In the introduction section of this review, problems were mentioned with regard to the comparability of different outcome studies of NE. Even after careful selection on the basis of predetermined criteria, comparing studies remained difficult. The main reason for this was the way test results were presented. In many cases, outcome of the children was reported in two or more categories, e.g., normal/mildly delayed/abnormal. One of the problems that arise due to such categorization, is the subjectivity involved in determining the cut-off points of the categories. Often, these cut-off points seemed to be chosen arbitrarily or not justified at all. Another issue that adds to this problem is the use of different tests of development, cognitive ability or behavioral problems. One can hardly compare, for example, the proportions of children with developmental delay in two studies when these are represented by a score below 85 on the Bayley Scales of Infant Development in one of the studies and by a score below 70 on the Griffiths Mental Developmental Scales in the other. To enable readers to compare studies, results of significance testing between subgroups of children with NE and comparison groups were added to the tables whenever available.
Children with a history of NE who suffer from CP at the age of 12 months are often also severely developmentally delayed. Estimates vary between 77% [12] and 90% [24]. Of the children without CP, those with mild NE have an excellent prognosis. Their intelligence, educational achievement and neuropsychological functions are comparable to healthy peers at least until middle childhood.
Children with severe NE appear to be impaired in every cognitive domain. They are less intelligent and perform worse at school and at neuropsychological tests than both healthy controls and children with mild or moderate NE.
As was expected, the children with moderate NE seem to form the most heterogeneous group. In general, these children are significantly less intelligent than children with mild NE and healthy peers, but their scores are still in the average range. And although these children are equally ready to go to school as other children at the age of 5.5 years, when measured in middle childhood, they perform less well in the domains reading, spelling and mathematics. On the basis of these results, it is possible that children with moderate NE fall behind in school, but do not show clear deficits at standardized testing. Another issue is raised by the relatively broad range of performance of children with moderate NE. Calculating mean IQ scores masks individual differences, which are large in groups of children with moderate NE. Therefore, means do not describe this group well, and it would be more informative to look at the distribution of scores. The description of neuropsychological functions in this review is based on six studies, in which five different tests were used in five different samples of children with moderate NE. These studies do not provide sufficient and comparable data to draw reliable conclusions about patterns of neuropsychological strengths and difficulties. The heterogeneous nature of the group of children with moderate NE raises the question whether these children really form one group. Marlow et al. [32] divided a group of children with moderate NE according to the severity of the symptoms of the NE. The group with the least severe neonatal symptoms showed a few discrete neuropsychological difficulties, but was more intelligent and performed better in school than the group with the most severe neonatal symptoms. Marlow suggested a dose-response effect of hypoxia within this group with moderate NE. This idea of a continuum of casualty with regard to morbidity has already been proposed by Low [29]. He suggested a theory of a critical threshold of asphyxia beyond which brain damage occurs. Beyond this threshold, a spectrum of minor and major deficits would occur. So, although the classification system of Sarnat [50] appears to be very useful for the prediction of the outcome of children with mild or severe NE, in case of moderate NE this qualitative system lacks predictive power. Future research should therefore focus on developing a reliable, more gradual model of prediction.
Another issue regarding prediction of outcome after NE also mainly concerns moderately affected children. It is still unclear whether NE has a general effect on intellectual abilities or leads to specific patterns of cognitive strengths and difficulties. This review shows that systematic, detailed research of neuropsychological function is still scarce. A few case reports, which are not included in this review, have shown specific impairments in episodic memory (memory for events), with relatively intact semantic memory (memory for facts) [21, 58]. However, these findings, which were based on small numbers of children, do not justify generalization to all children with NE and should therefore be replicated in larger, well-defined samples.
The introduction section of this review referred to possible associations between brain areas that are vulnerable to perinatal hypoxia and psychopathology, such as ADHD and autism. Long-term behavioral outcome of survivors of NE has received very little attention compared to neurological outcome and general intellectual ability. Only a few studies looked at the behavioral consequences of NE. Those studies found elevated rates of hyperactivity and autism in children with moderate NE. Evidence of hyperactivity and impulsive behavior has also been found in animal studies of PA [1, 60]. These findings suggest that, in addition to cognitive impairment, children with NE could be at risk of developing behavioral problems. Because of the clinical implications this would have, future research should include behavioral monitoring of all children with NE.
Long-term follow-up of children with NE has been advocated before, including the optimal ages of assessment of specific functions [48]. Recommendations include assessment of general cognitive and adaptive behavioral functioning at preschool age, and, in some cases, school achievement tests and a neuropsychological screening at school age. The authors of this review would like to go one step further in recommending standard screening for behavioral problems in all children with NE, regardless of their level of cognitive functioning. In addition, detailed neuropsychological assessment is recommended especially in all children with moderate NE.
From this review it is concluded that general intellectual, educational and neuropsychological outcomes are consistently positive for children with mild NE and negative for severely affected children. However, children with moderate NE form a more heterogeneous group with respect to outcome. On average, intelligence scores are below those of children with mild NE and age-matched peers, but within the normal range. Difficulties have been found in the domains reading, spelling and arithmetic/mathematics. So far, neuropsychological functioning of children with NE has received relatively little attention. The studies that were selected for this review have yielded ambiguous results in children with moderate NE. A few studies suggest elevated rates of hyperactivity in children with moderate NE and autism in children with moderate and severe NE. Therefore, behavioral monitoring is required for all children with NE. In addition, systematic, detailed neuropsychological examination is needed especially for children with moderate NE.
|
[
"neuropsychology",
"asphyxia neonatorum",
"behavioral problems",
"hypoxia, brain",
"hypoxia-ischemia, brain"
] |
[
"P",
"P",
"P",
"R",
"M"
] |
Bioprocess_Biosyst_Eng-4-1-2214824
|
Feasibility of an in situ measurement device for bubble size and distribution
|
The feasibility of in situ measurement device for bubble size and distribution was explored. A novel in situ probe measurement system, the EnviroCam™, was developed. Where possible, this probe incorporated strengths, and minimized weaknesses of historical and currently available real-time measurement methods for bubbles. The system was based on a digital, high-speed, high resolution, modular camera system, attached to a stainless steel shroud, compatible with standard Ingold ports on fermenters. Still frames and/or video were produced, capturing bubbles passing through the notch of the shroud. An LED light source was integral with the shroud. Bubbles were analyzed using customized commercially available image analysis software and standard statistical methods. Using this system, bubble sizes were measured as a function of various operating parameters (e.g., agitation rate, aeration rate) and as a function of media properties (e.g., viscosity, antifoam, cottonseed flour, and microbial/animal cell broths) to demonstrate system performance and its limitations. For selected conditions, mean bubble size changes qualitatively compared favorably with published relationships. Current instrument measurement capabilities were limited primarily to clear solutions that did not contain large numbers of overlapping bubbles.
Introduction
Accurate and representative bubble sizes and distributions are used to characterize biochemical processes containing gas-in-liquid dispersions, specifically processes for industrially important fermentation products. Quantification of bubble sizes and distributions during fermentation is important to establish mass transfer characteristics (based on gas–liquid interfacial area) when oxygen transport to cells across gas–liquid interfaces becomes a limiting factor. In these situations, there is a direct influence of bioreactor parameters that affect bubble size, such as agitation, on culture yields. Thus, measurements of bubble sizes and distributions are useful for biochemical process optimizations.
Direct application of bubble measurements in fermentation is based on linking bubble size to operational parameters, such as agitation and/or airflow rate, that influence gas–liquid volumetric mass transfer coefficients [1]. Such measurements can confirm that selected agitation/aeration conditions do not lead to impeller flooding and can quantify bubble distribution changes for (1) geometrically similar fermenters with scale up and (2) different fermenters of similar scale possessing different operating conditions, geometry, or bioreactor internals. Owing to the automated and fast (<2 min) nature of the Envirocam™ bubble measurement and analysis, it is a potential control tool for maintaining agitation/airflow rate set points to obtain desired bubble sizes and distributions for mass transfer, for gas hold up, or for minimizing bubble damage to cells. For example, utility costs for electricity to drive agitators or air compressors might be optimized further for large production scale fermenters.
Historical and currently available bubble measurement systems using photographic methods have been previously summarized, and the references cited within describe several applications of bubble analysis in clear solutions [2]. The accuracy, representation, and simplicity of bubble size measurements improve when these measurements are performed on-line and in situ rather than off-line using broth samples. A newly developed novel on-line and in situ bubble measurement device is described and evaluated.
Novel in situ probe measurement system
EnviroCam™ probe
The EnviroCam™ (Enviroptics; Colmar, PA) probe consisted of a hermetically sealed shroud constructed of 316 l stainless steel for wetted product-contact parts (Fig. 1). A silicone o-ring was used for the Ingold-style tank port insertion. The camera module itself attached to this shroud, and thus it was readily relocatable to other vessels without disturbing process integrity. There were no cracks or crevices since the shroud was entirely welded and helium leak-tested. The overall length of shroud was set at 6 in. for consistency with insertion lengths of other commercially available, in situ probes (such as pH, dissolved oxygen, or optical density probes). This length avoided interfering with fermenter internals, such as larger diameter hydrofoil impellers, but extended far enough into the well-mixed fermenter zone. Since bubble sizes and distributions vary according to their distance from high shear impeller zones, measurements were taken at a constant insertion length from the fermenter sidewall. External shroud attachments were minimized to permit location in tight areas, in some cases near platform structural steel supports. The captive retaining nut was obtained from Mettler-Toledo (Ingold; Bedford, MA) to match existing Ingold ports.
Fig. 1Environcam™ shroud diagram
Various prototype shrouds and camera modules were developed to evaluate different hardware options. A high-strength sapphire window was integrated into the 316 l stainless steel shroud via a gold brazed joint rather than epoxy to provide improved robustness. The sapphire window was constructed at a 75° angle to offset the 15° angle of the fermenter’s Ingold port. This angle resulted in the window being parallel to the vertical tank wall upon insertion. Owing to the low surface tension of sapphire, some small bubbles adhered to the window surface, particularly at the low agitation speeds used for animal cell cultivation. Using the 75° vertical orientation, as well as raising agitation speeds, reduced but did not eliminate this accumulation.
A ceramic disc backscreen provided a measured path length. Initially, light from a 150 Watt halogen lamp was reflected internally into the shroud backscreen via a high performance liquid light guide, but this arrangement did not result in uniform illumination. Next LEDs (red–orange 640–720 nm wavelength) were incorporated directly into the shroud backscreen for more uniform illumination, which produced bubble images with improved border definition. To minimize the impact of distortion in the depth and width of the measurement volume, a diffuser was installed in the backscreen. A calibrated reticle, consisting of two intrusion lines (180° apart with a gap of 7.5 mm), emanating from the circumference and heading towards its center, became the standard for in situ calibration. This reticle was located on the sapphire window itself rather than the backscreen to avoid interferences in opaque media.
With a backscreen present on the shroud, a defined path length or notch was introduced. Prototype shrouds were constructed to change the notch size between 1/4 and 1.0 in., but experiments were conducted with a 1/2 in. notch unless otherwise noted. Due to the small width of the support (3/8 in.), the orientation of the notch (i.e., up, down, sideways with the opening facing with or against agitator rotation) was deemed not to prevent larger bubbles from entering the measurement field nor influence measured bubble sizes or distributions. The externally located connector for the LED power supply provided confirmation of the internal notch position. Regardless of notch size, the bottom of the notch was located 1 in. from the insertion end of the shroud (Fig. 1). As notch size decreased, (1) the number of bubbles viewed per frame decreased, resulting in less overlap for high bubble volume broths, and (2) the path length decreased, providing more light for opaque solutions. In contrast, larger notch gaps may be required for lower bubble volume broths, such as animal cell cultivations.
EnviroCam™ imaging system
The EnviroCam™ imaging system is shown in Fig. 2. One imaging system, containing the camera module (about tenfold higher cost than shroud), could be used with multiple shrouds. Two high performance, 1.3 megapixel, 8 bit, monochrome cameras were tested, each of which attached directly to the shroud: (1) a monochrome camera (PL-A741, Pixelink; Ottowa, ON), equipped with a CMOS image sensor which operated at a shutter speed of 1/10,000 s (1/10,000 s = 100 μs selected), and (2) a monochrome camera (EC-1380, Prosilica; Burnaby, BC), equipped with a CCD image sensor (IXC-285 Exview, Sony; Tokyo, Japan) which operated at a shutter speed up to 1/100,000 s (1/25,000 s = 40 μs selected). Primary magnification of 2× (resulting in a system optical magnification of 30×) was implemented using one additional screw-on lens. The faster shutter speed associated with the CCD camera required additional light, which was sufficiently provided by the backlit LED geometry of the shroud. A partially telecentric 55 mm lens (Computar; CBC, Commack, New York) was selected for the CCD camera which reduced the viewing angle, but achieved magnification errors of <1% while providing high resolution and contrast with low distortion. The CCD camera became the preferred configuration consistent with the designs of other currently available photographic systems [2].
Fig. 2Environcam™ system diagram
The number of pixels [horizontal (H) × vertical (V)] was altered using user-defined region of interest (ROI) controls. Higher pixel numbers increased resolution but reduced the number of frames per second (fps). For the selected CCD camera, the hardware was set at 20 fps using 1,024 × 1,024 pixels with a minimum pixel size of 6.45 μm and a pixel depth of 8 bits (without additional magnification via screw-on lenses). This fps rate of 20 resulted in a time scale of about 25 s for the initial image scan of 500 frames, which was small relative to expected changes in bubble size characteristics during the time course of a typical fermentation. Bubble residence time through the measurement field (1/2 in. = 12.5 mm path length) was quantitatively estimated to be about 1.3–3.3 ms near the impeller blade tip (impeller tip speed of 3.8–9.4 m/s) and likely was up to an order of magnitude slower away from the impeller. Consequently, it was not necessary to increase the frame speed further above 20 fps, which corresponded to a characteristic measurement time of 50 ms/frame. The bubble residence time in the 1/2 in. notch measurement field was qualitatively determined to be 40–70 ms by comparing common features of subsequent frames. Thus, it was necessary to skip at least four to five frames to ensure bubbles were not counted more than once.
For the CCD camera, pixel size was set at 20 μm/pixel. With the 2× magnification added, and other geometric and positional factors considered, it was reduced to 7.5 μm/pixel. A small annulus of the viewing range (5 pixels wide) was discarded to reduce grey scale variations near the edges. The measurement area was 81 mm2, and the measurement volume was 993 mm3 (5.07 mm diameter, 5.07 mm height). A number-based hold-up estimate of 50–150 bubbles/cm3 for a bubble size range of 0.5–1 mm was calculated [2]. Thus, for the calculated measurement volume of nearly 1.0 cm3 about 50–150 bubbles were expected per frame image. The actual number of bubbles per image ranged up to 1,000 or more when larger numbers of smaller bubbles less than 0.5 mm were detected.
In comparison with these values selected above for the Environcam™ system, prior researchers have used 4–30 μm/pixel and 512 × 512 pixels for gas/liquid and liquid/liquid bubble/drop dispersion measurements [3–6]. For cell morphology measurements, 21 μm/pixel was typical [7], and 512 × 512 pixels was the most common for recent studies [2].
The remainder of the Environcam™ hardware system is shown in Fig. 2. The computer itself consisted of an Intel Pentium 4 class notebook PC (XP operating system) with 1 GB RAM (NEMA rating as required by the facility) and a graphic card minimum display resolution of 1,280 × 1,024 (with 32 bit color). The video output was the Firewire/IEEE 1394 interface, a universal interface that allowed direct connection of the camera to the PC laptop. A USB port was used for computer control of the external light source supplied by a Nema 4× power supply, which was powered only when images were being acquired by the camera. The lens mount was a standard C-mount adapter used for camera installation to lenses and to standard microscopes. A CD/DVD R/W drive, USB drive, and/or some other type of network interface was used for archiving images and/or data files, preferably with at least 80 GB of storage.
The operating temperature range was limited to 0–50°C for the camera module; consequently, it was not attached to the shroud during vessel sterilization. The shroud LED-operating temperature ranged up to 80–90°C, but its non-operating temperature ranged up to 120°C. All other shroud components were steam-sterilizable, including the glass diffuser. Thus, the shroud could be sterilized with the vessel if the LEDs remained unpowered. A LED-power supply kill switch based on a bimetallic temperature sensor was installed with a trip value of 90°C and reset value of 60°C. When it was powered, typically intermittently for 10 s per frame measurement cycle or continuously for video stream, the LED was required to be submerged in liquid as a heat sink for adequate cooling. LED lifetime was 100,000 h assuming the non-operating temperature remained less than 120°C; raising it a few degrees above this level for effective sterilization might sacrifice some lifetime, however. Shrouds were heat-tested using a 15–30 s temperature ramp from ambient to 130°C, held for 1 h, then returned to ambient temperature. No significant degradation, as measured by pixel light output, was observed after 50–60 temperature cycles. In addition, an actual sterilization was conducted successfully with the shroud in a pilot scale fermenter (180 l volume, 122°C, 40 min hold time). Based on this performance, shrouds were expected to withstand about 100 sterilizations of 45–60 min hold times for about a 3-year life span, assuming a 2-week batch length.
The 1951 USAF resolution target was used to evaluate the pure video resolution of the computer monitor, which was influenced by the quality of the video graphics card. Resolution was measured at 32 line pairs/mm, the reciprocal of which resulted in a resolution of 31 μm/pixel, significantly higher than the camera resolution of 7.5 μm/pixel. Thus, the accuracy of the image display did not diminish the accuracy of the photographs obtained.
The software platform was customized based on National Instruments’ (Austen, TX) LabVIEW Graphical Development Environment (version 8.2). The system’s main screen is shown in Fig. 3. The image analysis software had the following key features:Measured frames were included in analysis until the number of desired objects to be measured (a user input) was attained. Specifically, for a measurement time of 25 s (corresponding to 20 fps) which generated 500 frames, typically every tenth frame was skipped (i.e., measured frames were taken every 0.5 s) and 50 frames were measured to identify at least 500 bubbles. This number was consistent with other currently available measurement systems described in the literature [2]. This approach also was similar to skipping every 25th frame when measuring floatation cell moving aggregate sizes [8].The background was subtracted from the original image, which avoided repeatedly counting bubbles lodged on the sapphire window and to remove fixed blemishes in the backscreen. (This approach is similar to subtracting a background image without cells from the original image [9]). The background image was calculated by averaging all 500 frames obtained during the measurement period so that moving objects were filtered out to obtain a clear composite background image. This approach was speedy (800 ms for 500 frames). In addition, the background image was readily reconstructable from stored images if necessary. A user-selectable alternative algorithm was developed to construct a background composed of the brightest pixel of each of the 500 frames for each position, assuming that no pixel should be any brighter than a background pixel. The time to construct the background using either of these methods was similar.Edge enhancement techniques, based on a contrast threshold, were applied to convert grey images to binary black and white images so that the outside perimeters (or diameters) of bubbles were readily identifiable via object recognition.The analyzed image output was prepared using the calibrated reticle (measurement shown in red in Fig. 3) located on the shroud backscreen in the form of a binary mask. In addition, a blue reference grid of 1 × 1 mm squares (Fig. 3) was overlayed to permit the user to approximate bubble sizes on the computer screen.Depending on the application, bubbles of sizes greater than a designated cut-off (e.g., 2 mm in diameter), between 40 and this cut-off, and below 40 μm were identified. Both visual and numerical indications of the classification of objects were developed according to three tiers of predetermined rules:
Green circles represented the first level of filtering. Rules in this first tier were: single bubbles of sizes within the target measurement range, circularity cut-off based on a user-selected tolerance above 1.0 value, and discard of bubbles touching the border. The green circle was the best fit “circle” so there were slight inaccuracies around some of the circumference if the bubbles were not uniformly round. Larger or smaller circular bubbles outside the target measurement range were excluded.Yellow outlines indicated the second level of filtering typically using similar rules as the first level but with more relaxed, user-defined cut-offs. Additional rules also were added in this second tier. Up to three, more or less circular bubbles, that were touching but still individually discernable, were included in the count. Specifically, the ratio cut-off of cluster (i.e., more than one adjacent bubble overlapping) area to the calculated equivalent area for completely separated bubbles of the cluster was based on a user-selected tolerance below 1.0. Bubbles in this category were further differentiable using additional capabilities included in the LabVIEW software.Red outlines comprised a third level that indicated bubbles located by the system that did not pass filtering levels 1 or 2.White outlines comprised a fourth level of classification. This fourth tier contained non-circular blobs and large irregularly shaped or greatly overlapping bubbles. The percentage of viewing area occupied by blobs was estimated and measured frames were omitted based on a user-selected cut-off (e.g., if greater than a target percentage of 75% of the measurement area was comprised of blobs, the frame was discounted).The relevant statistical quantities were calculated (e.g., arithmetic, geometric, and/or Sauter mean diameters), and the appropriate histograms were displayed [2]. A customized SQL program extracted data into csv files (containing all raw data and analysis calculations). These csv files had a practical limit of including information about 65,000 bubbles, typically an amount greater than required for a single measurement cycle. Data was imported into an Excel template containing graphs and some statistical quantities as part of the measurement cycle (typically requiring 3 s to 2 min), and then into other statistical programs (e.g., Sigmaplot, LabView modules) for more advanced analysis should the user desire.The front panel image containing the configuration and user-adjustable controls was archived so that an identical set up could be reproduced in the future, if desired.Tiff files (still frames) and/or “avi” (video stream) files were archived, but avi files were reconstructable from Tiff files to save storage space. A video playback rate of 5 fps appeared appropriate. The actual measurement frames utilized for analysis were saved separately for ease of review. Minimum storage requirements for these unanalyzed frames were estimated at about 50 mB for each 500 frame measurement cycle (versus 4.5 GB for analyzed frames) with the entire analysis reconstructable from saved data. Thus, the ability to retrieve images and recheck/reanalyze results readily existed.For measurement frames, a review panel of the analysis sequence was created from archived files (Fig. 4). The analysis progression panels were reviewed manually to confirm accuracy.
To obtain an image of acceptable contrast, the user varied the camera shutter speed (length of exposure to light), brightness level (luminescence of image on LCD monitor), aperture (amount of light reaching camera lens), and gain (amplification of signal strength). Several parameters in the software were adjustable to ensure optimal image analysis, and the values selected could be archived. As a first step, threshold was varied automatically by the software to determine the value at which the maximum number of bubbles was identified. Subsequently, various filtering strategies available in the software were examined. FFT (fast Fourier transform) was found to potentially improve accuracy for measurements in water, but it was not implemented for the present work. Localized thresholding [10] was found to improve accuracy for bubble measurements in broths, and it was incorporated. The largest inscribable diameter was taken to calculate the bubble diameter [11].
Fig. 3Example main screen displayFig. 4Analyzed composite for a bubble frame (180 l DI water with 0.001% P2000, agitation 100 rpm, air flowrate 300 rpm) showing a original raw camera image, b binary image output with edge enhancement, c binary mask prior to application to image, d image after binary mask, e analysis output image containing colored circles identifying objects according to the three tiers of rules, and f instrument screen including histogram
Key features of the EnviroCam™ gas bubble measurement system are summarized in Table 1 in a format readily comparable to those assembled for prior and currently available optically based measurement systems [2].
Table 1Envirocam™ characteristicsCamera and mode of attachmentMagnification/calibrationIllumination/shutter or frame speedImage measurement and analysis methodCCD monochrome camera (Prosilica) probe inserted into an in situ Ingold-fitting shroud; 32 line pairs/mm30×/internal reticleLEDs in back screen (back lit)/20 fps; 1/100,000 s (1/25,000 = 40 ms selected)National Instruments’ (Austin, TX) LabVIEW Graphical Development Environment as basisMeasurement time per conditionNumber of objects per measurementMeasurement error/size rangeNumber of images (pictures/frames) per measurement5–25 s data acquisition; <2 min data analysis>500 (typically up to 10,000)<10% for monodisperse beads/60–2,000 μm50 (20–200 bubbles/picture)
Bead calibration
A calibration was conducted using beads of a known size distribution in applicable size ranges for bubbles. Calibration beads were measured individually by manually moving the bead past the notch or in a small volume mixed solution contained in a magnetically stirred beaker. Vortexing was minimized by reducing the stirring speed. Settling of the larger glass microsphere beads (e.g., specific gravity of 2.46 g/cm3) was reduced by performing the measurements in a perfluoropolyether (1,800 Da, specific gravity of 1.88 g/cm3, Fomblin 06/6 MFY06/6BB, Solvay Solexis; Bollate, Italy). In measurement frames, beads had a softer, lighter, less-contrasted outline when compared with bubbles, which slightly reduced the effectiveness of the optical imaging system without additional filters. Thus, measurement of beads was limited to sizes ≥60 μm, slightly higher than the lower limit for bubbles.
First, monodisperse spherical polyethylene calibration beads (Baltec; Balzers, Lichtenstein), 0.125 ± 0.002 in. (1.2% rsd) or 3.175 mm in diameter, were measured to evaluate the higher end of the expected Environcam™ measurement range. (Other reported options for calibration beads were expanded polystyrene beads with mean diameters of 5.47 and 3.1 mm and a density of 30 kg/m3 [12], 1.5 mm ball bearings [13], and red spherical particles of diameters 109 and 644 μm [7].) Using the external halogen lamp light source and the CMOS camera system, a value of 3,416 μm (single measurement) was obtained which was 6.7% higher than the standard’s value. Using the LED light source installed in the shroud backscreen and the CCD camera system, a value of 3,220 μm was obtained which was 3.9% lower than the standard’s value.
Next, high precision, NIST-tracable, glass micro-spheres (Whitehouse Scientific; Chester, UK) were selected in various monodisperse and polydisperse sizes. Distributions (in the format of count versus bin size) were fit to 4-parameter Sigmoidal or 4-parameter modified Gaussian distributions using Sigmaplot (Systat; San Jose, CA) (Table 2). These distribution equations better fit the bead measurement data than distribution equations used for bubbles [14, 15]. Monodisperse bead sizes, ranging from 20 to 600 μm in increments of at least 15 μm (corresponding to the instrument resolution), were tested (Table 3). Measured size values agreed reasonably well with the manufacturer’s data values. Measured ranges for 90% of the bead size range typically enveloped the manufacturer’s data, with differences most likely owing to the nature of the fitted distributions. The distribution of polydisperse beads over most of the Environcam™ measurement range (50–2,000 μm) also was measured and compared with the analyzed manufacturer’s distributions for three size ranges (Table 4). Reasonable agreements in the distribution shape were obtained. The accuracy of these Environcam™ measurements showed that differences in means were observed and standard deviations were lower. In contrast, other researchers using calibration beads for optically based instrument systems reported that means were identical to the standard’s values but standard deviations were higher (manufacturer: 109 ± 5 μm, measured: 109 ± 10 μm; manufacturer: 644 ± 13 μm, measured: 644 ± 24.8 μm) [7].
Table 2Distributions used to fit calibration bead measurement data (from Sigma plot software)DistributionProbability density functionModified Gaussian 4-parameter (G)Y = Yo + a exp[−0.5(abs(X − Xo)/b)2]Sigmoidal 4-parameter (G)Y = Yo + a/(1 + exp[−(X − Xo)/b])Table 3Comparison of manufacturer size (Whitehouse Scientific, Chester, UK) with measured size range for monodisperse beadsManufacturer’s size data (μm)Manufacturer’s data for 90% of beads within given range (μm)EnviroCam™ measured size (μm)EnviroCam™ data for 90% of beads within given range (μm)22.81 ± 0.7821.46–24.23Not able to be measured38.38 ± 0.5436.5–39.6Not able to be measured59.63 ± 1.057.1–62.266.3 ± 0.05 (G)58.8 ± 0.03 (S)53.4–64.3 (S)83.43 ± 0.8779.7–87.582.5 ± 0.11 (G)74.9 ± 0.09 (S)64.2–85.2 (S)98.10 ± 2.894.4–102.898.28 ± 0.82 (G)90.64 ± 0.27 (S)76.0–105.4 (S)155.8 ± 1.5151.4–163.1163.9 ± 0.21 (G)156.04 ± 0.2 (S)143.7–167.4 (S)200.9 ± 1.9196–206200.69 ± 0.49 (G)194.15 ± 0.07 (S)187.6–200.4 (S)258.6 ± 5.9251.4–265.6259.77 ± 0.16 (G)253.3 ± 0.07 (S)246–261 (S)297.9 ± 3.9289.7–309.3303.1 ± 0.05 (G)295 ± 0.07 (S)275.3–316.3 (S)361.6 ± 9.9344–376366.5 ± 0.3 (G)359.3 ± 0.86 (S)337.4–380.8 (S)405.9 ± 8.7396–419411.7 ± 0.07 (G)403.9 ± 0.2 (S)385.9–420.2 (S)589.0 ± 6572–615586.7 ± 0.4 (G)578 ± 0.09 (S)551.9–606.6 (S)Envirocam™ measurements fitted to modified Gaussian 4-parameter equation (G) and sigmoidal 4-parameter equations (S) using Sigmaplot software. All r2 values >0.99 unless otherwise noted. Beads of sizes 22 and 38 μm were not measurable owing to blurry edgesTable 4Comparison of manufacturer size (Whitehouse Scientific, Chester, UK) distribution with measured size distribution for polydisperse beadsBead distribution (μm) Size at fixed percentiles (μm)1025507590StandardEnviroCam™StandardEnviroCam™StandardEnviroCam™StandardEnviroCam™StandardEnvirocam™500–2,000796 ± 10744.5, 746.8936 ± 9849.3, 852.01,098 ± 111,069.6, 1,074.01,335 ± 321,232.0, 1,250.01,618 ± 401,442.5, 1,508.0150–650244 ± 4245.8306 ± 5304.6362 ± 5368.2424 ± 4423.8527 ± 18493.950–35094 ± 3.5104.7119 ± 1.2130.9151 ± 2.5166.5190 ± 4.8202.8237 ± 6.0252.8Envirocam™ measurements fitted to sigmoidal 4-parameter equation using Signmaplot software. Duplicate runs executed for 500–2,000 μm distribution
Gas bubble data analysis
The smallest bubble measurable was 30 μm in diameter, based on an expected maximum system resolution of 7.5 μm/pixel. A minimum of 2 pixels were needed to quantify the radii, and radii were used for diameter calculations based on initial ease of programming. Since the radius accuracy was ±1 pixel, the relative standard error for a 30 μm bubble was 50%, dropping to 25% for a 60 μm bubble. If required for other applications, smaller objects down to 20 μm might be measured using (1) two pixels to determine diameter or (2) using a back-calculated diameter based on equivalent surface area since only 2 pixels were required to define surface area.
Bubble diameter measurements, generated using the image analysis software, were compared with those generated manually for 225 objects from a single frame image taken using the CCD camera system and a 1/2 in. gap shroud. Results demonstrated that the greatest percentage error was observed with smaller bubbles in the diameter range of 38–57 μm. This error generally decreased with larger bubble diameters. Manually measured diameters were slightly longer than image analysis measurements greater than 99.7% of the time, most likely due to a small amount of shadowing around bubble edges. The measurement error was 0.1% for the calibration line itself. Coincidentally, for some actual bubble measurement conditions, the number of objects rose considerably for bubbles 38 μm in diameter. This rise may be caused by the greater measurement inaccuracy at this size; thus small changes in size were not detectable.
Since the minimum hardware resolution was 15 μm based on at least a 1 pixel change in radius, buckets in the distribution possessed 15 μm increments. This distribution resolution was consistent with the 25–90 μm value reported for an optical bubble measurement system with a 20× magnification [3, 4] and similar to the 15 μm value reported for a stationary flatbed scanner set up [7]. The system’s ability to measure bubble size changes due to typical differences in fermenter operating conditions was estimated based on this resolution. Specifically, an agitation rate increase from 100 to 150 rpm decreased the size of 300 μm bubbles to 184 μm, and an agitation rate increase from 100 to 125 rpm decreased the size of 300 μm bubbles to 230 μm (D α N−1.2; [2]). Both of these changes were substantially greater than the 15 μm resolution, and thus bubble size differences caused by agitation rate changes were believed detectable using this measurement system. For other operating conditions where the effect on bubble change was less pronounced [2], these differences were not expected to be as readily detectable.
Incremental and cumulative distributions were plotted, typically as number or cumulative percentage versus bubble size, respectively. Arithmetic, Sauter, and geometric means and standard deviations (as applicable) then were calculated according to published methods [2]. In addition, the 5% largest bubbles and 5% smallest bubbles were discarded and the arithmetic and Sauter means recalculated. Using these 90% cut-offs significantly reduced the skewing and variability impact of smaller numbers of larger bubbles in the size distributions on the arithmetic and Sauter mean diameter calculations (Fig. 5a, b). Owing to the nature of the geometric mean calculation, there was little effect when the 90% cut-off was used (Fig. 5c).
Fig. 5Arithmetic (a), Sauter (b), and geometric (c) means and standard deviations (as applicable) calculated using all the bubbles and for the 90% cutoff. Calculations conducted on 500 frames using every tenth frame for 50 frames and each calculation point covering 50 different frames, starting from successive initial frames
The total number of bubbles required to be analyzed per measurement greatly affected the measurement and data acquisition times (MAT and DAT, respectively), plus the data storage requirements. Based on a survey of published techniques [2], the target number was 500 bubbles per measurement condition. This value was confirmed based on determining that means and standard deviations changed only minimally (less than 4%) when bubbles numbers below and beyond 500 were analyzed (specifically 300, 400, 500, 750, 1,000, 1,500 and 2,000), also suggesting that as few as 300 bubbles were sufficient. The desired total bubble number target may be achieved based on a higher number of bubbles per frame (50) and a smaller number of frames (10), or a lower number of bubbles per frame (10) and a larger number of frames (50), with care taken not to count bubbles more than once in successive frames in either case.
Measurement reproducibility was evaluated using bubble measurement data from a 15,000 l fermenter by analyzing every tenth frame for 500 frames starting at the first frame until 2,000 bubbles were obtained, then re-analyzing every tenth frame starting from the second frame, then again, starting from the third through tenth frames. The relative standard deviations of the averages typically were under 7.5% for the 90% cut-off Sauter mean diameter and under 4% for the 90% cut-off arithmetic and geometric means. Thus, the sampling of frames used for analysis was representative of the total number of frames collected.
Discarding the blob area from the measurement was not felt to significantly influence bubble diameter means and distributions, nor impact subsequent predictions of volumetric mass transfer coefficients. The overall blob interfacial area per unit volume, a, was relatively low compared to that of the selected measured bubbles (e.g., for 5 cm blobs, a = 1/500,000 μm; for 5 mm bubbles, a = 1/5,000 μm) owing to the large diameters of the blobs. In addition, the percentage of larger bubbles (i.e., 2.5–5.0 mm diameter) was relatively low (<10%) for most typical agitator/sparger system set points and configurations. Their overall contribution to interfacial area also was low. Similarly, although bubbles below the 40 μm limit of detection had a large, collective, interfacial area, their small individual volume resulted in fast oxygen depletion by the broth, making them a less substantial source of oxygen supply [16]. Thus, accuracy for bubble sizes outside the target range of 40 μm–2 mm was considered less critical. Consequently, the maximum bubble size cut-off of 2 mm used for measurement appeared acceptable since larger bubbles tended to be non-spherical as previously described [5], and smaller bubbles (<2–2.5 mm diameter) in aqueous solutions previously were shown to behave as particles with rigid interfaces [17]. Consequently, significant numbers of irregularly shaped bubbles were not obtained within the target measurement range and the use of bubble diameter to approximate size was reasonably accurate.
Test systems
Water/media/broth
Liquid test media consisted of deionized water (DIW) and the following solutions, all prepared using DIW: 50 vol% glycerol (viscosity of 12–14 Pa s × 103 [1, 18] vs. 1.01 Pa s × 103 for water [19]), 0.01–2 ml/l (0.001–0.2 vol%) P2000 (polypropylene glycol 2000; Dow, Freeport, TX), 0.01–2 ml/l (0.001–0.2 vol%) antifoam C (Sigma), mono/di potassium phosphate (effect of pH at constant ionic strength of 0.037 M: 5 g/l monopotassium phosphate at pH 4.4 versus 6.4 g/l dipotassium phosphate at pH 8.8; effect of ionic strength at constant pH of 7.0: 0.037 M (5 g/l) versus 0.0037 M (0.5 g/l) monopotassium phosphate), 0.1–2.5 g/l (0.01–0.25 wt/vol%) cottonseed flour (size of 91% of particles <74 μm; Pharmamedia, Traders Protein; Memphis, TN), sterility medium [6 g/l yeast extract (Biospringer; Milwaukee, WI), 6 g/l cerelose (glucose monohydrate), 1 ml/l P2000], and low-protein animal cell culture media (LPKM, JRH Biosciences; Lenexa, KS). Varying viscosity by using a glycerol solution varied surface tension by only a few dynes/cm [20], but antifoam addition varied surface tension without appreciably changing viscosity. These solutions were selected to vary the physical and optical properties of the liquid phase sufficiently to detect changes in bubble size measurement ability. Temperature was 22°C, except for when its effect on bubble size was examined for set points of 15, 22 and 37°C (range of 22°C). Back-pressure was zero except when its effect on bubble size was examined (0.3–1.5 kgf/cm2).
To ensure that contamination did not cause changes in surface active properties [21], test media that supported growth were sterilized if the testing period was expected to be greater than 2–3 h. Also, since vessel cleanliness (i.e., residual surfactant concentrations) affected bubble density and size, all vessels were first rinsed thoroughly with hot water, and then a known amount of antifoam (0.001 vol% P2000) was added to DIW. These model systems formed the basis for initial tests of actual fermentation broths of microbial (filamentous bacterial culture, Amycolatopsis fastidiosa) and animal cell (suspension-adapted CHO) cultivations.
Fermenters equipped with open pipe/jet spargers
When performing measurements in agitated fermenters equipped with open pipe/jet spargers (1 in. opening at 180, 600, and 15,000 l scales; 0.7 in. opening at 1,500 l scale), it was apparent that under certain conditions the bubble distribution was bimodal. Some images consisted of smaller spherical bubbles along with very large irregularly shaped “blob” bubbles, presumably owing to gas entrainment from vortexing and possible impeller flooding at higher aeration rates relative to agitation rates. Quantification of the discarded blob area from each of these conditions assisted in identifying the onset of flooding conditions, and possibly was directly related to the gas hold up (even under conditions in which individual bubble diameters cannot be discerned).
Bubble size measurements were conducted at the 180, 600, 1,500, 15,000 l scales, spanning an 80-fold range in scale, all in geometrically similar fermenters. For some combinations of agitation and airflow rates, the bubble density was too high to clearly obtain individual bubble diameter measurements. This limitation to low agitation and airflow rates, and void fractions under 2%, was similar to that observed for other optically based systems [6, 22].
Qualitative pictures of the effect of agitation and airflow rate at the 180 l scale for a fermenter equipped with a Rushton impeller are shown in Fig. 6. As expected, higher agitation rates and thus power inputs (estimated at 0.24, 0.83, and 1.96 hp/1,000 l; [23]) created larger numbers of smaller bubbles and higher airflow rates (0.55, 1.1 and 1.67 vvm) increased the number and size of bubbles. Increases in both agitation and airflow rates increased bubble density often to the degree that measurement of individual bubbles was precluded.
Fig. 6Qualitative bubble size as function of agitation and airflow rate for open pipe sparger (1 in.) at the 180 l scale (0.001% P2000 in DIW)
Qualitative pictures of the effect of agitation and airflow rate at the 15,000 l scale for a fermenter equipped with a Rushton impeller are shown in Fig. 7a. Trends were similar to those obtained at the smaller scale, considering vvms (0.067, 0.13 and 0.2 vvm) and power inputs (estimated at 0.02, 0.12, and 0.36 hp/1,000 l; [23]) were considerably lower. The 90% cut-off Sauter mean diameter is shown by Fig. 7b, following expected trends with agitation and airflow rate. A comparison of the arithmetic, Sauter and geometric mean diameters, both with and without the 90% cut-off, is shown by Fig. 7c with expected trends generally observed. The relationship between bubble size and agitation rate at the 15,000 l scale was quantified for the 90% cut-off of the Sauter mean diameter as D α N−0.45, and compared to literature results, D α N−1.2 [2]. The lower dependence observed on agitation rate likely was due to the fact that the measured power inputs were lower than those used for the published correlation.
Fig. 7Analysis of bubble data at the 15,000 l scale, containing 0.001% P2000 in DIW, with an open pipe sparger (round 2 in. ring with four jets of id 1 in. each): a Qualitative bubble size as a function of agitation and airflow rates. b Bubble Sauter mean diameter (90% cutoff) as a function of agitation and airflow rates. c Arithmetic, Sauter and geometric mean comparisons as function of agitation rate at airflow rate of 2,000 lpm
At the 180 l scale, 50 vol% glycerol solution was tested using a fermenter equipped with Rushton impellers at constant agitation and airflow rates. The effect of agitation and airflow rate increases on bubble size and distribution was qualitatively similar between DIW (Fig. 6) and 50 vol% glycerol (Fig. 8). The number of smaller bubbles qualitatively was somewhat greater for 50 vol% glycerol, particularly at lower agitation rates, which was inconsistent with the expectation that bubble size increase with solution viscosity. However, at the same 100 rpm agitation rate, the power per unit volume was likely higher for the glycerol solution, since the impeller Reynold’s number, NRe, was ∼1.8 × 104 in the laminar flow range for 50 vol% glycerol and ∼2.2 × 105 in the turbulent range for DIW.
Fig. 8Comparison of bubble sizes in DIW and 50 vol% glycerol, containing 0.001% P2000, as a function of agitation and airflow rate at the 180 l scale
At the 180 l scale, various solutions (differing in ionic strength, pH, temperature, pressure) were tested using a fermenter equipped with Rushton impellers at constant agitation and airflow rates, but notable qualitative changes in bubble distribution were not apparent. As greater amounts of P2000 (beyond 0.001%) were added to DIW, observations became progressively more limited to lower agitation and airflow rates before becoming too dark due to overlapping bubbles. Gas hold up was indirectly measured for aeration and agitation conditions at the 180, 600 and 1,500 l scales, for conditions, which had high discarded area percentages (Fig. 9a, b). Hold up was estimated using the discarded area for each frame of the measurement and then averaging the results. Discarded areas were not calculated precisely as individual bubble areas, however, and standard deviations of average discard areas were around 50% at lower values, dropping to 10% at higher values. Trends in the values of discarded areas indicated that hold up, as well as bubble residence time, increased with greater agitation rates and airflow rates as observed by others [6]. In addition, as airflow rate increased, the impact of higher agitation rates on discard areas decreased suggesting impeller flooding, which at 0.5 vvm occurred for the 1,500 l scale but was not evident at the 600 l scale owing to lower gassed power draws. Specifically, higher discard areas were evident at the 600 l scale for hydrofoil versus Rushton impellers, consistent with greater measured gassed power decreases and gas hold ups [23, 24].
Fig. 9Discarded bubble area as a function of a agitation rate at constant airflow rate and b airflow rate at constant agitation rate for 180 l fermenter containing 0.001% P2000 in DIW
Fermenters equipped with ring spargers
Bubble measurements in fermenters with open pipe spargers exhibited several bubbles in each frame, often in swarms. In contrast, bubble measurements in fermenters with ring spargers represented the opposite extreme in which only a few bubbles were present in each frame. These latter tests were conducted using purified water in 75 and 750 l geometrically similar bioreactors, spanning a tenfold size range. Fermenters were equipped with A315 impellers and a ring sparger with holes drilled 1/32 in. in diameter on the top surface of the sparger ring. Only a few bubbles were observed regardless of notch orientation (left, right, down, top). It was believed that the fewer bubbles observed per frame were due partially to lower gas hold ups, typically 0.02 vvm, but also the relative spatial placement of the probe at the level of the sparger ring.
As the airflow rate was increased to its higher range values, more similarly sized bubbles were observed. In contrast, higher agitation rates caused more surface air entrainment; thus greater numbers of large, irregularly shaped bubbles were present in the frames. Higher airflow rates resulted in bubbles similar in nature to those observed at lower airflow rates, all governed by the holes in the ring sparger. In some cases owing to the slower agitation rates, smaller-sized (geometric mean of 95 vs. 200 μm) bubbles collected on surface of the shroud’s sapphire window, and these bubbles needed to be distinguished from the freely moving bubbles. Increases in silicone antifoam from 0.001 to 0.2% did not increase the low number of bubbles observed.
Use of an in situ bubble measurement system in fermentation
The ability of the Envirocam™ to measure in opaque solutions was examined using several model systems:
For the 180 l fermenter equipped with an open pipe sparger, the effect of Pharmamedia on bubble images was investigated. When 20 g/l Pharmamedia and 2 ml/l P2000 was added to DIW (100 μs shutter speed, 20 gain, 185 brightness and 8 aperture), the contrast of bubble edges decreased to an unacceptable level. The particles caused granularity on the screen and blurred bubble edges, making detection difficult using the 1/2 in. gap shroud. When 20 g/l Pharamedia and 0.5 ml/l P2000 was added to DIW using the 1/4 in. gap shroud to decrease path length (100 μs shutter speed, 0 brightness, 31 gain, 8 aperture), bubble contrast was improved, but the resulting opaque solution appeared still too high for reliable analysis.
In addition, there was limited ability to obtain bubbles with edges sufficiently sharp enough for accurate measurements when 50% diluted Amycolatopsis broth (initial dcw of 3 ± 0.5 g/l) was tested, despite raising settings to maximum values. The use of localized thresholding was attempted to sufficiently sharpen the bubble edges by reducing fuzziness for analysis, but it was only partially successful (Fig. 10a). Note that the bubble density is considerably lower for this image taken at 200 rpm and 300 lpm, compared with images taken in 0.001% P2000 at 200 rpm and 100 lpm (Fig. 6). This comparison demonstrated that the presence of broth potentially improves the agitation and aeration range over which the Envirocam™ can measure bubbles should the optical limitations of the broth on the measurement be mitigated.
Fig. 10Bubble photographs in fermentation broth: aAmycolatopsis fastidiosa broth with 1% P2000, 3 g/l dry cell weight diluted 1:1 with DIW, 180 l scale and open pipe sparger, 200 rpm, 300 lpm; b LPKM medium with 0.001% antifoam C, 40 l scale and ring sparger, 50 rpm, 4 lpm
For the 75 l fermenter with the ring sparger, images taken in cell-free LPKM medium (Fig. 10b) were compared with those taken in a 10-day CHO cell culture containing animal protein-free medium (containing 1 g/l Pluronic F68 and no antifoam) with about 7 × 106 cells/ml at 45% viability. Despite the low cell density of this broth, bubble edges remained fuzzy preventing accurate size analysis.
Summary and future considerations
A novel in situ bubble size and distribution measurement device was developed. The bubble measurement instrument design strategy permitted one camera module to be attached to a shroud, and thus one sensor (the camera) was able to be moved to multiple locations without disturbing the fermentation process. The small size and flexibility of the camera attachment permitted it to be readily relocatable. An Ethernet connection for the camera can further reduce the extent of the field hardware, and thus permit one camera system to be even more transportable.
The measurement system was tested at the pilot scale, both in clear and opaque model systems, which included fermentation broth. Experimental data for mean bubble size changes versus expected behavior qualitatively compared favorably with published relationships for selected conditions [2]. Quantitative comparisons were more difficult to establish owing to limitations in bubble size measurement capability at the higher agitation and airflow rate ranges at which these published correlations often were established.
At this time, application of the EnviroCam™ bubble measurement system appears limited to clear solutions that do not contain large numbers of overlapping bubbles. Different liquids (e.g., water, cottonseed flour, microbial broth, animal cell broth) possess different UV spectrum and light scattering properties, as well as varying surfactant properties, which influence bubble size and hold up. These differences suggest that some adaptation of the measurement system is necessary when moving from system to system. Specifically, agitation and aeration rate combinations which produce too many bubbles for measurement in a model DIW system may produce acceptable amounts of bubbles in a fermentation broth owing to changes in surfactant levels, but the ability to distinguish these bubble edges is diminished. Further expansion of the versatility and range of this instrument is the subject of future efforts, but key approaches being considered involve the further examination of available LabView filtering techniques to process bubble images and the use of smaller notch sizes.
|
[
"in situ",
"measurement",
"bubbles",
"size",
"distribution",
"fermenter",
"image analysis"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Intensive_Care_Med-3-1-2092441
|
High incidence of acute lung injury in children with Down syndrome
|
Objective Acute respiratory tract infection is a common reason for hospitalization in children with Down syndrome (CDS) and is characterized by a high morbidity. The severe course of disease in CDS may be related to a higher incidence of acute lung injury (ALI). This study evaluated the incidence of ALI and acute respiratory distress syndrome (ARDS) in mechanically ventilated CDS.
Introduction
Down syndrome (DS) is the most common chromosomal abnormality and occurs in 1 of 800 live births [1]. Although survival beyond the first year of life has increased considerably in recent decades, children with DS (CDS) still have a shorter life expectancy than those without DS [2]. Acute respiratory tract infection is a common reason for hospitalization in CDS and is characterized by high morbidity [3]. In children with DS it has been suggested to be related to pharyngeal incoordination predisposing to aspiration in combination with concomitant immunodeficiencies [4, 5]. High morbidity in CDS with respiratory disease may also be related to a higher incidence of acute lung injury (ALI). The aim of this study was to evaluate the incidence of ALI and acute respiratory distress syndrome (ARDS) in children with DS who were admitted to the pediatric intensive care unit (PICU) and needed mechanical ventilation. Preliminary data from this study were presented at the 35th Congress of the Society of Critical Care Medicine [6].
Methods
This study compared the occurrence of ALI/ARDS in two groups of patients requiring mechanical ventilation because of respiratory insufficiency. The first group consisted of all 24 consecutive children with genetically confirmed DS admitted to our PICU for mechanical ventilation between January 1998 and July 2005. The second group consisted of all 317 mechanically ventilated children without DS admitted to our PICU between January 1998 and January 2001. Baseline characteristics for CDS and the control group are shown in Table 1. The PICU at Emma Children's Hospital is a 16-bed, tertiary, multidisciplinary unit serving the greater Amsterdam area. Data on the incidence rate of ALI/ARDS in the second group have been published previously [7]. In both groups patients who were admitted for postoperative care directly following a surgical procedure, and those with a cardiac left to right shunt were excluded. Patients were retrospectively evaluated for ALI/ARDS during the second 24 h after admission according to the American–European Consensus Conference criteria [8]. Gas exchange criteria for ALI and ARDS were considered to be met if the PaO2/FIO2 ratio was less than 40.0 kPa for ALI and less than 26.7 kPa for ARDS in at least two consecutive measurements (> 8 h apart). A radiologist blinded to clinical information reviewed the chest radiographs for the presence of bilateral infiltrates. The presence of left heart failure was assessed based on echocardiographic results and/or clinical information. Disease severity on admission was expressed by the Pediatric Risk of Mortality (PRISM) II score, which depends on diagnosis and clinical parameters in the first 24 h of admission [9]. Sepsis was defined as systemic inflammatory response syndrome in the presence of clinical evidence for infection. Lower respiratory tract infection was defined as clinical infection with radiological evidence of alveolar consolidation. Upper respiratory tract infection was defined as infection of the oral and nasal airways, larynx, trachea, and/or bronchi without signs of lower respiratory tract infection.
Table 1Baseline patient characteristics and cause for respiratory failureDown syndrome (n = 24)Control group (n = 317)pMedian age (months; range)15.2 (0–127.2) 6.1 (0–217.7)0.39 cMales14 (58.3%)199 (62.8%)0.67 dMedian PRISM II score (range)13.5 (0–40) 9.0 (0–47)0.19 cSepsis 3 (12.5%) 52 (16.4%)0.78 eLower respiratory tract infection18 (75%)130 (41%)0.001 dUpper respiratory tract infection 1 (4.2%) 23 (7.3%)0.96 eCongenital heart disease 1 (4.2%) 22 (6.9%)0.99 eCNS disorders a 0 53 (16.7%)0.04 eOther b 1 (4.2%) 37 (11.7%)0.50 eaCNS disorders: convulsions, infections, asphyxia, contusion, cerebral
hemorrhage; bOthers, e. g., trauma, cardiomyopathies, electrolyte disturbances,
metabolic disease; c Mann–Whitney test; d Pearson's χ2 test; e Fisher's exact test
The comparability of patient characteristics for both groups was tested by means of a chi-square test or Fisher exact test for dichotomous data and a Mann–Whitney U test for continuous data that were not normally distributed. The association between DS and ALI/ARDS was expressed as an odds ratio (OR) with 95% Confidence Interval (CI). Other well known cause for ALI/ARDS and thus possible confounders for the incidence of ALI/ARDS among CDS were identified [8, 10] and evaluated with bivariate logistic regression analysis. Statistical significance was set at 5%.
Results
The criteria for ALI were met in 14 of 24 CDS (58.3%) and in 41 of 317 controls (12.9%; odds ratio 9.4, 95% confidence interval 3.9–22.6). The criteria for ARDS were met in 11 of 24 CDS (46%) and in 21 of 317 controls (7%; odds ratio 11.9, 95% confidence interval 4.8–29.8). There were no deaths in the CDS. In the control group ten patients with ALI died during admission to the PICU. The results of the logistic regression analyses are presented in Table 2. The odds for the incidence of ALI in CDS were hardly altered by adjusting disease severity (PRISM II score), lower respiratory tract infection, or sepsis.
Table 2Estimated odds ratio (OR) for ALI (Down/control) after adjustment for possible confounders (CI, confidence interval)Possible confounderOR95% CINone 9.43.9–22.6PRISM II score 9.63.9–23.6Sepsis12.24.9–30.6Lower respiratory tract infection10.84.3–26.9
We excluded patients who died in the first 24 h after admission. This might have caused a selection bias. Therefore PICU deaths during the first 24 h of admission were analyzed for ALI/ARDS: 36 in the control group, none of whom met the ALI/ARDS criteria. No CDS died in the first 24 h after admission. Thus the odds of developing ALI in CDS are about 9 times those in controls, and this cannot be attributed to the presence of confounders.
Discussion
This study found a very high incidence of 58% of ALI and 46% of ARDS in mechanically ventilated children with DS. This is significantly higher than the incidence of 13% of ALI and 7% of ARDS in the general pediatric population on mechanical ventilation in our unit. Likewise, Randolph et al. [11] reported an incidence of ARDS of almost 8% in a mechanically ventilated pediatric population.
The results of this study need to be interpreted with caution due to the retrospective design and the relatively low number of CDS. The results need to be confirmed in a prospective multicenter study in a larger cohort. Due to the small number of CDS admitted annually to our unit we included CDS for a longer period of time (1998–2005) than the control group (1998–2001). This may interfere with the comparability of the study groups. However, medical care on our unit did not change substantially during this extended period. In both study periods respiratory care was based on the same clinical protocols with a low tidal volume ventilation strategy. Identical ventilators were used. In addition, the disease severity (expressed as mean PRISM II scores) did not differ between 2001–2005 and 1998–2001. Therefore we have no reason to assume that the incidence of ALI in mechanically ventilated patients changed substantially during the 2001–2005 period.
Despite the high incidence of ALI/ARDS we found no mortality in our population of CDS. Others have reported a mortality rate of almost 5% in children with ARDS [11, 12]. Although this remains speculative, it is possible that CDS are more susceptible to progression to ALI/ARDS despite less severe underlying conditions, without concomitant mortality. In addition, there were only three patients with sepsis and none with trauma as underlying cause of ALI/ARDS in the CDS group, conditions associated with a high mortality rate. However, ALI/ARDS is associated with high morbidity in CDS. The need for ventilatory support was 21 ± 21 days in CDS with ALI vs. 8 ± 5 days in CDS without ALI (p = 0.03). Long-term effects of ALI on morbidity in CDS are unknown and remain to be defined by lung function measurements. The high mortality rate among patients with ALI in the control group might be explained by the relatively high number of patients (n = 4) with severe cerebral damage. This has been shown to be a strong predictor for mortality in ALI [13].
The mechanism leading to the high incidence of ALI in CDS remains uncertain. Initial ventilator settings have recently been shown to be a risk factor for the development of ARDS [14]. In this single-center study the ventilatory strategies for CDS did not differ from those in other patients and were based on the same low tidal volume protocol. Recently it has been shown that apoptosis or programmed cell death plays a pivotal role in the pathogenesis of injurious states of the pulmonary system such as of ALI/ARDS [15, 16]. The higher incidence of ALI/ARDS may be caused by an elevated rate of apoptosis in CDS. An enhanced level of apoptosis in DS has been shown in several cells including neurons [17, 18], thymocytes [19], and granulocytes [20]. It has been suggested that increased apoptosis in DS cells is related to an inability to deal with oxidative stress, leading to accumulation of reactive oxygen radicals [21]. One could speculate whether this is related to an imbalance in the antioxidant/oxidant status in patients with DS due to increased levels of superoxide dismutase 1, an important enzyme in the antioxidant pathway that is encoded for on chromosome 21 [22, 23]. To what extent this is associated with an increased susceptibility to develop ALI/ARDS deserves further investigation. From this point of view our findings are not only of scientific interest but may also be of clinical relevance. For example, ventilatory strategies or oxygen therapy might need to be reevaluated and adjusted for CDS.
In conclusion, we found an unexpected high incidence of ALI and ARDS in CDS. The explanation for these findings remains to be elucidated.
|
[
"down syndrome",
"acute respiratory distress syndrome",
"mechanical ventilation",
"epidemiology",
"cohort studies",
"child"
] |
[
"P",
"P",
"P",
"U",
"R",
"U"
] |
Theor_Appl_Genet-2-2-1544375
|
Coffee and tomato share common gene repertoires as revealed by deep sequencing of seed and cherry transcripts
|
An EST database has been generated for coffee based on sequences from approximately 47,000 cDNA clones derived from five different stages/tissues, with a special focus on developing seeds. When computationally assembled, these sequences correspond to 13,175 unigenes, which were analyzed with respect to functional annotation, expression profile and evolution. Compared with Arabidopsis, the coffee unigenes encode a higher proportion of proteins related to protein modification/turnover and metabolism—an observation that may explain the high diversity of metabolites found in coffee and related species. Several gene families were found to be either expanded or unique to coffee when compared with Arabidopsis. A high proportion of these families encode proteins assigned to functions related to disease resistance. Such families may have expanded and evolved rapidly under the intense pathogen pressure experienced by a tropical, perennial species like coffee. Finally, the coffee gene repertoire was compared with that of Arabidopsis and Solanaceous species (e.g. tomato). Unlike Arabidopsis, tomato has a nearly perfect gene-for-gene match with coffee. These results are consistent with the facts that coffee and tomato have a similar genome size, chromosome karyotype (tomato, n=12; coffee n=11) and chromosome architecture. Moreover, both belong to the Asterid I clade of dicot plant families. Thus, the biology of coffee (family Rubiacaeae) and tomato (family Solanaceae) may be united into one common network of shared discoveries, resources and information.
Introduction
Coffee is an important international commodity, ranking among the five most valuable agricultural exports from developing countries (Food and Agriculture Organization, http://apps.fao.org). Moreover, production and processing of coffee employs more than 25 million people worldwide (O’Brien and Kinnaird 2003). Despite its economic importance, coffee has received little attention with respect to molecular genetics and genomics research. As of December 2004, only 1,570 nucleotide and 115 protein sequences from coffee had been deposited in GenBank with the majority of those sequences derived from leaf ESTs. Many of the remaining sequences correspond to enzymes in the caffeine biosynthesis pathway—the most extensively studied pathway in coffee (Moisyadi et al. 1998; Ogawa et al. 2001; Mizuno et al. 2003; Uefuji et al. 2003).
Commercial coffee production relies mainly on two closely related species: Coffea arabica and Coffea canephora, accounting for approximately 70 and 30% of worldwide coffee production, respectively (Herrera et al. 2002). Although C. canephora accounts for a lower total proportion of the coffee market than does C. arabica, it is the main source for soluble coffee, which is consumed widely throughout the world. C. canephora is a diploid (2n=2x=22), outcrossing and highly polymorphic species native to central Africa, but which has expanded, through cultivation, especially to western Africa, Indonesia and Vietnam (Wrigley 1988). In contrast, C. arabica is believed to be a recently derived tetraploid (2n=4x=44) native to a small region of what is now Ethiopia. C. arabica is now grown widely throughout the world.
The goal of the current project was to increase the genetic and molecular knowledge of coffee through the generation and annotation of an EST database using high throughput single-pass 5′ sequencing of cDNAs derived from leaf, pericarp and seed tissues from a set of C. canephora varieties. Special emphasis was given to sequencing cDNAs from different stages of seed development, both to shed light on this important, but not well understood aspect of plant development and to capture as many genes as possible involved in determining the final chemical composition of seeds which constitute the commercial product. As a result, the EST database reported herein is, to our knowledge, the largest public database of seed-derived ESTs (White et al. 2000; Suh et al. 2003).
The EST database was used to derive a coffee unigene build, which was subsequently subjected to functional annotation. The coffee unigene build was compared with the model species Arabidopsis. Arabidopsis is the only fully sequenced dicot genome and hence provides a complete set of predicted genes against which to compare the EST-derived unigene set of coffee. However, coffee and Arabidopsis belong to different plant families (Rubiaceae and Brassicaceae, respectively) which are distantly related phylogenetically and which diverged from their last common ancestor approximately 94 million years ago (MYA) (Gandolfo et al. 1998; Crepet et al. 2004) (Fig. 1).
Fig. 1Dendrogram depicting phylogenetic relationships of coffee to other higher plant taxa (based on Chase et al. 1993)
The plant family most closely related to coffee in which extensive sequencing has been conducted is Solanaceae (Fig. 1). In this family, comprehensive EST databases have been developed for tomato, potato, pepper, eggplant and petunia (http://www.sgn.cornell.edu/) (Hoeven et al. 2002; Ronning and Stegalkina 2003; Lee et al. 2004). Both Rubiaceae and Solanaceae belong to the Asterid I clade of dicots, and based on existing fossil evidence, are thought to have diverged from one another approximately 50 MYA (Gandolfo et al. 1998; Crepet et al. 2004) (Crepet personal communication) (Fig. 1). The closer taxonomic affinities of coffee and Solanaceae (e.g. tomato) are paralleled by a number of striking botanic and genetic similarities, including the production of fleshy berries, a similar genome content (C=950 and 640 Mb for tomato and coffee, respectively) (Hoeven et al. 2002), similar basic chromosome number (x=12 for tomato and most other Solanaceae; x=11 for coffee) and similar chromosome architecture with highly condensed pericentric heterochromatin and decondensed euchromatin at the pachytene stage of meiosis (Rick 1971; Pinto-Maglio and Cruz 1998). For these reasons, the coffee unigene set was also compared against a series of Solanaceae EST-derived unigene sets.
Materials and methods
Library construction
Source of tissues
C. canephora was farm-grown in east Java. Plant tissue was frozen in liquid nitrogen, transported on dry ice and stored at –80°C until RNA extraction. The specified tissues (leaf, pericarp, young cherries with seed and pericarp mixed, middle stage seed and late stage seed) were collected from five different varieties Table 1). The maturation period of the C. canephora varieties under study is approximately 9–11 months, from pollination to ripening. The early stage cherry, middle stage seed and late stage seed were collected between 18 to 22 weeks, 30 weeks and 42 to 46 weeks after pollination, respectively.
Table 1Characteristics of the 5 cDNA libraries used to develop the coffee EST databaseLibrary nameTissueVarietiesAverage insert size, kbGood quality ESTsLeafLeaves, youngBP4091.5±0.68,942PericarpPericarp, all developmental stagesBP358, BP409, BP42, BP961, Q1211.4±0.58,956Early stage cherryWhole cherries, 18 and 22 week after pollinationBP358, BP409, BP42, Q1211.4±0.39,843Middle stage seedEndosperm and perisperm of seeds, 30 week after pollinationBP409, BP961, Q1211.4±0.310,077Late stage seedEndosperm and perisperm of seeds, 42 and 46 week after pollinationBP358, BP409, BP42, BP961, Q1211.4±0.39,096
RNA and mRNA isolation
Total RNA was extracted using phenol/chloroform (Rogers et al. 1999) and further treated with DNase I (RNase-free) and purified using an RNeasy Kit (Qiagen, Valencia, CA 91355). Messenger RNA was extracted from total RNA with PolyTrack mRNA Isolation Systems (Promega, Madison, WI 53711).
cDNA libraries
Directional cDNA libraries were constructed with 3–5 μg of mRNA with the ZAP-cDNA Gigapack III Gold Cloning Kit (Stratagene, La Jolla, CA 92037). The average insert length was estimated by PCR in 36 randomly selected clones from each library and ranged 1.2–1.5 kb (Table 1).
Sequencing
Bacteria, containing coffee cDNAs, were cultured in 384-well plates and cDNA inserts subjected to 5′ end sequencing at the BioResource Center at Cornell University (http://www.brc.cornell.edu). The average size of quality reads was 613 bp with a maximum of 1,037 bp.
Sequence quality processing
EST sequences were base-called and screened for vector sequences using PHRED software (Ewing et al. 1998). The longest stretch of overall high quality (PHRED score over 15 which corresponds to over 98% confidence) of each sequence was identified. PolyA repeats were trimmed to at most 20 bp and any sequence past the PolyA (mostly low quality sequence) was discarded. After the trimming, the sequences were screened against the E.coli K12 genome to remove any bacteria contamination. The remaining sequences were screened for minimum length (150 bp) and maximum allowed ambiguity (4%) and low complexity (60% of the sequence are of the same nucleotide, or 80% of same two nucleotides, which indicate error in sequencing).
Unigene assembly
Unigene sets were built by combining the sequences from all five coffee cDNA libraries. Clustering was performed using a program developed at the Sol Genomics Network (SGN at http://www.sgn.cornell.edu), which relied on a custom pre-clustering algorithm, and on the CAP3 program for contig generation (Huang and Madan 1999). The preclustering algorithm clustered sequences using a Smith Waterman type algorithm with initial word matching. The command line settings for CAP3 were as follows: -e 5000 -p 90 -d 10000 -b 60. The -e, -d and -b options are set such that the assembler disregards them or minimizes their effect. The -p option increases the sequence identity necessary on overlaps to 90 from a default of 75, which were found to be not stringent enough. Sequences were also checked for length, complexity and contamination. The builds were uploaded to the database, where each unigene was assigned a unique unigene ID.
Annotation
Protein prediction
The most likely coding frame and the corresponding transcribed peptide for each unigene were generated using ESTScan, a program that detects coding regions in EST sequences and corrects some sequencing errors (nucleotide substitution, deletion/insertion, erroneous stop codon, etc) (Iseli et al. 1999). As a hidden Markov Model (HMM) based program, ESTScan requires a training set of high quality coding sequences to generate a model. However, very few high quality coffee coding sequences are publicly available (less than 50 full length in GenBank). Therefore, we used as a training set of 483 nuclear genes from the closely related species, tomato (http://www.ebi.ac.uk/embl). Table 2 compares the coffee unigene assembly and translation (using ESTScan) with that of tomato, using the same training tomato set and parameters. The fact that the two give such similar results, gives confidence that the coffee EST-derived unigenes are being efficiently translated via the ESTScan software using the tomato training set (Table 2).
Table 2Comparison of the coffee and tomato EST databases derived from use of ESTScan calibrated with the same tomato training set (see Materials and methods for details)TomatoCoffeeTotal unigenes30,57613,175Average unigene length, bp774678Unigenes with coding regions96%95%Average length (bp) of predicated peptides569556Average ESTScan score409346
BLAST matches between coffee unigenes and other sequence databases
Sequence matches for coffee unigenes against other sequence databases were obtained using BLAST (Altschul et al. 1997) (see Results and discussion section for details of each BLAST analysis). The target databases were:The GenBank non-redundant (NR) protein and dbest dataset (NCBI, http://www.ncbi.nlm.nih.gov).The predicted Arabidopsis proteome and cDNA databases which were downloaded from the Arabidopsis Information Resource (TAIR, http://www.arabidopsis.org and ftp://ftp.arabidopsis.org/home/tair/).The Solanaceae EST-derived unigene sets, including tomato (Solanum. lycopersicum) (184,860 ESTs, 30,576 unigenes), potato (Solanum. tuberosum) (97,425 ESTs, 24,932 unigenes), pepper (Capsicum annuum) (20,738 ESTs, 9,554 unigenes), petunia (Petunia hybrida) (3,181 ESTs, 1,841 unigenes) and eggplant (Solanum melongena) (11,479 ESTs, 5,135 unigenes), all of which can be accessed at SGN (http://www.sgn.cornell.edu). To estimate ribosomal, chloroplast and mitochondrial contamination, the coffee unigene set was also screened against the Arabidopsis complete mitochondrial genome (NC_001284), tobacco ribosomal rRNA gene (5.8s AJ012367, 18s AJ236016 and 25s S52185) and the complete tobacco chloroplast genome (Z00044) sequences from GenBank.
Functional annotation based on predicted peptides
ESTScan-predicted coffee peptides were subjected to InterPro Scan annotation, which integrates the most commonly used protein signature databases (PROSITE, PRINTS, Pfam, ProDom, etc.) together with their associated scanning methods for protein domain analysis (Apweiler et al. 2001; Zdobnov and Apweiler 2001). Based on the domain annotation, GO accession of the unigenes were assigned using interpro2go conversion file from the GO consortium (http://www.geneontology.org, also available at http://www.ebi.ac.uk/interpro).
Functional categorization based on gene ontology
GO annotations were formatted for input into the GOSlim program and the output parsed to count the occurrence of each GO category. GOSlims are ‘slimmed down’ versions of the ontologies that allow a high-level view of gene functions. The GOSlim file and program were obtained from the Gene Ontology Consortium at http://www.geneontology.org.
Gene family analysis
The predicted protein sequences for the coffee unigene set and the Arabidopsis protein set were combined into a single file, formatted as a blast database using formatdb, and run with BLASTP (protein vs. protein sequence) against itself with option m8 for output. The resulting file was used as the input for the tribeML program. TribeML (Enright et al. 2002) formats the clusters, such that each cluster was given in a tab delimited file, one cluster per line. Simple scripts were used to parse the information to detect the largest gene families, coffee specific families, and families that showed large expansions in coffee.
Results and discussion
Generation of coffee EST database and unigene set
A total of 62,829 cDNA clones, derived from mRNA from five different tissues, were subjected to 5′ sequencing. After quality evaluation (see Materials and methods for details), the database was reduced to 46,914 high quality ESTs, averaging 613 bp in length. This corresponded to approximately 9,000 high quality ESTs from each of the five cDNA libraries (Table 1). These high quality ESTs were assembled into 13,175 unigenes with average length of 678 bp (ranging 150–2,714 bp), among which 7,272 (55%) were singletons and 5,903 (45%) were contigs (Fig. 2). Of the contigs, the majority (87%) was represented by two to ten ESTs while some (13%) were comprised of more than ten ESTs (Fig. 2). All trace files, sequences and derived unigenes/annotation can be found at the Solanaceae Genomics Network (SGN) website (http://www.sgn.cornell.edu) and will be submitted to GenBank at the time of publication.
Fig. 2Histogram depicting the distribution of EST content for all coffee unigenes. Numbers above bars equals the number of unigenes represented in each
Differentiating between paralogs and alleles
The coffee ESTs were derived from five C. canephora varieties (Table 1), however, unigene assembly for EST reads does not automatically distinguish between polymorphic alleles and highly similar paralogs. C. canephora is known to be a highly allogamous, heterozygous and polymorphic species, raising the possibility that some of the EST-derived unigenes might actually represent different allelic forms of the same gene. In order to evaluate how well paralogs were distinguished from allelic polymorphism in the unigene assembly, BLASTN (nucleotide vs. nucleotide sequences) was performed amongst all of the coffee unigenes and the results parsed to identify the best non-self matches. As a control, BLASTN was performed in a similar manner on a complete set of Arabidopsis predicted genes, including both coding and UTR regions obtained from TAIR (http://www.arabidopsis.org). Since the Arabidopsis database was derived from the complete genome sequence of a single, inbred stock (Columbia), the entire gene complement is present without the complication of allelic variation (initiative). For both coffee and Arabidopsis, pair-wise combinations for which the homology extended over more than 70% of the sequence were recorded. The sequence similarity for each of these combinations was plotted for both coffee and Arabidopsis (Fig. 3).
Fig. 3Plot depicting the sequence identify of the most similar match for each coffee unigene as compared with all other coffee unigenes. As a control, a similar analysis is shown for Arabidopsis genes (see Results for details)
The Arabidopsis plot showed two peaks, one with low identity (~87%) and the other with higher identity (over 99%) with 6.4 and 1.6% of genes falling into the low and high identity peaks, respectively. Like Arabidopsis, the coffee plot also showed two peaks, one with lower identity (around 91%) and the other with high identity (over 99%). These two peaks corresponded to 6.2 and 0.8% of the total unigenes, respectively. ESTs corresponding to ten pairs of coffee unigenes, from the >99% peak, were used as probes on genomic southern hybridizations to determine whether the matching pairs were truly duplicated in the coffee genome (paralogs) or rather allelic (single copy). For eight of the ten pairs, the paired ESTs hybridized the same single copy gene on southerns (data not show). Thus, a significant number (approximately 80%) of the unigenes in this peak are likely to be allelic. However, this category represented only a small portion of the coffee unigenes (0.8%). A similar experiment was performed with 11 ESTs from 11 pairs of unigenes in the second, lower homology peak (around 91% identity, see Fig. 3). In this case, the majority (8 out of 11) was determined to represent true paralogs (two or more copies in the genome) (data not shown). Thus for further discussions, it is assumed that the majority coffee EST-derived unigenes do in fact correspond to unique coffee genes.
Functional annotation of coffee EST-derived unigenes
Predicted coffee proteins
ESTScan (see Materials and methods for details) was able to identify protein-coding sequences in 12,534 coffee unigenes (95% of total unigenes), among which 1,515 (12%) were putatively full-length (starting with ATG and ending with a stop codon). Due to the cDNA library construction method, the unigenes were biased for the 3′ end—57% of the unigenes covered the 3′ end (ending with a stop codon) while only 36% covered the 5′ end (starting with ATG). Of the 5% of unigenes from which a protein sequence could not be predicted, 81% were singletons and the majority (97%) did not match to any Arabidopsis, GenBank non-redundant (NR) or Solanaceae unigene sequences, suggesting that they are not bona fide gene transcripts.
Protein domain annotation
Peptides translated from ESTScan predicted coding sequences were subjected to InterPro Scan (see Materials and methods for details) for protein domain analysis. A total of 1,678 distinct domains were identified in the database, which were present in 4,414 (35%) unigenes. Table 3 lists the 20 most represented InterPro domains found in the coffee genes, as well as comparative statistics for tomato and Arabidopsis genes. Coffee and tomato were very similar with regard to the most frequent protein domains; however, a number of striking differences were noted when comparing coffee and Arabidopsis. Most notably, the following five domains were much more common in coffee encoded proteins than in Arabidopsis: proline-rich regions, tyrosine protein kinases, glucose/ribitol dehydrogenases, ubiquitins and ubiquitin-conjugating enzymes. The latter two categories, involving ubiquitinization, suggest a more active and/or complex system for protein turnover in coffee versus Arabidopsis.
Table 3Twenty most abundant InterPro domains identified in coffee unigene set and comparative statistics for tomato and Arabidopsis genesInterPro accessionDescription% of unigenes (ranking)CoffeeTomato
Arabidopsis
IPR000719Protein kinase1.61.20 (1)3.0 (1)IPR000694Proline-rich region1.30.91 (4)0.003 (1763)IPR002290Serine/threonine protein kinase0.851.10 (2)0IPR001245Tyrosine protein kinase0.691.0 (3)0.15 (311)IPR008271Serine/threonine protein kinase, active site0.610.68 (5)2.6 (2)IPR000504RNA-binding region RNP-1 (RNA recognition motif)0.550.60 (6)0.59 (6)IPR001680G-protein beta WD-40 repeat0.490.51 (8)0.51 (8)IPR001611Leucine-rich repeat0.480.59 (7)0.59 (7)IPR002048Calcium-binding EF-hand0.360.34 (13)0.34 (13)IPR000379Esterase/lipase/thioesterase0.330.43 (10)0.43 (10)IPR001806Ras GTPase superfamily0.320.26 (22)0.43 (70)IPR003579Ras small GTPase, Rab type0.290.23 (27)0IPR0051232OG-Fe(II) oxygenase superfamily0.270.26 (21)0.47 (52)IPR000626Ubiquitin0.270.22 (32)0.40 (89)IPR002401E-class P450, group I0.270.46 (8)0.77 (24)IPR002347Glucose/ribitol dehydrogenase0.260.23 (28)0.33 (110)IPR001005Myb DNA-binding domain0.260.34 (15)1.34 (8)IPR005225Small GTP-binding protein domain0.260.24 (25)0.68 (27)IPR000608Ubiquitin-conjugating enzymes0.260.21 (34)0.19 (221)IPR007090Leucine-rich repeat, plant specific0.250.40 (12)1.07 (11)
Gene ontology annotation
Gene ontology (GO) annotations were made for each coffee unigene based on InterPro domain annotation (see Materials and methods for details). A total of 3,248 unigenes (25% of total unigenes and 74% of unigenes with InterPro domain match) could be assigned a GO annotation. In order to obtain an overview of representation of biological processes by the coffee unigene set, the GO annotations were mapped to the plant GOSlim, a group of selected higher-level categories of the GO ontology (see Materials and methods for details). Figure 4 depicts the biological process GOSlim categories for each of which at least 1% of the coffee unigenes were assigned. Metabolism is the most abundant category, both in terms of number of subcategories (14 subgroups) and number of unigenes in the subcategories (2,541 unigenes and ~80% of the GO annotated unigenes). In addition, signal transduction, cell growth and/or maintenance, response to stress and response to endogenous stimulus were among the most abundant categories.
Fig. 4Comparison of the gene ontology-based gene annotation categories for the coffee EST-derived unigene set, tomato EST-derived unigene set and the Arabidopsis proteome. Figure contains only categories in which more than 1% of the coffee unigenes were assigned. Categories for which coffee differs most significantly from Arabidopsis are shown in underline bold. (1) Cellular processes other than signal transduction and cell growth and/or maintenance. (2) Nucleobase/nucleoside/nucleotide and nucleic acid metabolism other than DNA metabolism and transcription. (3) Protein metabolism other than protein biosynthesis and protein modification. (4) Metabolism other than amino acid and derivative metabolism, biosynthesis, carbohydrate metabolism, catabolism, electron transport, lipid metabolism, nucleobase/nucleoside/nucleotide and nucleic acid metabolism and protein metabolism. (5) Cell growth and/or maintenance other than cell cycle and cell organization and biogenesis. (6) Physiological processes other than photosynthesis, response to stress, response to endogenous stimulus, response to external stimulus and metabolism
A comparison was made between the GOSlim biological process of Arabidopsis, tomato and coffee (Fig. 4). For both the tomato and coffee unigene sets, the GO annotations were based on InterProScan results and approximately 25% of both unigene sets were assigned a GO annotation. In Arabidopsis, the genes are of full length, giving a higher chance of finding functional domains. Moreover, extensive experimental research and manual annotation has been carried out in Arabidopsis, resulting in a higher proportion of genes with assigned GO annotation. Therefore 83% of the Arabidopsis genes are assigned GO annotations. No significant differences were observed in the annotated categories for coffee versus tomato—possibly reflecting their close taxonomic affinity. However, for a number of categories, coffee had significantly different proportions of genes than Arabidopsis. The categories with the largest significant differences (P<0.001, based on Chi-square test) are: carbohydrate metabolism, other metabolism, biosynthesis, catabolism, protein biosynthesis, protein modification and energy pathways. In all cases, coffee had a significantly higher proportion of genes in these categories than Arabidopsis (Fig. 4). It is interesting to note that many of these categories center around the synthesis, breakdown or modification of compounds. One of the hallmarks of coffee is its high-level diversity of primary and secondary compounds, which contributes to the sensory quality of brewed coffee beans. The Rubiaceae family in general contains some of the most diverse species with regard to secondary metabolism and is an especially rich source of alkaloids—a number of which have pharmacological and/or psychotropic properties (Kutchan 1995; Facchini 2001). In fact, the most widely used psychotropic drug, caffeine, comes mainly from coffee. One can speculate that this metabolic diversity is reflected in the relatively high proportion of coffee genes with putative functions related to metabolism.
In silico analysis of unigene expression
Complexity and uniqueness of different stages/tissues
In an effort to determine which tissues gave rise to the more redundant or more complex pools of gene expression, three library/tissue parameters were calculated: (1) the percentage of total coffee unigenes represented by each library/tissue; (2) the percentage of more highly expressed unigenes (represented by ten or more ESTs from the library) in each library/tissue; (3) the percentage of unigenes from each tissue/stage which are unique to that library/tissue (not found in any other cDNA library). The results from these calculations are depicted in Fig. 5. The pericarp and leaf libraries represented more unigenes, contributed to more unique unigenes, and contained fewer highly abundant unigenes than did the early cherry, middle seed and late seed libraries—perhaps reflecting the overall lower complexity of these latter stages/tissues (Fig. 5). Among the seed stages, the middle stage of development gave more novel genes than did the early or late stages of seed development (Fig. 5).
Fig. 5Characteristics of each coffee cDNA library in comparison to the entire coffee EST-derived unigene set. The total unigene and highly expressed unigene categories sum to greater 100% since the same unigene may contain ESTs from more than one library
Differential expression of genes across stages/tissues
The coffee cDNA libraries were non-normalized and deeply sampled, and thus could be used to compare gene expression amongst the five cDNA libraries using statistical method (Audic and Claverie 1997). In all the pair-wise comparison between the cDNA libraries, only a small proportion of the overall unigenes (approximately 5%) showed significantly different (P<0.05) expression levels between two or more libraries (Table 4). The leaf and the early stage cherry libraries showed the highest divergence in gene expression (752 unigenes which is 5.7% of total unigenes, showed significant differential expression) while the leaf and the pericarp libraries had the lowest number of differentially expressed unigenes (384 unigenes and 2.9% of total). Interestingly, the three stages of seed development showed as much differential expression, relative to each other, as they did to non-seed tissues (leaf and pericarp). These results provide strong evidence that throughout the approximately 11 months from pollination to mature seed, a continuously changing cascade of molecular events is unfolding in the developing seed.
Table 4Number of coffee unigenes showing significantly (P<0.05) different expression in pairwise comparisons of cDNA librariesLibraryPericarpEarly stage cherryMiddle stage seedLate stage seedLeaf384752548562Pericarp610458527Early stage cherry602728Middle stage seed585
Highly expressed genes
Table 5 lists the blast annotation, library composition and the best Arabidopsis/Solanaceae gene match for the 20 most highly expressed unigenes from the coffee EST database. Some of these highly expressed genes are common to and highly expressed in all plant species—a good example of this is Unigene 122071, which encodes the small subunit of ribulose bisphosphate carboxylase—a key enzyme in photosynthesis and carbon fixation (Table 5). However, a number of these highly expressed genes are unique to coffee, and many have expression patterns confined to particular tissue/stages, providing a glimpse into some of the potential unique aspects of coffee biology. A brief description of some of these is given below.
Table 5The 20 most highly expressed coffee unigenes: functional annotation and most similar Arabidopsis and Solanaceae homologsCoffee unigene#: annotationBest match (e value/score)EST countArbidopsisSolanaceae Unigene_speciesTotal ESTsLeafPericarpEarly stage cherryMiddle stage seedLate stage seed125230: putative 2s seed storage proteinND243065_tomato (e-103/238)1,219
7
15
21
1,037
139
120912: 11s seed storage proteinAt5g44120 (1e-88/324)228376_tomato (0/802)687
0
3
28
244
412
121707: unknown functionAt1g29050 (1e-139/489)246695_potato (e-163/283)324
2
3
1
149
169
120118: unknown functionAt5g59320 (2e-21/99.8)221585_tomato (e-134/475)292
0
0
3
58
231
124988: unknown functionNDND20458
84
55
1
6
120685: chitinaseAt5g24090 (2e-43/172)214596_tomato (1e-35/84.5)202
99
40
58
0
5
124158: photoassimilate-responsive proteinAt3g54040 (2e-36/149)196924_pepper (2e-39/138)182
1
1
2
150
28119890: unknown functionND204426_pepper (5e-07/52.8)183
0
0
0
183
0123265: ADP-ribosylation factorAt2g47170 (1e-99/359)238338_tomato (0/693)18258
14
65
21
24
124083: secretory peroxidaseAt4g21960 (e-153/537)196145_pepper (0/681)161
55
5
19
49
33
124911: metallothioneinAt5g02380 (0.32/32.3)207464_petunia (2e-06/51.0)163
40
65
30
11
17
119817: chitinaseAt3g12500 (e-103/373)248120_potato (e-148/521)148
0
22
0
0
126
124815: unknown functionAt3g29240 (1e-87/320)227940_tomato (e-146/517)145
2
0
1
0
142
122206: SAM synthaseAt2g36880 (0/711)270415_petunia (0/887)142
1
9
130
0
2
119460: WRKY4 transcription factorAt1g80840 (3e-75/279)237166_tomato (e-137/487)123
0
0
123
0
0
123045: unknown functionAt3g16000 (0.69/31.2)218824_tomato (90.36/33.1)123
81
17
22
3
0
120481: AdoMet synthaseAt4g01850 (0/723)243236_potato (0/886)108
19
27
31
20
11
121265: Mobl/phoceinAt5g45550 (e-119/425)196814_pepper (e-146/513)113
1
0
112
0
0
124791: plasmodesmal receptorAt5g15140 (1e-99/360)203764_pepper (8e-86/314)105
0
0
3
26
76122071: rubiso small subunitAt1g67090 (9e-70/260)207453_petunia (3e-89/297)99
76
8
9
6
0
BLAST match values are given in parenthesesBold numbers indicate library for which the highest number of ESTs were observed for each gene. Italic numbers indicate library for which the number of EST is significantly lower (P<0.05) than the highest
Seed storage protein genes
Unigene 125230: a putative 2S seed storage protein
Unigene 125230 is the most highly expressed gene across the entire coffee unigene set (1219 ESTs) and was the dominant transcript in the middle stage seed library, accounting for 10% of the ESTs at this stage (Table 5). This gene shows high homology to a tomato unigene derived from a developing seed cDNA library, but has no detectable homolog in Arabidopsis (Table 5). Other than the match with tomato, weak homology was also detected for 2S seed storage proteins from sesame, sunflower, and Brazil nut (in decreasing order of similarity). It is interesting to note that coffee, tomato, sesame and sunflower are fairly closely related taxonomically. All belong to the Asterid I/II clade of Eudicots (Fig. 1). This close phylogenetic relationship may explain why Unigene 125230 has homologous matches only in these species. Moreover, since Unigene 125230 shows homology to the 2S seed storage proteins in these related species, we conjecture that unigene 125230, its tomato unigene match, the sesame, Brazil nut and sunflower 2S storage protein gene all encode orthologous 2S seed storage proteins. This is the first time that a 2S seed storage protein has been identified in coffee or any Solanaceae species. Finally, a BLAST search of Unigene 125230 against the coffee unigene set revealed additional putative copies of the 2S seed storage protein. However, on close examination, all appear to be splicing variants or low quality sequences. Moreover, southern hybridization with a 2S cDNA probe on genomic DNA confirmed that the 2S gene is single copy in the coffee genome (data not shown).
Unigene 120912: 11S seed storage protein
Unigene 120912 is the second most abundant unigene, containing 687 ESTs (Table 5). This gene is preferentially expressed during middle and late stage seed development and shares high similarity (over 98% identity) with a previously cloned C. arabica 11S seed storage protein (Marraccini et al. 1999; Rogers et al. 1999). This unigene also has a highly significant match to the Arabidopsis 12S storage protein and to a tomato unigene derived from seed ESTs (Pang et al. 1988) (Table 5). Given these results, we conclude that unigene 12912 is allelic with the previously described 11S seed storage protein gene from C. arabica and orthologous to 11S/12S seed storage proteins in both tomato and Arabidopsis. A BLAST search of Unigene 120912 against the coffee unigene set revealed additional putative copies of the 11S seed storage protein. However, like the 2s seed storage protein (Unigene 124230), all appears to be results of alternative splicing or low sequence quality. Moreover, southern hybridization with an 11S cDNA probe on genomic DNA confirmed that the 11S gene is single copy in the coffee genome (data not shown).
Other seed-specific genes
Early stage seed development
Unigenes 122206, 119460 and 121265 were all highly expressed and specific to the early cherry stage. The early cherry library was derived from RNA from both pericarp and seed tissue while the pericarp library was derived from RNA coming from all stages of pericarp development. Thus, if the above genes were highly expressed in the pericarp of the early cherry, they should be present in the pericarp library as well. The fact that these genes showed little or no expression outside the early cherry stage, suggests that they are probably specific to early developing seed tissues and not pericarp tissues (Table 5). Unigene 122206 showed high homology to an Arabidopsis gene annotated as encoding the enzyme S-adenosyl-l-methionine (SAM) synthetase (Table 5). This enzyme synthesizes S-adenosyl-l-methionine from l-methionine and ATP and is often represented by multiple isozymes in plant species (Schroder et al. 1997). Thus, Unigene 122206 appears to be an SAM synthetase specific to early seed development (Table 5).
Unigene 119460 shows high homology to the highly conserved WRKY transcription factor family. The WRKY transcription factor is a large gene family having more than 70 members in the Arabidopsis genome (Dong et al. 2003). Previous studies showed that it is related to wounding, stress, pathogen infection and senescence in many plant species. In some recent studies, the WRKY protein family was found to be involved in sugar signaling in barley and seed development in Arabidopsis (Johnson et al. 2002; Sun et al. 2003). However, the function of the best Arabidopsis match to Unigene 119460 (At1g80840) has not been determined. Hence, understanding the function of this highly expressed, WRKY-like coffee gene awaits further study.
Unigene 121265 is highly homologous to a gene in Arabidopsis annotated as encoding a Mob1/phocein protein. Mob1/phocein proteins are found in virtually all eukaryotes. While they are conjectured to be involved in cell cycle control, there is still little experimental evidence demonstrating biological function (Pon 2004). Thus it seems premature to conjecture what role Unigene 121265 might have that is specific to the early development of coffee seeds.
Middle stage seed development
As described earlier, Unigene 125230 is a putative 2S seed storage protein with peak expression during middle seed development. Also showing preferential expression during this same stage were Unigenes 121707, 124158 and 119890. Unigene 121707 is a gene of unknown function with high homology matches both in Arabidopsis and Solanaceae EST-derived unigenes (Table 5). Unigene 124158 is homologous to an Arabidopsis gene classified as a photoassimilate-responsive protein, which is related to pathogenesis (Herbers 1995). Finally, Unigene 119890, which is also specific to middle stage seed development, is apparently a gene unique to coffee, which will be discussed more in the following section.
Late stage seed development
As discussed previously, Unigene 120912 corresponds to the 11S seed storage protein, which is largely expressed late in seed development. Other genes with preferential expression in late stage seeds are Unigenes 120118, 119817 and 124791. Unigene 119817 likely encodes a chitinase and is further discussed in the next section. Unigene 120118 shows high homology to genes in both Arabidopsis and Solanaceae EST-derived unigene sets; however, none have known function. Unigene 124791 gives a strong match to an Arabidopsis gene annotated as a plasmodesmatal receptor.
Two highly expressed genes with homology to chitinase
Unigenes 120685 and 119817 show high sequence similarity to a number of genes classified as chitinases in other organisms. Chitinases are a large and diverse class of proteins, some of which have been implicated in resistance to fungi in various plant species, including coffee (Rojas-Herrera 2002; Chen et al. 2003). The two unigenes differ in that Unigene 120685 is expressed in leaves, pericarp and early stage cherries, but not in mid or late stage seed development (Table 5). Unigene 119817, on the other hand, was found to be exclusively expressed in late stage developing seeds and pericarp tissue. As previously mentioned, early stage cherries contained both pericarp and seed tissues. The fact that Unigene 120685 was not found in the middle and late stages of seed development suggests that this gene may not be expressed in seeds, but rather in the maternally derived pericarp and leaf tissues. Based on these results, one can speculate that these two putative chitinase genes may be involved in pathogen defense in developing CHERRIES, with Unigene 120685 being expressed in early developing, post pollination pericarp and leaf tissues and Unigene 119817 being expressed primarily late in seed development, just prior to maturity.
Highly expressed genes unique to coffee
Unigene 124988
This highly expressed unigene had no significant matches in the Arabidopsis proteome, Solanaceae EST-derived databases, GenBank NR databases, or GenBank dbest. Moreover, the predicted protein encoded by Unigene 124988 has no recognizable domains, which might give clues to its function. ESTs for this unigene were detected in all five libraries, with highest expression being observed in the pericarp (Table 5).
Unigene 119890
Unigene 119890 also has no significant match in any of the tested databases, with the possible exception of a very weak match in the Solanaceae unigene sets (the best hit was from pepper with an e value of 5e-7, Table 5). Like Unigene 124988, its predicted protein has no recognizable domains. Unigene 119890 was highly and exclusively expressed in the middle stage of developing seeds (Table 5).
The fact that neither Unigene 124988 nor Unigene 119890 have counterparts in any other databases suggests that they may represent coffee-specific genes or genes that have been evolving at such a rapid rate that they no longer bear any recognizable homologies with proteins from other plants, including the closely related Solanaceous plants. We speculate that these genes may be related to chemical or morphological features unique to coffee.
Gene families unique or significantly expanded in coffee
The predicted protein sequences for the coffee unigene set and the Arabidopsis proteome were combined and organized into protein families using tribeML software (see Materials and methods for details). In most cases, the number of gene family members was higher in arabidopsis than in coffee (data not shown). However, this is to be expected since the entire Arabidopsis genome has been sequenced (thus revealing all genes), whereas in coffee the EST-derived unigene set is by nature incomplete and unlikely to contain all members of any given gene family. However, despite the incomplete nature of the coffee EST-derived unigene set, several cases was encountered in which coffee has more gene family members than does Arabidopsis. To further investigate this topic, a list was made of the gene families for which coffee has significantly more members than Arabidopsis (Table 6). We refer to these as coffee-expanded gene families. In addition, gene families found in coffee, but for which no counterparts exist in Arabidopsis were also compiled (Table 7). We refer to these as coffee-unique gene families. All gene families listed in both Tables 6 and 7 were subjected to manual examination to remove any members determined to be results of alternative splicing and/or poor sequence quality—both of which could inflate the number of putative gene family members in coffee. Hence, the differences in copy number reported are likely to be due to true differences in gene copy number and not artifacts of EST-based gene predictions.
Table 6Gene families expanded in coffee relative to Arabidopsis
Family ## Arabidopsis family member# Coffee family memberLongest coffee memberAnnotation266121122330Retrotransposon gag protein, class I180514124952Polygalacturonase isoenzyme 1 beta subunit with BURP domain632112123451Acidic endochitinase386210124158Photoassimilate-responsive protein38248119672Hypersensitive-induced protein, band 7 protein39427122791E-class P45048326120054Bet v I allergen62336119581Root hair defective protein1,18215126674Unknown function69525126974Tyrosine decarboxylase78325122423Unknown function1,11725119449Trypsin inhibitor KunitzTable 7Gene families unique to coffee in comparison to Arabidopsis
Gene family ## Family memberLongest memberSolanaceae hitAnnotation24327122956258190 potatoRetrotransposon gag protein, class II68711120121221585 tomatoThaumatin, pathogenesis related96510119718249401 potatoZn-finger, CCHC type974101202442610402 potatoDisease resistance protein (TIR-NBS-LRR class)8529119638225732 tomatoRetrotransposon gag protein, classs III360812199823671 tomatoDisease resistance protein1,0197124574222350 tomatoLeucine-rich repeat, disease resistance protein1,6077122216noneUnknown function1,6107130519noneUnknown function1,6767126264243065 tomatoUnknown function7086123769236157 tomatoABA/WDS induced protein1,8525120284213688 tomatoProline-rich region, extension-like protein2,3625122218237314 tomatoUnknown function2,4595124466267984 potatoLeucine-rich repeat, plant specific, receptor-related protein kinase
Coffee-expanded gene families
The most expanded gene family in coffee corresponds to a retrotransposon gag protein (Table 6). This result has two implications. First, the retroelement encoding this gag protein occurs at a higher frequency in coffee compared with Arabidopsis, although we cannot determine whether this difference is due to a true expansion of this element in coffee subsequent to divergence from Arabidopsis, or rather a loss of the element in the Arabidopsis lineage. Second, the fact that this retrotransposon gag protein element was discovered in an EST-database indicates that this particular retroelement is being transcribed in the coffee genome, and hence may represent an active retrotransposon.
Another gene family for which coffee has significantly more members than Arabidopsis, encodes proteins annotated in Arabidopsis as acid endochitinases and photoassimilate-responsive proteins (Table 6). As noted earlier, chitinases are associated with fungal resistance and are among the most highly expressed genes in coffee. The fact that chitinases are both highly expressed and represented by an expanded gene family in coffee may reflect a greater need for fungal resistance engendered both by the perennial nature of coffee and the fact that it is a tropical species for which a multiplicity of fungal pathogens is common. The reasons for the putative expansions of the other gene families listed in Table 6 remain for future studies to determine.
Coffee-unique gene families
Table 7 lists the top gene families (based on copy number), which occur in coffee, but not in Arabidopsis. Of the 15 gene families listed, four are of unknown function. For those that could be functionally annotated, five (45%) have putative functions related to disease resistance, such as TIR-NBS-LRR disease resistance proteins, LRR proteins and thaumatin pathogenesis-related proteins (Table 7). These finding are consistent with rapid evolution of genes/gene families related to disease resistance, likely driven by selection pressure from continuously changing pathogens and/or pathogens unique to the particular environments of a species (Meyers 1998; Michelmore and Meyers 1998). Also included in this list of coffee-unique gene families are two, which encode retrotransposon gag-proteins (Table 7).
Comparison of the coffee gene repertoire with that of Arabidopsis and Solanaceae species
Each coffee unigene was screened against the Arabidopsis proteome using BLASTX, which compares all six translated frames of the query with the target. In total, 21% of the coffee unigenes had no clear match in Arabidopsis (match score 0–30) (Fig. 6). These coffee genes, which lack an Arabidopsis counterpart, may have evolved rapidly since coffee and Arabidopsis diverged from their last common ancestor, and hence their counterparts in each genome are no longer recognizable by standard string searches like BLAST. To investigate this topic further, the 2,853 unigenes with no clear match in Arabidopsis (match score<30, see Fig. 6) were screened against the Solanaceae EST-derived unigene sets using TBLASTX which compares all six translated frames of both query (coffee unigenes) and target (Solanaceae unigenes, see Materials and methods for details) sequences. The family Solanaceae is much more closely related to coffee than is Arabidopsis (Fig. 1), and hence coffee genes without a match in Arabidopsis may well have recognizable orthologs in the Solanaceae. The results show that a large portion (90%) of the coffee genes that lack a match in Arabidopsis do have a match in Solanaceae at the same stringency criteria (match score>30). For many of these cases, however, the coffee-Solanaceae matches are still quite diverged (low match scores) suggesting that the proteins encoded by these genes have been evolving at a rapid rate in dicots.
Fig. 6Histogram showing match scores for each coffee unigene as compared with its best match in the Arabidopsis proteome
There were, however, some exceptions to this trend represented by coffee genes having no match in Arabidopsis, yet a strong match in Solanaceae. These latter cases may represent genes that have been under differential selection pressure since the time that the Rosids (which includes Arabidopsis) diverged from the last common ancestor with the Asterids (which includes Solanaceae and coffee) (Fig. 1). Alternatively, these genes may have been lost in the lineage leading to Arabidopsis, but retained in the Asterids. In either case, such genes may have functional roles that are, in part, responsible for the morphological, developmental and chemical characteristics that differentiate Arabidopsis from Rosid species (e.g. coffee, Solanaceae). In an effort to shed more light on this issue, the 20 coffee unigenes having the highest match scores with Solanaceae, but lacking a counterpart in Arabidopsis were parsed and annotated (Table 8). Seven of these coffee genes have no match either in GenBank NR or dbest hence their function is unknown. Another unigene (Unigene 125230, the 2s seed storage protein, see above discussion) had weak match to ESTs of sesame, a species also belonging to the Asterid I clade (Fig. 1). Thus, these eight genes (40%) appear to be specific to the tomato/coffee/sesame Asterid I clade (Fig. 1, Table 8). Elucidating their function may shed light on the molecular processes, which diversified early in the evolution of dicot species, especially those species belonging to the Asterid clade of dicots.
Table 8Coffee genes not found in Arabidopsis, but with conserved counterparts in tomato or other Solanaceous speciesCoffee unigeneSolanaceae EST-derived unigene matchScoreGenBank (non-redundant and dbest) best matchScoreAnnotation124978240871 tomato454Unknown function121324235756 tomato429gblCB686389.1 [Brassica napus]44Unknown function131820213100 tomato426gil50252229.1 [Oryza sativa]73Unknown function121542240321 tomato416refiNP_922676.1 [Oryza sativa]75Unknown function131934219759 tomato377embiCAE05735.1 [Oryza sativa]297TFIIH basal transcription factor p52 subunit121140236347 tomato320Unknown function131445225435 tomato320Unknown function125230243065 tomato238gil13183175 [Seasame indicum]452S albumin131030246364 potato213refINP_524404.1 [Drosophila melanogaster]110Phospyhatidyl inositol transfer protein120120237254 tomato202gbICF349465.1 [Rose]52Unknown function126635237314 tomato185Unknown function126575237314 tomato182Unknown function130675209387 petunia177gbICK093976.1 [Populus tremula]438Unknown function128020237314 tomato167Unknown function123615249253 potato163gbIAAO73272.1 [Oryza sativa]140Unknown function126432240551 tomato163gil34878866 [Rattus norvegicus]56Phosphatidylinositolglycan class N124384197378 pepper156gblCA815435.1 [Vitis vinifera]1,009Unknown function122126239632 tomato153Unknown function131601232010 tomato145gbICK229938.1 [Macaca mulatta]7440S Ribosomal protein S21119644237150 tomato143refINP_921250.1 [Oryza sativa]70HelicaseThe GenBanks Best match exclude those from Solanaceae, Coffea and Hedyotis (both members of the Rubiaceae family). Solanaceae EST-derived
Surprisingly, 8 (40%) of the 20 coffee genes having no match in Arabidopsis did have matches in species phylogenetically distant from both coffee and Arabidopsis, including two to non-plant species (Drosophila and rat) (Table 8). Five of these matches were to rice genes, which is a monocot and highly divergent from coffee, Arabidopsis and other dicot species (Table 8). The fact that all of these species diverged from Arabidopsis and coffee long before the latter diverged from each other suggests that these genes may have been present in the last common ancestor of Arabidopsis and coffee/Solanaceae, but subsequently lost in the Arabidopsis lineage.
Coffee genes share greater similarity to genes in tomato/Solanaceae than to Arabidopsis
As discussed earlier, coffee is much more closely related to the Solanaceae than to Arabidopsis (Fig. 1). Hence, Solanaceae species may be better models for coffee genomics than Arabidopsis. The results for fast-evolving genes, presented above, are consistent with this prediction. To further investigate this assertion, the degree to which each coffee unigene matched Arabidopsis versus Solanaceae was investigated. In doing this analysis, one has to keep in mind that the entire gene repertoire of Arabidopsis is known, whereas the EST-derived unigene sets for Solanaceae do not represent the entire gene repertoire of these species. We estimate that the combined EST-derived unigene sets of Solanaceae species represent as much as three-quarters of the Solanaceae gene content (Hoeven et al. 2002). Moreover, Arabidopsis genes are of full length, while Solanaceae EST-derived unigenes are not necessarily of full length.
To mitigate against these factors, each coffee unigene was first screened, via TBLASTX, against the Solanaceae EST-derived unigene databases. Only those coffee unigenes matching a Solanaceae unigene with a match score greater than 100 were compared with the Arabidopsis proteome. Moreover, the match to Arabidopsis was confined to only the portion of the coffee unigene matching a Solanaceae EST-derived unigene. The ratio of the Arabidopsis match score divided by the Solanaceae match score was used as a measure of the degree to which the Solanaceae homolog shows greater similarity to the coffee unigene than to the closest Arabidopsis homolog. The results of this analysis are shown in Fig. 7. The ratio of the Arabidopsis/Solanaceae match score averaged 0.75, indicating a significantly better match in most cases with Solanaceae versus Arabidopsis. Moreover, in more than 95% of the cases, the ratio was less than one, indicating that, in most instances, coffee contains genes significantly more homologous to Solanaceae than to Arabidopsis. For those cases in which the ratio was greater than 1 (5%), we cannot rule out the possibility that the true ortholog was missing in the Solanaceae EST-derived unigene sets, hence the comparison was between paralogs. These results, combined with the results on fast evolving genes presented earlier, provide clear evidence that the Solanaceae provide a much better model than Arabidopsis for genomic and biological studies in coffee. This is especially relevant as sequencing of the tomato genome is currently underway (http://www.sgn.cornell.edu/help/about/tomato_sequencing.html).
Fig. 7Ratio of highest Arabidopsis match score to highest Solanaceae match score for individual coffee unigene. The analysis restricted to coffee unigenes with a Solanaceae match score >100
Conclusions
Herein, we describe the development and analysis of a large EST database for coffee. The resulting 47,000 ESTs correspond to 13,175 unique genes (unigenes), a large portion of which are expressed during seed development—a stage important to coffee as a crop and one for which our understanding of molecular development is still rudimentary. To our knowledge, this is the largest public database for seed-derived ESTs. Hence, this EST database represents a new public resource, which can facilitate a better understanding of seed development, as well as genomic, molecular and breeding research in coffee. By comparisons with Arabidopsis and Solanaceous species, we have identified the two major seed storage proteins of coffee (2S and 11S) and demonstrated that these proteins are expressed at different times during seed development. Through in silico gene expression analysis, we have identified a number of highly expressed genes that show high specificity for different stages of seed development as well as for the pericarp tissue that surrounds the seeds. Many of these highly expressed genes are unique to coffee and/or the Asterid clade of higher plants. While the functions of most of these highly expressed, tissue/stage specific genes remain to be determined, the fact that they have been identified points the way to promoters, which can potentially be used to drive gene expression in specific stages/tissues of the coffee plant. Many of these genes are specific to defined periods of seed and/or pericarp development—both critically important for insect/pathogen resistance and in determining the quality of the coffee bean with respect to commercial coffee products.
Coffee, as a member of the family Rubiaceae, is distantly related to the model species Arabidopsis. A computational comparison of the coffee EST-derived unigene set with the sequence databases for Arabidopsis and Solanaceous species (e.g. tomato, pepper), indicate that the latter are much better genomic models for coffee than is Arabidopsis. These results are consistent with the fact that coffee and solanaceous species share very similar chromosome architecture and are closely related, both belonging to the Asterid I clade of dicot plant family. Moreover, the ability to identify orthologous genes between coffee and tomato opens the door to eventually developing detailed comparative maps for these two species and to the sharing of genomic and biological tools/discoveries—an outcome that should expedite research in both taxa.
|
[
"solanaceae",
"coffea canephora",
"seed development",
"rubiaceae",
"comparative genomics"
] |
[
"P",
"P",
"P",
"P",
"R"
] |
Purinergic_Signal-2-2-2096639
|
The CD38/CD157 mammalian gene family: An evolutionary paradigm for other leukocyte surface enzymes
|
Human CD38 is the mammalian prototype of a family of phylogenetically conserved proteins which share structural similarities and enzymatic activities involved in the production of an intracellular second messenger with calcium mobilizing effects. Engagement of CD38 by agonistic monoclonal antibodies and the CD31 ligand initiates a cytoplasmic signaling cascade involving tyrosine phosphorylation of the proto-oncogene c-cbl and of the extracellular regulated kinase 1 of 2 complex. Further requirements for signal transduction include a privileged localization within the cholesterol-rich areas of the plasma membrane and physical association with specialized surface receptors. CD38-mediated signals are crucial in heterotypic cell adhesion and migration as well as in the activation of proliferation/survival programs by normal and neoplastic cells. Here we review the most recent literature on this complex topic and attempt to formulate a single model reconciling the enzymatic and receptorial activities of CD38.
|
[
"calcium mobilization",
"cell surface enzymes",
"cell surface receptors",
"intracellular signaling"
] |
[
"P",
"R",
"R",
"R"
] |
Neuroradiology-4-1-2275774
|
Angiographic CT with intravenous administration of contrast medium is a noninvasive option for follow-up after intracranial stenting
|
Intracranial angioplasty and stenting (ICAS) is a therapeutic option in symptomatic intracranial atherosclerotic disease. Adequate follow-up examination is necessary to exclude in-stent restenosis. Conventional intraarterial digital subtraction angiography (ia-DSA) is the current gold standard, but it is an invasive technique and carries the risk of neurological complications. Angiographic CT (ACT) is a new technique that provides a volume dataset of the highest spatial resolution and high contrast resolution derived from a rotational acquisition of a c-arm-mounted flat-panel detector. The feasibility of ACT with intravenous administration of contrast medium (iv-ACT) for follow-up after ICAS is demonstrated. In two patients iv-ACT was performed as a follow-up examination 12 months after ICAS. High-resolution volume data from the rotational acquisitions were processed to provide delineation of the stent lumen as well as imaging of the brain parenchyma and vessels. In both patients the patency of the stent lumen was assessed successfully. In addition, all other brain vessels were displayed in a manner similar to their appearance on CT angiograms. The brain parenchyma was also adequately imaged in a manner similar to its appearance on CT images. We demonstrated the feasibility and diagnostic value of iv-ACT for follow-up imaging after ICAS. This new application has the potential to become the imaging method of choice after ICAS since it not only enables visualization of the patency of the stent lumen but also is minimally invasive and provides additional information about all brain arteries and the brain parenchyma.
Introduction
In symptomatic intracranial atherosclerotic disease refractive to medical treatment intracranial angioplasty and stenting (ICAS) is a therapeutic option increasingly used in specialized stroke centers. However, restenosis rates ranging from 8% to 30% have been reported, underlining the necessity for regular follow-up examinations to exclude significant restenosis [1–4]. CT angiography (CTA) and MR angiography (MRA) are known to suffer from considerable artifacts that often prevent assessment of the stent lumen. Therefore, current follow-up techniques include transcranial Doppler sonography and conventional intraarterial digital subtraction angiography (ia-DSA). However, diagnostic ia-DSA is an invasive technique with a rate of examination-related neurological complications ranging from 0.5% to 0.8% [5, 6]. In addition, ia-DSA is not offered on an outpatient basis at every medical center so patients have to be admitted for at least one night, significantly increasing the treatment costs. Furthermore, patient compliance with follow-up examinations may be a problem because the ia-DSA procedure is not very comfortable.
The new angiographic systems equipped with flat-panel detectors which allow rotational acquisition of volume data provide the possibility of CT-like low-contrast imaging [7]. The administration of even highly diluted contrast medium intraarterially gives good delineation of the vessels indicating that intravenous administration of contrast medium could be sufficient. We sought to combine this property with the possibility of nearly artifact-free imaging of small-vessel stents described previously [8].
Methods
DSA and angiographic CT (ACT) examinations were performed on an AxiomArtis dBA biplane angiography system equipped with flat panel detectors (Siemens Medical Solutions, Forchheim, Germany). The acquisition of rotational datasets was performed with the following parameters (20s-1k protocol): 20 s rotation, 538 projections, 220° total angle, no zoom (detector size 30×40 cm), CTDIweighted approximately 22 mGy (manufacturer’s information). Postprocessing of the image data to a volume dataset (ACT) was performed on a Leonardo workstation (DynaCT, InSpace 3D software, Siemens). The software includes the application of system-specific filter algorithms in order to correct for beam hardening, radiation scatter, truncated projections and ring artifacts. Postprocessing resulted in volume datasets each defined by a batch of about 400 slices in a 512×512 matrix. Voxel size was about 0.1×0.1×0.1 mm3. The ACT datasets were further processed to multiplanar reconstructed (MPR) slices with the thin-slice and thick-slice maximum intensity projection (MIP) technique parallel and perpendicular to the stent-supplied section of the vessel, respectively. The contrast medium used in the follow-up ACT examination was Imeron 350 (Bracco ALTANA Pharma, Konstanz, Germany). The contrast medium was injected to a volume of 100 ml into a cubital vein at a flow rate of 5 ml/s, and the start delay for rotational acquisition was 20 s. The patient was asked to close his or her eyes and to breathe calmly during the examination.
CTA was performed on a 16-slice CT scanner (Aquilion, Toshiba Medical Systems, Tokyo, Japan) with 0.5-mm slice collimation, intravenous injection of 60 ml Imeron 350 (Bracco ALTANA Pharma), and a CTDIvol of approximately 50 mGy [9]. Postprocessing was performed on a Vitrea II workstation (Vital Systems, Minnetonka, Minn.).
Case illustrations
Case 1
A 57-year old man suffering from brainstem infarction and recurring transient ischemic attacks of the vertebrobasilar circulation was admitted to our hospital. Transcranial ultrasonography and MRA revealed a high-grade stenosis of the basilar artery. The symptoms occurred despite double platelet inhibition (aspirin and clopidogrel) and ICAS was chosen as therapy. The procedure was performed under general anesthesia. A balloon-mounted Pharos (3.5/8 mm) stent (Micrus Endovascular, Renens, Switzerland) was placed and balloon-inflated to a diameter of 3.2 mm thereby reconstructing a diameter corresponding to about 90% of the normal vessel diameter (Fig. 1a,b). The successful deployment of the stent was documented by native ACT (Fig. 1c,d). No complications occurred after the intervention and the patient was discharged 3 days later.
Fig. 1Patient 1. a, b DSA with selective injection of contrast medium into the right vertebral artery shows the eccentric high-grade stenosis of the basilar artery (a). After stent placement (b) the lumen of the basilar artery is well reconstructed and considerable inflow from the contralateral vertebral artery can be seen. c, d Reformatted images in the sagittal plane from an ACT scan directly following the intervention reveal regular deployment of the stent (c MIP, 0.5-mm slice thickness; d MIP, 5-mm slice thickness)
An outpatient follow-up CTA was performed 6 months after the procedure. The stent lumen was not adequately visualized on thin-slice MPR (Fig. 2a). Therefore, admission for follow-up DSA was organized for 2 months later. The patient refused ia-DSA, so ACT with intravenous administration of contrast medium (iv-ACT) was offered as an alternative. Written informed consent for the examination as a novel application of a rotational acquisition was obtained. The examination was successful. Corresponding to the good low-contrast resolution of ACT described previously, the reconstructed images revealed full patency of the stent lumen (Fig. 2b,c) [10]. This result was highly correlated with the flow-rate measured by transcranial Doppler sonography that had remained normal from directly after the intervention until the 8-month follow-up.
Fig. 2Patient 1. a CTA image obtained at the 6-month follow-up is not of sufficient quality to reveal the stent lumen, and in particular the proximal and distal markers produce substantial artifacts (MIP, 1-mm slice thickness). b, c iv-ACT images obtained at the 8-month follow-up show an unchanged configuration of the stent in comparison to the examination performed directly after the intervention (see Fig. 1c,d). The contrast filling is more faint than on the CTA image but the patency of the stent lumen is visible much more clearly on the sagittal cross sections, and the artifact load is acceptable (b MIP, 0.5-mm slice thickness; c MIP, 5-mm slice thickness)
Case 2
An 82-year-old man was admitted suffering from progressive left-sided hemiparesis and impaired consciousness. CTA revealed a high-grade stenosis of the right middle cerebral artery (MCA) and a second middle-grade stenosis of the basilar artery (data not shown). CT-perfusion (CTP) showed a prolonged mean transit time in the whole right MCA territory but only small infarcted areas level with the basal ganglia, represented by lowered relative cerebral blood volume. Emergency ICAS was performed under general anesthesia. A balloon-mounted Driver (2.5/8 mm) stent (Medtronic, Minneapolis, Minn.) was successfully placed in the stenosis of the right M1 segment leading to a good morphological result and immediate restoration of normal blood flow (Fig. 3a,b). Proper stent deployment was documented by native ACT (Fig. 3c,d).
Fig. 3Patient 2. a, b DSA images with selective injection into the right internal carotid artery reveals a high-grade stenosis of the right middle cerebral artery (MCA) with reduced distal blood flow (a). After stent placement the blood flow is normalized (b). c, d Native ACT reformatted images in paracoronal orientation reveal regular deployment of the stent (c MIP, 0.5-mm slice thickness; d MIP, 5-mm slice thickness)
The patient was admitted for follow-up DSA 12 months after the procedure. Transcranial Doppler sonography showed a normal flow velocity in the right MCA. This patient also refused ia-DSA. Written informed consent was obtained in advance to perform iv-ACT and to omit ia-DSA if the iv-ACT was of diagnostic quality. The procedure was performed successfully and the reconstructed images revealed full patency of the stent lumen (Fig. 4a,b). All other intracranial arteries were visible at the highest resolution, revealing an unchanged configuration of the medium-grade stenosis of the basilar artery (Fig. 4c,d). The reconstructed brain parenchyma images showed already known residual infarction (data not shown).
Fig. 4Patient 2. a, b Iv-ACT images obtained 12 months after the intervention reveal an unchanged stent configuration and full patency of the stent lumen (a MIP, 0.5-mm slice thickness; b MIP, 10-mm slice thickness). Even small perforating arteries originating through the stent struts can be seen (barrow). c A transverse thick-slice (30 mm) MIP overview image gives a good impression of the insular MCA branches of both sides. d A second thick-slice (15 mm) MIP image oriented parallel to the basilar artery reveals a medium-grade stenosis (arrow) and an anatomic variant of the origin of the superior cerebellar artery
Discussion
In patients suffering from intracranial atherosclerosis, ICAS is a promising new therapeutic option [1, 3, 11]. To identify an in-stent restenosis a reliable imaging technique is needed. Here, DSA is the current gold standard, but it carries a 0.5% to 0.8% risk of permanent neurological impairment [5, 6]. Additionally, not every center is able to perform ia-DSA on an outpatient basis. The risk may be less with ia-ACT if the injection is performed in the aortic arch. However, this is still an invasive procedure.
In the study reported here we demonstrated that iv-ACT can reliably depict the lumen of small-vessel stents with high spatial resolution. As the low contrast resolution of ACT is comparable to that of conventional CT, a relatively low concentration of intravascular contrast medium is needed. All intracranial vessels can be viewed simultaneously in a high-quality image. CT-like images of the brain parenchyma are provided as well [10]. It therefore represents a new radiological imaging technique between CTA and ia-DSA.
Even though this new method needs to be evaluated in a larger number of patients, it possibly has many advantages for patients and physicians as a new option in the imaging of cerebrovascular disease: Examinations can be performed on an outpatient basis and the risk of neurological complications is lowered practically to zero. The examination can be performed quickly, in a time similar to that required for a CTA examination. The radiation exposure for the patient from a rotational acquisition is lower than that from a CTA examination (manufacturer’s information) and has been shown to be considerably lower than that from a single biplane DSA series [12].
Possible drawbacks of the iv-ACT technique first include movement artifacts. The patient has to lie nearly motionless for about 20 seconds, which requires a high level of compliance and is not always feasible in patients suffering from cerebrovascular disease.
The first stent we examined is a relatively new balloon-mounted stent designed for intracranial use (Pharos, Micrus Endovascular, Renens, Switzerland). All dedicated intracranial stents (Neuroform and Wingspan, Boston Scientific, Natick, Mass.; LEO, BALT, Montmorency, France; and Enterprise, Cordis, Miami Lakes, Fl.) have a similar strut size of about 0.06 mm [13]. In contrast, common coronary stents have thicker struts (e.g. Driver, Medtronic, Minn.: 0.09 mm) [14]. Our second patient demonstrates that the lumen of these stents can be visualized as well. This issue is not unimportant because until recently dedicated intracranial stents were not available. Therefore, the majority of stents implanted into intracranial vessels are still coronary stents.
Iv-ACT is a promising new technique for minimally invasive follow-up after intracranial stenting. Other possible applications can be envisaged, such as diagnostic and follow-up imaging before and after endovascular treatment of intracranial aneurysms or arteriovenous malformations.
|
[
"intracranial stenting",
"flat-panel",
"cerebrovascular disease",
"rotational angiography",
"angiographic computed tomography"
] |
[
"P",
"P",
"P",
"R",
"M"
] |
J_Fluoresc-4-1-2266787
|
Spectrofluorimetric Determination of Oxamniquine in Dosage Forms and Spiked Human Plasma through Derivatization with 1-dimethylaminonaphthalene-5-sulphonyl chloride
|
A sensitive, simple and selective spectrofluorimetric method was developed for the determination of oxamniquine (OXM) in pharmaceutical formulations and biological fluids. The method is based on the reaction between the drug and 1-dimethylaminonaphthalene-5-sulphonyl chloride (dansyl chloride) in presence of 0.5 M sodium carbonate (pH 10) to yield a highly fluorescent derivative that is measured at 445 nm after excitation at 335 nm. The different experimental parameters affecting the development and stability of the reaction product were carefully studied and optimized. The fluorescence concentration plot was rectilinear over the range of 0.02–0.2 μg ml−1 with a lower detection limit (LOD) of 0.007 μg ml−1 and limit of quantitation (LOQ) of 0.02 μg ml−1. The proposed method was successfully applied to the analysis of commercial capsules. The results obtained were in good agreement with those obtained using the official spectrophotometric method. Furthermore, the method was applied for the determination of oxamniquine in spiked human plasma, the mean % recovery (n = 4) is 97.77 ± 1.19. A proposal of the reaction pathway was presented.
Introduction
Oxamniquine (OXM), 1,2,3,4-tetrahydro-2[(isopropylamino methyl]-7-nitro-6-quyinoline methanol is an antischistosomal agent that is indicated for the treatment of schistosoma mansoni (intestinal schistosomiasis) infection (Fig. 1). It has been shown to inhibit DNA, RNA and protein synthesis in schistosomes. The oral bioavailability of oxamniquine is good and effective plasma levels are achieved in 1–1.5 h [1].
Fig. 1Structural formula of oxamniquine
Oxamniquine is the subject of a monograph in the USP (XXIII) [2] whereby a spectrophotometric method is recommended for its determination, whether in its pure form or in capsules. Several methods have been published for the determination of OXM, either per se or in pharmaceutical preparations and biological fluids. These methods include: spectrophotometry [3–5], non-aqueous titration [6], gas chromatography [7], HPLC [8–13], capillary electrophoresis [11], polarography [14], flow injection analysis [15], cyclic voltammetry [16] and fluorimetry [17].
All these methods are either insufficiently sensitive or tedious and require highly sophisticated and dedicated instrumentation [8–13]. This led us to study the reaction of oxamniquine with dansyl chloride to develop simple and sensitive spectrofluorimetric method for its determination in pharmaceutical preparations and biological fluids. Dansyl chloride is a useful derivatizing agent for primary amines, secondary amines, imidazoles and phenols, etc.. Several pharmaceutical compounds have been determined through this approach [18–24].
The reported spectrofluorimetric method for oxamniquine [17] is tedious and time consuming. It involves the use of 2-cyanoacetamide, a hazard reagent.
Experimental
Apparatus
The spectrofluorimetric measurements were recorded using ARF-1501 Shimadzu Spectrofluorometer, equipped with Xenon arc lamp.
Materials and reagents
All reagents and solvents were of analytical reagent grade
Oxamniquine pure sample was kindly provided by Pfizer (Sandwich, UK).Capsules containing 250 mg of Oxamniquine each (Vansil capsules) were obtained from commercial sources in the local market.1-Dimethyl aminonaphthalene-5-sulphonyl chloride (dansyl chloride), purchased from Sigma (St. Louis, USA). A stock solution containing 0.1% of dansyl chloride was freshly prepared in acetone and was further diluted with the same solvent to obtain 0.001% solution.Sodium carbonate (BDH, UK) 0.5 M aqueous solution (pH 10).Sodium hydroxide (BDH, UK) 1 M aqueous solution.Isobutylmethyl ketone (IBMK) (Merck, Darmstadt, Germany).Plasma was kindly provided by Mansoura University Hospital, and kept frozen until assay after gentle thawing.Methanol and diethyl ether (Merck, Darmstadt, Germany).
Standard solutions
A stock solution was prepared by dissolving 20.0 mg of OXM in 20 ml of acetone and 80 ml of 0.5 M sodium carbonate solution. This solution was further diluted with the same solvent mixture as appropriate. The standard solutions were stable for seven days when kept in the refrigerator.
General procedure
Aliquots of OXM standard solution were transferred into a series of 10 ml volumetric flasks. 0.7 ± 0.1 ml of 0.001% of dansyl chloride reagent was added, followed by 0.4 ml of acetone and mixed well. The reaction mixture was left for 30 min, and then completed to the mark with IBMK. the reaction mixture was allowed to stand for 10 min. The fluorescence intensity of the reaction product was measured at 445 nm after excitation at 335 nm. Blank experiment was carried out simultaneously. The corrected fluorescence intensity was plotted vs the final drug concentration (μg ml−1) to get the calibration graph. Alternatively, the corresponding regression equation was derived.
Applications
Procedure for commercial capsules
The content of ten capsules were emptied, and mixed well. A weighed quantity of the powder equivalent to 5.0 mg OXM was transferred into a small conical flask, and extracted with 2 × 10 ml of methanol. The extract was filtered into 25 ml volumetric flask. The conical flask was washed with several milliliters of methanol. The washing was passed into the same volumetric flask, and the combined extract was evaporated to dryness on a boiling water bath. The residue was dissolved and diluted to volume with a mixture of 5 ml of acetone and 20 ml of 0.5 M of sodium carbonate solution. Aliquots covering the working concentration range (cited in Table 1) was transferred into 10 ml volumetric flasks. The “General procedure” was then applied. The nominal content of the capsules was determined either from the calibration curve or using the corresponding regression equation.
Table 1Performance data of the proposed methodParameterProposed methodConcentration range (μg ml−1).0.02–0.2Minimum detection limit, LOD (μg ml−1).0.007Limit of Quantification , LOQ(μg ml−1).0.02Correlation coefficient ( r).0.9999Slope3,473.114Intercept0.574Sy/x1.273Sa7.733Sb8.965% Error0.28Sy/x = standard deviation of the residualsSa = standard deviation of the intercept of regression lineSb = standard deviation of the slope of regression line% Error = RSD%/√ n
Procedure for spiked human plasma
A stock solution containing 20 μg ml−1 of OXM was prepared. Control samples of plasma was spiked with different quantities of OXM to give a final drug concentration cited in Table 5. One molar NaOH (0.8 ml) was added to 1.0 ml of the spiked plasma and shaken genteelly. The solution was vortexed with 3 × 5 ml of diethylether for 2 min, then centrifuged at 2,500 rpm for 5 min. The resulting supernatant was evaporated to dryness under nitrogen at ambient temperature. The residue was dissolved and diluted to volume with a mixture of 5 ml of acetone and 20 ml of 0.5 M of sodium carbonate solution. Aliquots covering the working concentration range was transferred into 10 ml volumetric flasks. The recommended procedures were then applied. The nominal content of the drug was determined using the corresponding regression equation.
Results and discussion
Dansyl chloride was first introduced for the determination of some primary, secondary amines, imidazoles and phenols [18–20].
In recent reports. DNS was further used as a fluorogenic reagent for the determination of some pharmaceutical compounds [21–24].
In the present study, OXM was found to react with Dansyl chloride at pH 10.0 forming a highly yellow fluorescent derivative with λ maximum emission at 445 nm after excitation at 335 nm (Fig. 2).
Fig. 2Fluorescence spectra of A and B excitation and emission spectra of blank. Fluorescence spectra of A′ and B′ excitation and emission spectra of oxamniquine (0.2 μg ml−1) with dansyl chloride at pH 10.0
Optimization of experimental parameters
The spectrofluorimetric properties of the colored product as well as the different experimental parameters affecting the development of the reaction product and its stability were carefully studied and optimized. Such factors were changed individually while the others were kept constant. The factors include pH, concentration of the reagent, type of buffer, temperature, reaction time and dilution time.
Effect of pH
The influence of pH on the fluorescence intensity of the reaction product was studied. Maximum fluorescence intensity was obtained upon using mixture of acetone and 0.5 M sodium carbonate solution. The pH of the reaction mixture was found to be 10.0. pH 10 was found to be the optimum pH for dansylation because labeling of most amino acids, amines, imidazoles and phenols has been found to be optimal at pH 9.5–10.5 [25]. The rate of dansylation process was found to increase with increasing the pH value this is due to an increase in the rate of hydrolysis of dansyl chloride into dansyl hydroxide [25]. The latter shows strong fluorescence and hence interferes seriously in the determination. However, under the proposed chosen conditions and wavelengths used, there was no interference arising from any dansyl hydroxide formed, as indicated by the low fluorescence intensity of the reagent.
Effect of concentration of dansyl chloride
The influence of the concentration of dansyl chloride was studied using different volumes of 0.001% of the reagent solution. It was found that, the reaction of dansyl chloride with OXM started upon using 0.1 ml of the reagent in the presence of sodium carbonate (pH 10.0). Increasing the volume of the reagent, produces a proportional increase in the fluorescence intensity of the reaction product up to 0.6 ml and remains constant up to 0.8 ml. Therefore, 0.7 ± 0.1 ml of 0.001% of dansyl chloride solution was chosen as the optimal volume of the reagent (Fig. 3).
Fig. 3Effect of volume of dansyl chloride (0.001%) on the fluorescence intensity of the reaction product of oxamniquine (0.2 μg m−1) at pH 10.0
Effect of temperature
Increasing the reaction temperature higher than the room temperature would result in a subsequent decrease in the fluorescence intensity of the reaction product.
Effect of reaction time
Different time intervals were tested to ascertain the time after which the solution attains its highest fluorescence intensity. It was found that after 30 min, the reaction product reaches the highest fluorescence intensity (Fig. 4) and remains stable at room temperature for 20 min.
Fig. 4Effect of reaction time of on the fluorescence intensity of the reaction product of oxamniquine (0.2 μg ml−1) with dansyl chloride
Effect of diluting solvent
Different solvents were tried to dilute the reaction mixture through out the study. It was observed that isobutyl methyl ketone gave the highest fluorescence intensity. Dilution with 0.5 M sodium carbonate solution, water, acetone–water produced almost very week fluorescence and did not reduce the blank fluorescence intensity. While upon using isobutyl methyl ketone, the fluorescence intensity attained its highest value, this was attributed to the low fluorescence value of the reagent.
Effect of dilution time
Dilution times were tested to ascertain the time after which the solution attains its highest fluorescence. It was found that dilution with IBMK after 10 min, the reaction product reaches its highest fluorescence intensity (Fig. 5).
Fig. 5Effect of dilution time on the fluorescence intensity of the reaction product of oxamniquine (0.2 μg ml−1) with dansyl chloride
Analytical performance
Validation of the proposed methods The validity of the method was tested regarding; linearity, specificity, accuracy, repeatability and precision according to ICH Q2B recommendations [26].
Linearity
By using the above procedure, linear regression equation was obtained. The regression plots showed that there was a linear dependence of the fluorescence intensity on the concentration of the drug over the ranges cited in Table 1. Linear regression analysis of the data gave the following equation:
where F is the fluorescence intensity, C is the concentration of the drug in μg ml−1 and r is the correlation coefficient.
The limit of quantification (LOQ) was determined by establishing the lowest concentration that can be measured according to ICH Q2B [26]. The results are shown in Table 1. The limits of detection (LOD) were determined by establishing the minimum level at which the analyte can be reliably detected, and the results are also abridged in Table 1.
LOQ and LOD were calculated according to the following equation [26]:
Where σ: is the standard deviation of the intercept of regression line. S: is the slope of the calibration curve.
The proposed methods were evaluated for the accuracy as percent relative error (% Er) and the precision as percent relative standard deviation (% RSD) (Tables 1 and 2).
Table 2Application of the proposed and official methods to the determination of Oxamniquine in pure formParametersSpectrofluorimetric methodOfficial method [2]No. of experiments63Mean found (%)100.13100.09 ± SD0.680.87 % RSD0.680.87Variance0.4620.76Student’s t value0.081 (2.31)Variance ratio F test1.65 (5.41)Figures between parentheses are the tabulated t and F values respectively, at p = 0.05 [27]
Accuracy
To test the validity of the proposed method it was applied to the determination of authentic sample of OXM over the working concentration range. The results obtained were in good agreement with those obtained using official method. Using Student’s t test and variance ratio F test, [27] revealed no significant difference between the performance of the two methods regarding the accuracy and precision, respectively (Table 2).
The validity of the methods was proved by statistical evaluation of the regression lines, using the standard deviation of the residuals (Sy/x), the standard deviation of the intercept (Sa) and standard deviation of the slope (Sb). The results are abridged in Table 1. The small values of the figures point out to the low scattering of the points around the calibration line and high precision.
Precision
Repeatability
The repeatability was performed by applying the proposed methods for the determination of two concentrations of OXM in pure form on three successive times, and the results are listed in Table 3.
Intermediate precision
It was performed through repeated analysis of OXM in pure form, using the concentrations showed in Table 3 for a period of three successive days. The results are summarized in Table 3.
Table 3
Validation of the proposed method for the determination of oxamniquine in pure form
Sample concentration
% Recovery (repeatability)
% Recovery intermediate precision
0.08 μg ml−1
99.64
101.08
100.72
99.28
101.44
100.89
X′
100.60
100.42
± SD
0.91
0.99
%RSD
0.91
0.99
% Error
0.53
0.57
0.2 μg ml−1
99.74
101.05
100.44
100.53
101.29
98.95
X′
100.49
100.18
± SD
0.78
1.09
%RSD
0.78
1.09
% Error
0.45
0.63
Robustness of the method
The robustness of the method adopted is demonstrated by the constancy of the fluorescence intensity with the deliberated minor changes in the experimental parameters such as change in the volume of dansyl chloride (0.001%), 0.7 ± 0.1 ml, the change in reaction time 35 ± 5 min and the change in dilution time 15 ± 5 min. These minor changes that may take place during the experimental operation didn’t affect the fluorescence intensity of the reaction product.
Pharmaceutical applications
The proposed methods were then applied to the determination of OXM in its capsules. The methods were tested for linearity, specificity, accuracy, repeatability and precision according to ICH Q2B recommendations.
Specificity
The specificity of the method was investigated by observing any interference encountered from the common capsule excepients, such as talc, lactose, starch, avisil, gelatine, and magnesium stearate. These excepients did not interfere with the proposed method.
Accuracy
The results of the proposed methods were statistically compared with those obtained using the official method [2]. Statistical analysis [27] of the results, using Student’s t test and variance ratio F test revealed no significant difference between the performance of the proposed and reference methods regarding the accuracy and precision, respectively (Table 4).
Table 4Application of the proposed methods to the determination of oxamniquine in capsulesPreparations% RecoveryProposed methodOfficial method [2]Vansil capsules (250 mg oxamniquine/capsule)98.25101.0099.3899.8099.30 Mean found (%)± SD.99.55 ± 0.99100.56 ± 0.77 t value1.67 F value1.66The tabulated values of t and F are (2.37) and (6.59) respectively, at p = 0.05 [27]
Analysis of biological fluid
The high sensitivity of the proposed method allowed the determination of OXM in spiked human plasma. Oxamniquine is readily absorbed following oral ingestion, and a peak concentration in plasma occurs within about 3 h. The presence of food significantly delays absorption and limits the concentration achieved in plasma during the first several hours after administration. Urinary excretion is the major route of elimination in man [28]. Oxamniquine is given orally in a dose of 250 mg three times daily; this leads to a final blood level concentration of about 5 μg ml−1 i.e. higher than the upper limit of the working concentration range of the proposed method. The high sensitivity of the proposed method allowed the determination of OXM in spiked human plasma. The results are shown in Table 5.
Table 5Application of the proposed methods to the determination of oxamniquine in spiked human plasmaSampleAmount added (μg ml−1)Amount found (μg ml−1)% Recovery1-a-Plasma (inter-day precision)0.020.0195797.850.040.0385696.400.100.0975497.540.200.1985699.28 Mean97.77 ± SD1.191-b-Plasma (inter-day precision)0.200.191495.700.200.196998.450.200.194897.400.200.195597.750.200.197998.950.200.194597.25 Mean97.58 ± SD1.13Each result is the average of three separate determinations
The extraction procedure described by Woolhouse and Wood [7] was adopted here. The results are satisfactorily accurate and precise.
Precision The within-day precision was evaluated through replicate analysis of Plasma samples spiked with different concentrations of the drug. The percentage recoveries based on the average of four separate determinations were 97.77 ± 1.19, thus indicating the high precision of the method (Table 5).The inter-day precision was also evaluated through replicate analysis of plasma samples spiked with 0.2 μg ml−1 of drug on four successive days. The percentage recoveries based on the average of four separate determinations were 97.58 ± 1.13. The results are shown in Table 5.
Mechanism of the reaction
The stoichiometry of the reaction was studied adopting the limiting logarithmic method [29]. The fluorescence intensity of the reaction product was alternatively measured in the presence of excess of either dansyl chloride or OXM. A plot of log fluorescence versus log [dansyl chloride] and log [OXM] gave straight lines, the values of the slopes are 0.901 and 0.984 respectively (Fig. 6). Hence, it is concluded that, the molar reactivity of the reaction is 0.901/0.984, i.e. the reaction proceeds in the ratio of 1:1. Based on the observed molar reactivity of the reaction, and depending on the presence of secondary amino group and by analogy to similar reports dealing with the reaction of dansyl chloride with compounds containing secondary amino group, the reaction pathway proposed in Scheme 1 is presented.
Fig. 6Limiting logarithmic plots for the molar ratio. a Log F vs. Log [dansyl chloride]. b Log F vs. log [oxamniquine]Scheme 1Proposed reaction pathway between dansyl chloride and oxamniquine
Conclusion
The proposed method has the advantage of being simple, sensitive and suitable for routine analysis in quality control laboratory. Also, it is suitable for the determination of oxamniquine in spiked human plasma with minimum detection limit lower than the reported value. In addition, it could be applied to the determination of OXM in its pharmaceutical preparation.
|
[
"oxamniquine",
"dosage forms",
"spiked human plasma",
"1-dimethylaminonaphalene-5-sulphonyl chloride"
] |
[
"P",
"P",
"P",
"M"
] |
Purinergic_Signal-4-2-2377317
|
Calcium-dependent release of adenosine and uridine nucleotides from A549 cells
|
Extracellular nucleotides play an important role in lung defense, but the release mechanism and relative abundance of different nucleotide species secreted by lung epithelia are not well defined. In this study, to minimize cell surface hydrolysis, we used a low-volume, flow-through chamber and examined adenosine and uridine nucleotide concentrations in perfusate aliquots of human lung A549 cells challenged by 50% hypotonic shock. Adenosine triphosphate (ATP), adenosine diphosphate (ADP), adenosine monophosphate (AMP), and adenosine (Ado) were quantified in high-performance liquid chromatography (HPLC) analysis of fluorescent etheno derivatives, and uridine triphosphate (UTP) and uridine diphosphate (UDP) were measured using HPLC-coupled radioenzymatic assays. After the onset of hypotonic shock, ATP, ADP, UTP, and UDP in the perfusates increased markedly and peaked at approximately 2.5 min, followed by a gradual decay in the next 15–20 min; peak changes in Ado and AMP were relatively minor. The peak concentrations and fold increment (in parentheses) were: 34 ± 13 nM ATP (5.6), 11 ± 5 nM ADP (3.7), 3.3 ± 1.2 nM AMP (1.4), 23 ± 7 nM Ado (2.1), 21 nM UTP (>7), and 11 nM UDP (27). Nucleotide release was almost completely abolished from cells loaded with the calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA). Under isotonic conditions, elevation of intracellular calcium with the calcium ionophore ionomycin (5 μM, 3 min) also released nucleotides with kinetics and relative abundance as above, albeit less robust. ADP:ATP (1:3) and UDP:UTP (1:2) ratios in perfusates from stimulated cells were markedly higher than the cytosolic ratios of these species, suggesting that a nucleotide diphosphate (NDP)-rich compartment, e.g., the secretory pathway, contributed to nucleotide release. Laser confocal microscopy experiments illustrated increased FM1-43 uptake into the plasma membrane upon hypotonic shock or ionomycin treatment, consistent with enhanced vesicular exocytosis under these conditions. In summary, our results strongly suggest that calcium-dependent exocytosis is responsible, at least in most part, for adenosine and uridine nucleotide release from A549 cells.
Introduction
Extracellular nucleotides control a diverse range of physiological processes by interacting with a large group of cell-surface P2Y and P2X receptor families [1–3]. In the lungs, extracellular nucleotides regulate airway surface liquid homeostasis and mucociliary clearance by modulating epithelial ion and fluid transport as well as ciliary beating [4]. In alveoli, adenosine triphosphate (ATP) is a potent secretagogue that stimulates type II cell-surfactant secretion. Nucleotide levels on airway surfaces, measured in vitro and ex vivo, show dynamic changes due to the combination of basal and stimulated release and their rapid metabolism by several groups of membrane-associated and membrane-secreted soluble ectoenzymes. These ectoenzymes extend the signaling potential of ATP by converting it to adenosine (Ado), a ligand of the A1–3 family of G-protein-coupled receptors [1, 4–7].
Nucleotide release is stimulated by cell mechanical perturbations, such as shear stress, membrane stretch, medium change, hyposmotic swelling, and hypoxia [1, 8]. It is now recognized that mechanosensitive ATP release occurs from healthy cells via physiological processes, which do not involve cell damage. Numerous reports have suggested that the cystic fibrosis transmembrane conductance regulator (CFTR) and other members of the superfamily of ATP-binding cassette-transport proteins serve as a conductive pathway for ATP release or regulate an associated ATP channel. However, other groups using patch-clamp, lipid-bilayer and luminometry techniques have not found any detectable CFTR-mediated or CFTR-regulated ATP release (reviewed in [1]). Volume-regulated anion channels (VRAC) and voltage-dependent anion channels (VDAC, porins or maxi-Cl− channels) are known to have considerable permeability to cations and large organic anions, and several laboratories have implicated these channels in cell-swelling-induced ATP release and the autocrine regulation of cell volume. However, recent studies have demonstrated that conductive pathways, including VRAC, VDAC, and stretch-activated channels, are not involved in cell-swelling-induced ATP release from A549 cells [1, 9, 10]. Finally, connexin hemichannels also have been proposed to mediate ATP release, but their role in this process in a physiological setting is not always clear [1].
Whereas conductive release mechanisms are still being debated, exocytotic ATP release is well established in excitatory cells, blood platelets, and chromaffin cells. Furthermore, there is growing experimental evidence supporting such a release mechanism also by epithelial and other nonexcitatory cells. In particular, we recently demonstrated that cell-swelling-induced ATP release from A549 and 16HBE14o− epithelial cells and 3T3 fibroblasts is tightly correlated with intracellular Ca2+ elevations; it is abolished in cells loaded with the calcium (Ca2+) chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) or by low temperature, strongly suggesting the involvement of Ca2+-dependent exocytosis [10]. Whether release of other nucleotides from epithelial cells also involves Ca2+-dependent exocytosis remains incompletely explored. Such studies, however, are complicated by nucleotide hydrolysis at the airway surface, which makes it difficult to assess the magnitude and relative abundance of different nucleotide species released. Therefore, we used a flow-through chamber to minimize cell surface hydrolysis.
We found that hypotonic shock markedly increased ATP, adenosine diphosphate (ADP), uridine triphosphate (UTP), and uridine diphosphate (UDP) concentrations in perfusates, which peaked at approximately 2.5 min. Nucleotide release was almost completely abolished from cells loaded with BAPTA and, under isotonic conditions, could be evoked by elevation of intracellular calcium with the calcium ionophore ionomycin. High nucleotide diphosphate (NDP) concentrations in perfusates of stimulated cells suggested that an NDP-rich compartment, e.g., the secretory pathway, contributed to this release. Together with real-time FM1-43 fluorescence experiments, our results strongly indicate that calcium-dependent exocytosis is a major mechanism of adenosine and uridine nucleotide release from A549 cells.
Materials and methods
Cells
Human lung carcinoma A549 cells were grown in Dulbecco’s modified eagles medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 56 U/ml penicillin-G and 56 μg/ml streptomycin sulfate. All constituents of the culture media were from GIBCO-BRL (Burlington, ON, Canada). ATP efflux was measured from cell monolayers grown to confluency on 24 × 60-mm glass coverslips. Fura-2 calcium imaging and FM1-43 microscopy experiments were performed on cells grown on circular 15-mm diameter no. 1 glass coverslips.
Nucleotide efflux assay
To measure nucleotide efflux during hypotonic challenge with high temporal resolution, we used a custom-designed, low-volume (300 μl), flow-through chamber, as described previously [10]. Briefly, 24 × 60-mm glass coverslips with confluent cell monolayers of ∼500 cells/mm2 were mounted in the chamber and perfused with warm (37°C/in-line heater; Warner Instrument, Hamden, CT, USA) Ringer solution at the rate of 1.3 ml/min. The Ringer solution contained (in mM): 110.5 NaCl, 24 NaHCO3, 1.3 KCl, 1 MgCl2, 1 CaCl2, 2.5 Na2HPO4, 2.5 KH2PO4, 1.2 K2HPO4, and 10 glucose; pH 7.4 was maintained by bubbling with 5% CO2. After an equilibration period in isotonic solution (5–15 min), a 50% hypotonic solution was applied, and the perfusate was continuously collected for 30-s intervals during the initial burst of ATP secretion (0–5 min) and during 1 min elsewhere. Nucleotide concentrations in the samples were evaluated by high-performance liquid chromatography (HPLC0 analysis, as described below. For studies in absence of extracellular calcium, CaCl2 was omitted, and the solutions were supplemented with 0.1 mM ethyleneglycoltetraacetic acid (EGTA). The 50% hypotonic solution was prepared by appropriate reduction of salt concentration, whereas divalent cation concentrations were kept constant.
HPLC quantification of adenine and uridine nucleotides
UTP concentrations were quantified by UDP-glucose pyrophosphorylase-based reaction [11]. Briefly, 100-μl samples were incubated in the presence of 0.5 U/ml UDPglucose pyrophosphorylase, 0.5 U/ml inorganic pyrophosphatase, 1.6 mM CaCl2, 2 mM MgCl2, 25 mM hydroxyethylpiperazine ethanesulfonic acid (HEPES) (pH 7.4), and ∼100,000 cpm 1 μM [14C]glucose-1P. Incubations lasted 1 h at 30°C. Reactions were terminated by heating the samples at 95°C for 2 min. Conversion of [14C]glucose-1P to [14C]UTP was determined by HPLC (Shimadzu) via a Nova Pack C18 column and ion-pairing mobile phase. Radioactivity was measured on-line with a Packard Flo-One detector. Assay sensitivity was 1 pmol (3 nM in 100-μl samples).
UDP was quantitatively phosphorylated in the presence of [γ32P]ATP using nucleoside diphosphokinase (NDPK). Briefly, 100-μl samples containing 0.1 U/ml NDPK, 0.1 μCi 60 nM [γ32P]ATP, 1.6 mM CaCl2, 2 mM MgCl2, and 25 mM HEPES (pH 7.4) were incubated for 5 min at 30°C. Reactions were terminated by heating the samples at 95°C for 2 min. The resulting conversion of [γ32P]ATP to [γ32P]UTP was monitored by HPLC [12]. A calibration curve employing known amounts of UDP was chartered in parallel during each assay. This assay allows UDP quantification with sensitivity of 20 fmol (0.2 nM in 100-μl samples).
For etheno (ɛ) derivatization, samples (200 μl) were derivatized for 30 min at 72°C in the presence of 1.0 M chloroacetaldehyde and 25 mM Na2HPO4 (pH 4.0). The resulting fluorescent 1,N6-ethenoadenine derivatives were analyzed by HLPC (Waters) in a Hamilton PRP-×100 anion exchange column, as described previously [4]. ɛ-ATP, ɛ-ADP, ɛ-AMP, and ɛ-Ado were quantified with a sensitivity of 200 fmol (1 nM in 200-μl samples).
FM1-43 studies
Bulk exocytosis was quantified in A549 cells as changes in the fluorescence intensity of FM1-43 incorporated into the plasma membrane [13, 14] and recorded by real-time confocal microscopy. Cells were washed with Hank’s balanced salt solution + 20 mM HEPES + 2 mM MgCl2 and 1.6 mM CaCl2 (HBSSH) or with HBSSH devoid of calcium (HBSSH 0 Ca). They were mounted onto the stage of a Leica SP2 AOBS confocal microscope equipped with HCX Apo L63x NA 0.9 immersion Leica lens and a 488-nm Argon laser. Experiments were initiated by incubation of the cells with 3 μM FM1-43. Cells were incubated for at least 10 min, and hypotonic shock was applied by decreasing salt concentration 33% while maintaining calcium, magnesium, and FM1-43 concentrations constant. Alternatively, cells incubated with FM1-43 in HBSSH or HBSSH 0 Ca were stimulated with 5- to 10-μM ionomycin. Real-time recording was performed by laser scanning in thexz axes with a galvostage, initially every 10 s and then every 30 s for the time periods indicated in the figures. Overall fluorescence intensity changes associated with the plasma membrane were estimated by measuring the intensity value associated with each pixel through time. The entire apical membrane compartment displayed in a confocal plane and five random regions of basolateral and subapical domains were analyzed, normalized to basal values (time = 0), and averaged for each region. Cell swelling was estimated as a change of cell height in thexz plane at different time points, normalized to basal values, and averaged.
Fura-2 calcium measurements
To load Fura-2, cells were incubated (1 h, 37°C, 5% CO2) in physiological solution containing 25 μM Fura-2-AM + 0.02% Pluronic F127 and 2.5 mM probenecid. This was followed by 30 min deesterification period in physiological solution containing probenecid. The physiological saline solution consisted of (in mM): 140 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 glucose and 10 HEPES, pH 7.4, adjusted with sodium hydroxide (NaOH). Fifty percent hypotonic medium was prepared by reducing the salt concentration while keeping divalent cation concentration constant. For calcium imaging, coverslips with Fura-2-loaded cells were mounted in the imaging/perfusion chamber attached to the heated platform (Warner Instruments) on the stage of an inverted microscope (Nikon TE300). The cells were exposed to alternate (200 ms) illumination at 340 nm and 380 nm with a high-pressure mercury lamp (100 W) via interference filters (Chroma Technology, Brattleboro, VT, USA) mounted on a filter wheel (Sutter Lambda 10-C, Sutter Instrument, Novato, CA, USA) and a dichroic mirror (510/540 nm, Chroma Technology). Fluorescence images were recorded at 15- to 60-s intervals with the digital camera and stored for later analysis.
Chemicals
For calcium-imaging experiments, Fura-2-AM was obtained from Molecular Probes, Invitrogen Corp. (Burlington, ON, Canada). Probenicid, Pluronic F127 and all other reagents were from Sigma Aldrich (Oakville, ON, Canada).
Results
Kinetics of nucleotide release
Figure 1 shows an example of the time course of nucleotide release induced by 50% hypotonic shock. For clarity, the release of adenine and uridine nucleotides appears on separate graphs: a and b, respectively. The kinetics of release were remarkably similar for all nucleotides. After the onset of hypotonic shock, nucleotide concentration increased rapidly, peaking at ∼2.5 min, followed by a gradual decay in the next 10–15 min. The average peak values from several separate experiments are shown in Fig. 1c. Interestingly, ATP was the major species at the peak of stimulated release, whereas for basal release, Ado was the predominant species. The rank order of nucleotide abundance at the peak was: ATP > Ado ≥ UTP > ADP ≈ UDP > AMP. Relative increases of nucleotide conconcentrations at the peak were also the highest for nucleotide triphosphates (NTPs) and nucleotide diphosphates (NDPs) (ATP 5.6-fold, ADP 3.7-fold, UTP >7-fold, UDP 27-fold), whereas the increase was smaller for AMP and Ado (1.4-fold and 2.1-fold, respectively).
Fig. 1Transient nucleotide release from A549 cells induced by 50% hypotonic shock. Time course of adenosine (a) and uridine (b) nucleotide release observed in response to 50% hypotonic shock. A representative experiment is shown out of four performed under the same conditions. Hypotonic shock was applied at t = 0 min and was preceded by 15-min equilibration in isotonic solution. Basal (t = 0 min) and peak (t = 2.5 min) nucleotide concentrations detected in perfusates (c). Average values [± standard deviation (SD)] are from three to four experiments, such as in a and b, except for uridine triphosphate (UTP) and uridine diphosphate (UDP), which are from a single experiment. * Due to the limited sensitivity of UTP evaluations (≈3 nM), the basal level of UTP was found to be below the detection limit and was not shown
Role of intracellular calcium ([Ca+2]i)
To investigate the role of intracellular calcium ([Ca+2]i) in adenine nucleotide release, we tested the effect of the Ca2+ ionophore ionomycin. In the absence of hypotonic shock, exposure to 5 μM ionomycin for 3 min induced transient nucleotide release from A549 cells (Fig. 2a). This release reached a peak at approximately 1.5 min, i.e., slightly earlier and at somewhat lesser absolute peak amplitude compared with that induced by 50% hypotonic shock. Otherwise, the kinetics and relative nucleotide abundance were similar for both stimuli. Almost complete inhibition of swelling-induced adenine nucleotide release was observed for A549 cells loaded with the Ca2+ chelator BAPTA, compared with control untreated cells of the same batch tested in parallel experiments (Fig. 2b, c, respectively). Fura-2 fluorescence [Ca+2]i measurements confirmed a dramatic reduction of the [Ca+2]i response to 50% hypotonic shock in BAPTA-loaded cells (Fig. 2d). These experiments demonstrate that elevation of [Ca+2]i was required to trigger adenine nucleotide release from A549 epithelial cells.
Fig. 2Effect of intracellular calcium ([Ca2+i]) modulators on nucleotide release. (a). In the absence of hypotonic shock, application of 5 μM ionomycin resulted in transient nucleotide release with kinetics similar to that induced by hypotonic shock. An example, out of two similar experiments, is shown. Loading A549 cells with the Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA) almost completely inhibited hypotonic-stress-induced nucleotide release compared with controls that were run in parallel (b and c, respectively) (n = 2). Effect of BAPTA on the [Ca2+]i response (d). The two traces represent changes of the Fura-2 fluorescence ratio at two excitation wavelengths, λ340/λ380, in response to 50% hypotonic shock, applied at t = 0 min, in control and BAPTA-loaded cells. Note that the rapid peak of the [Ca2+]i response was abolished in BAPTA-loaded cells. Similar responses were observed in four experiments
FM1-43 fluorescence changes implicate vesicular exocytosis
Strong calcium dependence of nucleotide release may indicate the involvement of vesicular exocytosis. To further explore this possibility, bulk exocytosis was examined by real-time confocal microscopy in cells bathed in FM1-43. Fluorescence intensity associated with the plasma membrane increased rapidly by incorporation of the soluble probe into the membrane and remained almost constant after 5 min in nonstimulated cells. Plasma-membrane-associated fluorescence intensity rose rapidly (∼20 s) with hypotonic shock stimulation. This change was accompanied by increased cell volume seen as increment of cell height (Fig. 3a, b). Cell volume peaked at about 3 min, followed by a regulatory volume decrease in the next 10–15 min (data not shown). A similar plasma-membrane-associated FM1-43 fluorescence intensity increase was observed in the absence of hypotonic shock when the cells were stimulated with ionomycin in the presence but not in the absence of calcium in the bathing solution (Fig. 4). These data indicate that both hypotonic-shock- and ionomycin-stimulated bulk exocytosis in A549 cells.
Fig. 3Effect of hypotonic shock on FM1-43 surface membrane staining. Confocal FM1-43 fluorescence images of confluent A549 cells after 20 min of dye loading in isotonic solution, followed by 50% hypotonic shock stimulation (a). The images depict the xz scans of the cell monolayer at different time points. Note the increase in cell size and fluorescence intensity consistent with hypotonic-shock-induced cell swelling and heightened bulk exocytosis. Time course of FM1-43 plasma membrane fluorescence and cell height changes during hypotonic shock (b)Fig. 4Effect of ionomycin on FM1-43 surface-membrane staining. Confocal FM1-43 fluorescence images of confluent A549 cells prior to and after 4-min exposure to 5 μM ionomycin in the presence of extracellular calcium. (a). The images show the xz scans of the cell monolayer at different time points. Note the increase in fluorescence intensity after ionomycin stimulation, indicating enhanced exocytosis. Time course of FM1-43 plasma membrane fluorescence changes of unstimulated cells (Δ CON) and cells during ionomycin treatment in the presence (▪ ION) or absence (• ION 0 Ca) of extracellular calcium (b)
Discussion
In this study, we used the flow-through chamber and an etheno derivatization measurement technique to examine the kinetics of nucleotide release from A549 lung epithelial cells induced by hypotonic shock. We demonstrated that ATP, ADP, AMP, Ado, UTP, and UDP appeared in the perfusates with the same kinetics, peaking at ∼2.5 min (Fig. 1), which coincided with the peak of [Ca+2]i elevation evoked by hypotonic shock (Fig. 2d). Similar tight temporal correlation between the [Ca+2]i signal and ATP release, measured by luciferase-luciferin luminescence assay, was reported previously for A549, 16HBE14o− epithelial cells and NIH3T3 fibroblasts [10]. Such a close temporal association suggests that hypotonic-shock-induced elevation of [Ca+2]i is a trigger for the release of nucleotides. This was supported by the strong inhibitory effect of BAPTA-AM and the induction of nucleotide release in the absence of hypotonic shock solely by the rise of [Ca+2]i with ionomycin. These data demonstrate tight Ca2+ dependence of nucleotide release, and point to a Ca2+-dependent exocytotic mechanism. It should be noted that the bulk of [Ca+2]i changes shown in Fig. 2d likely represent superposition of spatially separated, distinct [Ca+2]i responses, one evoked directly by hypotonic shock and the other due to autocrine/paracrine actions of the released nucleotides on P2Y receptors. The latter effects might be minimized, at least partially, by continuous perfusion, which reduced nucleotide concentration in the bulk of the chamber far below the IC50 of P2Y receptor activation; however, this possibility requires further study.
Consistent with a Ca2+-dependent exocytotic mechanism, FM1-43 fluorescence studies revealed enhanced incorporation of the probe into the plasma membrane induced by hypotonic shock or ionomycin treatment. Increased plasma membrane-associated FM1-43 fluorescence intensity was a measure of the cumulative amount of membrane added by exocytosis [13]. FM1-43 fluorescence increased rapidly during the first 20 s of hypotonic shock (Fig. 3). Interestingly, the initial fluorescence increment exceeded the cell-surface increase estimated from xz confocal scans. In our previous study using a dual-image 3D cell reconstruction technique [15], we observed only an 11–30% surface increase of single substrate-attached A549 cells swollen in 50% hypotonic solution [10]. This difference could be attributed in part to different methods of cell-height and surface evaluation in that study, as well as differences in swelling responses of single isolated cells compared with confluent cell monolayer. However, a similar discrepancy between FM1-43 fluorescence, cell-surface and membrane electrical capacitance has been also reported by others, e.g., with pituitary lactotrophs, where the dense granules docked into the plasma membrane were intensely stained by FM1-43 in addition to the membrane added to the cell surface [16]. Therefore, the divergence of fluorescence and the cell-surface changes seen in our study with A549 cells may also result in part from FM1-43 staining of the lipophilic content of fused vesicles in addition to the staining of fused vesicular membranes. The FM1-43 dye is often employed to fluorescently marked surfactant-containing lamellar bodies in alveolar type II (ATII) cells [17, 18] and A549 cells, a model of ATII cells, also contain granules enriched in surface-active phospholipids [19]. However, the exact contribution of such granules to increased FM1-43 fluorescence in stimulated A549 cells was out of the scope of the present investigation and will require future direct study.
Vesicle exocytosis may be mechanistically associated with nucleotide release in two ways: nucleotides can be delivered to the extracellular medium as cargo molecules within exocytotic vesicles or, alternatively, secreted from the cytosol via vesicle-associated nucleotide-conducting channels or transporters transiently expressed at the plasma membrane as a consequence of vesicle-plasma membrane fusion. Analysis of the relative abundance of different nucleotide species that appear in the extracellular medium may help to distinguish between these two mechanisms. ATP and UDP sugars accumulate in the lumen of the secretory pathway up to 20- to 50-fold above their cytosolic concentrations, and luminal utilization of these molecules generates ADP and UDP. UDP is the major uridine nucleotide detected in ER/Golgi fractions [20] and, eventually, a UDP-selective apyrase (UDPase) converts UDP to urodine monophosphate (UMP) [21]. ER/Golgi ADP and UMP are exchanged for cytosolic ATP and UDP glucose via specific transporters [22–25]. As intraluminal nucleotides are not subject to the mechanisms that retrieve resident ER/Golgi proteins, they are predicted to remain within trafficking vesicles and to be released to the extracellular space from the secretory pathway. Our results suggest that this mechanism likely accounted for UDP release from A549 cells. Figure 1c illustrates that hypotonic stimulation promoted a sharp increase in UDP levels, with a UTP:UDP concentration ratio of 2:1. As the cytosolic UTP:UDP ratio is >10:1 [26], our results suggest, at least in part, a vesicular rather than a cytosolic source of UDP and, as a corollary, of ATP and ADP.
In summary, our results provide strong evidence for Ca2+-dependent vesicular exocytosis as a major mechanism of adenosine and uridine nucleotide release from A549 epithelial cells induced by hypotonic stress. Part of this release involves vesicles of the protein secretory pathway. Further investigations are needed to clarify the origin and contribution of other vesicular pools as well as the mechanisms and sources of intracellular Ca2+ elevations that evoke nucleotide secretion.
|
[
"hypotonic shock",
"exocytosis",
"lung epithelial cells",
"ca2+ dependence",
"nucleotide secretion"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Behav_Genet-3-1-2039802
|
Identification of Genetic and Epigenetic Variations in a Rat Model for Neurodevelopmental Disorders
|
A combination of genetic variations, epimutations and environmental factors may be involved in the etiology of complex neurodevelopmental disorders like schizophrenia. To study such disorders, we use apomorphine-unsusceptible (APO-UNSUS) Wistar rats and their phenotypic counterpart apomorphine-susceptible (APO-SUS) rats that display a complex phenotype remarkably similar to that of schizophrenic patients. As the molecular basis of the APO-SUS/UNSUS rat model, we recently identified a genomic rearrangement of the Aph-1b gene. Here, we discovered between the two rat lines differences other than the Aph-1b gene defect, including a remarkable cluster of genetic variations, two variants corresponding to topoisomerase II-based recombination hot spots and an epigenetic (DNA methylation) difference in cerebellum and (hypo)thalamic but not hippocampal genomic DNA. Furthermore, genetic variations were found to correlate with the degree of apomorphine susceptibility in unselected Wistar rats. Together, the results show that a number of genetic and epigenetic differences exist between the APO-SUS and -UNSUS rat genomes, raising the possibility that in addition to the Aph-1b gene defect the newly identified variations may also contribute to the complex APO-SUS phenotype.
Introduction
Schizophrenia is a neurodevelopmental disorder affecting nearly 1% of the world’s population (Jablensky et al. 1987), and is characterized by positive and negative symptoms (Kay and Opler 1987). The aetiology of schizophrenia and other related disorders, such as schizoaffective and bipolar disorder, is still unclear. Twin, family and adoption studies have suggested that complex interactions at the genetic and environmental level underlie the aetiology of schizophrenia (Gottesman 1991). It is thought that gene variations by themselves do not result in schizophrenia, but they can establish a predisposition status that, when combined with environmental stressors, may lead to schizophrenia pathogenesis. Numerous environmental factors, such as viral infections (Mednick et al. 1988), insufficient folate and methionine levels (Regland 2005), or repeated psychological stress (Goldstein 1987), can influence brain development of prenatal or early postnatal individuals with a genetic predisposition for neuropsychiatric disorders. Due to the heterogeneity in genetic and environmental interactions, most of the genes and pathways for schizophrenia and for other complex disorders are still unknown.
To get insight into the gene (or genes) that may be involved in schizophrenia pathogenesis, a rat model was developed with schizophrenia-like features. This model was based on the behavioural response of Wistar rats to the dopamine agonist apomorphine (Cools et al. 1990). The apomorphine-susceptible (APO-SUS) rat line displayed many features of psychopathology, with similar disturbances at the behavioural, physiological, endocrinological and pharmacological level as seen in schizophrenics (Ellenbroek and Cools 2002). For example, APO-SUS rats have a reduced prepulse inhibition and latent inhibition (Ellenbroek et al. 1995), display a higher plasma release of adrenocorticotropin (ACTH) and corticosteroids in response to novelty (Rots et al. 1995), are more sensitive to dopamimetic drugs (Ellenbroek et al. 2000), and have a higher susceptibility to inflammatory and infectious diseases when compared to apomorphine-unsusceptible (APO-UNSUS) rats (Kavelaars et al. 1997). We therefore wondered about the molecular-genetic basis underlying the APO-SUS/-UNSUS rat model and recently identified a genetic difference between the two rat lines (Coolen et al. 2005). Whereas APO-UNSUS rats harbour three gene copies of the γ-secretase component Aph-1b, APO-SUS rats have only one or two copies. This gene-dosage imbalance was due to an unequal crossing over event (nonallelic homologous recombination) between two direct repeats (a segmental duplication) within the Aph-1b locus. In addition, we observed a direct link between the Aph-1b genotypes and a number of phenotypic APO-SUS and -UNSUS characteristics (Coolen et al. 2005). Approximately 10 years after developing the APO-SUS and -UNSUS lines a second, independent breeding procedure was started that resulted in rats with features similar to those displayed by the original APO-SUS and -UNSUS rat lines (Ellenbroek and Cools 2002). Interestingly, the replicated rat lines also resulted in APO-UNSUS rats with three Aph-1b gene copy numbers and APO-SUS rats with only one or two gene copies (Coolen et al. 2005).
In the present study, we wondered whether genetic variations other than the Aph-1b gene-dosage imbalance may be present between the APO-SUS and -UNSUS rats, and whether epigenetic factors may be involved as well. Epigenetics has been defined as heritable changes in gene expression that do not occur by changes in the DNA sequence, but by modifications in DNA methylation and chromatin remodeling (Wolffe and Matzke 1999), or, in its widest sense, as any change in an organism that is not due to genetic factors (Van de Vijver et al. 2002). Increasing evidence suggests that epigenetic modifications play a role in disease susceptibility (reviewed by Jirtle and Skinner 2007). We used the arbitrarily primed-polymerase chain reaction (AP-PCR) fingerprinting technique (Welsh and McClelland 1990) to analyse the genomes and epigenomes (DNA methylation) of the APO-SUS and -UNSUS rats. Comparison of the AP-PCR fingerprints generated from the genomic DNAs of the two rat lines revealed genetic as well as epigenetic alterations and we conclude that, besides in the Aph-1b locus, a number of other variations are present in the APO-SUS and -UNSUS genomes and epigenomes.
Materials and methods
Experimental animals
The generation of the APO-SUS and -UNSUS rat lines with a high or low susceptibility for apomorphine, respectively, has been described previously (Cools et al. 1990). The present experiments were performed with male APO-SUS and -UNSUS rats belonging to the 32nd (original lines) and 18th (replicate lines) generation. At post-natal day 60 (PND60), APO-SUS and -UNSUS rats were sacrificed and the hippocampus, cerebellum and the combined thalamus/hypothalamus (further denoted as (hypo)thalamus) were isolated. To establish their apomorphine susceptibility, unselected male Wistar rats of the Nijmegen outbred population (PND60) were injected with apomorphine (1.5 mg/kg s.c.) and gnawing scores were measured in a gnawing box for 45 min, as described previously (Cools et al. 1990). Immediately following the measurements, the rats were sacrificed and the same tissues (hippocampus, cerebellum and (hypo)thalamus) were removed. All rats were bred and reared in the Central Animal Facility of the Radboud University Nijmegen under approved animal protocols and in accordance with institutional guidelines.
Arbitrarily primed-PCR
Genomic DNAs were isolated from hippocampus, cerebellum and (hypo)thalamus using standard procedures involving the use of proteinase K and phenol extraction. Two micrograms of genomic DNA were digested with 20 units of RsaI, 20 units RsaI in combination with the methylation-sensitive enzyme HpaII, or 20 units RsaI and MspI (MBI Fermentas) in a total volume of 40 μl at 37°C for 16 h. HpaII does not cut DNA if the internal cytosine (CCGG) is methylated, whereas MspI is insensitive to DNA methylation. Using such combinations of methylation-sensitive and -insensitive enzymes allows genome-wide screening for differences at the genetic level (single-nucleotide polymorphisms––SNPs, duplications, insertions, deletions and recombinations) as well as the epigenetic (DNA methylation) level. Restriction enzymes were heat inactivated by incubating the reactions at 65°C for 20 min. Digested DNA (100 ng) was amplified using AP-PCR (Welsh and McClelland 1990) with a single primer. PCRs were performed in a total volume of 25 μl containing 10 mM Tris-HCl (pH 8.3), 2.5 mM MgCl2, 50 mM KCl, 0.001% gelatin, 0.25% Nonidet P-40, 0.25% Tween-20, 200 μM each of the four deoxynucleotide triphosphates, ∼1 μCi of [α-32P]dCTP (3000 Ci/mmol, Amersham Corp.), 25 pmol of primer (AP-1: 5′-AACCCTCACCCTAACCCCGG-3′, AP-7: 5′-AACCCTCACCCTAAGGCGCG-3′, AP-777: 5′-CACTCCTCTACAAGGTGCCG-3′ or Topo: 5′-GCCTCCTTGCAGGTCTTT-3′), and 0.8 units of Taq polymerase (MBI Fermentas). Reactions were carried out in a thermal cycler (Perkin-Elmer) with five cycles of low stringency (94°C for 30 s, 40°C for 60 s, 72°C for 1.5 min), followed by 30 cycles of high stringency (94°C for 15 s, 55°C for 15 s, 72°C for 1 min). Two microliters of the PCR products were analysed on high-resolution 5% polyacrylamide gels under denaturing conditions (7 M urea) for 4–4.5 h at 70 W. Gels were dried and radiolabelled DNA was visualized by autoradiography at −70°C (CEA AB, Sweden).
Cloning and sequencing of AP-PCR fragments
AP-PCR fragments generated from APO-SUS and -UNSUS rat genomic DNAs were excised from the dried gels and incubated in 50 μl MilliQ at 80°C for 10 min. The eluted DNA (two microliters) was reamplified with the same primer as used for the AP-PCR to generate sufficient amounts of template for subsequent cloning. The reactions were carried out for 40 cycles of 94°C for 1 min, 55°C for 30 s, 72°C for 1 min, under the same conditions as described in the AP-PCR protocol (except that [α-32P]dCTP was not included). The PCR products were purified, cloned into the pGEM-T easy vector (Promega) and sequenced with a T7 or Sp6 primer according to the manufacturer’s instructions using the ABI310 machine (Applied Biosystems).
Sequencing and genotyping of chromosomal region 9q22
A 1948-bp fragment that harbours the nucleotide sequence corresponding to product 3 was obtained by PCR on genomic DNA derived from (hypo)thalamic tissue of an APO-UNSUS rat using forward primer 5′-GGGAAGCAACGCATCCTG-3′ and reverse primer 5′-CATATCAAAGCACCAAGTCCACAG-3′. The DNA was subsequently purified and directly sequenced using the ABI310 machine (Applied Biosystems). Genotyping of chromosomal region 9q22 was performed with PCR using primers specific for either the APO-SUS or APO-UNSUS genomic sequence. Briefly, PCRs were performed in a total volume of 20 μl containing 50 ng genomic DNA, 10 mM Tris-HCl (pH 8.3), 2.5 mM MgCl2, 50 mM KCl, 0.001% gelatin, 0.25% Nonidet P-40, 0.25% Tween-20, 200 μM each of the four deoxynucleotide triphosphates, 0.6 μM of each primer (FW: 5′-AACACTTGGACTCATTCTCACTGG-[G (SUS) or T (UNSUS)]-3′ and RV: 5′-CCTGGATGGAATGTTGACAC-[C (SUS) or T (UNSUS)]-3′), and 0.8 units of Taq polymerase (MBI Fermentas). Reactions were carried out at 94°C for 60 s, 58°C for 60 s and 72°C for 60 s for 35 cycles. Products were analysed on a 1% agarose gel.
Quantification and statistics
Quantification of AP-PCR products was performed using the Labworks 4.0 program (UVP BioImaging Systems, Cambridge, UK) and statistical evaluation was performed by means of an unpaired Student’s t-test.
Results
AP-PCR DNA fingerprint patterns of the APO-SUS and APO-UNSUS rat genomes and epigenomes
In order to identify differences between the genomes and epigenomes of APO-SUS and -UNSUS rats, we performed a comparative analysis of fingerprints of AP-PCR products generated from genomic DNAs of the two rat lines. Initially, genomic DNAs isolated from APO-SUS and -UNSUS (hypo)thalamic tissue and digested with RsaI in combination with the methylation-sensitive restriction enzyme HpaII (CCGG) was analysed using arbitrary primers AP-1, AP-7 or AP-777. These primers were selected from a total set of ten primers because they gave fingerprints with reproducible and discrete products (data not shown). Typical AP-PCR fingerprints obtained with the three selected arbitrary primers are shown in Fig. 1. With each arbitrary primer ∼30 chromosomal fragments were reproducibly amplified. DNAs digested with RsaI and the methylation-insensitive enzyme MspI served as controls to determine whether the observed differences were due to a differential methylation of the CCGG sequence or a genetic polymorphism in this sequence. AP-PCR analysis with primer AP-1 revealed 16 products corresponding to fragments without an HpaII site (“genetic fragments”) and 23 products corresponding to fragments containing an HpaII site (“epigenetic fragments”). Analysis with AP-PCR primer AP-7 showed 17 genetic and 11 epigenetic fragments, and with AP-777 primer 22 genetic and 6 epigenetic fragments.
Fig. 1AP-PCR analysis of genomic DNAs from APO-UNSUS (U) and APO-SUS (S) (hypo)thalamus. AP-PCR was performed with primers AP-1, AP-7 and AP-777 using genomic DNAs digested with RsaI and HpaII (H) or RsaI and MspI (M) as templates. Epigenetic products (methylation-sensitive and thus absent in the MspI lanes) are indicated by closed arrows and genetic products (methylation-insensitive and thus present in the MspI lanes) by open arrows
Genetic variations between the APO-SUS and APO-UNSUS rat genomes
Comparison of the genetic fingerprints generated with primers AP-1, -7 and -777 revealed three reproducible variations between the genomic DNAs from the original (F32) APO-SUS and -UNSUS rats, designated products 1, 2 and 3 (Fig. 2a). Product 1 was less prominent in the APO-SUS than in the APO-UNSUS rats, product 2 was found only in APO-SUS, while the level of product 3 was higher in the APO-SUS than -UNSUS rats. Interestingly, the three products were also present in the replicate (F18) lines and at the same levels, indicating that the replication of the APO-SUS and -UNSUS lines had resulted in a similar genotypic distribution. Next, digestions using RsaI in combination with MspI were used to examine whether the observed differences were due to a genetic or an epigenetic alteration. Following digestion with RsaI and MspI, products 1, 2 and 3 were still found, indicating that the presence of the three AP-PCR products was due to genetic differences (Fig. 2b).
Fig. 2Genetic variations in genomic DNAs from APO-SUS (S) and APO-UNSUS (U) (hypo)thalamus. Rats were from the original (F32) or the replicate (F18) rat lines. (A) Products generated by AP-PCR using primers AP-1 (product 1), AP-7 (product 2) or AP-777 (product 3) on genomic DNAs digested with the methylation-insensitive enzyme RsaI and the methylation-sensitive enzyme HpaII. ND = not determined. (B) Products 1, 2 and 3 generated by AP-PCR on genomic DNAs digested with RsaI and HpaII (H) or with the two methylation-insensitive enzymes RsaI and MspI (M). MspI products served as controls to determine whether products 1, 2 and 3 were due to differential methylation or to a genetic polymorphism. The fact that products 1, 2 and 3 were still present following MspI digestion indicates that they represent products without an HpaII site (“genetic fragments”). (C) Products generated by AP-PCR using a topoisomerase II binding site consensus sequence (5′-GCCTCCTTGCAGGTCTTT-3′) on genomic DNAs digested with the methylation-insensitive enzymes EcoRI (product 4) or MboI (product 5). Arrows indicate increased amounts of the AP-PCR products
We previously discovered that the Aph-1b gene-dosage imbalance between the APO-SUS and -UNSUS rats is the result of a DNA recombination event between the two Aph-1b genes. Furthermore, we identified the region in which the recombination occurred, namely in a region of 1106 nucleotides that is identical between the two genes and encompasses exon 5 (Coolen et al. 2005). In the present study, we decided to examine in detail the site of recombination and found a topoisomerase II binding site (5′-ACCCACCTGCTGGTGTCC-3′) in the DNA region harbouring the recombination site. Topoisomerase II binding sites (with the vertebrate consensus sequence 5′-RNYNNCNNGYNGKTNYNY-3′) (Spitzner and Muller 1988) are known to be hotspots where DNA recombination events occur easily (Craig and Nash 1983). We therefore wondered whether other topoisomerase II binding sites could have led to additional differences between the APO-SUS and -UNSUS rat genomes. Interestingly, using a primer based on the topoisomerase II binding site consensus for PCR analysis of genomic DNAs digested with EcoRI or MboI revealed two differences between the APO-SUS (n = 3) and -UNSUS (n = 2) rat genomes of the original (F32) lines, designated products 4 and 5 (Fig. 2c). Product 4 was present in APO-SUS but not in APO-UNSUS rat genomic DNA. In the replicate APO-SUS and -UNSUS lines (F18), the genomes of two APO-UNSUS rats did also not contain product 4, whereas it was present in two of the four APO-SUS rats tested. Product 5 was present in three of the four APO-UNSUS rats examined (in both the original and the replicate lines), whereas it was not observed in the seven APO-SUS rats tested (Fig. 2c).
Epigenetic variations between the APO-SUS and APO-UNSUS rat genomes
We then wondered whether, besides the five genetic variations, also epigenetic variations would be present between the APO-SUS and -UNSUS rat lines. AP-PCR analysis using primer AP-1 on RsaI- and HpaII-digested genomic DNAs from APO-SUS and -UNSUS (hypo)thalamus revealed one epigenetic variation, designated the E1-product (Fig. 3a). The difference was observed in both the original APO-SUS and -UNSUS rats as well as the replicated lines. An ∼1.4-fold reduced amount of the E1-product was observed in APO-SUS when compared with APO-UNSUS genomic DNAs (n = 12, P < 0.05). Since DNA methylation may be tissue specific, we decided to examine the E1-product in two other brain tissues. The level of the E1-product was ∼2-fold reduced in genomic DNAs from the cerebellum of APO-SUS compared to APO-UNSUS rats (n = 4, P < 0.01), while no difference in the E1-levels was found in the hippocampus of the APO-SUS and -UNSUS rats (n = 4).
Fig. 3Epigenetic variation in genomic DNAs from APO-SUS (S) and APO-UNSUS (U) (hypo)thalamus. Rats were from the original (F32) or the replicate (F18) rat lines. (A) Product E1 generated by AP-PCR using primer AP-1 on genomic DNAs digested with RsaI and HpaII. Arrow indicates a representative example of an AP-PCR product that did not show variable amounts using primer AP-1 (randomly chosen out of 38 products) and was used for normalization. (B) The AP-PCR E1-product generated from genomic DNAs from cerebellum, hippocampus and (hypo)thalamus. All rats were from different nests. The products indicated by the arrow were used for normalisation. Amounts for the epigenetic E1-product were significantly different between the APO-SUS and APO-UNSUS rats in the cerebellum (**P < 0.01; n = 4, plus s.e.m.) and (hypo)thalamus (*P < 0.05; n = 12, plus s.e.m.), but not in the hippocampus (n = 4, plus s.e.m.)
Chromosomal localizations of the genetic and epigenetic variations between the APO-SUS and -UNSUS rat genomes
To identify the locations of the genetic and epigenetic variations within the APO-SUS and -UNSUS rat genomes, AP-PCR fragments 1, 2, 3 and E1 were excised from the gel and each fragment was reamplified with the primer used for the original AP-PCR reaction. DNA sequence analysis of the amplified PCR fragments and database searches with the obtained nucleotide sequences revealed the chromosomal localisations of the four fragments; product 1 was located on chromosome 19q11, ∼3.3 kb upstream of the first exon of the GAIP-interacting protein, C terminus (GIPC) gene; product 2 was part of a repeat sequence located on chromosome 2q34; product 3 was located in the first intron of the myosin 1b gene on chromosome 9q22; the epigenetic variation E1 was located on chromosome 6q31, downstream of the Jun dimerization protein 2 (NP_446346.1) and upstream of the ATF-like basic leucine zipper transcriptional factor B-ATF (SF-HT-activated gene 2; XP_216745.2). Since product 3 was localized within a gene, we decided to analyse this chromosomal region in more detail. Sequence analysis of the DNA region corresponding to AP-PCR fragment 3 and its surrounding region revealed a remarkably high number of genetic variations: 10 of the 1948 base pairs analysed were different between the APO-SUS and -UNSUS genomes (Fig. 4). Comparison of the nucleotide sequences of the APO-SUS and -UNSUS DNA regions with the corresponding database sequence (geneID: 117057) revealed a 100% identity between the database and the APO-UNSUS sequences, indicating that the APO-SUS genome has diverged from the database sequence. Next, more APO-SUS and -UNSUS rats were genotyped for this region. All APO-UNSUS rats tested (n = 5) indeed showed the database sequence, whereas the APO-SUS rats tested (n = 5) all contained the relatively high number of variations in this chromosomal region.
Fig. 4Nucleotide sequence of chromosomal region 9q22, corresponding to product 3, in APO-SUS and APO-UNSUS rats. APO-SUS/-UNSUS genomic variations are indicated between brackets; the first nucleotide represents the APO-UNSUS sequence, the second nucleotide the APO-SUS sequence. The nucleotide sequences corresponding to the annealing sites of arbitrary primer AP-777 are underlined
The newly identified genetic and epigenetic variations, and apomorphine susceptibility in Wistar rats
We wondered whether the molecular-genetic difference between the APO-SUS and -UNSUS rats (the Aph-1b gene-dosage imbalance) would also be present in the Nijmegen outbred population of Wistar rats, or if the imbalance was generated during the breeding of the rat lines. PCR analysis of the Aph-1b locus in the 50 Wistar rats examined revealed in all cases the presence of three copies of the Aph-1b gene, suggesting that the reduction of Aph-1b copies in the APO-SUS rats had been induced during the breeding of the rats.
We next wondered whether the newly identified genetic and epigenetic variations in the APO-SUS and -UNSUS genomes were also induced during the breeding of the two rat lines, or if these variants are already present in Wistar rats. For this purpose, we performed AP-PCR analysis of (hypo)thalamic genomic DNAs from Wistar rats using primers AP-1, -7 and -777. The fingerprints revealed the presence of the APO-SUS as well as the -UNSUS variants of the genetic products 1, 2 and 3 (n = 8) and the epigenetic E1-product (n = 4) in Wistar rats (Fig. 5), indicating that the newly identified variations were not induced during breeding of the APO-SUS and -UNSUS lines, but were already present in the original Wistar population.
Fig. 5AP-PCR analysis (products 1, 2 and E1) and specific PCR (product 3) analysis of genomic DNAs from Wistar rats with low or high apomorphine susceptibility. (A) The presence or absence of products 1, 2 and 3 (genetic differences) was analysed in genomic DNAs from the (hypo)thalamus of Wistar rats with low (<10 gnaws in 45 min) or high (>500 gnaws in 45 min) apomorphine susceptibility. U: APO-UNSUS, S: APO-SUS. Lower panel: the amounts of the three products present in the Wistar rats were compared with the amounts found in the APO-SUS and -UNSUS rats. (■) the genotype of the genetic products 1, 2 or 3 in the Wistar rats with low apomorphine susceptibility is similar to the products 1, 2 and 3 genotype of the APO-UNSUS rats; (□) the genotype in the Wistar rats with low apomorphine susceptibility deviates from the APO-UNSUS genotype;
() the genotype in the Wistar rats with high apomorphine susceptibility is comparable with the genotype in the APO-SUS rat;
() the genotype in the Wistar rats with high apomorphine susceptibility deviates from the genotype in the APO-SUS rats. (B) AP-PCR analysis of the epigenetic E1-product (E1) on genomic DNAs from cerebellum and hippocampus of four Wistar rats. The product indicated by the arrow is an example of a product that did not show variable amounts and was used for normalisation of the E1-product. The amounts of the E1-products generated from the cerebellum or hippocampus were not different between the two Wistar rats with low gnawing scores (<10 gnaws in 45 min) and the two Wistar rats with high gnawing scores (>500 gnaws in 45 min)
To study whether the presence or absence of the genetic AP-PCR products observed in the APO-SUS and -UNSUS rats (products 1, 2 and 3) was linked to the apomorphine susceptibility of Wistar rats, we examined a group of five rats with low susceptibility to apomorphine (<10 gnaws per 45 min) and a highly susceptible group of three rats (>500 gnaws per 45 min). Product 1, which was less abundant in APO-SUS than -UNSUS rats, was not present in the three Wistar rats with high gnawing scores, but was also not found in two of the five Wistar rats with low gnawing scores. Of the eight rats examined, the APO-SUS-specific product 2 was found in two Wistar rats with a high and only one rat with a low degree of apomorphine susceptibility. The DNA region corresponding to product 3 was found twice as the APO-UNSUS sequence in the group consisting of the low-apomorphine-susceptible Wistar rats and once in the group of the high-apomorphine-susceptible rats (Fig. 5a). These results indicate that none of the three products was directly linked to the apomorphine susceptibility of the Wistar rats. However, apomorphine susceptibility may not be the result of only a single genetic variation, but rather of multiple genetic and epigenetic alterations. Interestingly, 67% of the three genetic products present in the Wistar rats with low apomorphine susceptibility were also found in the APO-UNSUS genome, and 78% of the three products in Wistar rats with high apomorphine susceptibility were APO-SUS-specific variants (Fig. 5a), indicating that a combination of the three genetic products may well be linked to apomorphine susceptibility. To study the link between the epigenetic E1-product and apomorphine susceptibility, two Wistar rats with low gnawing scores and two with high gnawing scores were examined for the amount of the E1-product in the cerebellum. Similar amounts of the E1-product as detected in APO-UNSUS rats were found in one Wistar rat with a low and in one Wistar rat with a high gnawing score. The amount of the E1-product detected in APO-SUS rats was found in one Wistar rat with low and in one Wistar rat with high apomorphine susceptibility, indicating that no direct link exists between the E1-product and apomorphine susceptibility. As observed in the APO-SUS and -UNSUS rats, no variation was detected in the level of the E1-product in hippocampal genomic DNAs from the four Wistar rats tested (Fig. 5b).
Discussion
In this study, we investigated the genetic and epigenetic background of the phenotypically well-characterized APO-SUS and -UNSUS rats. Unravelling the molecular basis of this rat model may help in our understanding of complex human neurodevelopmental disorders, since many of the characteristics of the APO-SUS rat line are also observed in schizophrenic patients (Ellenbroek and Cools 2002). We recently identified a gene-dosage imbalance in the Aph-1b locus of the APO-SUS and -UNSUS rats, leading to a reduced expression of the Aph-1b gene in APO-SUS rats and a segregation with a number of behavioural parameters (Coolen et al. 2005). Here, we used AP-PCR analysis as an approach to explore the presence of any other alterations in the two genomes and epigenomes. We indeed identified additional genetic and epigenetic variations, indicating that the Aph-1b locus might not be solely responsible for the observed phenotypes of the two rat lines and suggesting a multi-genetic and -epigenetic origin of the differences observed between the APO-SUS and -UNSUS rats. Importantly, we found similar AP-PCR patterns in both the original and replicate APO-SUS and -UNSUS lines. It is therefore highly unlikely that the observed variations between the APO-SUS and -UNSUS rat genomes were simply due to coincidence.
Detailed analysis of one of the newly identified genetic alterations (in the DNA region corresponding to product 3) revealed in APO-SUS rats a cluster of variations in the myosin 1b gene. Hence, this cluster appears to be a hotspot for genetic instability. Product 1 was located upstream of the GIPC gene encoding a protein interacting with membrane-associated and transmembrane proteins, including the dopamine receptors D2 and D3 (Jeanneteau et al. 2004). The genetic variation in this locus may therefore contribute to the differences in apomorphine susceptibility between the APO-SUS and -UNSUS rat lines. We also identified genetic alterations using a primer corresponding to a topoisomerase II binding site consensus. The rationale for this study was based on our present finding that a topoisomerase II binding site was located at the recombination site in the Aph-1b locus. The two newly identified, topoisomerase II-based genetic variations might point to a more general role for topoisomerase II binding sites in psychopathological mechanisms. We hypothesize that at these sites environmental factors, such as stress during early development, may cause an increase in the incidence of recombination and other mutagenic events, leading to brain dysfunction and affected behavior. Besides the genetic differences, one tissue-specific epigenetic variation was found between the APO-SUS and -UNSUS epigenomes. At present it is not clear what, if any, functional consequence should be attributed to the decrease in the methylation status of this CpG in the (hypo)thalamus and cerebellum, but not in the hippocampus, of APO-SUS relative to APO-UNSUS rats.
Earlier microarray analysis of mRNA expression in the hippocampus of the APO-SUS and -UNSUS rats (∼7000 full-length sequences and ∼1000 EST clusters) revealed that only Aph-1b was differentially expressed (Coolen et al. 2005). The newly identified variations open the possibility that for an explanation of the background of the rat model one has to consider more genes that operate together in a multi-genetic and -epigenetic setting with several susceptibility loci. Thus, based on our present findings, more than one locus may be responsible for the complex phenotype of APO-SUS rats. Recently, it has been established that the contribution of genetic modifiers is also of importance for the outcome of a phenotype, since they can modulate the severity of the affected phenotype and the phenotypic characteristics without having a clear effect on the normal situation (Nadeau 2001; Nadeau and Topol 2006). Hence, as part of the genetic and epigenetic background of the APO-SUS and -UNSUS rat lines, our newly identified variants can be genetic and epigenetic modifiers influencing the phenotypic expression of the model. Insight into the genetic and epigenetic background may provide diagnostic tools, and clues for mechanisms and pathways to explain complex disorders.
Besides the multiple variations in the APO-SUS and -UNSUS rat genomes and epigenomes, Aph-1b will presumably be a major player in the development of the complex APO-SUS phenotype, since it has a broad cellular effect via tissue-specific cleavage of many different substrates (Coolen et al. 2006b) and is functional already during early development (Coolen et al. 2006a). Remarkably, however, the number of Aph-1b gene copies was not linked to apomorphine susceptibility in Wistar rats. Whereas their apomorphine susceptibility varied, all unselected Wistar rats tested harboured three copies of the Aph-1b gene. Therefore, genetic and epigenetic factors other than Aph-1b will likely contribute to the susceptibility for apomorphine. We now indeed found a correlation between the newly identified genetic variations and the apomorphine susceptibility in the Wistar population, confirming our hypothesis that apomorphine susceptibility is caused by a number of genetic and epigenetic factors. A combination of the newly identified variations may have thus initially contributed to the degree of apomorphine susceptibility in the original Wistar rat population. Subsequent apomorphine injections to determine the susceptibility for the drug during the breeding of the APO-SUS and -UNSUS lines may have acted as an environmental stressor, triggering the Aph-1b recombination event at the topoisomerase II binding site only in rats with a high susceptibility for apomorphine. The induced gene-dosage imbalance of the Aph-1b gene, probably in combination with other genetic or epigenetic factors, could then have led to the complex phenotype observed in the APO-SUS rats.
In conclusion, the present findings suggest that psychopathological disturbances may be the result of multiple genetic as well as epigenetic factors. We infer that our newly identified variations are susceptibility loci for schizophrenia-like features in the rat and may give new insights into the genetic and epigenetic background of complex neurodevelopmental disorders.
|
[
"genetic and epigenetic variations",
"neurodevelopmental disorders",
"aph-1b",
"ap-pcr",
"apo-sus rat model"
] |
[
"P",
"P",
"P",
"P",
"R"
] |
Int_J_Biochem_Cell_Biol-1-5-1885942
|
Keratinocyte serum-free medium maintains long-term liver gene expression and function in cultured rat hepatocytes by preventing the loss of liver-enriched transcription factors
|
Freshly isolated hepatocytes rapidly lose their differentiated properties when placed in culture. Therefore, production of a simple culture system for maintaining the phenotype of hepatocytes in culture would greatly facilitate their study. Our aim was to identify conditions that could maintain the differentiated properties of hepatocytes for up to 28 days of culture. Adult rat hepatocytes were isolated and attached in Williams’ medium E containing 10% serum. The medium was changed to either fresh Williams’ medium E or keratinocyte serum-free medium supplemented with dexamethasone, epidermal growth factor and pituitary gland extract. The hepatic phenotype was then analysed using RT-PCR, immunohistochemistry, Western blotting and assays of liver function. Cells cultured in keratinocyte serum-free medium supplemented with dexamethasone, epidermal growth factor and pituitary gland extract maintained their phenotype for 3–4 weeks, based on expression of liver proteins, ureagensis and response to xenobiotics. In contrast, hepatocytes cultured in Williams’ medium E rapidly lost the expression of liver proteins after 3 days. Cells cultured in keratinocyte serum-free medium supplemented with dexamethasone, epidermal growth factor and pituitary gland extract maintained their expression of liver-enriched transcription factors (C/EBPα and β, HNF4α and RXRα) while expression was either lost or reduced in cells cultured in Williams’ medium E. These results suggest that keratinocyte serum-free medium supplemented with dexamethasone, epidermal growth factor and pituitary gland extract can maintain the hepatic phenotype for a prolonged period and that this is probably related to the continued expression of the liver-enriched transcription factors.
1
Introduction
The primary culture of hepatocytes from rodents constitutes an attractive model system for the study of liver function. However, a major limitation to such a model is the rapid and irreversible loss of differentiated hepatic functions in culture. De-differentiation is reflected not only in decreased liver-specific functions, but there is also an alteration of morphology: the cells flatten, depolarize, and lose many of the surface characteristics of normal hepatocytes in vivo. The mechanisms responsible for the loss of differentiated properties probably involve the downregulation of transcription factors involved in liver-specific gene expression (Padgham et al., 1993). The loss of differentiated functions has been attributed to the change in environmental conditions (extracellular matrix, hormonal conditions) following cell isolation (Padgham & Paine, 1993). The loss of the differentiated phenotype is most apparent in the rapid decline of total cytochrome P450s (CYPs) after isolation and in culture, particularly the CYP3A1 isoform (Padgham et al., 1993). Studies of liver function are therefore generally confined to the first few days of culture and this precludes longer-term studies. Conventional approaches to maintaining the differentiated properties of isolated hepatocytes in culture include supplementation of the medium with hormones such as dexamethasone (Dex) (Agius, Chowdhury, & Alberti, 1986; Enat et al., 1984), co-factors such as nicotinamide, pyruvate, DMSO and phenobarbital (LeCluyse, Bullock, Parkinson, & Hochman, 1996; Waxman, Morrissey, Naik, & Jauregui, 1990); the application of extracellular matrix components (Bissell, Arenson, Maher, & Roll, 1987; Gomez-Lechon et al., 1998) and co-culture with non-parenchymal epithelial cell-types (Rogiers & Vercruysse, 1993; Vallette et al., 1998).
In the present study we have introduced a new medium for the long-term culture of differentiated hepatocytes. This is keratinocyte serum-free medium (KSFM). It was tested both alone, and in combination with supplements (dexamethasone, EGF and pituitary gland extract), in comparison with the standard Williams’ medium E, and was used to maintain rat hepatocytes in culture for up to 28 days. We performed immunohistochemical, Western blotting and RT-PCR analysis of the cells under the different culture conditions. We chose hepatic markers which are known to be either rapidly switched off after hepatocyte isolation or are important for detoxification. Expression of a particular gene or protein is normally taken as an indication of intact liver function. However, it is difficult to know simply from expression of a gene or protein that the associated function remains intact. For this reason we also assayed for hepatocyte ureagenesis, glycogen synthesis and response to xenobiotics. Lastly, to gain some mechanistic insight we determined the expression of several liver-enriched transcription factors. Our results suggest that KSFM, in combination with dexamethasone, EGF and pituitary gland extract can maintain the liver phenotype for between 21 and 28 days and it is likely that the ability to do this depends on the sustained expression of liver-enriched transcription factors.
2
Materials and methods
2.1
Materials
Collagenase was obtained from Worthington Biochemical Corporation, dexamethasone, penicillin/streptomycin antibiotics, l-glutamine, Williams’ medium E and phenobarbital were purchased from Sigma Chemical Co. (St. Louis, MO). Fungizone (amphotericin B) and Keratinocyte serum-free medium (KSFM) and supplements, recombinant human epidermal growth factor and bovine pituitary extract at final concentrations of 5 ng/ml and 50 μg/ml, respectively, were obtained from GIBCO™. Sources of other chemicals and media were as described previously (Shen, Slack, & Tosh, 2000; Tosh, Shen, & Slack, 2002).
2.2
Rat hepatocyte isolation and culture
Male Wistar rats (270–330 g) were obtained from the Animal House of the University of Bath. Hepatocytes were isolated by the two-step collagenase perfusion technique as described previously (Tosh, Alberti, & Agius, 1988). Calcium-free EDTA perfusion medium was freshly prepared with the following composition: 0.05% (w/v) KCl (Fisher Scientific) in 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES, Sigma) buffer, 5 mM d-glucose (Sigma), 200 μM EDTA (Sigma) and 1/1000 v/v phenol red (Sigma) all in sterile phosphate buffered saline (PBS, pH7.4) The liver was perfused via the portal vein for approximately 10 min at a flow rate of 30–35 ml/min and then replaced with a collagenase-containing perfusion medium [composition: 20 mM HEPES buffer, 5 mM d-glucose, 1 mM CaCl2 (Fisher Scientific), 0.001% (v/v) phenol red (Sigma) all in sterile PBS (pH 7.4) and contained collagenase II (0.33 mg/ml, w/v)]. Approximately 150 ml was delivered to the liver for 10–15 min (with recirculation) at a flow rate of 30–35 ml/min. Following collagenase perfusion, the liver was removed and hepatocytes dissociated using fine forceps. The cells were filtered through a 70 μm filter. Cells were washed three times (50 × g for 2 min) in medium [20 mM HEPES buffer, 5 mM d-glucose, 1 mM CaCl2 (Fisher scientific), 0.001% (v/v) phenol red (Sigma) all in sterile saline at pH 7.4]. Using trypan blue (Sigma), the cell viability was determined to be approximately 85%. Approximately 3–5 × 105 rat hepatocytes were initially placed in 35 mm culture dishes in hepatocyte attachment medium (Williams’ medium E containing 10 U/ml penicillin, 100 μg/ml streptomycin, 2mM l-glutamine, and 10% foetal bovine serum). After 6–8 h, the attachment medium was removed and replaced with either serum-free Williams’ medium E or KSFM medium. Both media contained 10 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 50 μg/ml gentamicin, 100 ng/ml Fungizone with or without the indicated supplements. The composition with supplements was as follows: Williams’ medium E only (W), Williams’ medium E plus 1 μM Dex (WD), Williams’ medium E plus 5 ng/ml EGF (WE), Williams’ medium E plus 1 μM Dex plus 50 μg/ml pituitary gland extract (WDP), Williams’ medium E plus 1 μM Dex plus 5 ng/ml EGF (WDE), Williams’ medium E plus supplements (both 5 ng/ml EGF and 50 μg/ml pituitary gland extract) (WS) and Williams’ medium E plus supplements (5 ng/ml EGF and 50 μg/ml pituitary gland extract) plus 1 μM Dex (WDS); KSFM (K), KSFM plus 1 μM Dex (KD), KSFM plus 50 μg/ml pituitary gland extract (KP), KSFM plus 5 ng/ml EGF (KE), KSFM plus supplement (KS) and KSFM plus supplement plus 1 μM Dex (KDS).
2.3
Immunostaining and antisera
For immunofluorescence staining, cells cultured on glass coverslips were rinsed with PBS to remove any excess medium and then fixed with 4% PFA for 20–30 min. The cells were washed twice in PBS and kept at 4 °C prior to staining. Cells were immunostained as described previously (Shen et al., 2000). The coverslips were mounted with Gel/Mount aqueous mounting medium (Biomeda, Foster City, CA). The primary and secondary antibodies used for immunofluorescence staining are listed in Table 1. For the detection of connexin 32 protein, the cells were fixed in acetone/methanol (1:1, v/v) then incubated in 1× citrate buffer (Lab Vision Corporation, CA) at 37 °C for 30 min before blocking and staining.
2.4
Western blotting
Protein extracts from cultured rat hepatocytes were prepared by lysing pre-washed cells with lysis buffer [20 mM HEPES (pH 7.6), 150 mM NaCl, 1 mM EDTA, 2 mM dithiothreitol (DTT) and 1% (v/v) Triton X-100] containing a 1/100 dilution of protease inhibitor cocktail (Sigma) for 10 min on ice. The cell lysates were then centrifuged at 13,000 rpm at 4 °C and supernatants were stored at −80 °C until further use. The protein concentration of the supernatant was determined using the Bio-Rad protein assay reagent according to the manufacturer's instructions. Total protein (10–15 μg) was mixed with same volume of 2× sample loading buffer [125 mM Tris–HCl (pH 6.8), 4% (v/v) SDS, 20% (v/v) glycerol, 0.2 mM DTT and 0.02% (v/v) bromophenol blue] and denatured by heating to 100 °C for 5 min. The proteins were then separated in 10% or 15% Criterion™ pre-cast Tris–HCl polyacrylamide gel (Bio-Rad) and transferred onto BioTraceNT® nitrocellulose membrane (Pall Corporation, Pensacola, FL). The membrane was blocked with 5% (v/v) non-fat milk in 0.1% (v/v) PBS–Tween20 (Tween-20, was obtained from Sigma) at 4 °C overnight. The membrane was probed with primary and secondary antibodies (listed in Table 2) at room temperature for 2 and 1 h, respectively. Antibodies were diluted in 3% (v/v) non-fat milk in 0.1% (v/v) PBS–Tween and are listed Table 2. The signals were detected with the ECL™ Western blotting analysis system (Amersham) and developed on Hyperfilm™ (Amersham).
2.5
Qualitative and real-time RT-PCR
Total RNA was extracted from cultured rat hepatocytes using TRI reagent (Sigma, Poole, UK). The measurement of total extracted RNAs and reverse transcription were carried out as described previously (Li, Horb, Tosh, & Slack, 2005). Polymerase chain reactions containing the mixture of the same concentration of cDNA, 1.1X ReddyMix™ PCR Master Mix (ABgene, Surrey, UK) and 50 ng sense/antisense primers (listed in Table 3) were processed in a DNA thermal cycler using the following conditions: denatured at 94 °C for 1 min, amplification at 58 °C for 1 min and elongation at 72 °C for 1 min for indicated cycles. The samples were analysed in 1.2% agarose electrophoresis with 1 kb ladder marker (Invitrogen Life Technologies).
To obtain quantitative results, we performed real-time PCR using the Lightcycler system (Roche Diagnostics). The calibrators were the cDNA reverse transcribed from adult rat liver mRNA. The same amount of cDNA from calibrator and experimental samples were mixed with 1XSYBR@ Green Taq ReadyMix™ (Sigma), 50 ng sense and antisense primers and transferred into Lightcycler capillaries (Roche Diagnostics). The following conditions were used for the amplification of PCR products: denaturation at 95 °C for 30 s; amplification at 58 °C for 5 s followed by 72 °C for 20 s for 40–50 cycles; cool-down at 40 °C for 2 min. The fluorescence signal was detected at the same time point of each cycle. Data were presented as the normalised ratio, which is the target/reference ratio of the sample divided by the target/reference ratio of the calibrator by Lightcycler Relative Quantification software.
2.6
Periodic acid–Schiff's (PAS) staining
PAS staining was performed to detect glycogen. The cells were seeded onto coverslips and cultured for up to 4 weeks in KDS medium. The cells were then incubated with KDS plus 25 mM glucose for 24 h, fixed with 4% PFA and then permeabilised with 1% Triton X-100 at room temperature for 20–25 min. The cells were washed with tap water for 1–2 min and transferred to 1% periodic acid solution for 30 min. Next, the cells were washed with running tap water for 3 min and then incubated in the Schiff's reagent at room temperature for 30 min to develop. After washing in running tap water for 10 min, the slides were mounted in Gel/Mount mounting medium.
2.7
Urea cycle assays
Two colorimetric assays, designed to detect secretion of urea and activity of arginase, were performed on the cultured rat hepatocytes. The urea assay was based on a previously published protocol (Meng, Zhang, & Wu, 2004). Briefly, 100 μl of urea standards (0–50 μg/ml) and culture media collected over 24 h were incubated with 300 μl urease buffered solution (Sigma) at room temperature for 20 min. 600 μl of phenol nitroprusside, 600 μl of alkaline hypochlorite (both from Sigma) and 3 ml distilled water were then added, gently mixed and incubated at room temperature for 30 min. The absorbance of standards and samples were then measured at an optical density (OD) of 630 nm. Urea production was expressed as the amount accumulated in 24 h per culture dish (μg/dish/day).
The arginase assay was based on Corraliza, Campo, Soler, and Modolell (1994). Cultured rat hepatocytes were lysed in 0.1% (v/v) Triton X-100 containing a 1/100 dilution of protease inhibitor cocktail and shaken for 30 min. The lysate was then mixed with same volume of 25 mM Tris–HCl (pH 7)/5 mM MnCl2 and the enzyme was activated by incubation at 56 °C for 10 min. Activated lysate (25 μl) was then incubated with 25 μl of 0.5 M l-arginine at 37 °C for 1 h. The samples and the urea standards (0–500 μg/ml) were then incubated with 400 μl of an acid solution comprised of H2SO4:H3PO4:H2O at a ratio of 1:3:7 and 25 μl 9% (w/v, dissolved in 100% ethanol) iso-nitro-propiophenone (Sigma) at 100 °C for 45 min. The urea production was detected at an OD of 540 nm. For urea assay, the secreted urea was presented as the amount of urea per dish per day. The urea production from the hydrolysis of arginine by arginase was taken as a proportional representation of endogenous arginase activity in the arginase assay experiments. The results were normalised with total cellular protein and shown as the amount of urea produced per mg protein per day (μg urea production/mg total protein/day).
2.8
Image collection
Fluorescent images were collected using a Zeiss LSM 510 confocal microscope and figures compiled using Adobe Photoshop 7.0. For cell counting and PAS staining experiments, numbers of random fields were selected using the 40× objective lens of a Leica DMRB microscope. The cell numbers were visualised by 4,6-diamidino-2-phenylindole (DAPI) staining. Coverslips were incubated with DAPI (500 ng/ml) in PBS for 30 min at room temperature before being mounted on to slides in Gel/Mount mounting medium.
2.9
Statistical analysis
The data were expressed as mean ± standard deviation. Comparison of individual treatments was conducted using Student's t-test or one-way ANOVA analysis with Fisher's LSD pairwise comparison.
3
Results
3.1
Expression of hepatic differentiation markers in rat hepatocytes cultured in Williams’ medium E and KSFM
Despite numerous attempts, it has been very difficult to maintain hepatocytes in a well-differentiated state for more than a few days in culture without significantly changing the culture conditions. We therefore sought to develop a protocol for long-term maintenance of hepatic functions in vitro by the use of a simple medium. Several groups previously used KSFM for culture of oesophageal keratinocytes, corneal epithelial cells and foreskin keratinocytes (Andl et al., 2003; Chen, Chang, Lee, Javier, & Azar, 2002; Vallette et al., 1998). Moreover, Katsura et al. (2002) previously used KSFM medium along with serum to culture adult human hepatocytes (but in the absence of the EGF and PGE supplements). We tested the KSFM medium for maintaining the liver phenotype in primary rat hepatocytes in the absence and presence of EGF and pituitary gland extract. Adult rat hepatocytes isolated by the collagenase perfusion technique were allowed to attach in Williams’ medium E containing 10% serum and maintained for 6–10 h before changing to a serum-free defined Williams’ medium E or KSFM supplemented with Dex and/or EGF and pituitary gland extract. The initial measurements were taken for cultures at 72 h. Albumin expression was maintained in cells cultured in KSFM [either with Dex or EGF/pituitary gland extract (referred to as KD and KS, respectively, see Section 2)] in comparison to cells cultured in WD or WS (Fig. 1A). We then compared expression of other liver-specific markers including serum proteins (α1-antitrypsin, haptoglobin and transferrin), liver-enriched transcription factors (C/EBPα, C/EBPβ, HNF-4α and RXRα), enzymes associated with ammonia detoxification (GS and CPS) and Phase II metabolism (UDP-glucuronosyltransferase, UGT) in WS and KS. It was noted that KS maintained the hepatocyte heterogeneity in terms of populations of perivenous type GS and periportal type CPS-expressing cells (Fig. 1B). The results at 72 h showed that all the liver markers examined were maintained to a greater extent in KSFM-cultured conditions in comparison to Williams’ medium E (Fig. 1B and C).
To test the active component in the supplements, we added EGF and pituitary gland extract separately to the Williams’ medium E (Fig. 1C). Compared to EGF, pituitary gland extract had a greater effect on transferrin expression after 72 h of culture. However, we could not maintain the cells for more than 4–5 days, either with one or both the supplements. This result suggests that the KSFM medium itself must contribute to the maintenance of the hepatic phenotype, and it is not just an effect of supplements.
Vimentin is a cytoskeletal intermediate filament protein the expression of which has been associated with the de-differentiation of hepatocytes (Blaheta et al., 1998). We found that this was expressed in Williams’ medium E cultured isolated rat hepatocytes but was absent in KSFM-cultured cells after a 72-h culture period (Fig. 1D). This suggests that KSFM is preventing de-differentiation as well as maintaining differentiated functions.
Connexin 32 (Cx32) is the predominant gap junction protein expressed in hepatocytes (Kumar & Gilula, 1986; Nicholson et al., 1987). Gap junction intracellular communication plays an important role in regulating cell survival and apoptosis (Krysko, Leybaert, Vandenabeele, & D’Herde, 2005), cell differentiation and proliferation (Cheng et al., 2004) and tumourigenesis (Luebeck, Buchmann, Schneider, Moolgavkar, & Schwarz, 2005). We examined the expression of Cx32 in rat hepatocytes cultured for 72 h in Williams’ medium E or KSFM media. As shown in Fig. 2, gap junction-like arrays of Cx32 were only detected in hepatocytes cultured in KS and KDS (Fig. 2).
3.2
Rat hepatocytes cultured in KDS medium retain their differentiated properties for up to 28 days
Although the results for 3-day cultures with KSFM were encouraging, we wanted to know whether the medium could maintain the hepatic phenotype for longer culture periods. We began by examining the survival of hepatocytes under different culture conditions. Equal numbers of cells were seeded and cultured in six different KSFM culture conditions. The cell number at 24 h of culture was set as 100% and the percentages of hepatocytes at 1, 2 and 3 weeks were determined (Fig. 3A). In all cases there was a progressive loss of cells but the cells cultured in KDS exhibited the best survival rates, and we decided to use this combination for subsequent experiments. In addition, we also analysed the hepatic phenotype in the living cells by two different approaches. The first involved counting the number of DAPI-positive cells that also express liver proteins (Fig. 3B) or by immunostaining for the liver protein transferrin and collecting the corresponding transmitted light image to show the distribution of the cells (Fig. 3C). The results confirmed that more than 90% of the cells expressed liver proteins over 2 or 4 weeks of culture (Fig. 3B and C). Lastly, to determine whether KSFM prevents outgrowth of mesenchymal cells in long-term cultures, we co-stained for vimentin and transferrin after 2 and 4 weeks of KDS culture (Fig. 3C). At both time points we found only occasional vimentin positive cell (see insert in Fig. 3C) suggesting the KDS does indeed prevent outgrowth of mesenchymal cells.
We determined the expression of liver markers in KDS-cultured rat hepatocytes using RT-PCR (Fig. 4A), Western blotting (Fig. 4B) and immunostaining (Fig. 4C and D) and compared the results to fresh liver and freshly isolated hepatocytes (designated ‘pre’ and ‘post’). For RT-PCR, the primers were designed towards messages of detoxification enzymes and enzymes involved in amino acid metabolism. All the transcripts were highly expressed in KDS-cultured rat hepatocytes up to 3 weeks and most were still present after 4 weeks of culture (Fig. 4A). Using Western blotting, hepatic proteins including haptoglobin and the liver-enriched transcription factor hepatocyte nuclear factor 1α (HNF1α) were maintained without diminution for at least 2 weeks (Fig. 4B). Furthermore, we also demonstrated that the expression of several liver proteins was maintained for at least 3 weeks. These include apolipoprotein B (ApoB), transferrin, albumin and Cyp2E1, C/EBPα, RXRα and HNF4α (Fig. 4C and D). In general, the expression of both transcription factors and differentiation products was similar to those in fresh liver.
3.3
Glycogen storage and urea cycle activity is preserved in KSFM-cultured rat hepatocytes
To test whether hepatocytes cultured under different conditions maintained their functional capacity, we examined the potential to store glycogen and perform ureagenesis. We were able to detect glycogen in hepatocytes cultured with KDS for at least 4 weeks of culture (Fig. 5). We also investigated ureagenesis following culture of hepatocytes under KDS conditions. Assays were performed from 24 h up to 28 days of culture using either the detection of secreted urea (Fig. 6B), or by measurement of arginase activity (Fig. 6A). The results showed that both arginase activity and urea synthesis was generally maintained and are comparable to previously published values (Kang, Berthiaume, Nath, & Yarmush, 2004).
3.4
Hepatic metabolising enzymes are induced by phenobarbital treatment in KDS-cultured hepatocytes
Phenobarbital is widely used as an inducer of Phase I and II metabolising enzymes in liver. Following treatment by phenobarbital, the transcription of phase I (cytochrome P450s) and phase II (UGT) detoxification enzymes are activated through the binding of CAR-RXR (CAR, constitutive active receptor; RXR, 9-cis-retinoic acid receptor) heterodimers with the cis-acting element (DR4 motifs, such as NR1 and NR2 domain on the enhancer of Cyp2b10 gene) of target genes (Kakizaki et al., 2003). We initially examined the time course of expression of CAR and PXR (pregnane X receptor, an additional important nuclear receptor accounting for detoxification in hepatocytes) by RT-PCR in KDS-treated cells (Fig. 7A). The levels were well-maintained for 3 weeks and then started to drop off between 3 and 4 weeks of culture.
To examine the effect of phenobarbital on KDS-cultured rat hepatocytes, we analysed the mRNA expression of UGT, Cyp2B12, Cyp3A1 and Cyp7A1 by real-time RT-PCR. Rat hepatocytes were cultured for 7 or 14 days in KDS and then treated with and without phenobarbital for 3 days (total of 10 or 17 days culture). Fig. 7 shows that all the genes, including UGT and cytochrome P450s in the phenobarbital-treated cells, were upregulated 2–10 fold in comparison with control cells. The result demonstrates that hepatocytes cultured with KDS remain responsive to xenobiotics.
4
Discussion
Much effort has been devoted in the past in attempting to develop conditions suitable for the in vitro culture of rodent hepatocytes. Some culture media have been described for sustaining the differentiated state of hepatocytes but they only maintain function in the short term, i.e., 7 days or less (Michalopoulos, Bowen, Mule, & Luo, 2003; Miyazaki et al., 1998). Alternative approaches to the resolving the problem for human liver includes isolating and maintaining embryonic liver cells in culture or adult human hepatocytes (Lazaro et al., 2003; Runge et al., 2000). Our work here shows that KSFM medium with dexamethasone, EGF and pituitary gland extract is able to maintain the differentiated properties of rat hepatocytes for between 21 and 28 days. This statement is based on the sustained expression of liver-specific proteins (including liver-enriched transcription factors), the ability to induce drug metabolising enzymes, store glycogen and produce urea. The main benefit of KSFM medium is that it is relatively simple in comparison with some of the culture conditions reported by other groups. In previous studies, various supplements have been utilised to maintain hepatocytes in culture. These include growth factors (EGF and hepatocyte growth factor) (Michalopoulos et al., 2003; Miyazaki et al., 1998), differentiation-promoting chemicals (nicotinamide and DMSO) (Mitaka, 1998), metal ions (ferrous iron, copper, manganese or zinc) (Kojima et al., 1996), and hormones (e.g. glucocorticoid) (Shelly, Tynan, Schmid, Schutz, & Yeoh, 1989). In contrast, the present medium is relatively simple to prepare it is commercially available and the only supplements required are dexamethasone, EGF and pituitary gland extract. It is also worth noting that extracellular matrices (Rana, Mischoulon, Xie, Bucher, & Farmer, 1994) are not required. We have cultured hepatocytes in KDS media on collagen, fibronectin, matrigel and glass for 2 weeks and then immunostained for albumin. There were no differences showing that the KDS medium alone is sufficient to maintain albumin expression.
Liver-enriched transcription factors such as HNF1α, HNF4α and C/EBPα are important in the control of hepatic differentiation (Costa, Kalinichenko, Holterman, & Wang, 2003). Indeed over-expression of HNF4α by adenoviral infection prolongs the expression of liver-specific genes, cell viability and liver functions such as ammonia metabolism in cultured rat hepatocytes (Naiki, Nagaki, Asano, Kimata, & Moriwaki, 2005). In addition, C/EBPβ has been shown to be a master regulator for the formation of transdifferentiated hepatocytes from pancreatic-derived cells (Kurash, Shen, & Tosh, 2004; Shen et al., 2000; Tosh et al., 2002). When we examined the expression of the liver-enriched transcription factors we found that C/EBPα and β, HNF4α and RXRα expression were reduced within 3 days of culture with Williams’ medium E but were maintained when the cells were cultured in KDS. This observation suggests that loss of liver-enriched transcription factors is probably the cause of the loss of the differentiated hepatic phenotype.
Recently, it was suggested that loss of connexin expression might be used as a prognostic marker for hepatocellular carcinoma (Sheen et al., 2004). The presence of connexin proteins in liver is associated with normal liver function. The induction of Cx32 and Cx26 proteins in DMSO-containing primary mouse hepatocytes cultures was also observed and confirmed the importance of connexin proteins in differentiating hepatocytes (Stoehr & Isom, 2003). In the present study, Cx32 was shown to be expressed in hepatocytes maintained in KDS culture medium for up to 3 weeks (Fig. 4C). The persistent expression of Cx32 may provide further evidence of the efficacy of KDS as a medium for primary hepatocyte culture.
Although the KDS mixture does not constitute a defined medium, because of uncertainty about the precise composition of the pituitary extract, our results do show that KSFM, the EGF and the PGE all have a part to play in the maintenance of differentiated hepatocytes in culture. We note that PGE alone has previously been shown to maintain the differentiation status of mammary carcinoma cells (Kano-Sueoka et al., 1979; Platica et al., 1992).
In summary, we describe a simple, novel system for maintaining rat hepatocytes in culture and show that the differentiated phenotype can be maintained for long periods in culture. This method may provide a standard system for studies on drug metabolism and toxicity testing without the problem of overgrowth of mesenchymal cells. Primary hepatocytes are a better model for the liver in vitro than hepatoma cell lines, but they also provide a cheaper alternative and a more controllable system than the intact animal. The method may also have potential for reducing the number of animals used for this type of work. Finally, we believe that this long-term differentiated culture method could also pave the way for bioengineering applications such as the development of a bioartificial liver.
|
[
"liver-enriched transcription factors",
"hepatocyte culture",
"car, constitutive active receptor",
"w, williams’ medium e",
"ksfm, keratinocyte serum-free medium",
"dex, dexamethasone",
"egf or e, human epidermal growth factor",
"pge or p, pituitary gland extract",
"c/ebp, ccaat/enhancer-binding protein",
"hnf, hepatocyte nuclear factor",
"rxr, retinoid x receptor",
"pxr, pregnane x receptor",
"cyps, cytochrome p450 proteins",
"gs, glutamine synthetase",
"cps, carbamoylphosphate synthetase",
"cx, connexin",
"rt-pcr, reverse transcription polymerase chain reaction",
"ugt, udp-glucuronosyltransferase",
"dmso, dimethyl sulfoxide",
"keratinocyte serum-free media"
] |
[
"P",
"P",
"P",
"R",
"R",
"R",
"R",
"M",
"M",
"M",
"M",
"R",
"R",
"M",
"M",
"M",
"R",
"R",
"M",
"R"
] |
Appl_Microbiol_Biotechnol-4-1-2243254
|
Fumaric acid production by fermentation
|
The potential of fumaric acid as a raw material in the polymer industry and the increment of cost of petroleum-based fumaric acid raises interest in fermentation processes for production of this compound from renewable resources. Although the chemical process yields 112% w/w fumaric acid from maleic anhydride and the fermentation process yields only 85% w/w from glucose, the latter raw material is three times cheaper. Besides, the fermentation fixes CO2. Production of fumaric acid by Rhizopus species and the involved metabolic pathways are reviewed. Submerged fermentation systems coupled with product recovery techniques seem to have achieved economically attractive yields and productivities. Future prospects for improvement of fumaric acid production include metabolic engineering approaches to achieve low pH fermentations.
Introduction
Fumaric acid (Fig. 1) is a naturally occurring organic acid. It was first isolated from the plant Fumaria officinalis, from which it derives its name. Many microorganisms produce fumaric acid in small amounts, as it is a key intermediate in the citrate cycle. Fumaric acid is also known as (E)-2-butenedioic acid or trans-1,2-ethylenedicarboxylic acid. Sometimes the term “fumarates” is also used. In this review, this general term is not used to describe fumaric acid esters like dimethyl fumarate but only to describe the salts of fumaric acid (sodium fumarate, calcium fumarate).
Fig. 1Current applications of fumaric acid production. Miscellaneous include: lubricating oil, inks, lacquers, carboxylating agent for styrenebutadiene rubber, personal care additives. a Anonymous 2007; b Willke and Vorlop 2004
Currently, fumaric acid is produced chemically from maleic anhydride, which in turn is produced from butane. However, as petroleum prices are rising rather quickly, maleic anhydride as a petroleum derivative has increased in price as well (Anonymous 2007). This situation has caused a renewed interest in the fumaric acid production by fermentation that was operational during the 1940s (Goldberg et al. 2006), but was discontinued and replaced by petrochemical processes. The fermentation process is also interesting because it involves carbon dioxide fixation, as will be discussed later. Fungi are known for their organic acid-producing capability and have been used in fermentation processes for fumaric acid production.
Production by filamentous fungi of organic acids, including fumaric acid, has recently been reviewed (Magnuson and Lasure 2004; Goldberg et al. 2006). These reviews focused on the microorganisms with their associated metabolic pathways. In contrast, the current review focuses on development of fumaric acid production processes. After summarizing fumaric acid properties, applications, and production, metabolic pathways for fumaric acid production will be discussed only briefly before turning to fermentation performance of the most prominent fumaric acid-producing strains. Furthermore, different methods that have been studied to optimize fermentation processes will be mentioned, and future prospects for process development will be discussed.
Properties and applications of fumaric acid
Because of its structure (a carbon–carbon double bond and two carboxylic acid groups), fumaric acid has many potential industrial applications (Fig. 1). It can act as starting material for polymerization and esterification reactions (www.the-innovation-group.com/ChemProfiles/Fumaric%20Acid.htm; 10/06/07).
As raw material for polymerization, especially in the production of unsaturated polyester resins, maleic anhydride is currently preferred to fumaric acid because maleic anhydride [1.46–1.63 $/kg; (Anonymous 2007)] is cheaper than fumaric acid, the latter historically being around 10% more expensive than maleic anhydride (http://www.chemweek.com/inc/articles/t/2007/04/04/04/022.html). However, fumaric acid could be a better option for the polymer industry among other carboxylic acids or their derivatives because of its nontoxic nature. In addition, special properties like greater hardness in the polymer structure can be achieved when fumaric acid is used (www.the-innovation-group.com/ChemProfiles/Fumaric%20Acid.htm; 10/06/07). In addition to polymerization, there are two potentially new applications for fumaric acid and both require a different grade of purity of it. The first is as a medicine to treat psoriasis, a skin condition (Altmeyer et al. 1994; Mrowietz et al. 1998). Psoriatic individuals are unable to produce fumaric acid in their body (which is not the case in normal individuals) due to a certain biochemical defect that interferes with adequate fumaric acid production in the skin. Therefore, psoriatic individuals need to take orally fumaric acid in the form of fumaric acid monoethyl or dimethyl ester to treat their disease. The second potential application of fumaric acid is as supplement in cattle feed. Recent studies indicate that a large reduction in the methane emissions of cattle can be achieved (up to 70%), if this cattle receives fumaric acid-based additive as a supplement in their diet (Mcginn et al. 2004). This could greatly reduce the total emission of methane, as farm animals are responsible for 14% of the methane emission caused by human activity.
Fumaric acid, together with the related succinic and malic acids, has been identified as one of the top ten building block chemicals that can be produced from sugars via biological or chemical conversion (Werpy and Petersen 2004). Compared to these other dicarboxylic acids, fumaric acid has a low aqueous solubility [7 g/kg at 25°C; 89 g/kg at 100°C (Stephen 1965)] and low pKa values [3.03 and 4.44 (Lohbeck et al. 1990)], which are properties that can be exploited for product recovery.
Fumaric acid production by petrochemical routes
Fumaric acid is currently produced by isomerization of maleic acid, which is produced from maleic anhydride. Maleic anhydride, in turn, is industrially produced by catalytic oxidation of suitable hydrocarbons in the gas phase. Benzene used to be the predominant starting material, but oxidation of n-butane or n-butane–n-butene mixtures has become more important in recent years (Lohbeck et al. 1990). The butane oxidation reaction equation to maleic anhydride is: C4H10 + 3.5O2 → C4H2O3 + 4H2O.
In the most common maleic anhydride process, the catalyst is embedded in fixed-bed tubular reactors. The catalysts applied in this process are based on vanadium and phosphorus oxides. In this process, water is formed as a by-product, and a fairly small amount of it can be directly liquefied from the reaction gas by partial condensation. Organic solvents absorb maleic anhydride contained in the reaction gas. More than 98% of the anhydride can be absorbed in this way. The solvent–anhydride mixture is then subjected to fractional distillation to separate maleic anhydride from the solvent, and the latter is returned to the absorption column (Lohbeck et al. 1990). The thus obtained pure maleic anhydride is then hydrolyzed into maleic acid according to the reaction equation C4H2O3+H2O → C4H4O4.
Subsequently, the maleic acid is converted almost quantitatively by thermal or catalytic cis–trans isomerization into fumaric acid (Lohbeck et al. 1990). The most common catalysts used are mineral acids, peroxy compounds, and thiourea. The crude fumaric acid obtained in this way is purified by crystallization from water (Fig. 2). As will be shown later, a comparable purification may be used for fermentative fumaric acid production.
Fig. 2Plant for fumaric acid production from malic acid. a Isomerization vessel; b Centrifuge; c Dissolving tank; d Filter; e Crystallizer; f Dryer (adapted from Felthouse et al. 2001)
Production capacity of maleic anhydride in 2007 was about 1,807,000 ton/a (Anonymous 2007), with around 3% of this being used for fumaric acid production (http://chemicalmarketreporter.com/home/Default.asp?type = 17&liArticleID = 2015643...; 18/04/2007), corresponding to 90,000 ton/a.
Enzymatic conversion of maleic acid into fumaric acid
In the aforementioned chemical isomerization of maleic acid into fumaric acid, the conversion yield is restricted by the reaction equilibrium. This chemical conversion is occurring at high temperatures that causes formation of by-products from maleic and fumaric acids, and as a consequence, yields are below the equilibrium yields (Goto et al. 1998). This problem was the main motivation to look for a good enzyme that would facilitate isomerization, so that the equilibrium mixture would be obtained without formation of by-products.
It is known that the enzyme maleate isomerase (maleate cis-trans-isomerase) isomerizes maleic acid into fumaric acid. Microorganisms that produce maleate isomerase are Pseudomonas species (Otsuka 1961), Alcaligenes faecalis (Takamura et al. 1969), and Pseudomonas fluorescens (Scher and Lennarz 1969). However, maleate isomerase is unstable even at moderate temperatures (Takamura et al. 1969), and therefore, a heat-stable maleate isomerase is desired. Thermo-stable maleate isomerases derived from Bacillus stearothermophilus, Bacillus brevis, and Bacillus sp. MI-105 were reported by Goto et al. (1998), and according to the authors, the use of enzymes from these bacteria can improve the fumaric acid production process.
Furthermore, in a more recent study using Pseudomonas alcaligenes XD-1, high rates of conversion of maleic acid into fumaric acid (6.98 g L−1 h−1) were reported (NakajimaKambe et al. 1997). This organism does not normally accumulate fumaric acid, but by heat treatment of the cells (70°C for 1 h), they lost their fumarase activity, which otherwise would be responsible for a conversion of fumaric acid into l-malic acid. This heat treatment did not affect the maleate isomerase activity. In addition, when calcium ions were added during the conversion step, the thermal stability of maleate isomerase was increased. With the use of P. alcaligenes XD-1, the highest reported conversion yield, 95%, was achieved in the conversion of maleic acid to fumaric acid (Ichikawa et al. 2003). Immobilization of the heat-treated cells is currently under investigation.
Microbial production of fumaric acid
Fumaric acid production by fermentation was operated in the United States during the 1940s, but later, this process was discontinued and replaced by chemical synthesis from petrochemical feedstocks, the latter being explained in a previous section. Nevertheless, the continuous increase of the petroleum prices has brought back the interest in fumaric acid production by submerged fermentation (Goldberg et al. 2006). Fumaric acid production by fermentation using Rhizopus species has been patented occasionally (Waksman 1943; Kane 1943; Lubowitz and La Roe 1958; La Roe 1959; Goldberg and Stieglitz 1986). In 1989, DuPont patented an improved fermentation process producing carboxylic acids (fumaric, succinic, malic, and mixtures thereof) by controlling dissolved oxygen levels between 30 and 80% (Ling and Ng 1989).
In the field of fumaric acid production by fermentation, there are many aspects determining the productivity of the fermentation process, such as the applied microbial strain and its morphology, the use of a neutralizing agent, and the applied feedstock. Those aspects are reviewed and analyzed in more detail in the following sections.
Fumaric acid-producing strains
After the discovery of fumaric acid production in Rhizopus nigricans by Felix Ehrlich in 1911, Foster and Waksman (1938) screened 41 strains from eight different genera to identify high fumarate producing strains. The fumarate-producing genera identified were Rhizopus, Mucor, Cunninghamella, and Circinella species. Among these strains, Rhizopus species (nigricans, arrhizus, oryzae, and formosa) were the best-producing ones, yielding fumaric acid under aerobic and anaerobic conditions (Foster and Waksman 1938; Rhodes et al 1959; Kenealy et al. 1986; Cao et al. 1996; Carta et al. 1999). Table 1 lists these fungi, and as is shown, the R. arrhizus NRRL 2582 and R. oryzae ATCC 20344 gave the highest volumetric productivity, product titer, and product yield values (Gangl et al. 1990; Cao et al. 1996).
Table 1Literature data on fumaric acid production by different Rhizopus generaStrainFermenterSubstrateProduct Titer (g l−1)Yield (g g−1)Vol. Prod. (g l−1 h−1)Time (h)Final pHReferenceR. nigricans 45Shake flaskGlucose14.70.50–1686.5Foster & Waksman 1938Shake flaskGlucose20.00.660.25806.5Romano et al. 1967R. arrhizus NRRL 2582Stirred tankGlucose90.00.701.22726.0Rhodes et al. 1962Stirred tankGlucose107.00.822.00536.0Ng et al. 1986Stirred tankGlucose73.00.720.501475.5Gangl et al. 1990R. arrhizus NRRL 1526Shake flaskGlucose97.70.811.02966.0Kenealy et al. 1986Fluidized bedMolasses17.50.360.36486.0Petruccioli et al. 1996Stirred tankGlucose38.00.330.46825.5Riscaldati et al. 2000R. oryzae ATCC 20344RBCa plus AdsorptionGlucose92.00.854.25244.5Cao et al. 1996Stirred tankGlucose65.00.650.90725.0Cao et al. 1996RBCaGlucose75.50.753.78245.0Cao et al. 199710-l air liftGlucose37.80.750.81465.0Du et al. 1997Stirred tankGlucose35.80.600.90405.5Zhou 1999Bubble columnGlucose37.20.531.03365.0Zhou et al. 2002R. formosa MUCL 28422Stirred tankCassava bagasse21.3–––6.5Carta et al. 1999aRotatory biofilm contactor
Despite the fact that the aforementioned experimental studies involve fungi, the use of bacteria has been considered as well. Donnelly et al. (2001) have suggested using a Lactobacillus host strain that lacks the malolactate enzyme, fumarase, and fumarate dehydrogenase. Introducing the maeA gene for the malic enzyme from Escherichia coli would result in a pathway from pyruvate to malate and hence in a malic acid-producing mutant. Additional genetic engineering would result in fumaric acid production.
Genetic modification of microorganisms has hardly been explored for fumaric acid production but offers a potentially useful approach solution for improving yields and rates in fermentation.
Metabolic pathways to fumaric acid
Fumaric acid is primarily an intermediate of the citrate cycle, but is also involved in other metabolic pathways. In 1939, it was suggested that fumarate biosynthesis involved a C-2 plus C-2 condensation in Rhizopus species (Foster et al. 1949). The reactions in this pathway seemed to be similar to those of the glyoxylate bypass (Foster et al. 1949). Years after, the glyoxylate bypass mechanism was ruled out because the theoretical molar yield for this pathway of 100% was not in agreement with the experimental yield of 140% (Rhodes et al. 1959; Romano et al. 1967). However, the main evidence for rejecting the glyoxylate bypass mechanism is that the key enzyme of the glyoxylate pathway, isocitrate-glyoxylate lyase, was repressed when high glucose concentrations were present like in the experiments used for fumaric acid production (Romano et al. 1967).
It was discovered that a C3 plus C1 mechanism involving CO2 fixation catalyzed by pyruvate carboxylase under aerobic conditions explained the high molar yields in fumarate production (Overman and Romano 1969). This CO2 fixation leads to oxaloacetic acid formation (Osmani and Scrutton 1985), so that C4 citrate cycle intermediates can be withdrawn for biosynthesis during the growth phase under aerobic conditions. When nitrogen becomes limiting and the growth phase stops, the metabolism of glucose and CO2 fixation could continue and lead to an accumulation of C4 acids (Romano et al. 1967). The maximal theoretical yield in a nongrowth situation is 2 mol of fumaric acid per mole of glucose consumed, upon fixation of 2 mol of CO2 via reductive pyruvate carboxylation. However, if the reductive pyruvate carboxylation would be the sole pathway, no ATP would be produced for maintenance or transport purposes. Therefore, the citrate cycle is also active during fumaric acid production (Rhodes et al. 1959; Kenealy et al. 1986).
The CO2 carboxylation enzyme, pyruvate carboxylase, is known to be localized exclusively in the cytosol, together with NAD-malate dehydrogenase and fumarase (that are present in the cytosol and in the mitochondria), leading to fumaric acid synthesis in this cell compartment (Osmani and Scrutton 1985). Studies performed by Peleg et al. (1989) indicated higher activities of these enzymes (especially the cytosolic isoenzymes) during fumaric acid production. Kenealy et al. (1986) used mitochondrial inhibitors to investigate their role in fumarate accumulation and found no direct involvement of such inhibitors of the citrate cycle in fumarate production. However, carbon-labeling studies have demonstrated the simultaneous utilization of both the citrate cycle and the reductive pyruvate carboxylation pathways under aerobic conditions (Fig. 3). Besides the localization of fumarase isoenzymes, it was also found that addition of cycloheximide virtually eliminated the cytosolic isoenzyme and therefore caused a large decrease in the amount of fumaric acid produced by R. oryzae (Peleg et al. 1989).
Fig. 3Citrate cycle pathway and reductive carboxylation pathway leading to fumaric acid accumulation (adapted from Kenealy et al. 1986). The ratio between the two pathways is not 1:1 as suggested by this figure
The carboxylation of pyruvate has been studied in more detail for the analogous microbial succinate production, but the understanding of the different metabolic fluxes involved is still incomplete (McKinlay et al. 2007).
High-producing fumaric acid strains of Rhizopus not only produce fumaric acid but also other carboxylic acids like malic, lactic, acetic, succinic, and citric in smaller amounts than fumaric acid production (Rhodes et al. 1959; Carta et al. 1999). Sometimes, ethanol is also produced, like in the case of the high-producing strain R. oryzae ATCC 20344 (Cao et al. 1996). However, it was found that ethanol production can be reduced by a sufficient supply of oxygen to the culture (Cao et al. 1996).
Fumaric acid transport across the cell membrane
The transport mechanism of fumaric acid in fungi has not been studied yet. However, transport of l-malic acid and other dicarboxylic acids, including fumaric acid, has been studied in yeasts like Schizosaccharomyces pombe, Candida utilis, Candida sphaerica, and Hansenula anomala (Corte-Real et al. 1989; Corte-Real and Leao 1990; Saayman et al. 2000). These microorganisms may be comparable to fumaric-producing fungi. These studies showed that the initial uptake of malic acid was accompanied by disappearance of extracellular protons suggesting that the anionic form of the acid was transported by an accumulative dicarboxylate proton symporter. As fumaric acid seems to be a competitive inhibitor of l-malic acid uptake, it was suggested that fumaric acid uses the same import system.
On the other hand, these studies showed that undissociated dicarboxylic acid entered the cells slowly by simple diffusion. In addition, it was revealed that the rate of diffusion of the undissociated acid across the plasma membrane is subjected to opposite pH influences: an increase due to the relative increase of undissociated acid with decreasing pH and a decrease due to a decrease of the permeability of the cell membrane for the undissociated acid at decreasing pH (Corte-Real and Leao 1990). Increasing the number or activity of the dicarboxylic acid transporters could lower the intracellular fumarate concentration and could therefore have a positive effect on the production yield (van Maris et al. 2004).
Alternative fermentation feedstocks
Glucose has not been the only carbon source used for fumaric acid-producing strains. The use of xylose with immobilized R. arrhizus was studied, but the highest obtained volumetric productivity was only 0.087 g l−1 h−1 (Kautola and Linko 1989). Also, sucrose has been considered as a raw material in fumaric acid-production processes (Kautola and Linko 1989; Zhang et al. 2007), but sucrose has the disadvantage of being poorly metabolized by R. oryzae compared to glucose (Bulut et al. 2004).
Starch-containing materials were also screened as feedstock for the fumaric acid production. Potato flour was used as feedstock for R. arrhizus, and although fumaric acid was the main metabolic product, a volumetric productivity of only 0.35 g l−1 h−1 was achieved (Moresi et al. 1991). On the other hand, it was found that the highest yields of fumaric acid were obtained from acid hydrolysates of starch-based materials (Moresi et al. 2002). The same feedstock was used by Federici et al. (1993) in studies on the effect of agitation speed and applied neutralizing agent. The tested neutralizing agent was a combination of CaCO3 and KOH/KCO3, and it was concluded that comparable fumaric acid yields were obtained using these neutralizing agents and starch-based materials (Federici et al. 1993). Carta et al. (1999) optimized the fumaric acid production by R. formosa MUCL 28422 by using enzymatic hydrolysates of cassava bagasse, a high-starch-containing waste product. The productivity was not improved with respect to a fermentation where glucose was used as a feedstock, but the fact that these feedstocks are cheap and abundant makes these optimized fermentations economically attractive.
Nutrient requirements
Physical and nutritional requirements for fumaric acid fermentation have been studied in Rhizopus arrhizus (Rhodes et al. 1959; Rhodes et al. 1962). As Rhizopus species should enter a phase of limited growth during the fermentation, which can be achieved by nitrogen limitation, these studies showed that to achieve high yields during fumaric acid fermentation, the most critical parameter is the ratio of glucose to nitrogen. For example, a yield of fumaric acid on glucose of 85% (w/w) was obtained using an initial C:N molar ratio of 200:1 for R. arrhizus 2582. When nitrogen limitation is not desired, phosphorus limitation can be used instead (Riscaldati et al. 2000). Trace metal concentrations of 500 ppm, 4 ppm, and 100 ppb for Mg++, Zn++, and Fe++, respectively, were found optimal for the formation of small (1 mm) spherical pellets that produced high concentrations of fumaric acid (36 g L−1; Zhou 1999).
Because CO2 is needed for the oxaloacetate formation from pyruvate by pyruvate carboxylase, the addition of calcium carbonate (CaCO3), which is used in many cases as a neutralizing agent, seems also important as a CO2 source during the production phase of the fermentation. However, in case that no CO2 or carbonate is added, the complete catabolism of a mole of glucose via the citrate cycle provides 6 mol of CO2 that may be used for pyruvate carboxylation (see Fig. 3). In this case, the maximum theoretical yield would be 1.5 mol of fumaric acid per mole of glucose. This theoretical value is close to the value of 1.32 mol of fumaric acid per mole of glucose reported by Cao et al. (1996), who optimized a fermentation process without CO2 or carbonate feeding.
Neutralizing agents
Continuous neutralization of the pH has been necessary to obtain optimal yields in fumaric acid production by fermentation. Preferably, CaCO3 has been used as a neutralizing agent, but at the same time, this compound is causing viscosity problems due to the low aqueous solubility of calcium fumarate [21 g/l at 30°C (Gangl et al. 1990)] that precipitates as a fermentation product. Furthermore, the cells can interact with the precipitated product, as has been found for R. oryzae, resulting in a highly viscous suspension. This has a detrimental effect on the rate of oxygen transfer which can be achieved, and hence, the fermentation might fail due to oxygen-limitation problems. Therefore, replacement of CaCO3 by other neutralizing agents like Na2CO3, NaHCO3, Ca(OH)2, and (NH4)2CO3 has been studied by different authors (Gangl et al. 1990; Riscaldati et al. 2000; Zhou et al. 2002). However, these studies showed that the fumarate production rates are the highest when CaCO3 is used as a neutralizing agent.
Because of the high solubility of sodium fumarate, fermentative production of fumarate using Na2CO3 as neutralizing agent leads to cheaper downstream processing than when CaCO3 is used. This is due to the fact that the fermentation product, sodium fumarate, has a higher solubility than CaCO3, and hence, there is no need of heating to recover the fermentation product (see Fig. 4). In addition, cells can be reused (Gangl et al. 1990; Zhou et al. 2002). A similar situation was obtained when Rhizopus growth was limited by phosphorus so that (NH4)2CO3 could be used as neutralizing agent (Riscaldati et al. 2000). Nevertheless, it has been argued that a fermentation process without the use of neutralizing agents, and at the same time preventing product inhibition, will improve the economics of the general process (Gangl et al. 1990). However, when a high-yield process is developed without a carbonate as neutralizing agent, the required CO2 must be supplied by other sources. Fumaric acid fermentation systems without the use of neutralizing agents have been studied in the past and will be discussed in the following sections.
Fig. 4Flow-sheet for fumaric acid production via fermentation. a Formulation tanks containing glucose and nutrients; b Seed fermentor; c Production fermentor; d Filter; e Acidification tank; f Filter; g Rotary dryer (adapted from Gangl et al. 1990)
With respect to fumarate salts inhibition, Rhodes et al. (1962) reported that production of soluble sodium or potassium fumarates was inhibited when the concentration of fumarate reached values of 34–40 g L−1 using R. arrhizus as a producer strain. Nonetheless, Gangl et al. (1990) found that addition of sodium fumarate (71.3 g L−1) was not inhibiting the same strain, although the cells needed 35 h to adapt to the high sodium fumarate concentrations. On the other hand, the free fumaric acid does inhibit its own production because the accumulated protons in the production medium decrease the pH, thus exerting a progressive inhibitory effect on fumaric acid production (Riscaldati et al. 2000). At low pH, excreted fumaric acid will passively diffuse back through the plasma membrane of the fungus decreasing its intracellular pH, and due to this phenomenon, the fermentation fails. Proposed methods to enable carboxylic acid fermentation at low pH are: genetic engineering of acid tolerant organisms such as yeasts and in situ product removal techniques (Cao et al. 1996; Schügerl 2000; van Maris et al. 2004).
Morphology and oxygen-transfer problems
One of the difficulties of fermenting Rhizopus species is the morphology of these fungi. Rhizopus species tend to grow on the walls and on the stirrer of the reactor, and sometimes, clumps are formed. Therefore, the fermentation can suffer from oxygen limitation in particular when calcium fumarate is present. One way to solve this problem is to control the growth of the fungi and their morphology.
A way to minimize oxygen mass transfer limitation to the cells is to stimulate formation of small spherical cell pellets (Zhou 1999). Small pellets can reduce clump formation during fermentation, and even if CaCO3 is present, the viscosity of the broth can be reduced. Moreover, pellets can facilitate the performance of a biomass retention system. For R. oryzae, low initial pH values for the cultivation media favored pellet formation, and good fumaric acid yields have been reached (Zhou et al. 2000).
In another morphology improvement study, Cao et al. (1997) used a rotary biofilm contactor (RBC) as fermentor with immobilized R. oryzae to produce fumaric acid. CaCO3 was used as neutralizing agent during this experiment. During the fermentation, the discs with immobilized cells were rotating, moving the cells from the gas phase of the fermentor to the liquid phase and back again (Cao et al. 1997). When the cells are exposed to the air, high oxygen transfer rates can be reached, while the cells can take up substrate and excrete the produced fumaric acid when they are submerged. In this system, additional agitation was not needed. The volumetric productivity was very high, compared to the volumetric productivity of an equivalent stirred vessel fermentor setup (see Table 1). A disadvantage of this system could be the scalability potential of the RBC fermentor.
Immobilization techniques for Rhizopus species have been investigated to open the possibility of a continuous operation mode for fumaric acid production and to reduce oxygen transfer problems as well. Buzzini et al. (1995) searched for the most suitable support for cell immobilization in fluidized-bed reactors during fumaric acid fermentation. Cork, expanded polystyrene, expanded clay, and wood shavings were investigated. With 6 mm pieces of cork, the highest titer, 24.1 g fumaric acid l−1 in 144 h, was achieved, which is comparable to the titer of 37.7 g l−1 after 166 h under conventional submerged conditions (Buzzini et al. 1995). The performance of a semi-continuous process was investigated with R. arrhizus immobilized on 5 mm cubic particles made of polyurethane sponge (Petruccioli et al. 1996). These particles were used in repeated batch fermentations in a fluidized-bed reactor (48 h, eight times). In the optimized process, the fumaric acid titer was 12.3 g l−1, and the volumetric productivity was 0.256 g l−1 h−1.
Furthermore, as the oxygen mass transfer resistance through the boundary layer on the liquid side of the gas–liquid interface can affect the interfacial oxygen transfer from the gas phase to the liquid phase, pressure pulsation was applied in a stirred tank fermentor using R. oryzae (Zhou 1999) to reduce this resistance. Mass yield and volumetric productivity of fumarate were 70.1% and 0.99 g L−1 h −1, respectively, which were higher than for traditional stirred tank fermentations. The same organism was studied in an airlift loop reactor as a fermentation system for fumaric acid production (Du et al. 1997). Here, the airlift loop reactor with porous sparger produced favorable conditions for mass transfer, and also, higher yields and productivities were reached than in stirred tank fermentations.
Integrated fermentation and recovery of fumaric acid
Figure 4 presents a flow scheme proposed for fumaric acid production by batch fermentation (Gangl et al. 1990). Glucose and mineral salts are fed to the fermentor. The nitrogen source solution, (NH4)2SO4, is sterilized separately and fed to the seed fermentor. The harvested broth containing sodium fumarate (cells and trace amounts of Na2CO3) is filtered to remove the cells and then acidified by H2SO4 to pH 1. After acidification, fumaric acid precipitates out of the solution and is sent to a rotary dryer to be completely recovered (Gangl et al. 1990). When CaCO3 is used as a neutralizing agent instead of Na2CO3, additional heating after the fermentation was supposed to be required to dissolve calcium fumarate and the excess of CaCO3 that usually stick to the cells. This problem leads to a tedious and expensive downstream processing.
In the downstream processing field, recovery techniques have not so much been studied for fumaric acid producers in submerged cultivations as it has been done for related fermentation products like succinic acid (Zeikus et al. 1999), citric acid (Heinzle et al. 2006), and lactic acid (Joglekar et al. 2006). Recovery systems like reactive extraction and membrane electrodialysis have not been studied yet. However, simultaneous fermentation and adsorption have been studied to remove fumaric acid during its formation to control fermentation pH at desired values and avoid product inhibition (Cao et al. 1996; Zhou 1999).
Cao et al. (1996) used a rotary biofilm contactor (RBC) setup as fermentor for R. oryzae, in combination with an adsorption column. The produced fumaric acid was removed from the broth by the adsorption column in a recycle loop, reducing product inhibition and thus increasing the production rate and sustaining cell viability. Polyvinyl pyridine (PVP) anion exchange resin in the hydroxide form was selected as adsorbent because it yielded the highest loading capacity for fumaric acid (0.31 g g−1 dry wt). The RBC, coupled with the adsorption column, increased the fumaric acid productivity significantly to 4.25 g l−1 h−1 because the total fermentation time was much less than in traditional stirred tank fermentations (see Table 1). This volumetric productivity and the yield are the highest reported in literature. In this integrated system, the constant removal of the fumarate produced and the liberation of OH− from the adsorption column kept the fermentation pH at 4.5. The fumarate was desorbed from the adsorption resin by using 0.4 M NaOH so additional processing will be necessary to convert the sodium fumarate into neutral fumaric acid. Addition of extra CO2 was not necessary due to the fact that during the production stage, the biofilm was exposed to sterile air that was present in the headspace of the RBC enhancing the opportunity for CO2 fixation by the biofilm.
A novel product recovery process by an ion exchanger resin was developed by Zhou (1999). Fumarate was recovered from the fermentation broth without the use of neutralizing agents, while keeping the pH at 5, by cycling the broth over a column of a resin (Amberlite IRA-900 with OH- as counterion). After eluting the loaded column with ammonium hydroxide, ammonium fumarate solution was obtained. This was passed through a Y-zeolite column that retained the ammonium and liberated the fumaric acid. The ammonium hydroxide solution can be recovered and recycled by thermal regeneration of the zeolite (Zhou 1999). However, this integrated process did not surpass yield values obtained in a stirred tank fermentor when fumaric acid fermentation was controlled by CaCO3 under pressure pulsation (see previous section), but a higher productivity value of 1.09 g L−1 h−1 was reached (Zhou et al. 2000).
Economics of fumaric acid production
In a theoretical study, Gangl et al. (1990) compared a fermentation and a benzene-based petrochemical process with respect to economy. The fermentation process taken for this has been described in the beginning of the previous section. Although Na2CO3 was assumed as a neutralizing agent in this study, the productivity (2 g l−1 h−1) and yield (0.82 g g−1) of the fumaric acid fermentation with CaCO3 obtained in a previous study were used (Ng et al. 1986).
This economic evaluation showed that the fermentation route was less favorable than the petrochemical route (economic evaluation included upstream and downstream sections). In particular, the raw material costs were higher for the fermentation process than for the petrochemical route. It was concluded that the fermentation route could become competitive with the petrochemical route if oil prices were around 61 $/barrel. The latter price has been reached in the recent years (http://tonto.eia.doe.gov/dnav/pet/xls/PET_PRI_WCO_K_W.xls), while the productivity of the fermentation process has been improved significantly as is shown in Table 1. Besides, one can roughly calculate from Gangl’s study that the assumed sugar price was around 0.6 $/kg of glucose, which is high compared to the current price (see Table 2). Therefore, if the current glucose price is assumed in the calculations of Gangl’s study, one can expect lower raw material costs for the fermentation route. On the other hand, the petrochemical route has also been improved significantly, now using butane instead of benzene. This will be reflected in the current maleic anhydride price, which is shown in Table 2.
Table 2Comparison between petrochemical and fermentation route for fumaric acid productionParameterPetrochemical routeFermentation routeRaw materialMaleic anhydrideGlucoseReaction temperature (°C)90–10035Raw material price ($/kg)1.46–1.63 (Anonymous 2007)0.46aProduct yield (% w/w)112 (Lohbeck et al. 1990)85 (Cao et al. 1996)ahttp://www.ers.usda.gov/briefing/sugar/data.htm; for 4th quarter of 2006
The simple comparison of petrochemical and fermentation routes for fumaric acid production given in Table 2 suggests that the lower raw material price of the fermentative production might compensate the higher yields of the petrochemical production from maleic anhydride, and fermentation may become an economically viable alternative.
Conclusions and future prospects
Due to increasing prices of fossil feedstock, fermentatively produced fumaric acid could become a cheaper alternative to the petrochemically based maleic acid as unsaturated dibasic acid in polyester resins in the nearby future. Based on the available literature, the microorganism with the highest productivity and yield of fumaric acid appears to be R. oryzae. This organism produces fumaric acid via a combination of the citrate cycle and reductive pyruvate carboxylation. In addition, the fumaric acid production by fermentation can be improved if the use of neutralizing agents is reduced or avoided and the morphology of the fungi is optimized. However, avoiding the use of a neutralizing agent will lead to product inhibition; therefore, it would be useful if metabolic engineering is applied to achieve fumaric acid production in suspended, acid-resistant microorganisms such as yeast. Another way to solve product inhibition problems is applying in situ removal of fumaric acid during the fermentation.
|
[
"fumaric acid",
"fermentation",
"rhizopus species",
"product recovery",
"maleate isomerase"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
J_Headache_Pain-4-1-2386848
|
Topiramate in the prevention of pediatric migraine: literature review
|
Pediatric migraine is a disabling condition, which can cause a significant impact on quality of life. Currently, no drugs have been approved by the FDA for its preventive treatment. Our aim was to review the medical literature concerning the efficacy and tolerability of topiramate in the prophylactic treatment of migraine in children and adolescents. A total of five papers were reviewed: two randomized controlled trials (RCTs), a post-hoc subset analysis of adolescents who had been included in three RCTs carried out on adults and two open studies. Topiramate has been proven to reduce headache frequency and the accompanying disability. The frequency of side effects varied considerably among studies, the most frequent being weight loss, anorexia, abdominal pain, difficulties in concentrating, sedation and paresthesia. Since these adverse events, although often transitory, may be distressing for the child, we strongly recommend to assess the disability caused by the migraine episodes before deciding to initiate a prophylactic treatment. Nevertheless, dropout rates due to side effects in the studies were very low.
Introduction
Pediatric migraine is a disabling condition, which can cause a significant impact on quality of life. It can negatively influence the child’s school performances, social and family life and often causes school absenteeism. Migraine prevalence in childhood ranges from 2.7 to 10% [1]. About 3–5% of school-aged children suffer from migraine and this proportion gradually increases to 20% through adolescence. Initially, there is a slight male predominance; however, through adolescence, there is a shift toward female predominance, which remains through adulthood [2, 3].
Management of pediatric migraine includes lifestyle changes (in order to avoid foods, habits or enviromental factors that may trigger a migraine attack), the use of abortive medications and preventive measures, which can be either nonpharmacological or pharmacological. Guidelines addressing the indications for the use of preventive measures in children are lacking, but to date, most authors recommend preventive measures in children when headache frequency exceeds three to four episodes per month and/or the attacks are significantly disabling, as measured in scoring systems such as the Pediatric Migraine Disability Assessment Scale (PedMIDAS) [4, 5]. PedMIDAS is a validated six-item questionnaire based on the adult MIDAS tool [6] with developmentally appropriate changes and adjustments for childhood lifestyle. The questions deal with the impact of headache on school (school day absences, partial day absences, functioning at 50% or less ability in school), on household acitivities (ability to perform homework and chores) and on social functioning (including sports) [7, 8]. The assessment of migraine disability is particularly important in children to correctly evaluate and treat migraine.
Currently, no drugs have been approved by the Food and Drug Administration (FDA) for the preventive treatment of pediatric migraine. Topiramate is an antiepileptic drug, which has been approved by the FDA in the United States and in many other countries for the preventive treatment of migraine in adults and for the treatment of partial-onset seizures and primary generalized tonic–clonic seizures as add-on therapy in children as young as 2 years. Our aim was to review the medical literature concerning the efficacy and tolerability of topiramate in the prophylactic treatment of migraine in children and adolescents.
Materials and methods
We searched Pubmed (1966–July 2007) using the following keywords: “topiramate” + “headache” or “migraine” + “children” or “adolescents” or “pediatric.” We included randomized controlled trials (RCT), open-label trials and retrospective studies. We excluded case-reports and studies conducted on less than 10 patients. We classified the studies in accordance with the AAN classification of evidence for therapeutic intervention (Table 1) [9]. A total of five papers were reviewed.
Table 1AAN classification of evidence for therapeutic interventionClass I: Prospective, randomized, controlled clinical trial with masked outcome assessment, in a representative population.The following are required: (a) Primary outcome(s) is/are clearly defined (b) Exclusion/inclusion criteria are clearly defined (c) Adequate accounting for drop-outs and crossovers with numbers sufficiently low to have minimal potential for bias (d) Relevant baseline characteristics are presented and substantially equivalent among treatment groups or there is appropriate statistical adjustment for differencesClass II: Prospective matched group cohort study in a representative population with masked outcome assessment that meets (a)–(d) above OR a RCT in a representative population that lacks one criteria (a)–(d).Class III: All other controlled trials (including well-defined natural history controls or patients serving as own controls) in a representative population, where outcome is independently assessed, or independently derived by objective outcome measurement.Class IV: Evidence from uncontrolled studies, case series, case reports, or expert opinion.
Results
We found two RCTs, a post-hoc subset analysis of 51 adolescents (12–17 years) who had been enrolled in three pivotal RCTs of topiramate for migraine prophylaxis in adults and two open studies.
In a recent Class I randomized, double-blind, placebo-controlled trial [10], 44 children were randomized (in a 1:1 ratio) to receive either placebo or topiramate (titrated to 100 mg a day) for 12 weeks. Participants had a diagnosis of migraine without aura (according to 2004 International Headache Society criteria) [11] and a frequency of two or more headaches per month for 3 months before entering the study. Children with comorbid medical associations or who were already on migraine prophylaxis were excluded. Primary outcome measures were the reduction in migraine frequency and severity; secondary outcome measures included number of analgesics taken and the functional disability. Migraine monthly frequency decreased from 16.14 (±9.35) at baseline to 4.27 (±1.95) at the end of the study in the treated group as compared with a decrease from 13.38 (±7.78) to 7.48 (±5.94) in the placebo group (P = 0.025). The percentage of topiramate-treated subjects showing >50% reduction in monthly migraine days during the double-blind phase was 95.2% as opposed to 52.4% in the placebo group (P = 0.002). The Pediatric Migraine Disability Assessment Score (PedMIDAS) also decreased significantly (P = 0.003) as did school absenteeism (P = 0.002). Side effects were rated mild to moderate, did not interfere with daily activities or cause any dropouts and included weight-loss (mean loss of 0.3 kg) in 81% of patients, paresthesia (23.8%), loss of appetite (23.8%), lack of concentration (19%), sedation (19%) and abdominal pain (14.3%). This seems a well-performed trial in which the mean monthly migraine frequency reduction is clearly evident in the topiramate-treated group. Patients treated with topiramate suffered at least five episodes a month before entering the study (range 5–30); a preventive treatment was, thus, clearly indicated. The reduction in the PedMIDAS score indicates that the drug side effects (difficulties in concentrating, sedation) interfered with school, household or social activities to a lesser extent than the migraine episodes.
The other Class I, randomized, double-blind, placebo-controlled trial [12] was carried out on 162 children (age 6–15 years) who were randomized in a 2:1 ratio to receive either topiramate (which was titrated over 8 weeks to 2–3 mg/kg per day and maintained for 12 weeks) or placebo. Eligible criteria were the children with migraine with or without aura, weighing more than 20 kg, who experienced 3–10 migraine days/month for the 3 months prior to the screening and during the 4-week prospective baseline phase. Exclusion criteria included chronic migraine, analgesic overuse and previous failure of ≥2 adequately dosed migraine preventive medications. One hundred and fifty-seven subjects were included in the intention to treat (ITT) population, defined as randomized subjects who had received at least one dose of study medication and had at least one postbaseline efficacy assessment, and a total of 131 children completed the study (per-protocol population); of these, 108 were treated with topiramate. Primary outcome measure was the reduction of migraine days per month in each treatment group of the ITT population. During the double-blind-phase, relative to the 4-week prospective baseline phase, there was a reduction of 2.6 ± 2.6 migraine days per month in the topiramate group compared with a mean reduction of 2.0 ± 3.1 migraine days per month in the placebo group (P = 0.061). In the per-protocol population, the reduction in monthly migraine days (2.8 ± 2.4 as opposed to 2.2 ± 2.1 in the placebo group) reached statistical significance (P = 0.033). During the last 28 days of treatment, the mean number of monthly migraine was reduced by 3.1 ± 2.6, as opposed to 2.4 ± 2.8 in the placebo group (P = 0.023). In the treated group, 32% of patients experienced a ≥75% reduction in mean monthly migraine days compared with a 14% reduction in the placebo group (P = 0.02). The dropout rate for adverse events was 6.3% for the topiramate group and 4% for the placebo group. In the treated group, the most common adverse events were anorexia (13%); weight decrease (10.2%), with a mean loss of 1.4 ± 2.6 kg (as opposed to 0.7 ± 3.9 kg in the placebo group); abdominal pain (10.2%); paresthesia (8.3%); and somnolence (8.3%). Serious adverse events occurred in four topiramate-treated patients and included infection (n = 2), severe migraine (n = 1) and suicidal ideation (n = 1).
In this study, the mean monthly migraine frequency reduction seems less evident when compared to the previous study [10], probably because migraine frequency at baseline was lower (range 2–9 in the ITT topiramate population) and children with chronic migraine were excluded. An important aspect is that the reduction is more evident in the last 28 days of treatment, indicating that it may take a few weeks before treatment with topiramate reaches its maximum efficacy.
Winner et al. [13] performed a post-hoc subset analysis on 51 adolescents (12–17 years) who had been enrolled in three pivotal trials of topiramate for migraine prophylaxis in adults. Patients were to have had between 3 and 12 migraine attacks, and no more than 14 headache days per 28 days during the 3 months prior to the screening and during the 4-week prospective baseline phase. Also excluded were patients who overused analgesics or who had failed to respond to two or more adequate prophylactic treatments. Topiramate was administered at the dosages of 50, 100 or 200 mg a day for 26 weeks. When compared to baseline, there was a mean monthly migraine frequency reduction of 46, 63 and 65%, respectively, as opposed to a mean 16% reduction in the placebo group (P = 0.07, P = 0.02, P = 0.04, respectively). The most common adverse events (in the group treated with 100 mg a day) were as follows: paresthesia (38%), upper respiratory tract infections (23%), weight decrease (15%), abdominal pain (15%), anorexia (8%) and somnolence (8%). There were no dropouts in the group treated with topiramate. The incidence of adverse events was higher in the group treated with topiramate at the dosage of 200 mg, which did not appear to confer additional efficacy as opposed to the dosage of 100 mg a day.
One class IV study [14], assessing the efficacy of topiramate for pediatric migraine, included 97 children, 75 of which were reevaluated at a first follow-up visit (after 88.7 ± 35.7 days) and 41 at a second follow-up visit (after 203.1 ± 45.6 days). Included were children reporting more than three headaches per month. Topiramate was administered at the dose of 1.4 ± 0.74 mg/kg/day and headache frequency decreased from 16.5 ± 10 headaches/month to 11.6 ± 10.2 headaches/month (P < 0.001) at the first follow-up visit, with 43.1% of the patients experiencing a 50% or greater reduction in the number of headaches. By the second follow-up visit, headache frequency had decreased to 9.4 ± 8.4, with 56.1% of the patients experiencing a 50% or greater reduction in the number of headaches. Mean headache severity, duration and accompanying disability were also reduced. Disability was measured using the PedMIDAS score: a 50% reduction occurred in 48.6% of the patients at the first follow-up visit and in 62.5% of patients at the second follow-up visit. One quarter of patients complained of side effects, which declined in subsequent visits and included cognitive changes (12.5%), weight loss (5.6%) and sensory symptoms (2.8%). A possible bias of this study is that 50.7% of patients were on additional prophylactic medication (most often amitriptyline or divalproex).
In an open, prospective case series (class IV) study [15], topiramate was administered to 24 children with migraine, who had failed to respond to other prophylactics, at the mean dosage of 3.5 ± 1.7 mg/kg for 4 months. Authors report a reduction in the duration and in the intensity of headaches; headache frequency reduction, however, did not reach statistical significance. Adverse events (emotional instability, paresthesia, anorexia, asthenia, weight loss) were experienced by one-third of patients. In our opinion, the baseline headache frequency (3.6 ± 2.7 a month with a range of 1–12) and duration (2–6 h) did not necessarily constitute an indication for prophylactic treatment in all treated cases.
Unfortunately, the studies we reviewed were not homogeneous with regard to inclusion and exclusion criteria and with regard to outcome measures such as the “responder rate,” which was not considered in all studies. Furthermore, the degree of disability caused by the migraine episodes, which is essential in the decision to initiate prophylactic treatment, was not assessed in all studies.
Discussion
Management of pediatric migraine can be tricky, seeing as no pharmacological treatment is formally approved or indicated in the prophylactic treatment of migraine in children and adolescents. A recent review of the pharmacological treatment options in pediatric migraine [16] found only flunarizine, which is not available in many countries, to be effective in rigorous controlled trials. Authors also concluded that there is conflicting evidence regarding propanolol and trazodone, while data from uncontrolled studies suggest cyproheptadine, amitriptyline, naproxen and antiepileptic drugs (topiramate, valproic acid and gabapentin) to be effective. Since this review was published, two randomized, placebo-controlled trials evaluating topiramate in the preventive treatment of pediatric migraine have been carried out [10, 12].
Taken together, data from the papers we reviewed suggest that topiramate is effective in the preventive treatment of pediatric migraine. Furthermore, topiramate dosages, which seem to be effective in the treatment of migraine in children and adolescents (2–3 mg/kg/day) are much lower than those indicated for the adjunctive treatment of epilepsy (5–9 mg/kg/day) in children as young as 2 years.
As regards its safety and tolerability, serious adverse effects were very rare (one patient with suicidal ideation), though the risk of depression must be kept in mind. The percentage of patients experiencing side effects varied greatly among the studies, the most common ones being weight loss (which, in the RCTs [10, 12], was present in 81 and 10.2% of patients, respectively), anorexia (23.8 and 13%, respectively), abdominal pain (14.3 and 10,2%), difficulties in concentrating, somnolence/sedation (19 and 8.3%) and paresthesia (23.8 and 8,3%). Side effects tended to decline over time and dropout rates due to side effects were very low (range 0–6.3%). Since these adverse events, although often transitory, may be clinically significant and distressing both for the child and its family, the decision to initiate treatment must be carefully weighed and should take the degree of disability caused by the migraine episodes into account.
An important aspect, which emerged from some of the studies, is a significant decrease in school absenteeism due to migraine [10] and a reduction in the disability caused by migraine (as assessed by PedMIDAS), leading to an improvement in the childrens’ quality of life [10, 14].
In conclusion, topiramate seems to be a promising therapeutic option, though, clearly, further controlled trials are needed to confirm this data, as are studies comparing different drugs.
|
[
"topiramate",
"pediatric",
"migraine",
"children",
"headache"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Graefes_Arch_Clin_Exp_Ophthalmol-3-1-2082065
|
Characteristics of dynamic processing in the visual field of patients with age-related maculopathy
|
Purpose To investigate the characteristics of dynamic processing in the visual field of patients with age-related maculopathy (ARM) by measuring motion sensitivity, double-pulse resolution (DPR), and critical flicker fusion.
Introduction
Age-related maculopathy (ARM) is the most common cause of loss of central vision in industrialized countries [17, 26, 30, 33, 34, 62]. Almost all clinical assessments of the consequences of ARM for basic visual functions and quality of life refer to foveal or macular vision. However, a few studies have shown that retinal defects can occur beyond 10° eccentricity, which can be considered the maximum radial extent of the macula. Sunness et al., for instance, described histopathological anomalies outside the central retina, whereas sensitivity was unchanged in traditional static perimetry [55].
Curcio et al. described that, beyond general age-related photoreceptor loss, five out of six ARM-affected donor eyes showed cone and rod loss in parafoveal regions [12]. The latter can be expected to reduce dark adaptation performance, which was confirmed up to 25° eccentricity by Brown et al. [9]. In addition, ARM affects temporal aspects of visual information processing. Mayer et al. found reduced temporal resolution of ARM patients as measured by foveal flicker sensitivity with a stimulus of 2.8° diameter [40, 41]. Brown and Lovie-Kitchin showed reduced flicker resolution in the fovea and at 10° and 20° eccentricity and concluded that “the functional effects of ARM are not confined to the central retina, but affect a large region of the visual field” [11]. Falsini et al. found cone-mediated flicker sensitivity (CFS) losses in ARM by evaluating the focal electroretinogram (FERG) as a function of flicker modulation depth in the macula (9° radius) [13].
In summary, the results of these studies show age-related as well as ARM-related functional loss of dynamic processing. However, whether the loss is confined to the macula has not been definitively answered. A further question is whether a loss outside the macula is specific to temporal resolution or whether other temporal aspects of retinal processing, such as motion sensitivity, can also be affected beyond the macula in patients suffering from ARM. Furthermore, we asked whether the impairment is confined to the magnocellular system or also involves the parvocellular pathway.
Methods
Motion contrast threshold
Experimental setup
The experimental setup allowed measurements of peripheral motion perception up to 60° eccentricity under photopic conditions (luminance = 75 cd/m2). It consisted of a white, semicircular (180°) plastic screen (radius 90 cm, height 60 cm) with a rectangular opening in the center. A flat-screen LCD display was mounted behind the opening for the presentation of the test patterns. The subjects sat in front of the screen at a distance of 90 cm and monocularly fixated one of the fixation marks in primary view (fixation crosses with diameter = 4 cm and bar width = 0.5 cm). The marks were attached to the plastic background at 10°, 20°, 30°, 40°, and 60° eccentricity (Fig. 1). Fixation was monitored by a mirror mounted below the LCD display. Thus, the examiner could monitor the subject’s eye movements during the presentation of a stimulus. Additionally, the mirror could be moved horizontally to allow a clear view of the patient’s eye, even in extremely peripheral positions. We are confident that the examiner was able to detect saccadic eye movements of at least 5° amplitude, which is still only half of the distance between two measuring locations.
Fig. 1Experimental set up (top view). The subject’s chair was rotated to maintain primary gaze for all tested fixation marks. a LCD monitor screen. b Semicircular plastic screen. c Fixation crosses (diameter = 4 cm, bar width = 0.5 cm). d Chin rest for stabilizing the head and keeping the viewing distance constant. e Subject’s right or left eye
Monitor calibration
To achieve specified stimulus luminance and contrast, the monitor’s gamma curve (screen luminance vs. gray value) needs to be taken into account during stimulus presentation [4, 51, 52, 54]. We measured this for our monitor using Irtel’s PXL software to set the screen uniformly at the 256 possible gray levels and measured center screen luminance with a digital luminance meter (Gossen Mavomonitor G04068, Nürnberg, Germany) [25]. A second order function was fit to the luminance data by non-linear regression (r2 = 99%):
where L is center screen luminance in cd/m2 and g is the uniform gray value (range 0–255). These function parameters are incorporated in our stimulus presentation software (see below). Target contrast is specified in the Michelson measure, .
Design of the motion stimulus
The test software for motion perception was custom-developed for these experiments and consists of two separate modules, one for stimulus generation and another for running the actual test program. Since threshold measurements are much faster and more reliable using four-alternative forced choice than two-alternative forced choice tasks, we employed four directions of motion (up, down, left, and right) to be discriminated by the subject. The stimuli were plaids composed of two orthogonal Gabor patterns of 45° left and right tilt [15]:
where L and L0 are the luminances of pattern and background, respectively, is the square of radial diameter of the Gaussian envelope (the value where the amplitude has decreased to 1/e), and ϕ(t) is the spatial phase at time t (i.e. the shift within the envelope). For the plaids used here, we chose σ = 0.8, which results in about 1.5 visible cycles (Fig. 2).
Fig. 2Gabor stimulus with double sinusoidal modulation in spatial quadrature. The entire patch was stationary, while the plaid pattern moved within the envelope in one of four possible directions. In addition, the stimulus was temporally modulated by a Gaussian envelope function, so that it slowly appeared out of the medium grey background and then merged back into it
The subject’s task was to identify the direction of motion. Watson found that thresholds of motion detection and motion discrimination in healthy subjects are equal at sufficiently low spatial frequencies (e.g. 2 cpd) for both slow (1.5 Hz) and fast (12.4 Hz) movement of the test pattern [63]. Therefore, we used a low spatial frequency of 0.65 cpd and an intermediate speed of 5.71 °/s (corresponding to 3.75 Hz local luminance change) for the motion stimulus. The size of the test pattern at the viewing distance of 90 cm was 3.8°. To minimize attracting transient attention by abrupt pattern onset, the test pattern gradually appeared and disappeared by using a Gaussian temporal envelope [27, 36, 37]. Due to the small changes between frames in these stimuli, “tearing” of the image was extremely unlikely and was indeed never observed [16].
Adaptive algorithm
Monocular contrast thresholds for motion perception were measured along the horizontal meridian in the nasal and temporal visual field at 10°, 20°, 30° and 40° eccentricity and in the temporal visual field additionally at 60°. The initial contrast value for all eccentricities was 30% (1.48 log %-contrast). For each eccentricity, the threshold was determined twice. Two blocks of nine measurements each were taken and the arithmetic mean calculated from the two results. The adaptive algorithm for threshold determination was based on Kesten and was modified according to Kaernbach to allow the additional response alternative “I don’t know” [28, 29, 32, 58]. This paradigm, termed “unforced choice task” by Kaernbach, has been shown to be beneficial, especially for subjects who are unfamiliar with psychophysical testing procedures.
Kesten’s algorithm is of the staircase kind, i.e. the intensity of the next presentation depends on the previous response only, not on the entire history or a longer sequence of preceding responses [32]. After a correct response, the log contrast of the stimulus was decreased by one incremental step, and after an incorrect response, it was increased by 1.67 incremental steps (5/3). It thereby converged on the 62.5%-correct point (5/8), i.e. the point of inflection ϕ on the psychometric function (of log contrast) in a four-alternative forced-choice task. The ratio V=5/3=1:1.667 of up-down intensity changes results from Kesten’s algorithm given the number n of response alternatives (four) by
where ϕ is the probability at threshold or point of convergence,
with guessing rate γ = 1/n = 1/4. (c.f. Treutwein’s equation 16, the last term (Zn–ϕ) [58]). When the subject gave the indecisive response “I don’t know”, the intensity of the next stimulus was increased by one incremental step, as specified by Kaernbach’s extension which requires the specification of V as above [28]. The step width in Kesten’s algorithm is reduced by a factor of m/(m + 2) after a change in response category (correct-to-incorrect or vice versa) where m is the number of reversals (Treutwein’s eq. 16). Since Kaernbach’s extension requires the above-mentioned fixed ratio of up-down intensity changes, we used a step width reduction only every other reversal (i.e. m even). The Kesten rule is applied to log contrast, i.e. log Cn + 1= (logCn)±sn (where sn is the step width), which is equivalent to contrast being multiplied or divided by the antilog of step width.
Double-pulse resolution
Treutwein and Rentschler developed a hardware and software combination for the measurement of double-pulse resolution (DPR) which allows simultaneous measurement at multiple visual field positions [60]. The technique has been used in experiments on patients with multiple sclerosis and glaucoma, and in an extensive study characterizing the temporal properties of the visual field across the life span [46, 48–50]. The experimental setup allowed (limited by screen size) measurement of DPR up to 20° eccentricity, i.e. not as far into the periphery as for the motion contrast thresholds but well beyond the macular region.
Test setup
An analogue 15″ x-y-z display was used for stimulus presentation (Hewlett Packard model 1310, i.e. a CRT-display without a raster-scan generator), driven by a temporary buffer that stores the point coordinates and generates the control voltages for the display (“point plot buffer”; G. Finlay, Edmonton, Canada). Temporal control could be extended into the low microsecond range, so that the temporal resolution was better by a factor of 1,000 than in conventional setups [60]. A computer (PC) was used for running the experimental software that controlled stimulus generation, adaptive procedure, and data acquisition. The experimenter recorded the subject’s responses via the computer keyboard.
Stimulus characteristics
The tests were performed monocularly at 30 cm viewing distance with a diagonal screen diameter of 40 cm. The stimuli in Treutwein’s technique are eight squares of 1.15° × 1.15° visual angle (5 × 5 pixels) each, arranged in circles at 5°, 10° , and 20° eccentricity around a ninth square in the center [60]. Figure 3 shows the stimulus time course. Eight of nine squares are continuously present for 80 + γ + 280 ms. The target, one randomly selected stimulus, is switched off after 80 ms and then on again after a variable time gap γ. The subject’s task was to detect the gap and indicate the target position on the screen, which makes it a nine-alternative forced choice task. In supra-threshold trials, its location can be recognized as a short flicker of the target. A central fixation cross was presented continuously between stimulus presentations and was switched off 50 ms before the beginning of a new trial. The squares’ luminance was 215 cd/m2; subjects were light-adapted and the display background luminance was held constant in the low photopic range (luminance ∼15 cd/m2).
Fig. 3Double-pulse stimulus characteristics (figure modified after Poggel and Strasburger, originally from Treutwein and Rentschler [46, 60]. a, b Time line diagram of stimulus display. c Stimulus positions (5°, 10°, and 20° eccentricity)
Gap duration was varied by an adaptive thresholding algorithm (“YAAP”), based on maximum-likelihood psychometric function fitting [20, 59].
Flicker frequency analyzer
For measurement of foveal flicker fusion frequencies we used the “Wiener Testsystem”, a commercial computer-based diagnostic system with an emphasis on assessing driving fitness (Schuhfried, Moedling/Vienna, Austria). The system was chosen as a reference, because it is an established standard system in Germany used for psychological testing; due to its construction it allows foveal measurement only. Measurements were performed by presenting a circular flickering red light stimulus (1.2° diameter; luminance = 270 cd/m2; wave length 655 nm) on a white background in a tubular viewer. Thresholds were determined by the method of limits, i.e. the frequency of the flickering light is increased until permanent light is perceived and is then decreased until the light is once again perceived as flickering.
The subject indicated the perceived changes by a keystroke whereupon the critical frequency was recorded. The separate arithmetic means of the critical frequencies determined in the ascending and descending series are referred to as fusion frequency and flicker frequency, respectively. Each test cycle consisted of five training cycles, immediately followed by eight measurement cycles. For further analysis, the mean of the fusion frequency and the flicker frequency was calculated for each subject, which is referred to as the critical flicker fusion frequency (CFF).
Color perception
The Lanthony D-15 color vision test (desaturated) was used for the examination of color perception. The test consists of 15 colored chips, which are used to determine the subject’s capability to discriminate hues. At 50 cm viewing distance, every chip has a diameter of about 1.5°. The subject sorts the arbitrarily shuffled chips according to the arrangement in the color cycle. The chosen sequence was then documented in a test protocol and the color confusion score (CCS) and crossings over the color space were determined [1]. The measurement was done once under constant lighting conditions (color temperature ∼4000 K).
Subjects
We recruited three groups of subjects for this study: 14 subjects with ARM (19 eyes, mean visual acuity: 0.29, SE ± 0.048; ∼20/63), a control group of age-matched subjects with healthy eyes (14 subjects; 18 eyes, mean acuity: 0.65, SE ± 0.057; ∼20/32), and another control group of 7 young subjects (8 eyes, mean acuity: 1.25, SE ± 0.047; 20/16) (Table 1). ARM in 16 of the 19 eyes of patients were classified as “dry” and three as “wet”. Six of 19 eyes in the ARM group had received intra-ocular lens implants during cataract surgery, which was also the case for 14 out of the 18 eyes of the age-matched control group. Since in most cases the implant causes mild myopia (−0.5 dpt), the patients-though presbyopic-could observe the fixation mark comfortably without near correction. Thus, optical artifacts from eyeglasses were prevented, such as aberrations or interference from the glasses’ rims, which would otherwise occur at higher eccentricities.
Table 1Age and visual acuity of the subject groupsAgeVisual acuity NMeanSEMinimumMaximumMeanSEARM group1973.792.360900.29 (20/63)0.048Age-matched controls1871.51.560780.65 (20/32)0.057Young controls827.11.721341.25 (20/16)0.047Visual acuity is described in decimal notation and Snellen fractions
Visual acuity (VA) was measured by a Landolt ring test (Binoptometer, Oculus, Germany) at 90 cm viewing distance (VA = 1/ω′, where ω′ is the minimum angle of resolution in minutes). Based on these values, all subjects were capable of directing their gaze at the fixation mark without glasses. All subjects with ARM and the age-matched control group provided an up-to-date medical statement from an ophthalmologist, which appraised the condition of the retina and excluded other eye diseases (e.g. glaucoma, secondary cataract). All subjects were examined by their ophthalmologist with respect to clarity of the optical media, and none of them showed any evidence of changes in the visual pathways.
The complete examination (including the initial interview) took approximately 3 hours.
The ages of the subjects with ARM and the age-matched controls were sufficiently close; a t-test with a preceding test for normal distribution showed that the two groups did not differ significantly with respect to their age (p = 0.36). The study design had been approved by the ethics committee of the University of Munich and testing procedures were in accordance with the tenets of the Declaration of Helsinki. All subjects gave their informed consent for participation.
Results
Motion contrast threshold
Figure 4 shows the contrast thresholds for motion perception as a function of eccentricity in the visual field. Compared with the age-matched controls, the ARM group shows a distinct increase of contrast thresholds up to 20° in the nasal and up to 40° in the temporal field. Mann-Whitney U-tests of the contrast thresholds at 10°, 20°, 30°, and eccentricity on the horizontal meridian each showed significant differences between the two elderly groups (Table 2).
Fig. 4Mean contrast thresholds for motion perception of the subjects with ARM, age-matched controls, and young controls on the horizontal meridian. The broken lines show the increased contrast thresholds when stimuli were presented near the blind spot. The continuous lines show the regression (specified in the formula); omitted for the age-matched control group on the temporal side for clarityTable 2Results of Mann-Whitney U-tests of contrast thresholds (ARM group vs. age-matched control group); U is the statistical value of nonparametric Mann-Whitney test for comparison of independent samples based on ordinal ranks; z is the estimated statistical value of normalized distribution of U values; p is the significance levelEccentricityNasal fieldTemporal fieldUzpUzp10°16.00−4.71<0.00153.00−3.59<0.00120°69.50−3.090.00272.50−2.830.00530°148.50−0.680.4996.00−2.280.02340°145.50−0.250.80102.00−2.100.03660°−−−134.00−1.120.261
Double-pulse resolution
With respect to double-pulse resolution, the subjects with ARM showed strongly increased thresholds compared with the age-matched controls within the 20° visual field covered by the test area (Fig. 5). For better comparison with the motion sensitivity data, Fig. 6 also shows the double-pulse results on the horizontal meridian. Both the comparison on the two horizontal meridians and on the entire circles at the investigated eccentricities (5°, 10°, and 20°) show significant group differences. We performed a Mann-Whitney U-test at every position on the horizontal visual field meridian (Table 3).
Fig. 5Mean double-pulse resolution (DPR) of the ARM group, age-matched controls and young controls at 5°, 10°, and 20° eccentricity (stimuli positioned on concentric rings, Fig. 3). The 0°-value represents the foveal DPR in the 5°-ring measurement. Note that the rather small SEM error bars indicate a limited range of variation of valuesFig. 6Double-pulse resolution (DPR) of the ARM group, age-matched controls, and young controls along the horizontal meridianTable 3Results of Mann-Whitney U-tests at each eccentricity for the ARM group and the age-matched control group; U is the statistical value of non-parametric Mann-Whitney test for comparison of independent samples based on ordinal ranks; z is the estimated statistical value of normalized distribution of U values; p is the significance levelEccentricityNasal fieldTemporal fieldUzpUzp5°46.50−2.550.01146.50−2.560.01110°45.00−3.83< 0.00179.50−2.780.00520°76.00−2.890.00361.00−3.340.001
Critical flicker fusion frequency (CFF)
Figure 7 shows critical flicker fusion frequency (CFF) as a function of age for the three subject groups. There is a slight but significant (p < 0.05) loss of CFF with age. It amounts to 0.059 Hz per year of age (i.e. around 5 Hz for the entire life span), which accounts for 30% of the inter-individual variance in the healthy subjects. Age-independent interindividual variance (70%) thus far exceeds the age-related loss. More importantly here, however, the graph also shows the marked loss of performance in the subjects with ARM. On average, the CFF in the ARM group is lower by 5.47 Hz, or approximately 14%, in comparison to the age-matched controls. Note that one 90-year-old subject reached the level of young normal subjects, while all other subjects with ARM showed lower performance than all young controls.
Fig. 7Critical flicker fusion frequency as a function of age for the three groups. The regression line refers to the healthy controls
The analysis of variance (one-way ANOVA with group as factor and post-hoc test, Tamhane procedure) indicates a highly significant difference between the three groups (F = 25.16; df = 2; p < 0.001) which was caused predominantly by the highly significant difference between the age-matched control group and the ARM group (p<0.001).
Color perception
Performance in color perception is a sensitive indicator of differential impairment of the parvo- vs. the magnocellular system [31]. The result from the Lanthony D-15 color vision test (unsaturated) allows the calculation of the color confusion score (CCS), which was used for the statistical analysis of the results. A one-way ANOVA with subject groups as factor and subsequent post-hoc test (Tamhane procedure) showed highly significant differences (F = 25.285; df = 2; p < 0.001) between the three test groups. The difference between the age-matched control group and the ARM group was also highly significant (p < 0.001). Thus, the subjects with ARM clearly showed impaired color discrimination. The comparison of crossings over the color space also shows significant differences between the subject groups (F = 20.394; df = 2; p < 0.001; mean crossings: ARM = 5.81, age-matched controls = 1.78, and young controls = 0).
Discussion
Motion perception
Our findings document deficits of dynamic visual field properties in ARM in a retinal region that extends far beyond the macula. In a sense, the well documented loss of visual function (e.g. foveal acuity) in the macula of subjects with ARM seems just like the proverbial tip of an iceberg. To our knowledge, this is the first study of dynamic characteristics of the peripheral visual field up to 60° eccentricity in subjects with ARM. Functional studies on these patients in the past have either concentrated on foveal function only, or have used a large central stimulus that did not discriminate between foveal and nonfoveal function [5, 40–42, 57]. Only the study by Brown and Lovie-Kitchen has shown deficits in the temporal visual field up to 20° eccentricity [13].
There are only a few studies on dynamic processing characteristics in ARM. Mayer et al. measured flicker sensitivity only foveally and suggest that it can be a predictor of exudative ARM [40, 41]. Normal aging of visual function – from which ARM losses need to be distinguished – has received only limited attention [19, 45, 46, 61].
In the present study, we determined the contrast thresholds for motion perception far into the periphery of the visual field on the horizontal meridian, up to 40° on the nasal, and 60° on the temporal side. We found pronounced impairment in the ARM group up to 20° nasally and 30° temporally, and smaller impairments still further out on the temporal side. The macula is commonly described as having a radius of 10° [2, 22]. These elevated contrast thresholds for motion perception extend far beyond this eccentricity, in conflict with a widespread assumption that the effects of manifest ARM are restricted to the macula. The latter seems to be based on two facts. First, customary descriptions and classifications of the disease use morphological features that are visible by ophthalmoscopy or on fundus photographs [6–8]; and second, many experimental studies have tested vision in subjects with ARM only in central locations in the visual field as noted above, with a few notable exceptions [10, 11, 55].
In clinical practice, the assessment of the visual periphery is believed to play a minor role in ARM, for a lack of consequences in ophthalmic treatment and also because of the dramatic vision loss from damage to the central retina, immediately noticed by patients who report the loss of the ability to recognize fine spatial or contrast detail (e.g. small print). A loss of peripheral motion sensitivity, in contrast, is not easily recognized by the patient, possibly because the conscious (sustained) component of attention is normally directed towards central vision, whereas the transient (reflex-like) component of attention is not accessible by consciousness [44]. This phenomenon plays a well-known role in glaucoma which often goes unnoticed by the patient. Thus, the fact that patients do not complain about deficits in peripheral motion perception might lead the practitioner not to examine this retinal region in detail. Yet, practical consequences have to be considered. If diminished sensitivity to motion raises the possibility that a patient’s transient attention might not be attracted to a moving stimulus, a behaviorally valuable eye movement may not be performed and a potentially hazardous situation may not be recognized.
Additionally, the tests available in routine clinical practice do not provide the appropriate stimuli to detect such defects. For instance, Holopigian et al. examined peripheral vision in patients with early ARM using standard electrophysiological and psychophysical tests (dark adaptation curves, electro-oculograms and electro-retinograms [23]. They found that only few subjects with early ARM developed disease-related effects in this region when examined with standard clinical tests.
Double-pulse resolution (DPR) and critical flicker fusion (CFF)
The results of DPR revealed distinct and statistically significant differences between the ARM group and age-matched controls up to 20° eccentricity. Furthermore, the increased foveal thresholds of double-pulse resolution in subjects with ARM parallel the elevated foveal CFF thresholds, which also show a dependency on age and a significant increase in subjects with ARM.
Comparison of the age-matched with the young control group shows an age dependency which confirms earlier results [18, 46, 48, 61]. However, our results cannot be directly compared with the findings of Falsini et al. [13] who used a large homogeneous test field (18° diameter) that covered visual field regions with pronounced differences in functional characteristics.
The fact that reduced dynamic performance in our patients was found for two quite different methods of measurement emphasizes the presence of impaired temporal vision performance far beyond the macula and suggests a common underlying pathogenesis.
Color vision
The objective of the color vision test was to examine the relative involvement of the two neuronal subsystems of the primary visual pathway – magnocellular (M) and parvocellular (P) – in the previously obtained results. Both motion perception and temporal resolution are assumed to be predominantly conveyed by the M system, whereas color vision is predominantly mediated by the P system [31]. To simply find out whether color perception is intact, it is legitimate to use the color vision test foveally, or to allow the subjects to use the retinal area that serves them best. A (seemingly self-evident) conclusion that the deficits reported here may be caused by a specific damage to the M system can thus be rejected. Impaired color vision, as found here, shows that the P system is also affected.
The ARM group in our study showed a significantly lower foveal color vision performance than the age-matched control group. This would seem in conflict with the results of Frennesson et al. and Medina et al. who found no color vision deficits in subjects with ARM at an early stage [14, 42]. However, those studies used less sensitive tests with saturated colors, such as the Farnsworth D-15 test or the Farnsworth-Munsell 100 Hue test, whereas our findings are based on the more sensitive Lanthony D-15 test which uses unsaturated colors [14, 42]. Since (static) contrast sensitivity was not measured here, we cannot estimate how strongly the color vision findings were influenced by the expectedly reduced contrast sensitivity of these patients.
Consequences
Even though the perceptual limitations measured in this study are not immediately evident to subjects with ARM, their relevance for daily life should not be underestimated. Their importance is due especially to the fact that segments of the near periphery assist safe mobility in general and driving a car in particular [21, 39]. Visual perception of moving stimuli has two functions of great importance in daily life. First, we perceive a stable, stationary world in which we are moving, despite the fact that nearly all image components on the retina are moving [43]. Being able to distinguish between the effects of self-motion and those of moving objects requires the detection and analysis of relative motion [47]. Second, our attention is constantly redirected to interesting and potentially dangerous objects, which typically cause a reorientation of gaze. These functions were examined previously in connection with the distinction between sustained and transient attention [3, 38, 44].
Because of the impairment of the central visual field in ARM, motion perception with peripheral vision is of even greater importance for these patients. Measurements with dynamic test stimuli are especially important for the assessment of driving ability of visually impaired persons [24]. Marron and Bailey examined contrast sensitivity and visual acuity with regard to their importance for orientation and mobility [39]. Their results provide evidence that contrast sensitivity and the integrity of the visual field are of considerably greater importance for safe mobility than visual acuity. Studies from our own group show that dynamic peripheral sensitivity in a divided attention condition was the strongest predictor for safe driving in a group of healthy elderly drivers [53]. Recent studies have confirmed the importance of peripheral vision for driving by finding that contrast sensitivity and visual field limitations are the strongest predictors for mobility of visually impaired persons [10, 21, 35]. These studies demonstrate the importance of the assessment of visually guided performance regarding moving objects and contrast vision. In the tests at hand, these two tasks are linked. The growing possibilities for mobility of the elder generation with age-related visual impairment require examinations that take into account seniors’ new lifestyles and can help ensure safe navigation.
Another relevant connection can be made with communication and the ability to read speech from the lips of a conversation partner, which is a difficulty faced by many elderly with additional hearing impairment. Although a general notion of the role of vision in speechreading has been established [56], the importance of motion sensitivity for speechreading has not been sufficiently investigated to date.
|
[
"visual field",
"arm",
"temporal resolution",
"motion perception",
"contrast sensitivity"
] |
[
"P",
"P",
"P",
"P",
"P"
] |
Eur_Spine_J-2-2-1602206
|
Dorsal root ganglionectomy for pseudotumor of the L3 dorsal root ganglion: a rare case and a rare treatment
|
Dorsal root ganglia are oval enlargements on the dorsal nerve roots and contain the cell bodies of sensory neurons. Asymmetry of dorsal root ganglia may occur naturally, yet natural occurrence of gigantic dorsal root ganglion (DRG) is rare. The patient was 61-year-old woman who presented with atypical symptoms like neuropathic pain and urinary distention. Neuroimaging has shown left L3-4 far-lateral disc herniation and a gigantic L3 DRG. At surgery, the dural sheath of the ganglion had to be opened and a firm, yellow-colored abnormal tissue was exposed. The abnormal tissue considered to be a tumor of neural origin was gross totally excised and the patient’s symptoms ceased immediately after surgery. Histopathological examination of the specimen revealed nothing more than normal DRG morphology. At 4 months postoperatively, the patient is well with mild L3 hyperesthesia and hyperalgesia. Dural sheath opening in neurosurgery is not a routine practice. The sheath may need to be opened when surgeon suspects of a tumor, a free disc fragment and any inflammation within the ganglion. Operative morphology of a severely edematous but non-tumoral (pseudotumor) ganglion has not previously been documented.
Introduction
Small neoplastic lesions of the vertebral column and the spinal canal are often misdiagnosed as intervertebral disc disease owing to the commoner occurrence of the latter [1]. Rarely, a neoplastic lesion may coincide with a herniated intervertebral disc usually without a topographic relationship in between [8]. Concomittant appearance of a disk herniation and a neoplastic lesion compromising the same nerve root represent a unique situation [1]. We herein report the unique case of a pseudotumor of the L3 dorsal root ganglion (DRG) adjacent to a far-lateral L3-4 disc herniation.
Case history
A 61-year-old physician presented with excruciating pain in her left leg. The mild leg pain that has been around the left knee and around the posterior hip for some weeks suddenly became unbearable one afternoon. The patient came to emergency room 8 h after the onset of the pain. Upon further questioning, the patient confessed that she had not emptied her bladder since the pain started. Past medical history was positive for a distant episode of trigeminal neuralgia, cluster headache, and an episode of significant back and right leg pain 7 years ago and a recent pulmonary embolus. Past surgical history included total hysterectomy and bilateral salpingo-oopherectomy after a serious pelvic inflammatory disease and abscess formation within the uterine tube.
In the emergency room, the patient was in extreme pain. Any movement or touch would exacerbate the pain. Right leg-raising test was positive at 60 and left leg-raising test was positive at 20°. Power was normal in both legs, including knee extension. Patellar and the ankle reflexes were normoactive bilaterally. Abdominal examination for a distended bladder was fruitless because of softened subcutaneous fat. Lumbar MR showed disc degeneration and mild bulging at L3-4, L4-5 and L5-S1 levels yet at L3-4 level there was foraminal and a round extraforaminal disc herniation (Fig. 1a, b). Lumbar MR scan also verified a fully distended bladder (Fig. 1c). Considering vaque radiological findings in contrast to serious pain and urinary retention, the MR examination was extended to cover the whole cervico-thoracic and thoracic levels. There was no cord compression.
Fig. 1Axial T1-weighted (a) and T2-weighted (b) MR scans through L3-4 interspace show an extraforaminal lesion with questionable continuity with the disc space. The lesion is isointense on both T1- and T2-images. Please note that on T2-weighted axial image (b), the contralateral dorsal root ganglion (DRG) shows intermediate to high signal. c Parasagittal MR scan show L3-4 disk herniation obliterating the left neural foramen anteroinferiorly.Yet the left L3 nerve is of same width as the roots one above or below. d Parasagittal MR scan 5 mm lateral to section c shows that the enlarged DRG (arrows) appears to have been stuck in the distal neural foramen. Please note that the fat plane around the DRG is totally obliterated
The next morning, urinary retention re-occurred and an indwelling urinary catheter had to be placed. An ultrasound of the abdomen and pelvis was not contributory except identification of few benign adhesions secondary to previous pelvic surgery and a distended gall bladder. Meanwhile the left knee pain and the left sciatica got only a bit better but the patient described additional occasional electric shock like pain radiating medially from the left groin to the left knee. Lumbar computerized tomography (CT) to rule out a small bone chip originating from end-plate attached to a free disc fragment was done next. Axial CT scans showed the presence of a left-sided round hyperdense extraforaminal lesion at L3-4 level that was reported to be consistent with far-lateral disc herniation (Fig. 2a). Coronal reformatted CT images better demonstrated the lateral extent of the disc herniation (Fig. 2b). After an unsatisfactory 2-day narcotic treatment with codein 60 mg q6h, gabapentin 300 mg twice a day was started. On the third day of admission, a new pain emerged around the left inguinal area. By this time, the left patellar reflex diminished and left knee extension got 20% weaker. IV steroid (80 mg methyl prednisolon as the loading and 40 mg twice a day as the maintenance dose) was then started. Steroid alleviated a significant portion of the pain. Tapering steroid dose immediately caused return of intense pain, so surgery was considered as the last resort.
Fig. 2Axial CT image through the lower end-plate of L3 (a) and coronal reformatted CT (b) image show a round hyperdense extraforaminal lesion consistent with far-lateral disc herniation
The patient was informed to consent for the removal of far-lateral disc herniation at L3-4 level. After medial facetectomy and ligamentous removal, the left L4 nerve and the L3-4 disc space were exposed. The bulging disc was compressing the L4 root from its shoulder. After standard discectomy, the L4 root became mobile. Next, lateral disc space was emptied by angling the rongeur laterally. Yet, disc compression alone was far from explaining the dramatic clinical picture. After total removal of left L3 inferior facet and L4 superior facet and undercutting the L3 pedicle, the left L3 root was identified and traced laterally. L3 nerve was more or less normal sized yet L3 DRG was extremely different than normal. The DRG was 3–4 times bigger than a normal DRG. The L3 root was totally immobile (Fig. 3a). The DRG felt extremely gritty. The dural sheath of nerve root and the DRG was opened. A yellow-colored firm tissue came into view (Fig. 3b, c). There was good cleavage from the dural sheath and the underlying septum. The healthy ventral rootlets were pushed anteriorly and inferiorly by the mass (Fig. 3d). Overall, the mass lesion within the root sleeve was totally extirpated. Although no cerebrospinal fluid (CSF) was seen in the operative field, the dural sheath was sutured with interrupted prolene sutures and the suture was reinforced with tissue fibrin glue. On postoperative day 1, the urinary catheter was removed and the patient emptied her bladder spontaneously. The original pain was totally gone. The left knee function returned to normal. On third postoperative day, she complained of mild hyperesthesia and hyperalgesia over the left knee which did not affect her daily life. At 4 months postoperatively, she is leading a normal life with gabapentin 300 mg twice a day.
Fig. 3Operative microphotographs show that: a the left L3 root and the DRG are extremely swollen, b after opening the dural nerve sheath, the upper end of a yellow colored tumor is being dissected, c the lower end of the tumor is being dissected with better cleavage, d after the removal of the tumor normal appearing ventral motor rootlets come into view, e root size return to normal after suture closure of the dural sheath (black arrows)
The histopathological specimen was extensively studied and did not reveal anything other than normal DRG tissue with mature ganglion cells, axons and Schwannian stroma supporting the axons (Fig. 4).
Fig. 4Photomicrographs a with hematoxylin and eosin and b with neurofilament (NF) protein immuno-stain demonstrate axons (left) and ganglion cells dispersed in a cellular stroma consisting of bundles of elongated spindle cells (right) (×200 original magnification). This represents normal DRG morphology
Discussion
Dorsal root ganglia are oval enlargements on the dorsal nerve roots and contain the cell bodies of sensory neurons. Although both the DRG and the ventral motor root are surrounded by a common dural sheath, a thin fibrous septum separates both. The DRG are surrounded by fat within the intervertebral foramen. According to the baseline anatomic information that Hasegawa et al. [5] obtained from male volenteers using MRI, the mean width, length and height is 5.7×7.1×7.3 mm3 for L3 DRG, 6.2×8.4×8.2 mm3 for L4 DRG, 5.9×9.4×8.3 mm3 for L5 DRG and 6.2×11.2 mm2 for S1 DRG, respectively (data on the height of S1 DRG was not given in this study). The lower was the spinal level, the bigger was the ganglion size. In that study, the DRGs were located in lumbar intervertebral foramen in 92% of L1, 98% of L2, 100% of L3 and L4 and 95% of L5 levels. The S1 DRG was intraspinal in 79%. However analyzing patients presenting with low-back pain or sciatica, Hamanishi and Tanaka [4] found that 100% of L2 DRG, 48% of L3 DRG, 27% of L4 DRG and 12% of L5 DRG were located extraforaminally while 52% of L3 DRG, 72% of L4 DRG and 75% of L5 DRG were located intraforaminally.
Preoperative radiological diagnosis of spinal nerve tumors producing radiculopathy or masquerading as spinal and/or root pain has been facilitated by combined use of CT and MR. Intradural extramedullary tumors like schwannoma or extradural tumors like neurofibroma frequently affect the DRG and appear as dumbbell-shaped in 15% of cases [10, 14]. Diagnosis of tumors that had already enlarged the neural foramen is usually straightforward. However, a smaller pathology intrinsic to the root or the DRG will first be expected to cause an asymmetry as also depicted by MR scans in our patient. Yet asymmetry of DRGs may also occur naturally [4]. Hamanishi and Tanaka [4] found asymmetrical locations and sizes in 11% of the reviewed DRGs. Asymmetry at L3-4 level is rarer. Out of 104 patients that they reviewed, only one had DRG asymmetry at L3-4 level. Two cases who had asymmetrically gigantic DRGs similar to our case underwent surgery at different spinal levels [4]. In the first case, the lowered pedicle appeared to push down and kink the intraspinal portion of the L5 DRG. In the second case, the mass mimicking the gigantic DRG was histologically confirmed to be very old sequesterated nucleus pulposus at L5-S1 level. However, anteriorly compressed DRG could not be detected by MR scan at all [4].
As for MR signal intensity, schwannomas and neurofibromas will often appear iso- to hypointense on T1- and hyperintense on T2-weighted MR images [14]. The left L3 DRG in our case appeared isointense on both T1- and T2-weighted MR images. Intermediate signal on T2-weighted MR image usually suggests the presence of a tumor with numerous cellular and fibrous components and very little myxoid stroma. Unfortunately, paramagnetic contrast was not given in our case. One can argue that the preoperative radiological diagnosis of a tumor could have been possible if intravenous gadolinium was given. Although gadolinium enhanced MR imaging allows for better detection of intraneural edema with diffuse enhancement along the involved nerve roots, this characteristic probably would not have helped us in diagnosing a DRG tumor since DRG always enhance with contrast medium because of a less developed blood–nerve barrier anyway. Recently, a French group studying enhancement pattern on 180 healthy DRG also concluded that 100% demonstrated significant and homogenous enhancement [3]. Therefore, even if we had used intravenous paramagnetic contrast preoperatively, it would probably be impossible to differentiate between an intraradicular tumor and a healthy DRG.
It was only at surgery that the left L3 DRG was found to be extremely swollen and had to be opened. A nerve root and a DRG may become swollen because of both biochemical irritation of nucleus pulposus and the mechanical compression of the disc materal [2, 6]. The mechanical compression of the DRG can increase the endoneural fluid pressure within the DRG almost threefolds [11]. Yet since the dural sheath is not an elastic tissue, a DRG can hardly ever reach the gigantic size that we have observed. Morever, the far-lateral disc protrusion at L3-4 level observed at surgery (in contrast to the extent of the herniation as seen on CT and MR images) could not have exerted the force necessary to cause a significant endoneural fluid pressure rise within the left L3 DRG. Therefore, the adjacent far-lateral disc herniation probably just coexisted with the gigantic left L3 DRG of unknown origin and served a better purpose by attracting our surgical attention to this part of the spine. Not only the size of the involved DRG, but also its firm feeling suggested the presence of something unusual inside it. Although intraradicular disc herniations occur more commonly at L5-S1 and L4-5 levels [12], the surgeon’s impression before the microsurgical opening of the dural sheath was more inclined to a rare intradicular disc herniation rather than a tumor.
Dural opening in similar cases totally depend on surgeon’s experience and/or attitude or on the medicolegal restrictions of the preoperative consenting procedure. By opening the dural sheath of a nerve root or a DRG the surgeon not only will prolong the operative procedure but also will take the morbidity risk associated with CSF leak and difficulty with wound healing. Moreover, extra expense of tissue fibrin glue and/or external lumbar drainage may also trigger an hesitation before dural opening in the currently managed health care systems. If a gross pathology like a tumor or a disc fragment is not readily encountered between the rootlets, chances are histopathology will be the only diagnostic tool. In addition to a real mass lesion like tumor or disc material, inflammatory reactions within a DRG—although very rare—may also cause a tumor-like lesion. A team from Tübingen, Germany recently has reported a very similar case presenting with low back pain and progressive hypesthesia of the thigh [10]. MR imaging with contrast enhancement revealed an intraspinal–extradural mass at L1-2 level compressing the right L2 root. The surgical part of their case is almost identical to ours. The surgical team emptied the disc and the nerve was still compressed and appeared to be enlarged. In the end, they had to open the dural sheath and resect the indurated and swollen part of the ganglion. Histopathological diagnosis was ganglionitis characterized by loss of ganglion cells, increased number of nodules of Nagcotte and diffuse infiltration of T cells.
Ganglionectomy as performed for cancer pain or non-cancer pain (for failed back surgery syndrome or chronic lumbar radiculopathy) was originally described by Osgood et al. [9] and by definition is an extradural operation [13]. Wilkinson and Chan [15] in their recent publication advocated the use of microscope for the procedure. After exposure of the ganglion and surrounding plexus and compression of the venous plexus with cottonoids, the longitudinal incision path along the dural sheath is coagulated using low-current bipolar. Once the dural sheath is incised dorsally using a no. 11 knife, ganglion tissue will come into view. Young [16] described the DRG as a round, yellow, irregular structure. Wilkinson and Chan [15] described the ganglion as having an appearance similar to fat, with tiny “sparkles” within. The latter description is for ganglia from patients with failed back surgery syndrome or chronic lumbar radiculopathy. In rare circumstances like we had, differentiation between an healthy and a tumoral ganglionic tissue may be extremely difficult even with the operating microscope. At this stage, the surgeon has to decide between a biopsy or total removal of the abnormal looking ganglionic tissue. Biopsy may be safer but then there is the possibility of sampling error. If total removal of the DRG is aimed (sensory ganglionectomy), then the afferents and efferents of the ganglion are elevated first and then the ganglion is separated from the underlying septum. To lessen the likelihood of neuroma formation, the proximal end of the ganglion is either hemoclipped or cut with CO2 laser [7]. If leakage of CSF is not encountered (which usually is the case) a watertight closure of the perineurium is not necessary. Some even advocated to leave an autogenous fat graft in the surgically created defect [15]. As for complication of the procedure, deafferentiation pain has been exaggerated because it is usually transient and never reaches to the point of anesthesia dolorosa [13, 15]. Taha et al. [13] states that there is no uniformly beneficial treatment for deaffarentiation pain after ganglionectomy but our patient definitely benefited from brief use of gabapentin.
In conclusion, we reported the unique case of a pseudotumor of L3 DRG. The patient presented with atypical symptoms like neuropathic pain and urinary distention. Preoperative MR scans showed a gigantic left L3 DRG yet this asymmetry was masked by the presence of an adjacent ipsilateral far-lateral disc herniation. Attention to foraminal and extraforaminal part of the spine is of utmost importance in the evaluation of lumbar MR scans. Surgery was both diagnostic and therapeutic in this case. Before opening the dural sheath of a nerve root or DRG, a surgeon should realize that if a tumor or a disc fragment does not come out at once, then it may only be through histopathology that a proper diagnosis can be reached.
|
[
"dorsal root ganglion",
"ganglionectomy",
"magnetic resonance imaging",
"microsurgery"
] |
[
"P",
"P",
"M",
"U"
] |
Ann_Surg_Oncol-3-1-1914247
|
Long-Term Locoregional Vascular Morbidity After Isolated Limb Perfusion and External-Beam Radiotherapy for Soft Tissue Sarcoma of the Extremity
|
Background Isolated limb perfusion (ILP) with tumor necrosis factor alpha (TNF-α) and melphalan, followed by delayed surgical resection and adjuvant external-beam radiotherapy is a limb salvage treatment strategy for locally advanced soft tissue sarcomas. The long-term vascular side effects of this combined procedure were evaluated.
Optimum treatment strategies for locally advanced soft tissue sarcoma (STS) of the extremities have evolved considerably over the past 25 to 30 years. In 1982, Rosenberg et al.1 showed that amputations or exarticulations of the affected limb do not result in higher survival rates. This made preservation of the extremity and good limb function all the more important.2,3 An established limb salvage strategy is to perform a hyperthermic isolated limb perfusion (ILP) with tumor necrosis factor alpha (TNF-α) and melphalan, followed by delayed tumor resection and, in case of marginal or nonradical resection, by adjuvant external-beam radiotherapy (EBRT).4,5 Although ILP, in which high doses of chemotherapy are regionally delivered, is a technically demanding operative procedure, amputation necessitated by peri- or postoperative surgical complications occurs in <1% of cases.4 On the other hand, patients undergoing an intentionally limb-saving treatment with ILP still have a short- or long-term risk up to 40% of loss of the affected limb due to massive necrosis of the tumor and overlying skin resulting in a soft tissue deficit, late local recurrence, or critical leg ischemia.5
In ILP, cytotoxic agents are used, which cause a local toxic effect in the perfused extremity. After investigation of several perfusion agents for STS, the two-drug regimen of TNF-α and melphalan turned out to be highly effective.4 The pathological tumor microvascularization is the primary target for TNF-α, resulting in coagulative and hemorrhagic necrosis of the tumors.6 Olieman et al.7 demonstrated by angiography findings of specific destruction of tumor vessels after perfusion with TNF-α, leaving the normal vasculature unchanged. The surgical procedure of ILP, however, exposes the cannulated blood vessels to intravascular manipulation, by means of arteriotomy, venotomy, and the insertion of catheters. Consequently, a 1% to 10% risk of developing acute vascular complications, such as thrombosis at the arteriotomy site and deep-vein thrombosis (DVT), has been described in the literature.8–10 Moreover, chemotherapy and atherosclerosis risk factors could have a synergistic role on the development of arterial stenosis.11 Adjuvant EBRT leads to a high incidence of early and late morbidity. Long-term vascular side effects, in particular arterial stenosis, can be induced by radiotherapy.12–14
Because little is known with respect to the late vascular morbidity of the combined procedure of ILP and EBRT, a retrospective study was conducted to evaluate the long-term locoregional vascular side effects of ILP, followed by delayed surgical resection and adjuvant high-dose EBRT.
PATIENTS AND METHODS
From 1991 to 2003, a total of 73 patients with a median age of 54 (range, 14–80) years and with locally advanced STS underwent 77 perfusions with a combination of TNF-α and melphalan, with (n = 19) or without (n = 58) interferon gamma at the Division of Surgical Oncology of the University Medical Center Groningen, University of Groningen. The perfusion technique has been previously extensively described.15,16 Perfusion was followed by delayed local resection of the tumor remnant, and in cases of marginal or microscopically positive resection margins, high-dose postoperative EBRT was considered.
In 2005, 39 patients were still alive and in follow-up. Information about their past and current oncological and vascular status was obtained from the follow-up notes in the medical records. Seven patients were unable to participate in the study because of severe or advanced morbidity or comorbidity (n = 2) or because of nonmedical reasons (n = 5). Consequently, 32 patients, 14 men (44%) and 18 women (56%), with a median age of 47 (range, 14–71) years, were included (response rate, 82%). They completed a vascular checklist that asked questions about their medical vascular history, and they were screened for risk factors of vascular disease, such as smoking, hypertension, and diabetes. A blood sample was drawn to determine the cholesterol spectrum, which was used as an indicator of risk for atherosclerosis. All patients underwent a noninvasive arterial and venous vascular work-up that comprised duplex ultrasonography and vascular pressure measurements. The ankle-brachial index (ABI) and pulsatility index (PI), which are indications of vascular blood flow velocity, were measured to indicate the arterial status of the limb. For evaluation of veins, ultrasonography was performed to detect DVT and insufficiency. All investigations were performed by the same examiner (J.L.).
Five patients had a STS of the upper extremity (16%) treated with axillary perfusion. The remaining 27 patients had a STS located in the lower limb (84%) and were treated with iliac (n = 13, 41%), femoral (n = 5, 16%), or popliteal (n = 9, 28%) perfusion. There were 27 primary STS and 5 recurrent STS. Twenty-four patients (75%) received adjuvant 60 to 70 Gy EBRT in fractions of 2 Gy. Adjuvant systemic chemotherapy was provided to seven patients. Definite histopathological classification of tumor (one embryonal rhabdomyosarcoma and one extraossal osteosarcoma) was the reason in two patients; three patients were treated with chemotherapy because they participated in an European Organization for the Research and Treatment of Cancer trial (EORTC 62931). Two patients received chemotherapy in a palliative setting for distant metastasis. All patients were treated after informed consent was obtained according to institutional guidelines. Twelve different histological types of STS were distinguished. The pathological grade of the tumor was scored following the criteria of Coindre et al.,17 and the stage of the tumor was assigned according to the American Joint Committee on Cancer.18 Patient and tumor characteristics are summarized in Table 1.
TABLE 1.Characteristics of patientsAge (y)SexDiseaseLocalizationHistologyGradeAJCC stageLevelEBRTAmputationFollow-up (mo)18FPrimaryUpper legEmbryonal rhabdomyosarcoma33IliacNN15944FRecurrentPopliteal fossaMyxoid liposarcoma11IliacYN15243MPrimaryUpper legSynovial sarcoma33IliacYN15118MPrimaryLower legMyxoid chondrosarcoma23PoplitealYY14948FPrimaryUpper legWell differentiated liposarcoma11IliacYN14356FPrimaryProximal tibiaPNET33IliacYY13750FRecurrentFootPUS33PoplitealNN13525FPrimaryPopliteal fossaSynovial sarcoma23PoplitealYN13244MPrimaryUpper legMyxoid liposarcoma11IliacYN13124MRecurrentLower legSynovial sarcoma23PoplitealYN12937FPrimaryElbowMPNST23AxillaryYN12148MPrimaryPopliteal fossaMyxoid liposarcoma23IliacYN11163FRecurrentLower legPUS23PoplitealNY10737MPrimaryUpper legMyxoid liposarcoma11IliacYN10558MPrimaryUpper legPUS23IliacYN9871FPrimaryUpper legLeiomyosarcoma11FemoralNN9445FPrimaryUpper legPUS33IliacNN8256FPrimaryLower legPUS33PoplitealYN6163MPrimaryPopliteal fossaMyxoid chondrosarcoma3FemoralYN5037FPrimaryKneeLeiomyosarcoma11FemoralYN3628MPrimaryPopliteal fossaSynovial sarcoma23FemoralYN3657FPrimaryUpper legPUS33FemoralYN3542FPrimaryKneeSynovial sarcoma23IliacYY3247FPrimaryArmPUS33AxillaryYN3058FPrimaryLower legPUS33PoplitealYN2627MPrimaryArmEpithelioid sarcoma34AxillaryYN2671MRecurrentArmMyxoid fibrosarcoma11AxillaryNN2456FPrimaryArmExtraosseal osteosarcoma33AxillaryNN2214MPrimaryLower legPUS33PoplitealYN2065FPrimaryFootSynovial sarcoma23PoplitealNY1863MPrimaryUpper legSynovial sarcoma33IliacYN1771MPrimaryUpper legRhabdomyosarcoma33IliacYN17AJCC, American Joint Committee on Cancer; EBRT, external-beam radiotherapy; PNET, primitive neuroectodermal tumor; PUS, pleomorphic undifferentiated sarcoma; MPNST, malignant peripheral nerve sheath tumor.
The results were statistically analyzed by the Wilcoxon rank test and the Mann-Whitney U-test. Values of P ≤ .05 were considered significant. SPSS 12.0.1 for Windows statistical software was used (SPSS, Chicago, IL).
RESULTS
Before ILP, none of the patients had a history of cerebrovascular insufficiency, coronary artery disease, peripheral arterial occlusive disease, or any other vascular disease. Risk factors for cardiovascular disease, hypertension, diabetes, smoking, obesity, hypercholesterolemia, and familiar arterial occlusive disease were analyzed (Table 2).
TABLE 2.Risk factors for atherosclerosisRisk factorn%Dutch population (%)Hypertension10319.9a,bDiabetes mellitus4133.1bSmoking29.6b Smokers928 Former smokers1031BMI BMI >25.0175346.5b BMI >30.041310.9bHypercholesterolemia (total cholesterol >6.5)61914cFamiliar cardiovascular disease1444NABMI, body mass index; NA, not available.a Range, 5% to 45%.b From http://www.cbs.nl/.c From http://www.hartstichting.nl/.
Two patients underwent vascular reconstruction after en-bloc tumor resection with the artery. One patient underwent a reconstruction of the femoral artery with an infragenual femoral-popliteal bypass, and the second patient underwent a popliteal artery reconstruction with an autologous vein graft. A DVT in the lower leg was diagnosed in the postoperative period in three patients (7.7%).
Nine (23%) of the 39 alive patients underwent a lower limb amputation at different levels after ILP with or without EBRT. No amputations of the upper extremity were performed. In two patients with microscopically involved resection margins, an amputation of the affected limb was indicated because adjuvant radiotherapy was contraindicated. One patient underwent amputation due to tumor recurrence 18 months after ILP and EBRT. Amputation was performed for critical leg ischemia due to atherosclerosis in two patients at 110 and 125 months after ILP and EBRT. Because of severe obstruction of the crural vessels, reconstructive therapy was not suitable.
The other four amputations were performed for extensive postperfusion necrosis of the involved tumor site (three patients at 0, 8, and 12 months after ILP with or without EBRT) and arterial occlusion without any vascular reconstructive possibilities (one patient 15 months after ILP and EBRT).5[A1] Figure 1 illustrates treatment, outcome, and reasons for amputation. Five of the 32 patients included in this study underwent amputation and could therefore only partially undergo vascular follow-up.
FIG. 1.Flow chart illustrating treatment, outcome, and reasons for amputation. EBRT, external-beam radiotherapy; AMP, amputation; NI, not included in this study.
Arterial Vascular Disease
With a median follow-up of 88 (range, 17–159) months, no patient experienced peripheral arterial occlusive disease. With duplex ultrasonography, however, two patients showed a complete arterial occlusion at the level of cannulation (6%). One of them underwent an iliac perfusion and EBRT because of a STS of the lower leg in 1993. After 10 years, she had signs of critical leg ischemia, which required a transfemoral amputation. She is now free of pain and can walk with a well-functioning prothesis. The other patient showed an axillary occlusion 10 years after an axillary perfusion, followed by resection and EBRT at the forearm. She did not have any symptoms of arterial occlusive disease, and her arm functions normally. Duplex ultrasonography showed good collateral flow. Two other patients (6%) showed an arterial stenosis at cannulation level at 22 and 30 months after axillary perfusions, followed by EBRT at the upper arm and axilla in one and systemic chemotherapy in both patients. They had normal arm function without any signs of ischemia.
The arteries at resection level were also evaluated. One patient with a fossa poplitea STS had complete occlusion of the superficial femoral artery 6 years after ILP and EBRT, with open arteries below knee level. A femoral-popliteal bypass was considered but was not performed because of the absence of any symptoms. At the time of evaluation, the ABI was .50; the patient did not have any complaints besides compression ulcers on his feet, which responded well to conservative therapy. The patient who needed an autologous vein graft of the popliteal artery has a stenosis at this level without clinical signs of arterial occlusive disease.
Twenty-two of the total 32 patients included underwent ABI measurements; 22 had their PI determined at the femoral level and 20 at the popliteal level. Five patients were excluded because of amputation of the limb; another five were excluded because of a STS located in the upper limb. Of the remaining 22 patients, 18 received postoperative EBRT for the above-described indications.
In 21 of 22 patients, the ABI (normal, ≥1.00) in the treated leg was .91 or more. As mentioned above, one patient had an ABI of .50 without any complaints. ABI measurements in the involved leg (median, 1.02; range, .50–1.20) showed a significant decrease compared with the contralateral leg (median, 1.09; range, .91–1.36; P = .001). Seven patients showed a decreased femoral PI (normal range, 5–10). Six of them underwent an iliac perfusion (86%). Popliteal PI (normal range, 6–12) was decreased in six patients. Perfusion at a more proximal level was performed in four of these patients (67%). When compared with the contralateral leg, PI was lower in the treated leg in 17 of 22 patients at femoral level (median, 6.30; range, 2.1–23.9 vs. median, 7.35; range, 4.8–21.9; P = .011) and in 19 of 20 patients at popliteal level (median, 8.35; range, 0–21.4, vs. median, 1.95; range, 8.0–32.6; P < .0005) (Table 3). There were no differences in objective vascular side effects between men and women, between patients who did or did not undergo EBRT, or between patients <50 years old versus older patients (Table 4). In patients with a follow-up period of >5 years, there was more often a decrease in ABI, compared with patients with shorter follow-up (13 of 14 patients vs. 4 of 8 patients, P = .024). With a longer follow-up period, there was also more often a decrease in PI at the femoral level (13 of 14 patients vs. 3 of 7 patients, P = .024), but not at the popliteal level (12 of 12 patients vs. 7 of 8 patients, P = .221) (Table 5). No patient required vascular intervention during follow-up.
TABLE 3.Results of ankle-brachial index (ABI) and pulsatility index (PI) measurements in lower extremities of 22 patientsIndexTreated extremity, median (range)Contralateral extremity, median (range)P valueABI1.02 (.50–1.20)1.090 (.91–1.36).001PI femoral6.30 (2.1–23.9)7.35 (4.8–21.9).011PI popliteal8.35 (.0–21.4)10.95 (8.0–32.6)<.0005TABLE 4. P values of differences between subgroups in decrease of ankle-brachial index (ABI) and pulsatility index (PI) for 22 patientsCharacteristicSexFollow-up after EBRTAgeFMYesNo>5 y<5 y<50 y>50 yNo. of patients1111184148139ABI.619.241.024.963––––PI femoral decrease.619.907.024.963––––PI popliteal decrease.269.937.221.414––––EBRT, external-beam radiotherapy.TABLE 5.Differences in decreases in measurements according to length of follow-upDecrease in perfused legFollow-up, n (%)P value>5 y (n = 14)<5 y (n = 8)ABI13 (93%)4 (50%).024PI femoral13 (93%)4 (50%).024PI popliteal12 (100%)a7 (88%).221ABI, ankle-brachial index; PI, pulsatility index.a n = 12; data for two patients were not available.
Venous Vascular Disease
Many patients had some symptoms of venous insufficiency, such as lower limb pain, feelings of heaviness and discomfort, night palsies, paresthesia, and edema. One patient (3%) wore compression stockings because of superficial vena saphena magna insufficiency, and one patient (3%) had unilateral superficial varicosis in the treated leg. Another three patients (9%) showed uni- or bilateral deep insufficient venous vasculature by duplex ultrasonography, which in all cases was located at the level of resection. One DVT (3%) and one partial venous obstruction (3%) were found, both at the resection level. Venous vascular obstruction was not seen at the level of cannulation.
DISCUSSION
ILP with TNF-α and melphalan, with or without adjuvant EBRT, has become an approved treatment strategy for locally advanced STS of the extremities after the publication of the results of an European multicenter trial performed in the 1990s.4 Because our center was one of the first to use this treatment modality, we now have had more than 10 years’ experience of ILP used to treat STS. This gives us the opportunity to evaluate the long-term treatment-related morbidity. Critical leg ischemia approximately 10 years after ILP, necessitating amputation in two patients, led to an interest of analyzing the vascular complications of ILP with or without EBRT.5
Despite a complete destruction of tumor vasculature, ILP does not seem to have any effect on the macrovasculature of the perfused limb.7,19,20 However, because atherosclerosis risk factors and chemotherapy could play a synergistic role in the progression of arterial stenosis, patients who already have severe atherosclerotic disease are less suitable for ILP.11,19,20
In our study, many patients had one or more risk factors for atherosclerosis. In the study population, there seemed to be more patients with diabetes or hypercholesterolemia than in the standard Dutch population. The smoking rate is similar (28% vs. 30%); however, another 31% used to smoke (Table 2).
Acute vascular complications after ILP have been described in the literature. Deep-vein thrombosis (DVT) occurs in 1.7% to 10% of cases in the immediate postoperative period.10 In our studies, three patients (7.7%) had a successfully treated DVT after ILP. DVT seems to be extremely difficult to diagnose because the postoperative effects of ILP itself—a warm, swollen edematous limb—mask the classical symptoms. To prevent this undesired complication, anticoagulant therapy is recommended. Thrombosis at the arteriotomy site has also been mentioned in earlier studies. It was not encountered in this study but should be treated by prompt trombectomy.9,21 In case of repeated ILP in the same limb, perfusion through more proximal arterial and venous sites is recommended because fibrosis at the previous operative site does not allow insertion of the cannulas at the same level. All patients in this study underwent only one ILP.
Long-term effects after ILP are mainly functional and consist of edema, stiffness, functional impairment, and muscle atrophy.21–23 Late vascular morbidity after ILP without EBRT has not been previously mentioned in other studies. In our study, one partial arterial occlusion (3%) was found. In contrast, long-term effects of EBRT have extensively been described in the literature. In a study of Butler et al.,12 three types of vascular damage have been described. At first, thrombosis occurs within 5 years after EBRT; a second fibrotic occlusion is seen within 10 years; and finally, a predisposition to the development of atheroma together with periarterial fibrosis is associated with a latent interval of 20 or more years. Rijbroek et al.24 mention claudication after a latent period. This is masked by the other effects such as pain, functional impairment, and edema. According to Kalman et al.,13 long-term vascular side effects of EBRT resemble the natural process of atherosclerosis, which is a combination of direct damage, periarterial fibrosis, and occlusion of vasa vasorum. The EBRT accelerates this process, sometimes 10 to 27 years after treatment. Little is known about the combined effects of ILP and EBRT. Vrouenraets et al.22 described one brachial arterial occlusion, but this was only 4 months after treatment. Olieman et al.16 described the occurrence of necrosis after ILP but did not find any increased tumor-related morbidity after ILP followed by EBRT. In another series, ILP was performed after EBRT, but it did not lead to more local toxicity or complications.25
In our study, amputation was performed in nine patients (23%). This is higher than the 9% to 21% described in the literature.4,26,27 However, amputations due to long-term complications like critical leg ischemia, which was performed in two patients (5.1%), have not been described in other studies. Furthermore, three complete (9%) and one partial (3%) arterial occlusion were found after the combined procedure.
Fortunately, these patients did not have invalidating symptoms of arterial occlusive disease. In most cases, vascular pressure measurements and duplex ultrasonographic evaluation showed statistically lower values compared with the contralateral extremity, but these were still within the normal range. This could explain the absence of symptoms.
The PI value normally increases progressively from aorta to ankle. A decrease in PI indicates stenosis in a more proximal vascular segment.28 In our study, six of seven patients with an abnormal femoral PI underwent iliac perfusion. This implies a vascular reaction at the level of cannulation. Furthermore, seven patients showed a decrease in popliteal PI compared with the femoral PI, which implies that vascular damage is present at a more proximal (i.e., cannulation or resection) level. However, with the present data, it is not possible to support this thesis because of the relatively small study population, and the absence of symptoms and other abnormal outcomes of the vascular work-up.
It is important to realize that blood vessels lose their elasticity with the years and are subject to the normal process of atherosclerosis. Furthermore, unilateral peripheral arterial occlusive disease is common. However, the fact that lower rates of ABI and PI were found in the treated leg in almost all patients and that no marked difference was found between younger and older patients makes the assumption of treatment-induced differences plausible.
Because only four patients with a lower leg STS were not treated with adjuvant EBRT, there are still no clinically relevant data about the late vascular damage of the single procedure of ILP. Thijssens et al.29 showed that adjuvant EBRT seems beneficial in all patients, so it might be impossible to evaluate the sole effects of ILP on the vascular system in the future.
Major late vascular complications after the combined procedure, like critical leg ischemia necessitating amputation, were found in two patients in this study. One patient had no risk factors for peripheral arterial occlusive disease besides smoking. He had had intermittent claudication 8 years after ILP and EBRT, and a normal ABI of 1.09. His condition did not improve after 2 years of conservative therapy and reconstruction was impossible, so he underwent an amputation 125 months after the initial limb salvage treatment strategy. The second patient, who had diabetes mellitus and familial cardiovascular disease, experienced no complications the first few years after ILP and EBRT. After 6 years, she experienced a pathological proximal tibia fracture due to radiation-induced osteonecrosis in the previous surgically treated area and needed to wear a brace. She lived for 3 years with a disabling function impairment of the knee and then developed signs of critical leg ischemia. An upper leg amputation was performed 110 months after the initial treatment. She now walks with a well-functioning prothesis.
The importance of leg preservation has been mentioned before. Even though most patients still are satisfied with the initial limb-saving treatment that they underwent, amputations due to major late complications with severe morbidity make the rationale of limb salvage worth reconsidering.
Recently, one patient was successfully treated with thrombectomy for an acute popliteal artery occlusion. This patient, who underwent an ILP at femoral level followed by EBRT for a STS in the fossa poplitea 5 years before, had no complaints of peripheral arterial occlusive disease during his vascular work-up for this study 10 months earlier. ABI (1.03) and femoral PI (6.3) were normal; popliteal PI was low (5.3). With duplex ultrasonography, some atherosclerosis in the femoropopliteal section was seen. This patient has diabetes and hypertension, is obese (body mass index >30), and was a former smoker, so he was especially at risk for developing atherosclerosis.
This study shows that despite the importance of preventing major late vascular complications after the combined procedure of ILP and EBRT, a routine noninvasive vascular work-up does not seem to add value to normal follow-up. It is more important to realize that patients after the limb salvage treatment strategy of ILP, followed by delayed tumor resection and EBRT, are at risk for late vascular complications. This risk is substantially raised when personal risk factors such as smoking, diabetes, hypertension, obesity, and hypercholesterolemia are present. Because symptoms of vascular complications can be masked by the normal effects of the treatment itself, the possibility of vascular morbidity should be considered if the patient has any complaint. Furthermore, all patients must be convinced of the importance of reducing their additional personal risk factors for atherosclerosis.
In conclusion, objective measurements show a time-related decrease (>5 years’ follow-up) of ABI and femoral PI in the treated leg, usually without subjective complaints. ILP followed by delayed resection and EBRT for a locally advanced STS can lead to major late vascular morbidity that requires amputation.
|
[
"vascular",
"perfusion",
"sarcoma",
"complications",
"radiation"
] |
[
"P",
"P",
"P",
"P",
"U"
] |
J_Occup_Rehabil-3-1-1915628
|
Effects of Ambulant Myofeedback Training and Ergonomic Counselling in Female Computer Workers with Work-Related Neck-Shoulder Complaints: A Randomized Controlled Trial
|
Objective: To investigate the effects of ambulant myofeedback training including ergonomic counselling (Mfb) and ergonomic counselling alone (EC), on work-related neck-shoulder pain and disability. Methods: Seventy-nine female computer workers reporting neck-shoulder complaints were randomly assigned to Mfb or EC and received four weeks of intervention. Pain intensity in neck, shoulders, and upper back, and pain disability, were measured at baseline, immediately after intervention, and at three and six months follow-up. Results: Pain intensity and disability had significantly decreased immediately after four weeks Mfb or EC, and the effects remained at follow up. No differences were observed between the Mfb and EC group for outcome and subjects in both intervention groups showed comparable chances for improvement in pain intensity and disability. Conclusions: Pain intensity and disability significantly reduced after both interventions and this effect remained at follow-up. No differences were observed between the two intervention groups.
Introduction
Work-related musculoskeletal complaints in the upper extremity are common among workers in Western industrialised countries. In the Netherlands, about 15% of the working population report complaints in the neck, shoulders and arms [1] defined as pain, numbness, or tingling, resulting in loss of productivity, sick-leave or even disability. In 1998, about 8% of all Dutch employees were absent from work due to work-related musculoskeletal complaints, and 2% of the employee population was absent from work for more than four weeks [2]. These complaints impose a substantial economic burden in compensation costs, lost wages, and productivity. Apart from the individual suffering, the financial costs within member states of the European Union associated with musculoskeletal complaints are high [3].
Work-related musculoskeletal complaints are multifactorial in origin and involve biomechanical, psychosocial, and individual components [4–6]. As a consequence different intervention approaches exist. Most often interventions address alterations of the physical work environment by adjustments of the work station and/or education about working posture according to ergonomic principles [7–10]. There are several studies reporting on the effectiveness of ergonomic approaches [11, 12] and it is commonly accepted that properly designed work stations are prerequisites for healthy working. However, in spite of attention to ergonomics musculoskeletal complaints remain a considerable problem [13–15] and this is the rationale for the search for innovative interventions.
A relatively new intervention approach addressing neck-shoulder complaints is myofeedback training based on the Cinderella-hypothesis [16]. The Cinderella-hypothesis is one of the most influential hypotheses explaining the process of development and persistence of pain in low intensity jobs [17] like computer work, and states that lack of sufficient muscle relaxation is a crucial factor in this process. Continuous muscle activity, even at low intensity levels, may result in homeostatic disturbances of the activated motor units due to affected blood flow and removal of metabolites [17]. Several studies have found an association between absence of moments of complete muscle rest and myalgia, especially for the commonly affected descending part of the trapezius muscle (e.g. [18–20]). Warning subjects when their muscle relaxation is insufficient could thus contribute to recovery, and this is the rationale for the Cinderella-based myofeedback training. This approach is different from traditional myofeedback training in which feedback is provided when muscle activity exceeds a certain level thereby aiming at decreasing muscle activation (e.g. [21, 22]). As the Cinderella-hypothesis suggests however that muscle relaxation is more relevant than muscle activation, a myofeedback-based intervention aiming at increasing muscle relaxation may be more beneficial.
The Cinderella-based myofeedback system [23] consists of a harness incorporating dry surface electrodes and a feedback unit. The system is ambulant and subjects can wear the harness under their clothes during working days. The harness is connected to a feedback unit worn at the waist which provides feedback by means of vibration and a soft sound when muscle relaxation is insufficient. A feasibility study using this equipment for four weeks in a group reporting work-related neck-shoulder complaints resulted in significantly decreased levels of pain intensity and also changed muscle activation patterns. These effects were still at hand four weeks after the myofeedback training ended [23]. However, as the study did not contain a control group and only investigated the lasting effects of the intervention for a short follow-up period (4 weeks), further research is needed.
As proper ergonomics are indispensable for healthy working, the myofeedback training is preferably applied in combination with an intervention approach aimed at improving ergonomics. Thus, the present study aimed at investigating the immediate and lasting effects of Cinderella-based myofeedback training including ergonomic counselling, compared to ergonomic counselling alone, on pain intensity and disability in females with work-related neck-shoulder complaints. As interventions focusing on multiple factors have shown to be related to decreased incidence of complaints [24] it was hypothesised that 4-weeks myofeedback training including ergonomic counselling would be more effective in reducing pain intensity and disability than the intervention based on ergonomic counselling alone.
Methods and materials
Design and subjects
A randomized controlled trial was performed to assess the effect of 4-weeks Cinderella-based myofeedback training combined with ergonomic counselling (together referred to as Mfb), compared to ergonomic counselling alone (EC), on pain intensity and disability. Measurements were performed prior to intervention but before randomisation (Baseline), immediately after four weeks of intervention (T0), and at three (T3) and six (T6) months after the intervention.
Participants were recruited in Sweden (area of Göteborg) and the Netherlands (area of Enschede) between March 2003 and June 2005. Computer workers like job counselors (Sweden) and (medical) secretaries (Sweden and the Netherlands) were approached by telephone and announcements, and volunteers were subsequently sent a screening questionnaire [25] which was developed within the EU-funded NEW project (Neuromuscular assessment in the Elderly Worker) [26]. Subjects eligible for participation were symptomatic female computer workers, predominantly over the age of 45 as the prevalence of complaints is especially high in this age category [27], working for at least 20 hours a week, and reporting perpetuating work-related musculoskeletal complaints in the neck and/or shoulder region for at least 30 days during the last year.
Subjects were excluded when they reported pain in more than 3 body regions, when they suffered from severe arthrosis or joint disorders, when they were using muscle relaxants, or when reporting other complaints in the upper extremity not related to (computer) work.
Power calculation, based on the results of Hermens and Hutten [23], indicated that at least 35 subjects should be included in each intervention group (estimated proportion of subjects showing an effect on pain intensity was set at 0.6 in the Mfb group and 0.2 in the EC group; 1−β = 0.90; α = .05). Block randomization was used to assign subjects to either Mfb or EC: When a new group of subjects started the intervention, half of them were assigned to Mfb and half of them were assigned to EC. The study was approved by the local Medical Ethics Committees and all participants gave their informed consent prior to participation.
Interventions
The interventions were provided by three different therapists: One physiotherapist in Sweden and two health scientists in the Netherlands. They were thoroughly trained and practiced together prior to the start of the study to ensure that they would provide as identical interventions as possible. The character of the intervention made blinding of the therapists and the subjects to the intervention impractical. To prevent from information bias, subjects were informed that the aim of the current study was to compare the effects of two interventions and that there was no evidence favouring one of these interventions.
Ergonomic counselling (EC)
Subjects received four weeks of intervention during which they kept a diary of activities and pain intensity scores. During the four weeks period they were visited weekly by their therapist. The first visit comprised an ergonomic workplace investigation by means of the risk inventory of Huppes et al. [28]. This checklist contains questions to evaluate work tasks, working hours, work load, work station, and working methods. Based on the outcome, possible improvements were discussed with the subject. With regard to the work station, the focus was primarily on modifying the existing work station rather than providing new equipment. The remaining visits were used to further discuss the ergonomic aspects, the consequences of possible ergonomic adjustments etc., according to a manual to guarantee a uniform intervention.
Myofeedback (Mfb)
Subjects randomized to the Mfb group received Cinderella-based myofeedback training on top of EC. A two-channel ambulant myofeedback system combined with a harness incorporating dry surface Electromyography (sEMG) electrodes to enable a stable recording of upper trapezius muscle activity was used. The harness was connected to a sEMG processing and storage device (see Picture 1).
Picture 1
Myofeedback system: Harness and processing/storage unit
The sEMG signal was amplified (15×), band pass filtered between 30 and 250 Hz, sampled at 512 Hz, digitized (22 bits ADC), and smooth rectified with removal of the low frequency components. Embedded software provided the detection and calculation of muscle rest, expressed as the sEMG parameter Relative Rest Time, which was defined as the percentage of time in which Root Mean Square was below a threshold (10 μV) for at least 0.12 s. This threshold was based on the noise level of the myofeedback system including mounted electrodes at the skin. Sensory feedback by means of vibration and a soft sound was provided after each 10 s interval when the relative duration of muscle relaxation in that particular interval was below 20%. The choice for a 10 s interval was based on the results of Voerman et al. [29], and the 20% threshold was chosen based on the work of Hägg and Åström [30]. Subjects were instructed that they should respond to the feedback by relaxation, which could be reached by slightly depressing the shoulders, or by sitting down quietly with the eyes closed, the hands in the lap while breathing deeply. Another relaxation strategy subjects were instructed was to maximally elevate the shoulders for three seconds to build up muscle tension and then to let loose this tension. This relaxation was anticipated to contribute to recovery of the Cinderella motor units. When there was no adequate response to the feedback, i.e. relaxation, the duration of the feedback signal progressively increased.
Immediately after Baseline, subjects assigned to the Mfb group were given the myofeedback device and they were explained the working mechanism and background of the myofeedback training. Subjects wore the system for four weeks, for at least eight hours a week (distributed over two hours a day and two days a week as a minimum) while performing their regular work. During the weekly visits the sEMG-data from last week were scrutinized and discussed to give the subject insight in their relaxation patterns and to identify possible situations of concern. This procedure was facilitated by means of the diary.
Assessment of pain intensity and disability
Pain intensity in the neck, shoulder (left and right), and upper back at time of the measurement was assessed by means of four Visual Analogue Scales [31]. Subjects were instructed to rate their subjectively experienced level of pain intensity at that moment for each body region. The Visual Analogue Scale consists of a 10 cm horizontal line with ‘no discomfort at all' at the left and ‘as much discomfort as possible' at the right endpoints of the line. Psychometric properties of the Visual Analogue Scale have been shown to be sufficient [32–34].
The level of subjectively experienced disability was assessed with the Pain Disability Index, a self-rating scale that measures the impact of pain on the abilities to participate in life activities [35]. The Pain Disability Index contains 7 items, one for each domain, i.e. (1) family and home responsibilities, (2) recreation (hobbies, sports, and leisure time activities), (3) social activity (participation with friends and acquaintances), (4) occupation (activities partly or directly related to working), (5) sexual behavior (frequency and quality of sex life), (6) self care (personal maintenance and independent daily living), and (7) life-support activity (basic life-supporting behaviors). Answers were provided on a categorical 11-points scale with ‘not disabled’ and ‘fully disabled’ at the extremes. In a chronic pain population, psychometric properties of the Pain Disability Index appeared to be sufficient [36].
Analysis
Data inspection showed non-normal distributions for all of the variables, and data were therefore log transformed, enabling parametric analyses. Analyses were performed both on the group as well as the individual level.
For each subject, VAS scores for the neck, left shoulder, right shoulder, and upper back were summed and averaged resulting in a combined neck-shoulder pain intensity score. Differences between the two interventions, i.e. Mfb and EC, for VAS and the Pain Disability Index were investigated using a linear mixed-model analysis technique for repeated measurements. The following variables were included in the model as fixed factors: Time of measurement (i.e. Baseline, T0, T3, and T6), intervention type (i.e. Mfb and EC), and study group (i.e. Sweden and the Netherlands), including two-way interactions. The factor study group comprises variance due to possible (socio)demographic differences as well as variance caused by the different therapists, organizations, and job characteristics in the two countries. The factor subject was included in the model as a random factor. Paired and independent samples t-tests were used for Post Hoc analysis.
At the individual level the percentages of subjects improving and the odds for improvement in both intervention groups were investigated. First, individual differences were calculated between Baseline and T0, Baseline and T3, and Baseline and T6 for pain intensity in the neck-shoulder region (i.e. the averaged score on the VAS for neck, left and right shoulders, and upper back) and Pain Disability Index. These differences were then dichotomized into ‘clinically relevant improvement’ or ‘equal/deteriorated.’ For the Visual Analogue Scale the cut-off point for a clinically relevant improvement was defined at 13 mm which corresponds to the upper limit of the confidence interval of clinically relevant changes in Visual Analogue Scale scores in acute and non-acute pain patients as reported by Kelly in 2001 [37]. This cut-off point was also the result of the study of Todd [34]. Cut-off points for clinically significant changes in the Pain Disability Index scores have not been provided in literature. For a comparable measure, the Neck Disability Index which is a 10 items scale with 6 answering options resulting in scores varying from 0 to 50, the cut-off point was set at 5 which is 10% of the maximum score [38]. Using this as a starting point, a clinically relevant change was defined as a change of ≥10% of the maximum score of 70 of the Pain Disability Index, i.e. ≥7 units of the Pain Disability Index score.
Logistic regression analyses providing Odds Ratios were performed to investigate whether the two intervention groups differed in terms of chances for clinically relevant improvements in pain intensity and disability. Three different models were investigated: Model I represents the crude Odds Ratio describing only the relationship between intervention type and odds for improvement without adjustment for confounding factors. The factors study group and baseline pain intensity/baseline disability were assumed potentially confounding factors. To correct for these factors, two additional models were built as extensions of Model I: Model II incorporates study group as confounding factor, and Model III is an extension of Model II incorporating also the factor baseline pain intensity/baseline disability level. For Models II and III interaction effects were additionally included in the analysis, to study effect modification, but these remained only included when significant. −2 Log Likelihood tests were used to select the best model. Odds Ratios including 95% Confidence Intervals were calculated and presented for each of the models.
Statistical Package for Social Sciences 11.5 was used for statistical testing and alpha was set at .05 for statistical significance.
Results
Subjects
Seventy-nine female computer workers with neck-shoulder complaints were included in this study: Forty-two subjects were assigned to the Mfb group and 37 to the EC group. Mean age was 52.0 (SD 5.8 years) in the Mfb group and 50.7 (SD 5.5) years in the EC group. Mean duration (days between Baseline and T0) of the interventions was 37 (SD 8) days for the Mfb group and 36 (SD 7) days for the EC group. Forty-one out of the 79 subjects were recruited in Sweden and the remaining 38 in the Netherlands. The number of subjects at Baseline, T0, T3, and T6 and the number of drop-outs are shown in Diagram 1.
Diagram 1Flow chart subject recruitment and randomisation
Drop-outs did not differ in age, weight, height, BMI, and pain intensity scores (Visual Analogue Scale), and disability (Pain Disability Index) from those fulfilling the intervention (p > .08).
Table 1 provides an overview of sociodemographic characteristics of the subjects in the Mfb and EC groups. Thirty-eight subjects reported complaints in both the neck and shoulder, while 41 subjects reported complaints either in the neck or in the shoulder at the time of recruitment. Comparison of the characteristics between the Mfb and EC groups indicated that subjects assigned to the Mfb group reported more years within the same job compared to the EC group (p < .05).Table 1Characteristics of subject populationMfb (n = 42)EC (n = 37)Sociodemographics Body mass index25.2 (3.9)25.2 (3.7) Side dominance% Right-handed95%97% Living situation% Living alone16.710.8 Working hours per week32.8 (7.8)32.8 (8.3) Working hours per week since (in years)17.0 (11.4)*12.0 (8.9)*Complaints Trouble in neck last yearYes92.991.9 Trouble shoulders last yearYes, in both shoulders36.632.4Yes, in the right shoulder36.643.2Yes, in the left shoulder12.213.5 Trouble in upper back last yearYes66.748.6 Was work performance affectedYes31.718.9*Significant at the .05 level.
Analysis at group level
Pain intensity
Baseline (geometric) mean VAS scores for each region separately were 27 mm (neck), 15 mm (left shoulder), 14 mm (right shoulder), and 13 mm (upper back) in the Mfb group and 24, 14, 19, and 18 mm in the EC group respectively.
Figure 1 shows a box plot of the VAS score in the neck-shoulder region at Baseline, T0, T3, and T6 for the Mfb and the EC group. A clear decrease was observed at T0, T3, and T6 compared to Baseline, although at T6 pain intensity seemed to increase slightly in both groups.Fig. 1Box plot for averaged VAS score of pain intensity in the neck-shoulder region at Baseline, T0, T3, and T6 for both the Mfb and the EC group
Mixed linear modelling showed that pain intensity in the neck-shoulder region significantly changed over time (F = 12.08, p ≤ .01), without additional effects for the type of the intervention (F = 1.54, p = 0.22), study group (F = .48, p = .49), or interaction effects (F ≤ .87, p ≥ .35). Post Hoc comparisons revealed that the VAS score was significantly reduced at T0 (t = 4.37, p < .01), T3 (t = 5.10, p < .01), and T6 (t = 3.54, p < .01) compared to Baseline but also the reduction between T0 and T3 was significant (t = 2.85, p = .01).
Disability
Figure 2 shows a box plot of the Pain Disability Index score at Baseline, T0, T3, and T6 for the Mfb and the EC group. A comparable pattern to what was observed for pain intensity was found, with decreased disability levels at T0, T3, and T6 compared to Baseline.Fig. 2Box plot for the Pain Disability Index score at Baseline, T0, T3, and T6 for both the Mfb and the EC group
Disability levels significantly changed over time (F = 17.68, p < .01) and were significantly different between the two study groups (i.e. Sweden and the Netherlands) (F = 5.30, p = .02). No additional effects were found for intervention type (F = .86, p = .35) nor the interaction terms (F ≤ 1.97, p ≥ .12). Post Hoc comparisons showed that subjects in the Swedish study group had lower Baseline values and reported reductions in disability only between Baseline and T0 (t = 2.20, p = .04) with a trend for reductions at T3 (t = 1.89, p = .07) and T6 (t = 1.81, p = .08). Subjects in the Dutch study group showed a significant decrease at T0 (t = 3.26, p < .01), T3 (t = 3.58, p < .01), and T6 (t = 3.51, p < .01) compared to Baseline.
Analysis at the individual level
Pain intensity
Figure 3 gives an overview of the percentage of subjects in the Mfb and EC groups showing an improvement in pain intensity in the neck-shoulder region. About half of the subjects showed a clinically relevant improvement in pain intensity in the neck-shoulder region. Slightly more subjects in the Mfb group showed improvements compared to the subjects in the EC group immediately after the intervention period (T0) and at 6 months follow-up (T6).Fig. 3Percentage of subjects showing improvements in averaged VAS score of pain intensity in the neck-shoulder region at T0, T3, and T6 compared to Baseline for the Mfb and EC group [without correcting for confounding factors]
Crude Odds Ratios (Model I) for improvement showed higher odds for clinically relevant improvement in the Mfb group at T0 and T6 compared to Baseline, while between T3 and Baseline subjects assigned to EC were more likely to show improvements (see Table 2). However, Odds Ratios were not significant (p ≥ .36) also not when corrected for potential confounding factors (p > .19). −2 Log Likelihood tests indicated that Model III, adjusting for the factors study group and baseline pain intensity/disability level, superimposed the Models I and II (−2 LL; p < .05). As no significant interaction effects were found (p > .08) these were excluded from the final models as presented in Table 2.Table 2(Adjusted) odds ratios (95% confidence interval) for improvement in pain intensity in the neck-shoulder region for the Mfb group compared to the EC groupModel IModel IIModel IIIVAS1.561.593.42Baseline–T0(.60–4.05)(.71–4.14)(.55–21.11)VAS.78.78.40Baseline–T3(.29–2.08)(.29–2.09)(.08–1.99)VAS1.311.311.04Baseline–T6(.47–3.60)(.47–3.64)(.29–3.77)Model I: Odds Ratio Crude.Model II: Odds Ratio Adjusted for factor study group.Model III: Odds Ratio Adjusted for factor study group and VAS at Baseline.
Disability
An overview of the percentage of subjects in the Mfb and EC groups showing an improvement in disability is provided in Fig. 4.Fig. 4Percentage of subjects showing clinically relevant improvements in Pain Disability Index at T0, T3, and T6 compared to Baseline for the Mfb and EC group [without correcting for confounding factors]
Immediately after the intervention period about twice as many of the subjects in the Mfb group showed clinically relevant improvements in disability compared to the EC group. This share of subjects increased somewhat in both groups after three months and then showed a small decline after six months where about one third of the subjects of the EC group and half of the Mfb group showed clinically relevant improvements in disability.
Subjects assigned to the Mfb group had higher odds for improvement in disability at T0, T3, and T6: At T3, the odds for improvement in disability was 2.77 times higher in the Mfb group compared to the EC group which was significant (Model I; 95% CI 1.00–7.65; p = .05). When corrected for confounding variables, however, Odds Ratios for improvement were still higher for the Mfb group compared to the EC group, but this was not significant (p ≥ .057) except for Model II for changes between Baseline and T3 (p = .04). Baseline disability levels significantly affected Odds Ratios: Model III was generally better than the models I and II (−2 LL; p < .05). As no significant interaction effects were found (p > .08) these were excluded from the final models as presented in Table 3.Table 3(Adjusted) odds ratios (95% confidence interval) for improvement in disability the Mfb group compared to the EC groupModel IModel IIModel IIIPain disability index2.702.721.48Baseline–T0(.97–7.54)†(.971–7.60)†(.39–5.62)Pain disability index2.772.991.64Baseline–T3(1.00–7.65)*(1.03–8.65)*(.34–7.97)Pain disability index2.542.611.48Baseline–T6(.88–6.82)†(.91–7.47)(.37–5.88)Model I: Odds Ratio Crude.Model II: Odds Ratio Adjusted for factor study group.Model III: Odds Ratio Adjusted for factor study group and Pain Disability Index at Baseline.*p ≤ .05.†.05 ≤ p ≤ .10.
Discussion
This randomised controlled trial investigated the effects of 4-weeks ambulant myofeedback training combined with ergonomic counselling in subjects with work-related neck-shoulder complaints relative to ergonomic counselling alone, on pain intensity and disability. The effects were evaluated immediately after the interventions, and at three and six months follow-up. Mean pain intensity and disability levels significantly reduced after both interventions (i.e. Mfb and EC). The effects were clinically relevant in a large part of the subjects: About 30–50% of the subjects showed clinically relevant improvements in pain intensity and/or disability. No difference was found for the effect of the intervention on outcome and after correction for confounding factors subjects in both intervention groups did not differ with regard to chances for clinically relevant improvements in pain intensity and disability.
Several studies have shown the beneficial effect of myofeedback training on pain reduction [21, 39–43] although Faucett and colleagues [22] reported changes in muscle activity rather than changes in pain intensity. The feedback approach used in these studies is different from the Cinderella-based myofeedback approach in that the traditional feedback method provides feedback when muscle activation is too high [21, 22, 42, 44–47], rather than when the time the muscle has relaxed is too short [23]. Two previous studies applying the Cinderella-based myofeedback training in subjects with work-related complaints [23] and subjects with a whiplash associated disorder [48] reported reduced pain intensity and disability levels. Compared to these studies, baseline pain intensity values were generally lower and the cut-off points for clinically relevant changes were set to higher levels in the current study, but the results indicate a comparable percentage of subjects reporting clinically significant reductions in pain intensity, i.e. between 35 and 50% of the subjects. This consistency in results adds to evidence for the effectiveness of (Cinderella-based) myofeedback training on pain intensity and disability in musculoskeletal neck-shoulder complaints.
In line with existing literature (e.g. [8, 10, 49, 50]) also subjects in the ergonomic counselling group reported reduced pain intensity and disability. It was however hypothesised that a treatment approach including both myofeedback training and ergonomic counselling (Mfb) would be more effective than treatment comprising only ergonomic counselling (EC) as interventions focusing on multiple factors have shown to be related to decreased incidence of complaints [24]. In the current study the two types of interventions did not differ in terms of outcome. This is in line with findings of other studies (e.g. [8, 10, 51–53]) that concluded that occupational interventions have generally comparable effects, although these studies did not include a myofeedback intervention. Newton-John and others [43] compared myofeedback with cognitive-behavioural therapy and a waiting list control group and found that both interventions showed favourable outcome in terms of pain intensity and disability compared to the waiting list control group without a difference between the two interventions.
There are possible explanations for the absence of differences between the two groups. One concerns the presence of subgroups in which the intervention is beneficial. Results showed that the effect is clinically relevant in about 30–50% of the subjects. Main question to be addressed here is whether, and how, these subjects can be characterised in terms of cognitive-behavioural characteristics and whether this characterisation can contribute to predict outcome of the intervention. This could substantially improve the efficiency and effectiveness of the interventions. Further, differences between the two groups may have been masked by using rather generic outcome measures. For instance, coping and patient-rated parameters may better represent the difference in outcome between different kinds of interventions [54]. Investigation into the specific working mechanisms of both interventions could clarify this. Finally, initial VAS and disability levels were low in both the Mfb and the EC group, especially in the study group from Sweden. This results in a smaller potential for improvement (floor effect) and as a result a smaller chance for finding differences between the Mfb and EC groups.
Methodological considerations
This study did not include a non-intervention or placebo control group (for instance randomly administered feedback) which makes it hard to control for non-specific effects like regression to the mean or the Hawthorne effect [55]. This effect was already described in 1933 by Mayo, and is reported as a significant positive effect without causal basis in the theoretical motivation for the intervention, but is related to the effect on the participants knowing themselves to be studied in connection with the outcomes measured. An argument against such effect is that although it might occur in connection with the intervention (i.e. at T0) it is not likely that this effect would remain [56] after three to six months. In addition, the study population also contained subjects with chronic complaints (i.e. duration >6 months) who received a variety of treatments in the past, which were not successful. It is likely to expect that any non-specific effects of treatment would already have occurred during past treatments and that this effect in the current study would thus be small. Furthermore, results from previous studies learned that pain reduction in myofeedback-trained subjects were higher compared to a wait-list control group [43], a no intervention group [46], or subjects receiving a placebo [57]. Furthermore, an attention-only group showed no changed outcome in terms of disability and pain intensity [58].
The present study included a rather specific subject population: Participants were females, predominantly over the age of 45, still at work, and characterised by relatively low pain intensity and disability levels. Interpretation and extrapolation of results to other populations therefore requires caution, but as comparable effects of Cinderella-based myofeedback training have also been shown in mixed, younger subject populations [23, 48] and in a sample of patients who were on sick leave because of their neck-shoulder complaints [48], generalisation of findings might be legitimate. The subjects were selected based on self-reported complaints rather than a clinical evaluation. In a previous study [60] with female computer users above the age of 45, applying the same inclusion criteria as the present study, it was found that in subjects with neck-shoulder complaints the following clinical diagnoses were most prevalent: Trapezius myalgia (38%), tension neck syndrome (17%), and cervicalgia (17%). These diagnoses were found in 60% of the subjects reporting complaints. There were cases with supraspinatus tendinitis, frozen shoulder, and biceps tendinitis, but these were less common. This general pattern of clinical signs is likely to be at hand also in the present study. It could be hypothesised that myofeedback may be particularly relevant and helpful in cases with muscular pain syndromes.
Despite extensive standardisations, the recruitment of subjects in two different study groups (Sweden and the Netherlands) resulted in heterogeneity of the subject population with regard to age, working hours, seniority, and working posture. To correct for this, the factor study group was considered a confounder needed to control for during analysis. This reduced the power of the study. An additional likely confounding factor is the compliance of the patient and therapist to the intervention. This is an often uncontrolled and thus potentially confounding factor in occupational intervention studies [59] and it is known that changes in knowledge and skills do not necessarily result in a behavioural change. In the Mfb group, the compliance was partly controlled as the number of hours the system was worn was recorded by the system itself, but the compliance to ergonomic knowledge and skills is hard to register objectively. This definitely is a challenge in future occupational intervention studies.
The drop-out rate in the Mfb group was higher than in the EC group (i.e. nine compared to five), which was probably related to the myofeedback system itself. Some subjects found the system inconvenient and disturbing in daily working activities. This may have suppressed the effect of the intervention. Improvement of the current device in terms of usability is therefore required to optimise the myofeedback training.
Conclusions
Four-weeks of intervention significantly reduced pain intensity and disability, and this effect remained after three and six months follow-up. Myofeedback training combined with ergonomic counselling is thus beneficial for female computer workers over the age of 45, reporting pain and disability in the neck-shoulder region but no evidence was found favouring myofeedback training combined with ergonomic counselling over ergonomic counselling alone. Future research may aim at identifying possible subgroups of patients in which the interventions are especially beneficial. This may enhance the efficiency and the effectiveness of the interventions. Finally, as non-specific effects may have interfered with outcome, future studies could include a placebo control group for more insight in the specific effects of ambulant myofeedback training combined with ergonomic counselling.
|
[
"myofeedback training",
"work-related neck-shoulder complaints",
"neck-shoulder region"
] |
[
"P",
"P",
"P"
] |
Doc_Ophthalmol-4-1-2235911
|
ISCEV guidelines for clinical multifocal electroretinography (2007 edition)
|
The clinical multifocal electroretinogram (mfERG) is an electrophysiological test of local retinal function. With this technique, many local ERG responses, typically 61 or 103, are recorded from the cone-driven retina under light-adapted conditions. This document specifies guidelines for performance of the test. It also provides detailed guidance on technical and practical issues, as well as on reporting test results. The main objective of the guidelines is to promote consistent quality of mfERG testing and reporting within and among centers. These 2007 guidelines, from the International Society for Clinical Electrophysiology of Vision (ISCEV: http://www.iscev.org), replace the ISCEV guidelines for the mfERG published in 2003.
Introduction
The electroretinogram (ERG) is a mass potential, the result of the summed electrical activity of the cells of the retina. Full-field electroretinography is a well-established clinical technique for evaluating global retinal function [1]. The multifocal ERG (mfERG) technique was developed to provide a topographic measure of retinal electrophysiological activity. With this technique, many local ERG responses, typically 61 or 103, are recorded from the cone-driven retina under light-adapted conditions. In 2003, the International Society for Clinical Electrophysiology of Vision (ISCEV) published guidelines for recording the mfERG [2]. These were guidelines, not standards, to allow for further research before standards were set.
Although the mfERG now has been used for more than 10 years to aid in the diagnosis of diseases of the retina, ISCEV decided that it is still premature to set specific standards. However, to take into consideration recent developments in technology and practice, this document provides revised guidelines for recording clinical mfERGs. These guidelines will be reviewed periodically, consistent with ISCEV’s practice.
Description of multifocal electroretinography
The mfERG technique is a method of recording local electrophysiologic responses from different regions of the retina. Electrical responses from the eye are recorded with a corneal electrode as in conventional, full-field ERG recording. However, the nature of the stimulus and the form of the analysis differ. These differences allow a topographic map of local ERG activity to be measured. For the basic mfERG described here, the retina is stimulated with an array of hexagonal elements, each of which has a 50% chance of being illuminated every time the frame changes (Fig. 1). Although the pattern appears to flicker randomly, each element follows the same pseudo-random sequence of illumination with the starting point displaced in time relative to other elements. By correlating the continuous ERG signal with the sequence of on- and off-phases of each element, the local ERG signal is calculated. Although we will refer to these local ERG signals as mfERG responses, it is important to keep in mind that they are not direct electrical potentials from local regions of retina, but rather they are a mathematical extraction of the signal. Further, because the stimulation rate is rapid, the waveform of the local mfERG response can be influenced both by preceding (‘adaptation effects’) and subsequent stimuli (‘induced effects’), as well as by the responses to light scattered on other retinal areas.
Fig. 1(a) Representative hexagonal mfERG stimulus array with 61 elements scaled with eccentricity. Roughly half of the elements are illuminated at any one time. (b) Same as in panel A for an array with 103 elements
The typical waveform of the basic mfERG response (also called the first-order response or first-order kernel) is a biphasic wave with an initial negative deflection followed by a positive peak (Fig. 2). There is usually a second negative deflection after the positive peak. These three peaks are called N1, P1 and N2, respectively. There is evidence that N1 includes contributions from the same cells that contribute to the a-wave of the full-field cone ERG, and that P1 includes contributions from the cells contributing to the cone b-wave and oscillatory potentials. Although there is some homology between the mfERG waveform and the conventional ERG, the stimulation rates are higher for the mfERG and, as noted above, the mfERG responses are mathematical extractions. Thus, the mfERG responses are not technically “little ERG responses”. Therefore, the designations ‘a-wave’ and ‘b-wave’, used for the full-field ERG, are not appropriate to describe features of the mfERG waveform.
Fig. 2Diagram of an mfERG response to show the designation of the major features of the waveform
Basic technology
Electrodes
Recording electrodes
Poor or unstable electrode contact is a major cause of poor quality records. It is important to follow the recommendations concerning fiber, foil, loop and contact lens electrodes in the full-field ERG Standard [1] and Pattern ERG (PERG) Standard [3]. In particular, electrodes that contact the cornea, or nearby bulbar conjunctiva, are required. In addition, good retinal image quality and proper refraction is desirable.
Reference and ground electrodes
Proper application of suitably conductive electrodes is essential for reliable mfERG recordings. Follow the recommendations made in other ISCEV Standards [1, 3].
Electrode characteristics, stability and cleaning
Follow the recommendations in the ISCEV full-field ERG and/or PERG Standards [1, 3].
Stimulation
Stimulus source
Until recently, the mfERG stimuli were most commonly displayed on a cathode ray tube (CRT), i.e. a monitor. CRT monitors are being rapidly replaced with other devices such as liquid crystal display (LCD) projectors, arrays of light emitting diodes (LEDs) and organic LEDs (OLEDs). These alternative modes of stimulation can affect the amplitude and waveform of the mfERG making it essential to specify the type of display when reporting results.1
Frame frequency
A CRT frame frequency of 75 Hz has been used widely. (LCD displays use 60 Hz.) Use of different frequencies can substantially alter the amplitude and waveform of the mfERG response. Whatever frame frequency is used, normative values for normal healthy subjects need to be determined separately for that frequency. Further, it is essential to specify the frame frequency when reporting results.
Luminance and contrast
In general, the luminance of the stimulus elements for CRT displays should be 100–200 cd/m2 in the lighted state and low enough in the dark state to achieve a contrast of ≥90%. This means that the mean screen luminance during testing will be 50–100 cd/m2. Although higher luminance levels can be used, the ability to detect local defects may be decreased due to the effects of stray light. In addition, the luminance requirements may differ if a non-CRT display is used. In any case, the surround region of the display (the area beyond the stimulus hexagons) should have a luminance equal to the mean luminance of the stimulus array.
Calibration
As with other electrophysiologic signals, luminance and contrast affect the recorded signal and it is important to calibrate the stimulus following ISCEV guidelines [4]. The luminance of the dark and the light stimulus elements should be measured with an appropriate calibrator or spot meter. Many monitor screens are not of uniform brightness over the entire screen. While some variation is to be expected, a variation of greater than 15% is considered unacceptable. Some commercial systems are equipped to calibrate the display. If this ability is not present, we urge manufacturers to provide instructions for calibration of their devices.
Stimulus parameters
Stimulus pattern: The typical mfERG display is a hexagonal stimulus pattern scaled in size to produce mfERG responses of approximately equal amplitude across the retina. Thus, the central hexagons are smaller than the more peripheral ones. Different patterns may be useful in special cases (e.g. equal size hexagons for patients with eccentric fixation). These guidelines cover only the typical stimulus pattern, scaled to produce approximately equal size responses for healthy control subjects.
Flicker sequence: Most commercial mfERG instruments use an m-sequence to control the temporal sequence (between light and dark) of the stimulus elements. An m-sequence, in which the elements can change with every frame, is recommended for routine testing. Different sequences, or the inclusion of global light or dark frames, have been suggested for specialized applications, but they are not a part of these guidelines, and should not be done in exclusion of a “standard” mfERG for routine clinical purposes.
Stimulus size and number of elements: The overall stimulus pattern should subtend a visual angle of 20–30° on either side of the fixation point. The stimulus region can be divided into different numbers of hexagons. The most frequently used patterns have 61 or 103 hexagons, with 241 hexagons occasionally used. The choice depends on balancing the need for good spatial resolution and a high signal-to-noise ratio, while minimizing the recording time. (See discussion below under clinical protocol).
Fixation targets: Stable fixation is essential for obtaining reliable mfERG recordings. Central fixation dots, crosses and circles are available with most commercial systems. The fixation targets should cover as little of the central stimulus element as possible to avoid diminishing the response. However, the examiner should always verify that the patient can see the fixation target. When the fixation targets are enlarged for low-vision patients, care should be taken not to obscure regions of interest. For example, a larger central cross will lead to a smaller central response simply due to occlusion of more of the stimulus.
Recording, analyzing and presenting results
Amplifiers and filters
The gain of the amplifier should produce recognizable signals without saturation. Appropriate band-pass filtering removes extraneous electrical noise, without distorting waveforms of interest. For a “standard” mfERG, the high pass cutoff can range between 3 and 10 Hz and the low pass cutoff between 100 and 300 Hz. Filter settings, even within the ranges suggested, will influence the response waveform. Thus, the filter settings should be the same for all subjects tested by a given laboratory, as well as for the norms to which they are compared. Line-frequency or notch filters should be avoided.
Signal analysis
Artifact rejection: Because blinks and other movements can distort the recorded waveforms, commercial software usually includes ‘artifact rejection’ algorithms to eliminate some of these distortions. When applying an artifact rejection procedure after the recording, care should be exercised to assure that clinically important aspects of the waveform are not being modified. In any case, artifact rejection procedures should be specified when reporting mfERG results.
Spatial averaging: In order to reduce noise and smooth the waveforms, some commercial programs allow the averaging of the response from each stimulus element with a percentage of the signal from the neighboring elements. Spatial averaging can help visualize the mfERG signal in noisy records. However, spatial averaging may obscure small, local changes or the borders of regions of dysfunction. Thus, it should be used with care and specified when reporting results. Further, the default conditions of commercial software should be examined, as spatial averaging may be a default condition.
Displaying results
Trace arrays: All commercial programs can produce an array of the mfERG traces (Fig. 3a, b). This display of the results is useful for visualizing areas of abnormality and for comparing the mfERG results to visual fields from perimetry. The trace array is the basic mfERG display and should always be included in the report of clinical results. Commercial software typically allows for these responses to be displayed with either a visual field view or retinal view as the frame of reference. The frame of reference should be indicated. Note that in Fig. 3, each of the 61 (panel a) or 103 (panel b) responses are shown in approximately correct topographic representation. Adding contours to show the loci of different eccentricities, as in Fig. 4, can help in determining whether the regions of low amplitude correspond to regions of the visual field that show a loss of sensitivity.
Fig. 3Sample mfERG trace arrays (field view) with 61 elements (panel a) and 103 elements (panel b). (c, d) The 3-D response density plots (field view) associated with panels a and bFig. 4The mfERG trace array (left panel, field view) and the probability plot from standard automated perimetry (right panel) for a patient with retinitis pigmentosa. The contours for a radius of 5 and 15° are shown. The light gray, dark gray, and black squares indicate statistically significant field loss at the 5, 1 and 0.5 percent levels, respectively
Group averages: Commercial programs allow for the averaging of groups of responses from the trace arrays. This can be helpful for comparing quadrants, hemiretinal areas, normal and abnormal regions of two eyes, or successive rings from center to periphery. The latter is particularly useful for patients who have diseases that produce approximately radially symmetric dysfunction. In addition, responses from stimulus elements associated with a local area of interest can be averaged for comparison with a similar area in an unaffected eye or to data from control subjects. Figure 5 shows the result of grouping the responses in Fig. 3b by rings. The most commonly used display is response density (Fig. 5a), in which the responses from the elements in each ring are summed and then divided by the area of these elements. Figure 5b and c are alternative representations of the data in which summed responses from each ring are shown (panel c) or are normalized so that they have the same amplitude (panel b).
Fig. 5The mfERG responses in Fig. 3b were grouped by concentric rings and summed to yield the ‘Summed’ responses in panel c. These summed responses are divided by the area of the elements of the ring for the ‘Response Density’ responses (panel a) and normalized so each has the same amplitude for the ‘Normalized’ responses (panel b)
Topographic (3-D) response density plots: The 3-D plot (Fig. 3c, d) shows the overall signal strength per unit area of retina. This display can sometimes be useful for illustrating certain types of pathology. In addition, the quality of fixation can be assessed by observing the location and depth of the blind spot. However, there are major dangers in using the 3-D plot in assessing retinal damage. First, information about the waveforms is lost. Thus large, but abnormal, or delayed responses can produce normal 3-D plots. Second, a central peak in the 3-D plot can be seen in some records without any retinal signal (see, Appendix: Artifact recognition examples for electrical noise and weak signals). Finally, the appearance of the 3-D plot from a given recording is dependent on how the local amplitude is measured. Whenever 3-D plots are presented, the method used to measure local amplitude should be identified, and the corresponding mfERG trace array also should be presented.
Signal extraction: Kernels—This document is aimed at the general mfERG user and only describes the basic response, the first-order kernel. Higher order kernels, particularly the second-order kernel, occasionally are reported, and used in special applications.
Clinical protocol
Patient preparation
Pupils
The pupils should be fully dilated and pupil size noted.
Patient positioning
Subjects should sit comfortably in front of the screen. Relaxation of facial and neck muscles will reduce artifacts from muscles; a headrest may be helpful. The appropriate viewing distance will vary with screen size, in order to control the area (visual angle) of retina being stimulated.
Fixation monitoring
Good fixation, both central and steady, is essential. Thus, fixation should be monitored, preferably by the use of monitoring instrumentation available on some units. When this option is not available, careful direct observation may be employed.
Refraction
Although there is some evidence that the mfERG is unaffected by moderate blurring of the retinal image in healthy individuals, we recommend refraction for optimal acuity. On some commercial machines, a manual adjustment of the viewing optics is possible. Alternatively, lenses can be placed in a holder positioned in front of the eye. In the latter case, the viewing distance must be adjusted to compensate for the relative magnification of the stimulus. Also care must be taken to avoid blocking the view of the stimulus screen by the rim of the lens or the lens holder and thus creating an apparent scotoma.
Monocular versus binocular recording
Recording is typically done with monocular stimulation. Those who record binocularly should be aware that signals can be altered by misalignment of the eyes.
Adaptation
Pre-adaptation (before test): Subjects should be exposed to ordinary room lighting for at least 15 min prior to testing. Longer adaptation times may be needed after exposure to bright sun or bright lights such as those used for fundus photography or retinoscopy.
Room illumination: Moderate or dim room lights should be on and ideally should produce illumination close to that of the stimulus screen.
Stimulus and recording parameters
Stimulus Size
The stimulus should subtend 20–30° of visual angle on either side of the fixation point.
Number of elements
A display containing 61 or 103 elements should be used.
Duration of recording
A total recording time of at least 4 min for 61 element arrays, or 8 min for 103 element arrays, is recommended, although these times might be adjusted by experienced laboratories according to clinical needs. The overall recording time is divided into shorter segments (e.g. 15–30 s) so that subjects can rest between runs if necessary and also so that a poor record (from noise, movement or other artifacts) can be discarded and repeated without losing prior data.
Trade-offs
Various manipulations will affect the signal-to-noise ratio (SNR) of the responses. In particular, decreasing the size (increasing the number for a fixed stimulus field size) of the stimulus elements and decreasing the duration of the recording will decrease the SNR of the responses. While decreasing the number (increasing the size) of elements will increase the SNR, it will decrease the spatial resolution of the test. In general, conditions with larger (i.e. 61) elements and a shorter recording time (e.g. 4 min) are easier for the patient and suitable for a general screening of macular function. On the other hand, conditions with 103 elements and a longer recording time (e.g. 8 min) are useful for assessing foveal function and mapping the outline of retinal defects. Very small elements (such as a 241 hexagon array) may sometimes be helpful for diseases with very small or irregular effects. Repeat recording is recommended to confirm small or subtle abnormalities.
Further, the choice of electrode type will also influence the SNR of the responses. For example, bipolar corneal contact electrodes yield recordings with the highest SNR. Thus, longer recording times, repeat measurements and/or fewer stimulus elements are necessary to obtain comparable SNRs when a foil or fiber electrode is used.
Data reporting
Mode of display
Trace arrays: It is essential to show the trace array when reporting mfERG results (see Fig. 3a, b). These arrays not only show topographic variations, but also demonstrate the quality of the records, which is important in judging the validity of any suspected variations from normal. Trace lengths of 100 ms or more should be used for these displays. (It is hard to detect interference from line frequency and/or kernel overlap in shorter trace lengths.).
Group averages: Arranging responses by groups can be a useful way to summarize the data. Concentric rings of traces, from the center outward, are most commonly used. Regions with fundus pathology can be averaged together if desired. Most laboratories report response density (Fig. 5a).
Three-dimensional plots: These should be used with caution and only when accompanied by trace arrays (see above). The 3-D plots (Fig. 3c, d), without accompanying trace arrays, can be misleading (see Appendix). Note that if fixation is steady and central, a clear depression due to the blindspot should be present and located in the appropriate place.
Measurements calibration marks: Calibration marks must accompany all traces or graphs. This will enable comparisons among patients or within a patient on sequential visits.
Measuring mfERG amplitude and timing: The N1 response amplitude is measured from the starting baseline to the base of the N1 trough; the P1 response amplitude is measured from the N1 trough to the P1 peak (see Fig. 2). The peak times (implicit times) of N1 and P1 are measured from the stimulus onset. Measurements of group averages should routinely include the N1 and P1 amplitudes and peak times.
Commercial software provides measures of the overall amplitude and timing of the mfERG traces. There are various procedures for measuring amplitude (e.g. trough-to-peak amplitude), latency (e.g. response shifting, response stretching, time to peak), or overall response waveform (e.g. scalar product, root-mean-square (RMS)). A description of these techniques is beyond the scope of these guidelines. However, it should be noted that when a template is needed (e.g. for scalar product measures), the template should be formed from age-similar control data obtained from that laboratory.
Color scales: The use of color scales is optional; care should be taken when reproducing color images on a gray scale as the luminance sequences may not be in the proper order.
Normal values
Each laboratory must develop its own normative data. Variations in recording equipment and parameters make the use of data from other sources inappropriate. Because electrophysiologic data are not necessarily described by a normal distribution, laboratories should report the median value rather than the mean, and determine boundaries of normality. The mfERG, like the full-field ERG, is somewhat smaller in amplitude in older individuals and in those with highly myopic eyes. Although these effects are generally not large, they can be important in the evaluation of some patients. In any case, age-adjusted normative data is recommended.
Reporting of artifacts and their resolution
Reports should indicate any problems with the recording such as movement, head tilt, poor refraction capability, poor fixation, etc. that might affect reliability and interpretation. Also, indicate explicitly any artifact reduction procedures or post-processing maneuvers used to prepare the data. This should include the type and number of artifact rejection steps, the spatial averaging with neighbors (noting the extent and number of iterations), and any other averaging or filtering procedures.
|
[
"multifocal electroretinogram",
"electroretinogram",
"clinical guidelines"
] |
[
"P",
"P",
"R"
] |
Purinergic_Signal-3-4-2072924
|
Activation kinetics of single P2X receptors
|
After the primary structure of P2X receptors had been identified, their function had to be characterized on the molecular level. Since these ligand-gated ion channels become activated very quickly after binding of ATP, methods with adequate time resolution have to be applied to investigate the early events induced by the agonist. Single-channel recordings were performed to describe conformational changes on P2X2, P2X4, and P2X7 receptors induced by ATP and also by allosteric receptor modifiers. The main results of these studies and the models of P2X receptor kinetics derived from these observations are reviewed here. The investigation of purinoceptors by means of the patch clamp technique following site-directed mutagenesis will probably reveal more details of P2X receptor function at the molecular level.
Primarily, P2X receptors are ligand-gated ion channels. This ion channel family, also known as ion channel-coupled receptors, mediates the fastest known signal transduction from extracellular messengers to the intracellular environment. This signal transduction mechanism is so simple (i.e., includes few steps) that it becomes fully activated in the low millisecond range. The best known biological structures using this approach are neuronal synapses in which receptors for the classic transmitters acetylcholine, glutamate, gamma-aminobutyric acid (GABA), glycine, or serotonin transmit excitation from one neuron to another in less than 1 ms [1]. This rapid activation kinetics of the ligand-gated ion channels is too fast to become analyzed by whole-cell current recordings or even slower fluorescent dye-based measurements of changes of the intracellular ion composition. The main reason for the low speed of these recording techniques compared to the receptor kinetics is the rate-limiting solution exchange at the extracellular side of the cell membrane where the ligand binding site is located. These obstacles in the investigation of ligand-gated channels can be overcome by the single-channel patch clamp technique where tiny parts (patches) are excised out of the cell membrane containing at best only one receptor. In such preparations, solution exchanges can be achieved within less than 1 ms [2]. Apart from the high time resolution, this technique has a resolution of ionic currents in the pA (picoampere) range sufficient to observe the ionic flow through and the gating (opening and closing) of single ion channels. With this technique, ion channels can be characterized and identified according to their conductance, macroscopic kinetics (time course of activation, i.e., opening of channels after ATP application, and of deactivation, i.e., closing of channels after ATP withdrawal), and microscopic kinetics (i.e., the characteristic mean times channels spend in closed or open configurations). The ultimate goal of the kinetic analysis is the development of a kinetic (normally Markovian) model of closed and open states which describes the agonist concentration dependency of the macroscopic and microscopic kinetics. Additionally, single-channel analysis may be able to discriminate whether various effects of agonists and antagonists are caused by changing agonist binding or efficacy or by altering the ion channel permeation behavior [3].
Additionally, patch clamping allows the separation of currents flowing through the ligand-gated channels from contaminating currents mediated by conductances activated downstream by secondary signals. In the case of P2X receptors, such signals may be changes in the intracellular concentration of Ca2+, Na+, K+, or H+ ions which can permeate the channel pore. The next steps in signal cascades may be changes of the cell volume with effects on the cytoskeleton, oxidative stress, and furthermore activation of kinases, phosphatases, or transcription. Therefore, investigating the time course of opening and closing of P2X receptors may help to understand signaling cascades they activate. It may also detect functional changes induced by receptor mutations more precisely than whole-cell current recordings or measurements of P2X receptor-induced changes of the intracellular Ca2+ concentration or even later effects.
Soon after the introduction of the patch clamp technique, ATP-activated single channel events were discovered and analyzed. They were first described in rat sensory neurons [4] and cultured chicken myoblasts [5]. Later, different ATP-activated single ion channel currents were reported in smooth and cardiac muscle, glands, and other neurons where single-channel conductances between <1 and 60 pS were measured [6].
More detailed information about P2X receptor kinetics was obtained from measurements of the rise time of ATP-induced whole-cell currents which, as typical for ligand-gated receptors, decreased with increasing agonist concentration in the range of 5–25 ms [7, 8]. A more sophisticated kinetic analysis of P2X receptor function was performed in bullfrog dorsal root ganglion cells. The concentration dependence of the activation and deactivation time course as well as of the steady-state current during prolonged ATP application could well be described by a linear Markov model where the independent binding of three ATP molecules to equal binding sites leads to the opening of the channel which in the triliganded state is open 25% of the time:
Here, the rate constants are given in s−1 and β/α = 1/4 [9]. Remarkably, the trimeric structure of cloned P2X receptor channels described much later [10–12] implies the necessity of binding of three ATP molecules and therefore coincides well with this model.
Investigations of the microscopic kinetics of native P2X receptors were used to characterize aspects of the gating behavior of the channels in distinct preparations but were not developed into kinetic models [5, 13–18].
After cloning the P2X receptor subtypes, the kinetic analysis of heterologously expressed distinct P2X receptors became principally possible. However, the collection of enough single-channel data for establishing a kinetic model turned out to be too laborious or even seemed to be impossible on receptors which desensitize quickly (within milliseconds) and recover slowly (within minutes) [19, 20]. Therefore, models for P2X1 [21] and P2X3 [22] receptors have been developed which, like the model developed by Bean [9], are based on whole-cell recording and include additional desensitized receptor states.
For the slowly desensitizing P2X4 receptor, the effect of external Mg2+ ions was investigated on the single-channel level. The blocking effect of Mg2+ could be partially ascribed to the reduction of the mean open time [23]. A more detailed analysis of the single-channel kinetics of P2X4 receptors revealed three different mean open times and five mean closed times following activation of the receptor by 0.3 μM ATP, where desensitization is slow. Hence, it was concluded that a kinetic model should have at least three open and five closed states [24].
A more comprehensive investigation of the single-channel kinetics under steady-state conditions (i.e., during prolonged application of constant extracellular ATP concentrations where the distribution of conformational states of receptors is not changing with time) was performed on the very slowly desensitizing P2X2 receptor [25]. It should, however, be mentioned that P2X2 receptors may exhibit substantial fast desensitization in excised patches [26]. The following model was developed:
In contrast to Bean’s model [9] the three ATP binding sites are not equal in this case. Specifically, the second binding step possesses a higher association rate constant and a much higher dissociation rate constant compared to the first. Both together result in a decreased affinity. This could be interpreted in such a way that the binding of ATP at the first site induces a conformation change that leads to an easier accessibility of a second binding to ATP. But this second binding is weaker, possibly due to the repulsion of both anionic ATP4− molecules. Negative cooperativity due to ATP repulsion was also suggested for the human P2X7 receptor [27]. The addition of an open (O6) and a closed state (C7) to a simple C–C–C–C–O model was necessary to describe openings of the P2X2 receptor which occurred as single openings or as bursts. This model was however unable to explain P2X4 receptor kinetics of multichannel patches, and therefore a positive cooperativity toward opening of P2X4 receptor channels was assumed [28].
The P2X7 is another non-desensitizing P2X receptor which proved to be suitable for kinetic analysis [29]. A relatively simple model was used to describe both the ATP dependence of microscopic (open and shut times, open probability of single channels) and macroscopic kinetics (activation and deactivation time course) of the main gating mode measured by a combination of single-channel recording and an ultrafast solution exchange system:
The model could be simplified compared to that describing the kinetics of the P2X2 receptor, since the open time distribution could be described in most cases by one exponential component and the channel did not display a bursting behavior. The much lower association rate constants compared to the P2X2 model are owed to the low ATP affinity of the P2X7 receptor. The kinetics of a second gating mode of single P2X7 receptors were about four times slower but could principally also be described by this model.
The assumption of only two ATP binding sites was sufficient for the approximation of the measured P2X7 receptor kinetics. This is a further deviation from the models mentioned above. The finding that P2X7 receptors are putative trimers like other P2X receptors and therefore contain probably three ATP binding sites seems to be contradictory to the P2X7 receptor kinetics model. As yet, a comprehensive explanation cannot be given but there are indications for a third binding site that induces very small single ion channel currents (Riedel et al., unpublished).
Mean currents calculated by this model displayed an approximately monoexponential activation and deactivation time course with time constants in the range of 10–25 ms. It was much simpler than the kinetics of several measured P2X7 receptor-dependent whole-cell currents [11]. This indicates that under whole-cell conditions either other ion currents may become secondarily activated or the biophysical properties of single P2X7 receptors themselves become less complex due to washout of an essential receptor component. However, even in cell-attached recordings, in which the intracellular environment is preserved, the single-channel current measurements at P2X7 receptors gave indication neither of more complex activation kinetics nor of a dilatation of the channel pore which is assumed to occur during prolonged ATP application [11]. Moreover, these single-channel current characteristics are indistinguishable from those measured in human lymphocytes [16], corroborating the view that the native human P2Z receptor is a genuine P2X7 receptor. Detailed analysis of the permeation characteristics [30] revealed a stable single-channel conductance even in external Na+-free and low Ca2+ solutions during long-lasting application of large ATP concentrations where a permeability of the cell membrane to large organic cations was induced in whole-cell preparations [11, 31, 32]. Instead, the permeation behavior and the calculated pore size of about 8.5 Å was similar to P2X1 and P2X2 receptors [33, 34]. Accordingly, it can be concluded that the apparent pore dilatation of P2X7 receptors observed in macroscopic current recordings [32] has no equivalent at the single-channel level, suggesting that the P2X7 receptor-induced permeability increase is not due to pore dilatation but reflects the induction of additional conductances mediated by other proteins secondary to P2X7 receptor activation. As recently reported, such additional conductance may be established by pannexin [35, 36], which can be activated by intracellular Ca2+ ions [37] and membrane stretch [38]. Similarly, single-channel recordings from P2X2 [33] and P2X4 receptors [24] do not indicate any pore dilatation which is assumed to occur in whole-cell preparation [39–41].
Instead of changing the permeation behavior, extracellular Na+ removal was found to drastically change the P2X7 ion channel kinetics [30]. The open probability, the mean open time as well as the activation and deactivation time constants considerably increased when Na+ was replaced by other monovalent cations. This effect was highly specific to Na+, suggesting that Na+ is not only a permeating ion but also a modulator of P2X7 receptor gating. The fact that the P2X7 receptor-induced increase of the cell membrane permeability to organic cations is promoted in Na+-free media [42, 43] can therefore be explained by an enhanced P2X7 receptor activation. This results in an increased Na+ and Ca2+ influx that may activate further permeation pathways.
It turned out that the kinetic model for P2X7 receptors [29] could easily explain the potentiating effect of Na+ removal, assuming that binding of Na+ to an extracellular site reduces the rate constant from O4 to C3 from 200 s−1 to 8 s−1. Although this is an allosteric effect not interfering with ATP binding, the concentration-response curve is shifted leftwards by Na+ removal, giving rise to a reduced apparent ATP binding constant. Such an apparent increase of a binding site affinity due to an increased efficacy has already been described by Colquhoun [44]. In this respect, single-channel recording has again proven to be the (possibly only) method to quantitatively measure and distinguish effects on single proteins of drug binding and the following conformational changes considered as efficacy. The model explains on the molecular level the known stimulating effect of Na+ substitution on effects mediated by P2X7 or P2X7-like native receptors [11].
Under conditions of hypoxia or tissue injury, ATP and K+ are secreted into the extracellular space [11]. Therefore, in metabolically compromised tissues, a replacement of extracellular Na+ by K+ may enhance the efficacy and potency of ATP on P2X7 receptors. Furthermore, ATP-mediated cell depolarization increases the driving force for K+ efflux and lowers the EC50 value of Na+ since the site seems to be located in the electrical field of the membrane [30]. In this way, K+ efflux may reinforce the ATP effect on P2X7 receptor-expressing cells, and even low concentrations of ATP may activate the P2X7 receptor to a substantial extent. However, whether such a mechanism is of relevance in vivo remains to be established.
In the future, the combination of single-channel recordings and molecular engineering methods will reveal more details concerning the function of P2X receptor-dependent ion channels. To this end, it may help to dissect and understand early and late steps of the signaling cascades which are started with the binding of ATP and are followed by biological processes known to be influenced by P2X receptors like contraction, secretion, perception, proliferation, and apoptosis [45]. Furthermore, patch clamp recording will continue to describe the effects of P2X receptor modulators on the single-channel level (for review of the investigations already performed, see North [11]).
|
[
"kinetics",
"p2x4",
"p2x7",
"model",
"single channel",
"purinergic receptor",
"voltage clamp"
] |
[
"P",
"P",
"P",
"P",
"P",
"M",
"M"
] |
Neurochem_Int-2-1-2194163
|
Control of excitatory synaptic transmission by capsaicin is unaltered in TRPV1 vanilloid receptor knockout mice
|
Several studies have shown that capsaicin could effectively regulate excitatory synaptic transmission in the central nervous system, but the assumption that this effect is mediated by TRPV1 vanilloid receptors (TRPV1Rs) has not been tested directly. To provide direct evidence, we compared the effect of capsaicin on excitatory synapses in wild type mice and TRPV1R knockouts. Using whole-cell patch-clamp techniques, excitatory postsynaptic currents (EPSCs) were recorded in granule cells of the dentate gyrus. First, we investigated the effect of capsaicin on EPSCs evoked by focal stimulation of fibers in the stratum moleculare. Bath application of 10 μM capsaicin reduced the amplitude of evoked EPSCs both in wild type and TRPV1R knockout animals to a similar extent. Treatment of the slices with the TRPV1R antagonist capsazepine (10 μM) alone, or together with the agonist capsaicin, also caused a decrease in the EPSC amplitude both in wild type and TRPV1R knockout animals. Both drugs appeared to affect the efficacy of excitatory synapses at presynaptic sites, since a significant increase was observed in paired-pulse ratio of EPSC amplitude after drug treatment. Next we examined the effect of capsaicin on spontaneously occurring EPSCs. This prototypic vanilloid ligand increased the frequency of events without changing their amplitude in wild type mice. Similar enhancement in the frequency without altering the amplitude of spontaneous EPSCs was observed in TRPV1R knockout mice.
1
Introduction
Capsaicin, the pungent ingredient of red peppers, severely affects pain sensation, inflammation or hyperalgesia. Systematic studies aiming to elucidate the effects of capsaicin revealed that this alkaloid primarily targets sensory fibers of the C type, where it activates a member of the transient receptor potential (TRP) channels, TRPV1 vanilloid receptors (TRPV1Rs) (Szallasi and Blumberg, 1999; Szolcsanyi, 2004). These receptors, cloned by Caterina et al. (1997), are non-selective cation channels gated by heat, low pH or endogenous ligands, such as anandamide (Tominaga et al., 1998; Zygmunt et al., 1999; Caterina et al., 1999; Smart et al., 2000). In addition, capsaicin could also affect the operation of both voltage-gated sodium and calcium channels (Balla et al., 2001; Lundbaek et al., 2005; Kofalvi et al., 2006), indicating that some of the capsaicin effects might not be linked to TRPV1Rs.
In contrast to the well-established function of TRPV1Rs in the periphery, its role is much less obvious in the central nervous system. Using autoradiography or immunohistochemistry, TRPV1Rs were shown to be present in several brain regions, including cortical structures (Acs et al., 1996; Mezey et al., 2000; Roberts et al., 2004; Toth et al., 2005; Cristino et al., 2006). Importantly, the specificity of signals in two of these reports has been confirmed in TRPV1R knockout mice (Roberts et al., 2004; Cristino et al., 2006), strongly arguing for the existence of TRPV1Rs in the CNS, yet their subcellular (synaptic or extrasynaptic) localization remains to be determined by high resolution electron microscopy. The functional role of TRPV1Rs in distinct brain regions was addressed by electrophysiological experiments. In the hippocampus, for example, an increase in paired-pulse depression of field potentials after application of capsaicin or anandamide has been noticed, and this effect was found to be sensitive to TRPV1R antagonists (Al-Hayani et al., 2001; Huang et al., 2002). In addition to the investigation of capsaicin effects on field potentials, other studies have examined the activation of putative TRPV1Rs on synaptic transmission more directly. These studies have found that in distinct parts of the brain, glutamatergic, but not GABAegic synaptic communication could be controlled by capsaicin, an effect that was also antagonized by TRPV1R antagonists (Sasamura et al., 1998; Hájos and Freund, 2002; Marinelli et al., 2002, 2003; Xing and Li, 2007). Interestingly, excitatory postsynaptic currents (EPSCs) evoked by electrical stimulation were found to be depressed after application of capsaicin, whereas the same treatment significantly increased the occurrence of spontaneous EPSCs without affecting their amplitude.
To reveal whether the effect of capsaicin on synaptic glutamate release is indeed mediated by TRPV1Rs, we investigated the properties of EPSCs in dentate granule cells after bath application of this prototypic vanilloid ligand both in wild type and TRPV1R knockout mice.
2
Experimental procedures
Experiments were carried out according to the guidelines of the institutional ethical code and the Hungarian Act of Animal Care and Experimentation (1998. XXVIII. section 243/1998). Wild type and TRPV1R knockout mice of both sexes (20–77 days old, C57BL/6J strain) (Davis et al., 2000) were used. Mouse genotyping was performed on tail DNA. Neo PCR primer sequences were the following: NeoF 5′-CCGGCCGCTTGGGTGGAGAGG and NeoR 5′-products on 300 bp (targeted allele), and TRPV1Rs (VRF1 5′-CATGGCCAGTGAGAACACCATGG and VRR2 5′-AGCCTTTTGTTCTTGGCTTCTCCT) products on 150 bp (wild type allele). Amplification reactions were carried out in 25 μl total volume with presence of 1% dimethyl-formamide, 0.2 μM primer each, 0.2 μM dNTP, 1.5 mM MgCl2 (93 °C for 15 s, 58 °C for 15 s, 72 °C for 45 s, for 30 cycles). Amplification products were analyzed by agarose gel electrophoresis on 1.2% agarose gels. An example for the results of genotyping of a litter is shown in Fig. 1.
The animals were deeply anaesthetized with isoflurane followed by decapitation. After opening the skull, the brain was quickly removed and immersed into ice-cold cutting solution containing (in mM: NaCl 126, KCl 2.5, NaHCO3 26, CaCl2 0.5, MgCl2 5, NaH2PO4 1.25, glucose 10) bubbled with 95% O2/5% CO2 (carbogen gas). Thick horizontal slices (300–350 μm from mice) were prepared using a Leica VT1000S Vibratome. The slices were stored in an interface type chamber containing ACSF (in mM: 126 NaCl, 2.5 KCl, 26 NaHCO3, 2 CaCl2, 2 MgCl2, 1.25 NaH2PO4, and 10 glucose) at room temperature for at least 1 h before recording.
Whole-cell patch-clamp recordings were obtained at 34–36 °C from granule cells in the dentate gyrus visualized by infrared videomicroscopy (Versascope, Marton Electronics, Canoga Park, CA). Patch electrodes were pulled from borosilicate glass capillaries with an inner filament (1.5 mm o.d., 1.12 mm i.d.; Hilgenberg, Germany) using a Sutter P-87 puller. Electrodes (∼3–6 MΩ) were filled with a solution containing (in mM) 80 CsCl, 60 Cs-gluconate, 3 NaCl, 1 MgCl2, 10 HEPES, 2 Mg-ATP, and 5 QX-314 (pH 7.2–7.3 adjusted with CsOH; osmolarity 275–290 mOsm). Excitatory postsynaptic currents (EPSCs) were recorded at a holding potential of −65 mV. Slices were perfused with ACSF containing 70–100 μM picrotoxin to block inhibitory neurotransmission. The solution was bubbled with carbogen gas at room temperature and perfused at a flow rate of 2–3 ml/min in a submerged type chamber. To evoke EPSCs, the stimulating electrode was placed in the stratum moleculare of the dentate gyrus. Pairs of electrical stimuli separated by 50 ms were delivered via a theta glass pipette (Sutter Instrument Company, Novato, CA) filled with ACSF at 0.1 Hz using a Supertech timer and isolator (Supertech Ltd., Pécs, Hungary, http://www.superte.ch). Access resistances (between 4 and 18 MΩ, compensated 65–70%) were frequently monitored and remained constant (±20%) during the period of analysis. Signals were recorded with an Axopatch 200B (Molecular Devices, Sunnyvale, CA), filtered at 2 kHz, digitized at 6 kHz (National Instruments PCI-6024E A/D board, Austin, TX), and analyzed off-line with the EVAN program (courtesy of Prof. I. Mody, UCLA, CA).
The drugs were perfused until the maximal effect was reached (usually 3–4 min). The effect of drugs on evoked EPSCs was calculated as follows: control EPSC amplitudes in a 2–3 min time window were compared to those measured after 5–6 min drug application for the same period of time. Only those experiments were included that had stable amplitudes at least for 10 min before drug application. The paired-pulse ratio was calculated from the mean amplitude of the second EPSCs divided by the mean amplitude of the first EPSCs. For spontaneously occurring EPSCs, the amplitude and the inter-event interval for individual events were calculated and medians of their distributions were compared before and after 5 min of capsaicin application. After each experiment, the tubing made of Teflon was washed with ethanol for 10 min and with ACSF for 15 min. For comparison of data, Wilcoxon matched pairs test or Mann–Whitney U-test were used in STATISTICA 6.1 (Statsoft, Inc., Tulsa, OK). Data are presented as mean ±S.E.M.
Picrotoxin was purchased from Sigma–Aldrich, while (E)-capsaicin and capsazepine were obtained from Tocris. Both drugs were dissolved in DMSO giving a 100 mM stock solution, which were stored at 4 °C.
3
Results
The effects of the prototypic TRPV1R agonist capsaicin on EPSCs evoked by focal stimulation of fibers in the stratum moleculare were measured in dentate granule cells of wild type mice and TRPV1R knockouts. Similar to what we found earlier (Hájos and Freund, 2002), bath application of 10 μM capsaicin significantly reduced the amplitude of EPSCs (by 36.6 ± 6.1% of control) in wild type mice (control: 143.8 ± 19.3 pA; capsaicin: 88.6 ± 12.1 pA; n = 6; p < 0.02; Fig. 2A and B). In TRPV1R knockout mice, a similar significant reduction was observed after capsaicin application, the amplitude of EPSCs was suppressed by 31.6 ± 4.1% of control (control: 146.9 ± 34.2 pA; capsaicin: 103.8 ± 26.9 pA; n = 6; p < 0.02; Fig. 2A and B). The inhibitory effect of capsaicin on the amplitude of EPSCs was indistinguishable in wild type mice and TRPV1R knockouts (p > 0.1).
Next, we tested the effect of 10 μM capsazepine, a TRPV1R antagonist, on excitatory synapses. We found that bath application of this drug also significantly reduced the EPSC amplitude (by 34.9 ± 5.4% of control) in wild type animals (control: 165.1 ± 35.5 pA, capsazepine: 112.4 ± 28.2 pA; n = 5; p < 0.05; Fig. 3A and B). When we co-applied 10 μM capsaicin together with 10 μM capsazepine, the amplitude of evoked EPSCs was similarly decreased (by 33.3 ± 9.4% of control; control: 149.5 ± 21.2 pA; capsaicin + capsazepine: 99.6 ± 19.6 pA; n = 4; p < 0.05). In TRPV1R knockouts, suppression of the EPSC amplitude was comparable to that seen in wild type mice (i.e., by 30.2 ± 1.5% of control after capsazepine treatment; control: 205.2 ± 7.3 pA; capsazepine: 143.2 ± 6.4 pA; n = 4; p < 0.05; Fig. 3A and B). Similarly, the treatment of slices with a mixture of capsaicin and capsazepine reduced the EPSC amplitude by 38.1 ± 6.2% of control (control: 211.8 ± 4.4 pA, capsaicin + capsazepine: 131.3 ± 13.5 pA; n = 3; p < 0.05), just like in the wild types. These results suggest that capsazepine, as well as capsaicin alone can reduce the amplitude of EPSCs independent of TRPV1Rs, and their effects are not additive.
By a comparison of the paired-pulse ratios of evoked EPSCs, we next examined whether capsaicin and capsazepine affect excitatory synapses presynaptically or postsynaptically. If glutamate release is altered, then the paired-pulse ratio should change. If the reduction in EPSC amplitude is not accompanied by changes in the paired-pulse ratio, then the conductivity of glutamate receptors should be modified by the drug treatment. Therefore, we first investigated the effect of capsaicin on paired-pulse ratio in wild type mice and TRPV1R knockouts. After drug application, the ratio significantly increased to 121.3 ± 3.2% of control in wild types and to 132.5 ± 9.9% of control in knockouts (n = 6 each, p < 0.02). Comparable to these findings, capsazepine treatment also caused a significant increase in the paired-pulse ratio both in wild types (129.7 ± 10.1%; n = 5; p < 0.05) and knockouts (126.4 ± 8.8; n = 4; p < 0.05). Thus, the effects of both capsaicin and capsazepine appear to be presynaptic, reducing glutamate release from excitatory terminals both in wild type and TRPV1R knockout mice.
In further experiments, we investigated how capsaicin alters the properties of spontaneous EPSCs (sEPSCs). In wild type mice, bath application of capsaicin significantly increased the occurrence of spontaneous events (i.e., reduced the inter-event interval by 32.5 ± 11.8% of control, control: 0.43 ± 0.23 s; capsaicin: 0.2 ± 0.05 s; n = 5; p < 0.04, Fig. 4A and B) without changing their amplitude (control: 16.6 ± 2.1 pA; capsaicin: 16.1 ± 2.7 pA; n = 5; p > 0.1; Fig. 4A and B). Similarly, capsaicin also elevated the frequency of sEPSCs in TRPV1R knockouts, since the inter-event interval was reduced by 29.3 ± 7.4% of control (control: 0.16 ± 0.04 s; capsaicin: 0.12 ± 0.04 s; n = 6; p < 0.02; Fig. 4A and B). Similar to those observed in wild types, the amplitude of synaptic events did not change (control: 19.2 ± 3.6 pA; capsaicin: 17.7 ± 3.3 pA; n = 6; p > 0.1; Fig. 4A and B). The comparison of the decrease in the inter-event interval of sEPSCs between wild type mice and TRPV1R knockouts showed no difference (p > 0.1). These results provided further evidence that capsaicin affected synaptic glutamate release in wild type and TRPV1R knockout mice to a similar degree, in a similar manner.
4
Discussion
Electrophysiological data presented here strongly suggest that capsaicin actions on excitatory synaptic transmission are not mediated by TRPV1Rs, at least in the dentate gyrus. Our previous observations already raised this possibility (Hájos and Freund, 2002), as we have shown that the suppression of the amplitude of EPSCs after the second application of capsaicin was indistinguishable from that seen after the first application. This observation was not consistent with the known desensitization properties of TRPV1Rs upon repeated capsaicin application (Dray et al., 1989; Docherty et al., 1991; Caterina et al., 1997). Our results seem to contradict those pharmacological data, where the effect of capsaicin on synaptic transmission has been found to be fully blocked by antagonists specific for TRPV1Rs (e.g., capsazepine or iodo-resiniferatoxin) (Al-Hayani et al., 2001; Marinelli et al., 2003). Here we found that, in adult mice, capsazepine also effectively reduced the amplitude of EPSCs, similar to that seen after capsaicin application. In line with these data, a study by Kofalvi et al. (2003) has shown that glutamate release from synaptosomes prepared from the hippocampus of adult rats could be significantly suppressed by capsazepine. These findings are in contrast with our published results (Hájos and Freund, 2002), where capsazepine could antagonize the effect of capsaicin on EPSC amplitude recorded in slices from rats of P15-22. We repeated the experiments with capsazepine in young rats and found that this drug on its own could substantially enhance the amplitude of EPSCs (unpublished observations), which is in sharp contrast observed in adult animals (present study; Kofalvi et al., 2003). Thus, it seems likely that during development the molecular target of capsazepine changes its effect on synaptic transmission, or the binding site(s) of capsazepine might alter. We therefore propose that, in the hippocampus of young rats, the suppression of EPSC amplitude by capsaicin is counterbalanced by the enhancement caused by capsazepine, therefore no reduction in glutamate release can be observed.
As to the presynaptic mechanism of capsaicin actions, a recent study showed that iodo-resiniferatoxin as well as capsaicin (both applied in μM concentrations) could markedly reduce the high K+-induced Ca2+ entry (Kofalvi et al., 2006). Since transmitter release is highly sensitive to Ca2+ entry, one might assume that glutamate release from excitatory terminals in the dentate gyrus could be affected with a similar mechanism. In the present study we found that capsaicin, capsazepine, or co-application of the two, reduce the amplitude of EPSCs to a similar extent, suggesting that capsazepine might also decrease Ca2+ entry at the same site, where capsaicin acts.
Data from other laboratories, as well as our own results, showed that capsaicin could reduce the amplitude of evoked EPSCs, while it increased the frequency of spontaneous EPSCs without changing their amplitude (Hájos and Freund, 2002; Marinelli et al., 2002, 2003). These unconventional effects seem to suggest that the target molecule of capsaicin regulating glutamatergic transmission could be located at the presynaptic axon terminals, an assumption that is also supported by the observed increase in the paired-pulse ratio (present study). If capsaicin modulated postsynaptic glutamate receptors, the amplitude of sEPSCs should have also been altered, which was not the case (present study; Marinelli et al., 2003). The question arises how capsaicin could reduce the amplitude of evoked EPSCs, while in the same time increases the frequency of sEPSCs. One explanation might be that if capsaicin triggers release of retrograde messengers from postsynaptic neurons that directly promote fusion of vesicles irrespective of Ca2+ concentration within the terminals, this could enhance the action potential-independent glutamate release (i.e., the majority of sEPSCs under our circumstances). Although TRPV1Rs have been shown to be located in the dendrites and somata of hippocampal principal cells (Cristino et al., 2006), these receptors are unlikely to be involved in the enhancement of sEPSC frequency, since the same effect was observed in the TRPV1R knockout mice. Further studies should clarify the possible mechanisms underlying the capscaicin-induced changes in excitatory neurotransmission.
In summary, our data presented here suggest that TRPV1Rs may not be the target molecules for capsaicin regulating glutamatergic synapses in the hippocampal network, although they might have a function during pathological conditions like fever or osmotic changes in extracellular space (Caterina et al., 1997; Gavva et al., 2007; Liu et al., 2007).
|
[
"excitatory synapses",
"granule cell",
"dentate gyrus",
"glutamate",
"transmitter release",
"brain slices"
] |
[
"P",
"P",
"P",
"P",
"P",
"R"
] |
Anal_Bioanal_Chem-4-1-2262919
|
Aptamers as molecular recognition elements for electrical nanobiosensors
|
Recent advances in nanotechnology have enabled the development of nanoscale sensors that outperform conventional biosensors. This review summarizes the nanoscale biosensors that use aptamers as molecular recognition elements. The advantages of aptamers over antibodies as sensors are highlighted. These advantages are especially apparent with electrical sensors such as electrochemical sensors or those using field-effect transistors.
Introduction
With the completion of the Human Genome Project, we have a better understanding of disease-related biomaterials such as DNA, RNA, proteins and small biological molecules. Such knowledge can impact strongly on human health, because the identification of a disease-specific marker at an early stage of a disease can greatly improve the clinical success rate, and improve disease-related mortality. For early detection of disease, it is necessary to locate minute amounts of disease-related proteins, peptides or other small molecules in a patient’s bloodstream or body fluid. Nanotechnology combined with biotechnology is expected to provide paradigm-breaking solutions in this area, as these technologies are extremely sensitive [1]. Already, nanoscale materials such as quantum dots [2] and metallic nanoparticles [3, 4], as well as nanodevices such as the nanocantilever [5], and nano field-effect transistors (FETs) [6–9] show potential for detecting tiny molecular signals.
In this review, we concentrate on nanoscale electrical biosensors with aptamers as molecular recognition elements. Biosensors are devices that can be used to detect the presence of a target analyte. As shown in Fig. 1, biosensors are built up from a sample delivery component, a molecular recognition subsystem, a transducer changing the binding or reaction of target molecules into a measurable physical signal (mass, charge, heat or light) and a signal-processing unit. We are particularly interested here in electrical transducers, because they provide the most promising solutions for point-of-care disease detection. Biosensors using electrical transducers are rapid and convenient, and they do not require heavy optical instruments or expensive measuring devices.
Fig. 1A biosensor. MIP molecularly imprinted polymer
Aptamers as molecular recognition elements
For the highly sensitive detection of a target analyte, the selectivity of the recognition element is vital in sensor performance. Until recently, immunosensors using antibody–antigen recognition were employed for biodetection. However, highly sensitive and stable biosensors may use different recognition elements. In the 1980s, functional nucleic acids that can specifically bind and regulate viral or celluar proteins with high affinity were discovered [10]. The major breakthrough for functional nucleic acids occurred in 1990. First, the presence of small functional RNA molecules (TAR aptamers) that can inhibit viral replication [11] was revealed; they can effectively bind and inhibit the activity of viral protein, thereby raising the possibility that they may be used as therapeutic agents. In the second study in 1990, Tuerk and Gold [12] established the in vitro screening process termed "systematic evolution of ligand by exponential enrichment" (SELEX), to identify the RNA ligands with affinity for T4 DNA polymerase and various organic dyes [12]. Such nucleic acids were termed "aptamers" [from the Latin aptus (fit or appropriate)] by Ellington and Szostak [13].
The SELEX process is a technique for screening a very large library of oligonucleotides with random sequences by iterative cycles of selection and amplification. Figure 2 shows a schematic diagram of the SELEX process [14]. The starting pool must be large enough to generate a high probability of producing a desired aptamer (normally around 1015 different sequences). The initial library of random-sequence oligonucleotides, which is obtained through combinatorial chemical synthesis, is incubated with a target of interest. Oligonucleotides showing affinity for the target are partitioned by affinity chromatography or filtration, and then amplified by PCR (for DNA libraries) or reverse transcription PCR (for RNA libraries) to create a new pool enriched in those oligonucleotides having a higher affinity for the target. As this cycle of selection and amplification is repeated, the abundance of the high-affinity oligonucleotides increases exponentially. Negative selection and counterselection are often employed in order to remove aptamers which bind to supports and molecules similar to the target, respectively. Typically after eight to 15 cycles, cloning and sequencing of the enriched library are carried out, revealing the sequence of oligonucleotides highly specific to the target. Once the sequence information has been obtained, the desired aptamer can be readily produced by chemical synthesis. For a given target molecule, DNA aptamers and RNA aptamers can be generated. However, they are quite different in sequence and three-dimensional structure since DNA lacks the 2′ hydroxyl group of the RNA. For example, both the DNA version of the RNA aptamer for ATP and the RNA version of the DNA aptamer for ATP do not recognize ATP [15].
Fig. 2The systematic evolution of ligand by exponential enrichment (SELEX) process [14]
In Table 1, we summarize the advantages of aptamers over antibodies and the yellow-colored rows detail special advantages of aptamers as recognition elements [16]. First, aptamers possess high target affinity. It has been generally recognized that aptamer affinity is comparable to or even higher than that of antibodies, even though this is not always true. Unlike antibodies, aptamers can distinguish between targets even if the targets are in the same family. For example, anti-caffeine aptamers have lower affinity (by a factor of 104) for theophylline, despite the two compounds differing by only a single methyl group [17]. Because aptamers are selected using an in vitro selection process (SELEX), they are cost-effective, have uniform activity, and antitoxin aptamers can be readily synthesized. Antitoxin antibodies are more difficult to prepare, since they are cultured in animal cells. The immobilization of the recognition element can be crucial in biosensors, and this is easier with aptamers than with antibodies, because chemical modification of nucleic acids is simple and straightforward compared with that of antibodies. One of the biggest advantages of aptamers lies in their high stability. Being nucleic acids, aptamers are highly stable, and therefore have unlimited shelf lives. Also, because aptamers can undergo reversible changes of conformation with variations in temperature or salt concentration, aptamer-based sensors are potentially recyclable. The major limitation of aptamers (especially RNA aptamers) as molecular recognition elements is degradation by nucleases. However, it was shown that the problem can be overcome through chemical modifications of the sugar at the 2′-position [18] or use of mirror-image analogs [19]. Recently chemically modified nucleic acid aptamers for in vitro selections have been reviewed [20].
Table 1Advantages of aptamers over antibodies [16]
As aptamers show high target affinity and high target specificity, they can be used as therapeutic agents, in affinity columns and as recognition elements in sensors detecting a wide range of targets. There are virtually no limits of targets for aptamers; small molecules [14], proteins [15] and antibiotics [21]. The first use of an aptamer as a molecular recognition element in sensors was reported in 1996, when fluorescent-tagged aptamers were employed for optical detection of human neutrophil elastase [22]. Since then, various transducers employing aptamers for recognition have been reported. These include the quartz crystal microbalance [23], the surface plasmon resonance technique [24], evanescent-wave-induced fluorescence [25], microcantilevers [26], electrochemical sensors [27–35] and FETs fabricated with single-walled carbon nanotubes (SWNTs) [36, 37]. In this review, we summarize recent advances in electrochemical and FET types of aptamer sensors, focusing on the high sensitivity and recyclability of the aptamer-containing devices.
Electrical aptamer sensors
Electrochemical aptamer sensors
Typical electrochemical sensors operate by reacting with an analyte of interest and producing an electrical signal proportional to the analyte concentration. Many papers have reported the fabrication of electrochemical aptamer sensors where specific binding of an analyte to an aptamer on an electrode can produce an electrical signal [27–35]. Some authors use sandwich assays, employing a secondary aptamer for signal enhancement [28, 29]. However, a technique that is label-free and fast in response is favored especially for electrical types of biosensors. A difficulty with sandwich assays has recently been removed by Xiao et al. [31], who developed a “signal-on” architecture. Earlier electrochemical aptamer sensors used a “signal-off” architecture [30], in which the binding of the target analyte resulted in a decrease in the signal. In such cases, no more than 100% of the signal can be suppressed by target binding, and hence the dynamic range of the sensor is limited. Also, it is difficult to distinguish a real binding event from a false signal originating from contaminants or sensor degradation. To fabricate the signal-on electrochemical aptamer sensor, Xiao et al. introduced a short DNA sequence tagged with methylene blue (MB) that can hybridize with an aptamer. A single-stranded DNA aptamer, which binds to thrombin, was chosen as a model system. Thrombin is a proteolytic enzyme which facilitates blood clotting by converting fibrinogen into fibrin and is also regarded as a tumor marker in the diagnosis of pulmonary metastasis [38]. Figure 3 shows a schematic diagram of the signal-on aptamer sensor. As shown, the sensor is constructed by immobilizing a thiolated thrombin aptamer to a gold electrode, and an MB-tagged partially complementary DNA is added to form a double-stranded complex. In this configuration, owing to the large persistence length of double-stranded DNA, the MB moiety is far away from the sensor surface, and only a small Faradaic current is observed. The persistence length, defined as the distance over which the direction of a polymer segment persists, provides a measure of elasticity in polymer chains. In the case of double-stranded DNA, the persistence length is about 50 nm, or 150 bp [39]. Since the thrombin aptamer is much shorter than 150 bp [40] (the recognition part is only 15 bp in length), the complex would be in a “standing” position. When thrombin is introduced to such a complex, the thrombin aptamer binds to thrombin, and the MB-tagged single-stranded DNA approaches the electrode surface, improving the electron transfer with an increase of current. The persistence length of a single-stranded DNA is 1.5 nm in 2 M NaCl solution, and 3 nm in 25 mM NaCl solution [41].
Fig. 3Thrombin binding with an engineered aptamer. MB methylene blue. (Reprinted with permission from [31], copyright 2005 American Chemical Society)
In the work of Xiao et al., a threefold signal gain was observed with 260 nM thrombin, while only a 40% signal decrease was observed using a signal-off sensor. The detection limit of the sensor thus rose to 3 nM as shown in Fig. 4. Such a high sensitivity is comparable even with that given by sensors using amplification by labeling [28, 29] or optical sensors using evanescent-wave-induced fluorescence anisotropy [25]. Table 2 summarizes the electrochemical aptamer sensors available for thrombin.
Fig. 4Electrochemical signal from the signal-on aptamer sensor. (Reprinted with permission from [31], copyright 2005 American Chemical Society)Table 2The detection limits of an electrochemical aptamer sensor for thrombin Detection limitCharacteristicsReference13 nMUses an MB-tagged "holding" DNA sequence[31]280 nM/3.5 nMPeroxidase-labeled thrombin gives an 80 nM detection limit. Biotin-labeled aptamer with horse radish peroxidase labeled streptavidin gives a detection limit of 3.5 nM[27]31 μMSandwich assay using two aptamers with different binding sites. Secondary aptamer labeled with glucose dehydrogenase[28]410 nMSame as for 3, but pyrroquinoline quinine glucose dehydrogenase is used for labeling[29]56.4 nMMB-tagged thrombin aptamer (signal-off sensor)[30]60.1–0.15 nMThrombin aptamer with ferrocene moiety (square-wave voltametry or chronopotetiometry used for the measurement)[32]711 nMMB-intercalated thrombin aptamer[25]MB methylene blue
Owing to the unique three-dimensional structure of an aptamer, it is not even necessary to use "holding" DNA sequences. Baker et al. [34] reported the rapid, label-free detection of cocaine in both adulterated samples and biological fluid. The cocaine aptamer engineered by Stojanovic et al. [42] was used in this work. This aptamer is thought to form a partially unfolded structure, with only one of its three double-stranded stems intact, when target molecules are absent. In the presence of target, the aptamer folds into the cocaine-binding three-way junction structure of Fig. 5a. With this unique approach, Baker et al. reported the detection of 500 μM cocaine in biological fluid, even when contaminants were present. The ability to detect cocaine in undiluted biological fluids (blood, for example) is crucial for simple forensic detection technology. The binding of small molecules to an aptamer is highly reversible, so it is possible to regenerate the sensor. Baker et al. [34] could regenerate the sensor to a degree of 99% by simply washing a used sensor with buffer.
Fig. 5a Binding of coccaine with an MB-tagged aptamer. b Binding of platelet-derived growth factor (PDGF) with an MB-tagged aptamer. (a Reprinted with permission from [34], copyright 2006 American Chemical Society; b reprinted with permission from [35], copyright 2007 American Chemical Society)
Regeneration of an aptamer sensor is also possible after protein binding. Lai et al. [35] reported the detection of picomolar levels of platelet-derived growth factor (PDGF) directly in blood serum. PDGF is a protein that regulates cell growth and division, and is very important for blood vessel formation (angiogenesis). While generally undetectable in normal cells, PDGF is overexpressed in a variety of disease such as atherosclerosis, and human tumors including sarcomas and glioblastomas [14, 43]. As shown in Fig. 5b, specific binding of PDGF to the aptamer decreases the distance between MB and the electrode and consequently increases the electrical signal. The sensor could be also regenerated by 4-min incubation in 10% (w/v) sodium dodecyl sulfate.
In this section, we have shown that engineered aptamers, employed as recognition elements in electrochemical transducers, can greatly improve sensor performance. It is an advantage of aptamers over antibodies that aptamers can be engineered to either enhance or activate signals. Recently, aptazymes (aptamers with catalytic activities) have been reported, which combine an enzymatic role with a recognition function [44]. Carefully designed aptazymes can be developed as sensors that do not require labeling. An electronic (electrochemical) sensor approach with carefully designed aptamers offer potential for highly sensitive, selective and reusable sensor platforms.
Aptamer sensors with carbon nanotube field-effect transistor transducers
Since the first report on the biosensing ability of Si nanowire in 2001 [45] and the development of the chemical sensor operation of carbon nanotubes in 2000 [46], extensive research efforts have been initiated aimed at developing nanotube-based or nanowire-based sensors. Owing to the extremely large surface-to-volume ratio of one-dimensional nanostructures, it is possible to develop sensors with exquisite sensitivity. This is especially true in the case of carbon nanotubes, where all constituent atoms are at the surface. Any small disturbance of or adsorption to the surface could result in a large change in electrical conductance. Already, biosensors fabricated with semiconducting nanowires or nanotubes have been used as sensors that can detect a single virus [6], small molecules [47] or minute amounts of proteins in solution [7–9].
FET sensors measure a change in the surface potential occurring during the binding between an analyte and a recognition element. Figure 6 shows a schematic diagram of the working principle of carbon-nanotube-based FET sensors. FET sensors “feel” the charges when target molecules bind with probe molecules immobilized on the sensor surfaces. Such an effect is often referred to as an “electrostatic gating effect” since the bound target molecules behave as an additional gate electrode. In the case of carbon nanotubes, SWNTs behave as p-type semiconductors in ambient atmosphere. In p-type FETs, positive gate voltage decreases the conductance (depletion) and negative gate voltages increase the conductance (accumulation) from the device. Therefore, if positively charged molecules bind with probe molecules on SWNT-FETs, this event appears as a decrease of conductance, while binding of negatively charged molecules appears as an increase of conductance in SWNT-FETs as denoted in Fig. 6.
Fig. 6The working principle of field-effect transistor (FET) sensors based on carbon nanotubes
Using FETs for immunosensing was first proposed by Shenk [48] in 1978. It was suggested that FETs could be used to detect surface polarization owing to the formation of an antibody and an antigen. This concept failed, however, because small ions in the solution caused a screening effect inhibiting detection of the changes [49]. Figure 7 shows the charged semiconductor surface in solution. As shown in Fig. 7, when a charged surface (this might be a metal electrode, a semiconductor or a charged dielectric) is brought into contact with an ionic solution, the surface charge potential is dissipated by redistribution of counter ions in the solution. The domains where the variation of electric potential occurs are called electrical double layers, and the depth of an electrical double layer can be determined by calculating the Debye screening length. The Debye screening length is the scale over which mobile charge carriers screen the electric field. In an electrolyte, the Debye screening length can be defined as follows:
Fig. 7Electrical double layer at the sensor surface
Here the Bjerrum length lB is defined as
and is 0.7 nm at standard pressure and temperature. kB is the Boltzmann constant, k is defined as 1/4πɛ0, ρi is the number density of ions and the zi values are valencies of various ions.
The Debye screening length, therefore, is proportional to the inverse square root of the ion number density, i.e., to ionic strength. For pure water, the Debye screening length is about 1 μm, while it is only 0.3 nm in 1 M NaCl solutions. In 1 M NaCl solution, therefore, two ions or charged species separated by 1 nm do not interact with each other. If we consider the sensor surface, any interactions occurring outside the Debye screening length cannot be detected using FET sensors. The biggest disadvantage of FET sensors is that they may not work with blood samples or body fluids, because the ionic strength of such physiological solutions is about 150 mM, which yields a Debye screening length of about 1 nm.
Until recently, FET-type sensors have employed an antibody as the recognition receptor. However, the typical size of an antibody is about 10–15 nm [50, 51]. In that case, the binding of an antigen with an antibody occurs far outside the Debye screening length in physiological solutions. Molecular recognition elements with high specificity, selectivity and small sizes are essential for applications of FET-type sensors. In this respect, aptamers are highly suitable. Figure 8 shows the typical size ranges of antibodies and aptamers [52]. Because aptamers are mostly short nucleic acids ranging from 10 to 60 bp in length, the binding of target molecules with aptamers can occur inside the electrical double layer, even in 20–50 mM salt solutions.
Fig. 8Size comparison of an aptamer with an antibody. IgG immunoglobulin G. (Reprinted from [52], copyright 2006 Elsevier)
Such a biggest merit of aptamers in FET sensors was first reported by So et al. [36] using a carbon nanotube biosensor with thrombin aptamers as molecular recognition elements. They also demonstrated the sensitivity and selectivity of aptamers in SWNT-FETs. Anti-thrombin aptamers (5′-GGT TGG TGT GGT TGG-NH2-3′) were immobilized onto the surface of the SWNT-FET, fabricated using a patterned growth technique, with carbonyl diimidazole–Tween 20 as the linking molecules (Fig. 9a). The authors observed an abrupt decrease in conductance upon reaction with thrombin, while almost no changes were observed with elastase (Fig. 9b). Elastase is a member of the thrombin family and has almost the same molecular weight and isoelectric point as thrombin. Moreover, the sensor could be regenerated by simple washing with 6 M guanidine hydrochloride solution. The authors regenerated the sensor more than five times without significant loss of activity. As shown in Fig. 9c, the authors measured the sensitivities of SWNT-FET sensors with aptamers as recognition elements. Before each measurement, the sensor was "reset" by washing it with guanidine hydrochloride solution. The lowest detection limit was around 10 nM thrombin, and the sensor signal was saturated at about 100 nM thrombin.
Fig. 9a Single-walled carbon nanotube (SWNT) FET sensor with aptamers as molecular recognition elements. b Selectivity of an anti-thrombin aptamer-functionalized SWNT-FET. c Sensitivity of an anti-thrombin aptamer-functionalized SWNT-FET. The sensor was reset by washing it with 6 M guanidine hydrochloride solution before each measurement. CDI carbonyl diimidazole. (Reprinted with permission from [36], copyright 2005 American Chemical Society)
Maehashi et al. [37] have recently addressed the size issue of recognition elements in FET-based sensors. In their work, immunoglobulin E (IgE) aptamer (5′-NH2-GCG CGG GGC ACG TTT ATC CGT CCC TCC TAG TGG CGT GCC CCG CGC-3′) and monoclonal anti-IgE antibody (IgE-mAb) were immobilized on carbon nanotubes, with the aid of the linking molecule 1-pyrenebutanoic acid succinimidyl ester. Figure 10 a and b shows the responses of the aptamer-functionalized and antibody-functionalized sensors, respectively, when they were exposed to known concentrations of target IgE in 10 mM phosphate-buffered saline (PBS). As shown in Fig. 10, the aptamer-functionalized sensor exhibited small changes with only 250 pM target protein molecules, while a negligible change in conductance was observed in the antibody-funtionalized sensor even with 140 nM IgE. In 10 mM PBS, the Debye screening length would be roughly 3 nm, and the binding of IgE with the aptamer would be safely inside the electrical double layer. With IgE-mAb, however, there is little hope that binding can be detected in the buffer solution used for the experiment, because the antibody molecule is about 10 nm long [53].
Fig. 10The sensor signal measured from a an aptamer-functionalized SWNT-FET and b an antibody-functionalized SWNT-FET. IgE immunoglobulin E, PBS phosphate-buffred saline. (Reprinted with permission from [37], copyright 2007 American Chemical Society)
As discussed, aptamers may allow a wider range of analytes in FET sensors owing to their small size. SWNT-FET sensors with aptamers as recognition elements showed high sensitivity and selectivity, and could be readily regenerated. Highly sensitive carbon nanotube sensors combined with small, economic, highly selective and stable aptamers could provide cost-effective point-of-care testing devices.
Concluding remarks
This review has focused on the advantages of electrochemical and FET sensor types with aptamers as recognition elements. Through use of aptamers for recognition, no labeling is required for electrochemical sensors, and signal-on architecture that is only possible with aptamers has made possible the sensitivity improvements. Also, as pointed out already, sensors can be recycled because of the reversibility of aptamer configurations. Alternatively, bound proteins may be simply washed off, without damaging the aptamers. In immunosensing, by contrast, it is practically impossible to remove bound antigens from antibodies without damaging the antibodies, because both antibodies and antigens are proteins in nature. Aptamers have proved their superiority over antibodies in nano-FET sensors. When no signal can be measured with an antibody–antigen pair because of the large size of the antibody, the binding of small aptamers to various targets can occur inside the electrical double layer where the nanotube can "feel" the change.
Owing to the rapid aging of society, there is great demand for paradigm-breaking biosensors to detect the onset of disease with speed, convenience and accuracy. In a few more years, electrical nanosensors, with engineered aptamers, will become prominent in the market. They will be small in size, high in sensitivity and competitive in price.
|
[
"aptamers",
"nanobiosensors",
"electrochemical aptamer sensors",
"field-effect transistor aptamer sensors"
] |
[
"P",
"P",
"P",
"R"
] |
Community_Ment_Health_J-2-2-1705498
|
Mental Health and Substance Abuse Services Preferences among American Indian People of the Northern Midwest
|
ABSTRACT
INTRODUCTION
A 2001 supplemental report on mental health to the Surgeon General describes the current United States public health perspective as a “population-based approach (that) is concerned with the health of the entire population, including its link to the physical, psychological, cultural, and social environments in which people live, work, and go to school (Dept of Health and Human Services (DHHS), 2001).” This statement reflects a concept of holistic healing in a time of growing acceptance and utilization of alternative or complimentary health services beyond those available within professional medical systems. Rodenhauser (1994) notes that although psychiatry as a specific medical field shows increasing cultural awareness regarding clinical sensitivity within the academic literature, there has been little to no evidence of cultural factors found within mainstream journals for health administrators. In addition, very little research is available concerning the co-utilization of “alternative” and western medical approaches, let alone an examination of this overlap within specific cultures (Gurley et al., 2001; Rhoades & Rhoades, 2000; but see Novins, Beals, Moore, Spicer, & Manson, 2004). An understanding of these services may be especially important within an American Indian cultural context where traditions often involve complex ceremonies and beliefs that require consideration far beyond that of mainstream religion (Rhoades & Rhoades, 2000).
The purpose of this study is to examine factors that influence the choice between traditional cultural and western mental health and substance use associated care among American Indians from reservations in the northern Midwest. Specifically, we examine preference in terms of perceived effectiveness and actual utilization of traditional vs. Western-based service outlets.
Barriers to Health Services
American Indians and Alaskan Natives have been found to report highest rates (12.9%) of frequent mental distress among all ethnic groups, including Hispanics (10.3%), African Americans (9.7%), Whites (8.3%), and Asian/Pacific Islanders (6.1%) (Centers for Disease Control (CDC), 1998). The 2000–2001 National Household Survey on Drug Abuse found American Indians/Alaskan Natives to be second only to multi-ethnic individuals to have received mental health treatment in the past year. At the same time, American Indian/Alaska Native respondents were second highest among all ethnic groups who had not received mental health care to report perceived unmet service needs (Barker et al., 2004). A similar pattern is found for substance abuse where again, American Indians/Alaskan Natives are second to multi-ethnic persons in terms of prevalence of illicit drug or alcohol dependence (Substance Abuse and Mental Health Services Administration (SAMHSA), 2002).
A majority of Indian people within the United States receive physical and mental healthcare from Indian Health Services (IHS) (Gone, 2004). IHS’s Indian Health Care System served around 1.6 million American Indians/Alaska Natives residing on or near reservations in 2003 (IHS, 2003). Over one-third of service utilization within this system involves mental health and social service related concerns (IHS, 2002). Although Indian Health Services is meant “to raise the physical, mental, social, and spiritual health of American Indiana and Alaska Natives to the highest level” (IHS, 2002), a variety of barriers may impede American Indians’ ability to receive adequate health care. Specific to mental health care delivery, and often overlapping with the characteristics of healthcare in general, Rodenhauser (1994) characterizes barriers into two major groups: (1) those within the current medical system, including under-funding and staffing issues, inconsistent services, insensitivities, fear of litigation and non-compliance with organizational requirements, and staff burnout; and (2) those inherent to American Indian/Alaska Native cultures, such as fear, high rates of accidents/violence, intrinsic conditions (i.e. denial, helplessness and hopelessness), and mistrust of government agencies.
Traditional Help Seeking
Research and anecdotal evidence suggests that many American Indians report health related help-seeking via culturally traditional outlets. For example, Marbella, Harris, Diehr, Ignace, and Ignace (1998) surveyed 150 American Indian adult patients at the Milwaukee Indian Health Center and found that 38% had sought help from a traditional healer in addition to physician-administered care. Furthermore, 86% of those interviewed said that they would consider seeking a traditional healer in the future. In instances where traditional healer’s and physician’s advice differed, patients said they would rely on their healer’s advice 61.4% of the time. In a separate study including respondents from 2 different American Indian tribes, Novins, Beals, Moore, Spicer, and Manson (2004) found traditional healing options to be prevalent and important to both tribes, whether used alone or in combination with biomedical services.
Although traditional beliefs and practices can guide some American Indians through mental and physical illnesses and healing, there is disagreement concerning the degree of connection between traditional and professional medical services at an institutional level. A number of researchers discuss a gap wherein the processes of traditional and mainstream professional healing should be linked (Marbella et al., 1998; Mohatt & Varvin, 1998; Novins, Beals, Sack, & Manson, 2000; Novins, Fleming, Beals, & Manson, 2000). Conversely, others feel that formal funding for traditional practices is difficult and perhaps inappropriate. Payment to a traditional healer by a patient may be compromised in a situation where billing must occur via a third party. Issues surrounding payment are complicated further considering the need to establish credentials for employees of government funded agencies like IHS (Rhoades & Rhoades, 2000). It is also important to consider the heterogeneity of the over 300 different tribal or language groups from which American Indian individuals belong (Beauvais, 1998). Tribal differences extend to various opinions regarding an integration of the two health care systems (Rhoades & Rhoades, 2000; see also Novins et al., 2004).
Despite some documentation on the preference of traditional vs. western medical treatment among American Indians who receive both, we found little empirical evidence that directly tests the factors that affect this choice (but see Gurley et al., 2001; Novins et al., 2004; Robin, Chester, Rasmussen, Jaranson, & Goldman, 1997). In addition, there is a lack of empirical research surrounding American Indian mental health/substance use related services utilization (Manson, 2000), and what work has been done typically focuses on Indian children (i.e. Novins, Harman, Mitchell, & Manson, 1996; Costello, Farmer, Angold, Burns, & Erkanli, 1997; Novins et al., 2000). This research begins to address such gaps in the literature by examining predictors of service preferences among a sample of American Indian adults.
Hypotheses
In a review of the literature, Manson (2000) includes differences related to culture (i.e. acculturation vs. enculturation) as important predictors of formal mental health service utilization. Elsewhere, higher levels of social support (here measured as community support) have been shown to be related to receiving a treatment recommendation for substance use problems (Novins et al., 1996); in turn, being recommended for treatment may increase actual service utilization. In addition, rates of mental health treatment have been found to be higher for adults with a lower perceived health status (Barker et al., 2004).
Based on these findings and coupled with the barriers to care discussed earlier, we hypothesize that higher levels of enculturation, experiencing discrimination, and living on a reservation will be positively related to perceived effectiveness or preferences for informal services and negatively related to formal services. In addition, self-reported physical health is hypothesized to be negatively related to preferences of both service types (Barker et al., 2004). We also predict that higher levels of social support will be positively related to use of both informal and formal service outlets. Lastly, we control for the effects of age, gender, education, and employment on service utilization preferences.
METHODS
Sample
These data were collected as part of the “Healing Pathways Project,” a 3-year lagged sequential study currently underway on four American Indian reservations in the Northern Midwest and five Canadian First Nation reserves. Because of national differences in health care systems, the data presented here includes only that from U.S. reservations. The data are from wave one of the study collected on two U.S. reservations from February through October 2002, and wave one on a second pair of U.S. reservations collected from February through October 2003. The reservations share a common cultural tradition and language with minor regional variations in dialects. The sample represents one the most populous Native cultures in the United States and Canada.
The project was designed in partnership with the participating reservations and reserves. Prior to the application funding, the research team was invited to work on these reservations, and tribal resolutions were obtained. As part of this agreement, the researchers promised that participating reservation names would be kept confidential in published reports. An advisory board was appointed by the tribal council at each participating reservation and is responsible for advising on difficult personnel problems, questionnaire development, reading reports for respectful writing, and assuring that published reports protected the identity of the respondents and the culture. Upon advisory board approval of the questionnaires, the study procedures and questionnaires were submitted for review and approved by an Institutional Review Board. All participating staff on the reservations were approved by the advisory board and were either tribal members themselves or non-members who are spouses of tribal members. To ensure quality of data collection, all interviewers underwent special training for conducting pencil-and-paper and computer-assisted personal interviewing for diagnostic measures, including feedback sessions. In addition, all of the interviewers completed a required human subject’s protection training that emphasized the importance of confidentiality and taught procedures to maintain the confidentiality of data.
Each tribe provided lists of families of enrolled children aged 10–12 years who lived on or proximate to (within 50 miles) the reservation or reserve. We attempted to contact all families with a target child within the specified age range. Families were recruited via personal visits from Native interviewers during which they were given an explanation of the project, a gift of wild rice, and an invitation to participate. After agreement to participate and later completion of interviews, each participating family member received $40 for their time. This recruitment procedure resulted in an overall response rate of 79.4%.
Sample Characteristics
The sample for this analysis is made up individuals from the U.S. reservations only and consisted of 865 parents/caretakers (264 males and 601 females) of tribally enrolled children aged 10–12 years. Fathers/male caretakers ranged in age from 21 to 70 years with an average age of 42 years; mothers/female caretakers ranged in age from 17 to 78 years with an average of 39 years.
Measurement
To assess the perceived effectiveness of health service providers, respondents were given a series of questions asking them how effective various service outlets would be if ever there were to have (1) an emotional problem, or (2) a substance abuse problem. The response categories range from not at all effective to extremely effective. Traditional/informal services include family, talking to an elder, sweat lodge, pipe ceremony, offering tobacco, traditional healer, traditional ceremony, and healing circle. Formal services include Indian Health Service (IHS), doctor, psychologist, social worker, counselor, psychiatrist, chemical dependency counselor, and nurse. Respondents were asked to evaluate the perceived effectiveness of formal services both on and off the reservation. For multivariate analysis a mean effectiveness score was computed for three categories of services: (1) traditional, (2) formal on reservation, and (3) formal off reservation, all within each realm of service need (emotional and substance abuse). Each mean score has a range from 1 to 5, with higher scores indicating greater perceived effectiveness.
Enculturation is a latent construct assessed by three basic elements: (1) participation in traditional activities, (2) identification with American Indian culture, and (3) traditional spirituality (see Costello et al., 1997; Whitbeck, McMorris, Hoyt, Stubben, & LaFromboise, 2002). The enculturation scale has high internal consistency (Cronbach’s α = .79), and the standardized version of the scale is used in these analyses.
Perceived discrimination is measured with an 11-item scale designed to tap a range of potential types and sources of discrimination. Respondents were asked to report how often they had experienced specific instances of discrimination. Those items included in the scale were; how often someone said something derogatory or insulting because of your race; how often a store owner, sales clerk or person working at a business has treated you in a disrespectful way because of your race; how often the police have hassled you because of your race; how often you have been ignored or exclude from an activity because of your race; how often someone has yelled a racial slur or racial insult at you; how often someone threatened physical harm because of your race; how often someone suspected you of doing something wrong because of your race; how often you have been treated unfairly because of your race; how often you have encountered whites who didn’t expect you to do well because of your race; how often someone discouraged you from trying to achieve an important goal because of your race; and how often you have been treated unfairly in the courts because of your race. The four response categories range from never to always, with the mean of all 11 items used in these analyses. The scale has a high internal reliability with Cronbach’s α of .90.
Social support is measured with a 17-item scale in which respondents were read statements regarding community and neighborly cohesiveness and conflict. Statements include: this is a close knit community; people around here are willing to help their neighbors; there are adults in this community that children can look up to; there are long standing family grudges in the community; the community is split by politics; the community is split by alcohol or drugs; and so on. Respondents were asked to answer true or false to each statement. Items indicative of community conflict were reverse coded, and then all items were summed so that higher values represent higher community support. Cronbach’s α for this scale is .86.
Health status was assessed by response to the question “In general, how well would you say your health is?” Response categories ranged from excellent (5) to poor (1). Values are coded so that higher scores indicate better health.
Several control variables are included in the present analysis. Gender is a dummy variable coded as male = 0 and female = 1. Education is a five category item ranging from less than high school to an advanced degree. Employment is a dichotomous variable indicating if the respondents are employed full-time. Those with full-time employment are coded as 1. To assess the difference between those enrolled tribal members living on and off the reservation respondents were asked if they currently lived on the reservation at the time of their interview. Those who lived off of the reservation were given a value of 1.
RESULTS
Descriptive Characteristics
The adults who comprised this sample had an average educational level falling between a high school diploma and at least some secondary school experience (mean = 2.4; SD = .87), and more than half (57%) were employed full-time. Very few of the adults lived off of the reservation (12%) at the time of our interviews. The self-reported health status of the adults in our sample averaged between ‘good’ and ‘very good’ (mean = 3.3; SD = 1.0), and the overall level of reported social support in the communities was 8.4 (SD = 4.1). Enculturation is a standardized variable in these analyses (mean = 0; SD = 1), while the mean level of perceived discrimination was 1.5 (SD = .69), indicating that the average response across all of the discrimination questions fell between ‘never’ and ‘a few times.’
Among the dependent variables concerning service preferences among respondents, the highest mean scores were found for informal/traditional services for both mental health (mean = 3.06; SD = 1.06) and substance use (mean = 3.09; SD = 1.15) related issues. Next highest were mean preference scores for on-reservation formal services (mental health = 2.51; SD = 1.01; substance use = 2.53; SD = 1.09), followed lastly by mean preference ratings for formal off-reservation care (mental health = 2.32; SD = 1.03; substance use = 2.34; SD = 1.09).
Perceived Effectiveness of Services
Figure 1 illustrates the perceived effectiveness of services that are commonly used for mental health or substance abuse problems. The reported percentage indicates those respondents who felt the services would be very or extremely effective. There is a clear distinction of the perceived effectiveness between informal or traditional services and more formal services. Seventy-one percent of respondents felt that talking to a family member would be very or extremely effective. Nearly 60% felt talking to a tribal elder would be highly effective for dealing with mental health or substance problems. More traditional practices/services such as offering tobacco and praying, seeing a traditional healer, traditional ceremonies, healing circle, sweat lodge, and pipe ceremony all were rated between 30% and 50%. Speaking to a counselor on the reservation was the highest rated professional service with 32.9% of respondents indicating it would be very or extremely effective. Most of the on reservation professional services rated higher than services off the reservation but lower than more traditional services. Those services that were seen as least effective were those professional services located off of the reservation.FIGURE 1Perceived Effectiveness of Services
Use of Informal and Formal Services
Table 1 reports the coefficients from ordinary least squares regression analyses for predictors of respondent perceptions of the effectiveness (i.e. preferences) of informal traditional and formal services for substance abuse and mental health concerns. Because this sample contains some cases where two adult reporters were interviewed within a household, reported coefficients are based on standard errors that have been adjusted to account for the potential bias of nested designs (analyses performed using STATA 7.0). The same sets of variables significantly predict informal service preference for both mental health (MH, Model 1A) and substance use (SU, Model 1B) related problems. For both models, females were more likely to prefer informal services (MH: β = .06, p < .05; SU: β = .05, p < .10), as were the employed (MH: β = .07, p < .05; SU: β = .07, p < .05). For both Model 1A and 1B, higher rates of social support (MH: β = .13, p < .001; SU: β = .13, p < .001), higher enculturation scores (MH: β = .49, p < .001; SU: β = .44, p < .001), and higher reports of perceived discrimination (MH: β = .09, p < .01; SU: β = .11, p < .01) were all positively associated with informal service preferences.TABLE 1OLS Regression Models Predicting Perceived Effectiveness (Preferences) of Mental Health and Substance Use Related ServicesMental HealthSubstance Use1A2A3A1B2B3BInformalFormal (on res.)Formal (off res.)InformalFormal (on res.)Formal (off res.)BβBβBβBβBβBβAge − 0.003 − 0.030.010.05 − 0.002 − 0.020.0010.010.010.07+0.000.003Gender (female = 1)0.150.06*0.130.06+0.040.020.130.05+0.050.002 − 0.03 − 0.01Education0.040.030.140.12**0.210.18***0.040.030.170.13***0.190.15***Employment0.140.07*0.080.04 − 0.01 − 0.010.150.07*0.020.01 − 0.06 − 0.03Live off reservation0.170.050.180.060.280.09*0.070.020.270.08*0.470.14***Physical health status0.010.010.030.030.020.02 − 0.02 − .010.020.010.010.01Social support0.030.13***0.030.13***0.020.08*0.040.13***0.050.17***0.020.09***Enculturation0.520.49*** − 0.05 − 0.05 − 0.05 − 0.050.510.44*** − 0.05 − 0.05 − 0.08-.07*Discrimination0.140.09**0.110.07+ − 0.07 − 0.050.190.11**0.170.1** − 0.09 − 0.06Constant2.371.281.731.101.77R20.290.050.060.250.060.07Adjusted R20.280.040.050.240.050.06+p < .10; *p < .05; **p < .01; ***p < .001 (two-tail test).Sample size across models ranges from n = 803 to n = 822 after listwise deletion.
Moving to predictors of formal service preferences on reservations (Models 2A & 2B), higher levels of social support were a significant predictor in both the MH (β = .13, p < .001) and SU (β = .17, p < .001) models, as were higher rates of perceived discrimination (MH: β = .07, p < .10; SU: β = .10, p < .01) and education (MH: β = .12, p < .01; SU: β = .13, p < .001). Age is significantly associated with formal on-reservation service preferences only for substance use problems (β = .06, p < .10), whereas females are more likely than males to prefer on-reservation formal care in the case of mental health problems (β = .06, p < .10).
Models 3A and 3B illustrate the coefficients for predictors of formal off-reservation service preferences. Both models contain similar findings: higher education (MH: β = .18, p < .001; SU: β = .15, p < .001), living off of reservation lands (MH: β = .09, p < .05; SU: β = .14, p < .001), and higher levels of social support (MH: β = .08, p < .05; SU: β = .09, p < .05) were all positively associated with higher preferences for formal off-reservation services. In addition, Model 3B shows that those who scored higher on levels of traditional enculturation were less likely to perceive formal off-reservation care as effective in terms of substance use related problems (β = − .07, p < .05).
Across all of the models in Table 1, the independent variables explain proportionately more of the variance for the informal service models than those with formal services as the dependent variable. Adjusted R2 values for these models indicate that the predictors explain 28% and 24% of the variance in informal service preferences for MH and SU problems, respectively. These values for formal reservation-based care are 4% for MH and 5% for SU problems. The models for formal off-reservation care explain 5% of the variance in preferences for MH problems and 6% of the variance for SU problems.
DISCUSSION
The adults in our sample felt that more culturally traditional services (such as family or a traditional healer) would be more effective than either professional services on the reservation, or professional services off the reservation. Off reservation services were perceived to be least effective. These findings support literature that suggests American Indians generally place considerable trust in traditional practices (i.e. Marbella et al., 1998). The finding that mainstream professional services are perceived as less effective than traditional informal services and some of the on-reservation professional services underlies the previous discussion surrounding issues of trust and appropriateness of Western care in relation to American Indian cultures. It is also possible that a crisis-oriented care system may not be suited to adequately deal with mental health and substance abuse needs (Oetting & Beauvais, 1990–1991).
The multivariate analyses supported several of our hypotheses. Among the control variables, females were slightly more likely than males to report higher preference for both (MH & SA) types of informal care, as well as for on-reservation MH care. Though not explicitly hypothesized, these results correspond with previous help-seeking research that has found women more likely than men to view care as an appropriate remedy for psychological issues (Horwitz, 1987). Higher educational levels and employment were generally related to higher effectiveness ratings for services in this study, also corroborating previous research (see, for example, Horwitz, 1999, p. 66).
As expected, those who are more enculturated were more likely to utilize informal traditional services than those who are less enculturated. In addition, in the case of services for substance use disorders, higher levels of enculturation were negatively associated with preferences for formal off-reservation care (Model 3B). Those who lived off the reservation were more likely to use formal services, especially in terms of off-reservation care. Consistent with past research (Novins et al., 1996) and as hypothesized, perceived social support was positively associated with higher perceptions of effective care across all types of services. Our hypothesis that higher rates of perceived discrimination would be associated with preferences for informal or traditional care is supported here. We found similar effects for discrimination and formal on-reservation care. These findings are congruent with Rodenhauser’s (1994) description of the barriers to dominant-culture care in terms of mistrust of government agencies, all stemming from a history of colonization and continued institutional racism experienced by many American Indian peoples (see, for example, Duran & Duran, 1995). Overall, the results of this study highlight a preference for informal or culturally based care, especially among the enculturated.
The differences in the proportions of variance explained across each of the models in Table 1 warrant discussion. These analyses include only one measure of cultural identity: enculturation. As discussed, this measure was significantly associated with ratings of informal service preference, but did little in terms of predicting formal service preferences. It may be that an important predictor of formal services is acculturation, or perhaps even biculturalism. Future work might include such measures to further our understanding of how cultural identity affects service preference and utilization (see Novins et al., 2004).
Limitations
Although we interviewed a broad range of American Indian adults on multiple reservations that are dispersed geographically across two Midwestern states, these results pertain to a single culture and capture variations within this culture. We believe the findings represent the culture well, but they cannot be generalized to other Native cultures. A second caution regarding the sample is that it is made up of parents and caretakers of children aged 10–12 years. This could reflect a selection bias in that parents/caretakers may be more likely to utilize services and to be currently mentally healthy and alcohol and drug free than individuals with no children. Finally, all of our measures are based on self-reports. We did not have access to service utilization rates from local clinics or reports of services use from traditional healers.
Policy Implications
These findings have important policy implications for those who provide health services to American Indian people. Cultural traditions are very much alive on U.S. reservations and are preferred methods of healing for mental health or substance abuse problems. This should be taken into account when designing health services systems either by providing alternative services onsite or through creatively engaging informal traditional services. Simply acknowledging informal service providers by asking the patient if they have seen a traditional healer and then contacting him or her for an opinion would be an important step. Health providers could include key community spiritual leaders and healers on health advisory boards or create health partnerships that would encourage mutual referrals. Perhaps the place to begin would be to respectfully seek the advice of elders on decreasing barriers between the two.
Actual integration of services is already taking place on many reservations, but implementation is not always straight forward (Marabella et al., 1998; Mohatt & Varvin, 1998; Rhoades, & Rhoades, 2000). Issues about selection of traditional healers, and how to reimburse them such as whether they should charge fees for services and coverage for third party payments must be worked out culture by culture. Some view such integration as a potential acculturation risk.
Changing the current short comings in mental health delivery first involves acknowledging that parallel systems exist in some cultures. The impetus for change may have to come from formal services providers who should demonstrate respect for traditional healers, invite their advice, and consult with them if patient’s give permission. Clinicians may need to be proactive in offering their patients this option rather than waiting for the patient to request it. Creating this sensitivity may require specialized training for services providers on American Indian reservations. Grant funding groups may want to consider supporting model programs that bring together the two healing modalities.
Future Research
There is much to be done to increase American Indians’ confidence in health services systems. One avenue for this would be to work more closely with informal traditional services that are trusted. We need control group trials of innovative services models that cross over between formal medical services and traditional approaches to healing to assess the efficacy of combining the approaches. Continued failure to acknowledge the strong preferences for traditional ways is to ignore a valuable health resource.
|
[
"mental health",
"substance abuse",
"american indians",
"service utilization"
] |
[
"P",
"P",
"P",
"P"
] |
Matern_Child_Health_J-2-2-1592144
|
Family Planning Services: An Essential Component of Preconception Care
|
Family planning services are necessary for the widespread adoption of preconception care for two reasons. First, preconception care is more likely if pregnancies are planned, and family planning services encourage pregnancy planning. Second, family planning services usually include counseling, and counseling provides an opportunity to discuss the advantages of preconception care. However, the potential of family planning services to promote preconception care is limited by underutilization of these services and inadequate attention to preconception care during family planning visits. This article suggests ways to reduce these problems.
If all women of reproductive age, or at least those at elevated risk, are to benefit from preconception services, the use of family planning services must be increased and the content of such services expanded. Family planning services are essential for preconception care for at least two reasons. First, in the absence of such services, pregnancies will occur that have not benefited from preconception care. Preconception care during the reproductive years is dependent on women and men planning their pregnancies, not only in respect to their timing but also to health-related factors that would maximize the chances for a healthy pregnancy and a healthy infant. In the absence of such care, offered by family planning services, many pregnancies will not benefit from preconception advice. Second, family planning counseling provides an opportunity for promoting and providing preconception care. At the same time women and their partners are receiving advice about family planning, they can also receive instruction about the range of activities that lead to healthy pregnancies and healthy infants.
Unfortunately, several factors prevent family planning from reaching its maximal potential for preconception care. These include underutilization of family planning services and inadequate attention to preconception counseling during family planning visits.
Underutilization of family planning services
Two types of data point to underutilization of family planning services with implications for decreasing opportunities for preconception care: the percentage of women actually seeking family planning services and the percentage of unintended pregnancies.
Visits for family planning
The 2002 National Survey of Family Growth (NSFG) reported that 41.7% of women 15 to 44 years of age received at least one family planning service from a medical care provider in the 12 months prior to the interview [1]. This percentage is not as alarming as it might appear at first glance, because some of the women who did not seek family planning services already were pregnant, seeking to become pregnant, or infertile because of sterilization or other reasons. Nevertheless, this rather low percentage suggests that some women are not planning their pregnancies either deliberately or because they are experiencing problems obtaining family planning services.
Pregnancy planning
It is undoubtedly true that some women do not wish to plan the timing of their pregnancies. In many cases, these are married women who have religious objections to family planning. In Women of Crisis, one woman states “you don't think of life that way – of having children that way. You don't sit down and say you can afford to have a certain number of boys and girls; you have your children and try to do the best you can to be a good parent” [2]. This attitude is probably not widely held, rather most women and their partners have some preferences in terms of the timing of births. This may be expressed in terms of age, marital status, or attainment of some educational, career, or economic goal. Despite this, a large percentage of pregnancies are unintended, including those that were experienced earlier than wanted (mistimed) or those that were not wanted at the time they occurred or at any future time (unwanted).
According to the 2002 NSFG, 30.8% of all women 15 to 44 years of age had experienced an unintended birth at some time in their lives and, in the five years before the survey, 20.8% had had a mistimed birth and 14.1% an unwanted birth [3]. Using data from the 1982, 1988, and 1985 NSFGs, as well as abortion data, Henshaw estimated that in 1994, 49.2% of all pregnancies were unintended [4]. Employing a different set of questions to assess wantedness, the Pregnancy Risk Assessment Monitoring System (PRAMS) noted that in 1999 the percentage of unintended pregnancies resulting in live birth ranged from 33.7% to 52% across the 17 reporting states [5].
Although the relationship between use of a family planning method and pregnancy planning is sometimes tenuous - an analysis of the 1995 NSFG found that 30.9% of the women who stated that their pregnancies had resulted from a contraceptive failure nevertheless classified the pregnancy as intended [6]. These data suggest that if preconception care were to be practiced widely, a larger percentage of women will need to seek family planning services to avoid unplanned pregnancies.
Reasons for underutilization
Many reasons have been suggested for the underutilization of family planning services, including cost, availability, and limited contraceptive methods [7, 8].
Financial issues
The possible patient payment sources for family planning services include private health insurance, Medicaid, or out-of-pocket payments. The March 2005 Current Population Survey found 24.5% of women 18 to 20 years of age, 30.6% of those 21 to 24, 21.8% of those 25 to 34, and 17.1% of those 35 to 44 to be uninsured [9]. If these uninsured women want to receive family planning services, they must rely on their own funds or seek care from facilities that provide services free of charge or on a sliding fee scale. These include publicly supported facilities such as health department family planning clinics, community health centers and public hospitals, and facilities that combine public and private funding sources, such as Planned Parenthood centers, women's clinics, school-based health centers, and not-for-profit hospitals. Public funds for family planning services in such facilities come primarily from Medicaid and Title X (the family planning act). These funds are inadequate to meet the need. For women whose eligibility is due to a pregnancy, eligibility for all medical services, including family planning, ends 60 days postpartum, except in states with a waiver to extend the period. Title X tries to fill in the gaps in coverage but it is chronically underfunded. Even for those with employer-based health insurance, coverage for family planning services is not universal. In 2003, 93% of health plans offered an annual ob/gyn visit, 88% covered oral contraceptives; 87%, sterilization; and 72%, all five reversible contraceptives. HMOs were more likely to offer contraceptives and sterilization than conventional plans, PPOs or POSs [10].
Availability
Family planning facilities need to be easily available both physically and psychologically. This means that they should be located where they can be reached easily by private or public transportation, be open days and hours that are convenient for potential users, offer the full range of contraceptive methods, provide educational materials in the languages of their patients, and employ personnel who can speak the language of the users and whose attitudes and behaviors are warm, friendly, and culturally sensitive. A recent four-state survey found that distance to a publicly financed family planning facility was not associated with teen or unintended pregnancies, suggesting that geographic availability is no longer a problem for family planning facilities [11], although it still is for abortion services [12]. The same study found major differences in terms of the other availability factors, such as limited hours, few contraceptive choices, and lack of translators [13]. Such problems may be partially responsible for inadequate contraceptive use.
Limited contraceptive methods
Although the number of family planning methods has increased in the past few decades, many women and men have difficulty finding one with which they are comfortable. The 2002 NSFG gave respondents a list of 19 family planning methods from which to choose. Among women 15–44 years of age, 61.9% currently were using a contraceptive method: the pill-18.9%; female sterilization-16.7% and the condom-11.1%. About 4% were using the least effective methods, including periodic abstinence and withdrawal. Among women who had intercourse in the three months before the interview and who were not sterile, pregnant, postpartum, or seeking pregnancy, 7.4% were not using a contraceptive - an increase from the 5.2% in the 1995 NSFG. (Of those at risk for an unintended pregnancy, 89.3% were currently using a method.) [1]. It is important to recognize that contraceptive use data are based on self-reports and provide no indication of how consistently or correctly the methods are being used. Failure rates are high possibly because of inconsistent or improper use. In one study of the 1995 NSFG, 9% of women experienced a contraceptive failure within one year of starting to use a reversible method of contraception: 7% of those on the pill, 9% of those relying on a male condom, and 19% of those practicing withdrawal [14].
After female or male sterilization, hormonal methods are considered the most reliable form of contraception, but many women remain concerned about their safety. The side effects of many hormonal methods, especially Depo-Provera, discourage many women from their use [7, 8]. Further, the need to take a pill daily even in the absence of frequent intercourse creates problems for many women. Finally, the absence of a male contraceptive other than sterilization and the condom is a major barrier to effective and ongoing contraception.
The availability of one effective family planning method, emergency contraception (meant as back-up protection, not as a primary method), is currently limited by the refusal of the federal Food and Drug Administration to allow it to be sold over the counter and by the refusal of some pharmacists to fill prescriptions for this medication.
Inadequate attention to preconception counseling
Research is just beginning on the availability of preconception care overall although certain conditions, such as folic acid supplementation and counseling of diabetic women, have been studied extensively. The little existing evidence suggests that preconception care is not routinely integrated into family planning services, whether provided by obstetricians or family medicine specialists in private practice or HMOs, or by personnel in family planning facilities [15, 16]. Under such circumstances, the potential of this service delivery system for improving pregnancy and infant health through preconception care is not yet being realized.
Approaches to increasing utilization
Several evaluations indicate that the utilization of family planning services can be increased. California's Family Planning, Access, Care, and Treatment Program (Family Pact) provided contraceptive services to low income, medically indigent women increased use and reduced the numbers of unintended pregnancies [17]. State family planning waivers that expand Medicaid coverage for women beyond the 60 day postpartum limit have had a similar impact [18, 19]. Programs such as these should be expanded to more states, but they address only financial barriers. Attention should also be paid to other access-related problems, such as hours and days that facilities are open and language problems. Additional outreach also needs to be directed to low literacy and immigrant populations [20].
Approaches to improving content
Moos [15] reviewed many of the activities that will be essential if preconception care is to be integrated into health care generally. The American College of Obstetrics and Gynecology recently published a Committee Opinion, “The Important of Preconception Care in the Continuum of Women's Heath Care” [21]. Here the focus is on adding preconception counseling to visits for family planning. The recent publication by the Centers for Disease Control and Prevention of ``Recommendations to Improve Preconception Health and Health Care - United States'' [22] should also lead to more and better preconception counseling.
In all likelihood, it may be most difficult to change patterns of family planning practice among physicians in private practice. In contrast, those in HMOs have the potential for exposure to more education about the need for such counseling. The establishment of standards for family planning services by the Health Plan Employer Data and Information Set (HEDIS) might also accelerate change. However, standard setting and financial and other incentives may be most effective when used with community health centers and health department and other publicly-funded sites of family planning services - as the women who use seek family planning at these sites may be those most in need of preconception care. Because such sites are funded by federal, state, and local governments, these units can insist that preconception care be integrated into their family planning services.
Conclusions
Progress towards preconception care for all women will only be possible if a larger percentage of women and men plan their pregnancies. At present, although visits for contraceptive advice and methods provide an excellent opportunity for counseling about ways to achieve healthy pregnancies and healthy infants through preconception care, this potential is not being achieved. Increased use of family planning and increased attention to preconception care within family planning services may require that additional resources be devoted to public and professional education and to service delivery standards and financing.
|
[
"family planning",
"preconception care",
"pregnancy",
"access",
"intendedness"
] |
[
"P",
"P",
"P",
"P",
"U"
] |
Eur_Arch_Otorhinolaryngol-4-1-2217622
|
Changes in CMDP and DPOAE during acute increased inner ear pressure in the guinea pig
|
During and after an increase of inner ear pressure, induced by injection of artificial perilymph, the 2f1 − f2 and f2 − f1 distortion products (DPs) in cochlear microphonics (CM) and otoacoustic emissions (OAE) were recorded in the guinea pig. An inner pressure increase of ∼600 Pa gave only small changes in CMDP and DPOAE. Along with a decrease in f1 amplitude, a small decrease in amplitude of the 2f1 − f2 and a small increase in the f2 − f1 were measured in CM. This matches a shift from a symmetrical position of the operating point for hair cell transduction, leading to an increase in even-order distortion and a decrease in odd-order distortion. Similar, a decrease in 2f1 − f2 DPOAE was expected. This might be the case at the generation sites but this effect was then more than compensated for by a better middle ear transfer, accounting for the increase of 0.4 dB of the 2f1 − f2 DPOAE amplitude. In conclusion, changes of overall inner ear fluid pressure have minor effects on cochlear function. This is a relevant finding for further understanding of diseases with changed inner ear fluid volumes, as Ménière’s.
Introduction
The cochlea is a non-linear mechano-electrical transducer that transforms sound pressure waves into electrical information, the so-called mechanoelectric transduction (MET). The reverse electromechanical transduction is performed by the outer hair cell (OHC) which mechanically influences the vibration of the cochlear partition [7]. When the cochlea is stimulated by two tones, distortion products (DPs) can be measured in cochlear microphonic potentials (CMDP) and in the external auditory meatus as otoacoustic emissions (DPOAE). Both microphonic and acoustic DPs are considered to be reflections of OHC electromotility in the cochlea. Obviously, it is of interest to find a relation between these two functional modalities of the cochlea and the OHC in particular.
In this study the most prominent DPs, the 2f1 − f2 and f2 − f1, have been observed while increasing inner ear pressure. This was done by rapid injection of artificial perilymph into the scala tympani of the guinea pig’s cochlea. Measurement of specific DPs may describe changes in gain and operating point of the non-linear cochlear amplifier. Furthermore, transfer characteristics of forward and backward transmission may be clarified by using CMDP and DPOAE [17]. These data could help gaining insight into pathophysiological mechanisms involved in pathologically increased inner ear pressures.
When pressure variation is in the normal physiological range, this does not influence cochlear function as measured by compound action potentials in the guinea pig [2]. In humans, during activities like coughing or sneezing, perilymphatic pressure may undergo substantial fluctuations up to 10 mmHg (1.4 kPa).
An increased pathologic inner ear pressure may be the result of an increased intracranial pressure since the perilymphatic space is linked to the subarachnoid space via the cochlear aqueduct. Human OAE have been proven to be very sensitive to modifications of intracranial pressure [10]. Another clinical example of increased inner ear pressure or volume with an accompanying change of cochlear function is the endolymphatic hydrops in patients with Ménière’s disease.
In a previous study [11], we measured solely the 2f1 − f2 DPOAE’s in response to microinjection of artificial perilymph into the cochlea. Large changes of ∼500 Pa in overall inner ear pressure gave only a small mean change of 0.7 dB in 2f1 − f2 DPOAE amplitude. The magnitude of change could (partly) be explained by a change in oval window stiffness. In this study, the same experimental setup was used, with the addition of the simultaneous recording of the 2f1 − f2 and f2 − f1 CMDP. We attempt to establish a better understanding of increased inner ear pressure effects on DPs in both CM potentials and OAE’s.
Materials and methods
Experiments were performed in eight guinea pigs (Harlan Laboratories, The Netherlands; body weight 350–450 g) with a positive Preyer reflex. Animal care and use were in accordance with the principles of the declaration of Helsinki and approved by the animal experiment committee (protocol number 3047/3103).
General anesthesia was induced by intramuscular administration of ketamine/xylazine (60/3.5 mg/kg). Maintenance doses of the anesthetic were administered every hour. Muscle relaxation was obtained with succinylcholine (2.5 mg/kg). The animals were artificially ventilated through a tracheostoma (Columbus Instruments, model 7950). Body temperature was maintained at 38°C with a heating blanket. Heart rate was monitored by skin electrodes placed on both sides of the thorax. The animal’s head was kept in a stationary position by means of a steel bolt fixed to the skull with dental cement. Following a retroauricular incision, the bulla and external auditory canal were exposed. Subsequently, the bulla was opened equalizing middle ear pressure to normal air pressure [17]. By opening of the bulla, the round window was exposed. Through the round window membrane, the tip of a double-barreled micropipette was inserted into scala tympani. The double-barreled micropipettes were drawn from borosilicate glass (1.5/0.84 mm diameter per barrel) and the tips were beveled (Narishige EG-40). The total tip diameter was around 60 μm, which is a compromise between a low enough flow resistance for fluid injection and tip smallness. One barrel of the pipette was used to measure inner ear pressure (WPI 900A micropressure system). Through the other barrel, artificial perilymph [9] was injected with a constant flow rate by applying a controllable pneumatic pressure to the barrel end. The injected volume was measured as the displacement of the fluid meniscus in the pipette, for which the inner diameter is precisely known (0.84 mm). The fluid injection rate was calculated as the total injected volume divided by the total injection time.
During the experiment the DPOAE’s were continuously measured using an Etymotic ER-10C DPOAE probe system. The two primary frequencies were set at 6 kHz (f1) and 7.4 kHz (f2), meaning a f2/f1-ratio of 1.23. The low-level intensities were set at, respectively, 65 dB SPL (L1) and 55 dB SPL (L2). Full details of the measuring equipment are described elsewhere [11, 12].
Cochlear microphonic potentials were recorded with two differential Ni–Cr electrodes with a diameter of 50 μm. One electrode was placed in scala tympani through the bony wall of the basal turn of the cochlea. The other was placed in scala vestibuli next to the oval window. The signals were routed via an amplifier to a spectrum analyzer and two lock-in amplifiers (SRS, model SR830 DSP). The lock-in amplifiers displayed the amplitude of the 2f1 − f2 and f2 − f1 CMDP which were referenced to an electronic 2f1 − f2 and f2 − f1-DP derived from the original primaries.
A measurement contained a sequence of ten microinjections of artificial perilymph of 10 s each with pauses of 40 s. The rate of microinjection was 50 nl/s. The repetitive microinjection was controlled with a precision electronic timer (Stanford DG535). During an experiment, National Instruments LabVIEW® was used for recording of the following output signals: amplitude and phase of the 2f1 − f2 DPOAE, 2f1 − f2 CMDP, f2 − f1 CMDP, f1 CM, and inner ear pressure. Calculations and fits to relevant portions of the obtained recordings were made off-line with an appropriate software package.
Results
Successful repeated sequences of injection with a rate of 50 nl/s (n = 10) were performed in all guinea pigs (n = 8). The mean steady state inner ear pressure measured prior to injection of artificial perilymph was 205 Pa.
To determine the rise and fall times of the curves from the measured signals, proper fits were obtained with a simple exponential function: In some cases (Figs. 1c, 2c, 4c) the curves needed an extra term in the formula for a proper fit: According to Wit et al. [13] time constants τ were calculated from the fits.
Fig. 1a Averaged changes (n = 8) in inner ear pressure (Pa) during and after injection of 0.5 μl of artificial perilymph into scala tympani. The gray area depicts the injection period. Standard error bars are shown at t = 5, 15, and 40 s. b The shape of the mean inner ear pressure curve at the start of injection (time axis corresponds with Fig. 1a). The recording (filled circles) is fitted (broken line) with an exponential function, yielding a time constant τ. c The same is done for inner ear pressure recovery directly after injectionFig. 2a Averaged changes (n = 8) in f1 CM amplitude (μV) during and after injection of 0.5 μl of artificial perilymph into scala tympani. The gray area depicts the injection period. Standard error bars are shown at t = 5, 15, and 40 s. b The shape of the mean f1 CM amplitude curve at the start of injection (time axis corresponds with Fig. 2a). The recording (filled circles) is fitted (broken line) with an exponential function, yielding a time constant τ. c The same is done for f1 CM amplitude recovery directly after injection
In Fig. 1a, the averaged recorded inner ear pressure during and after injection of 0.5 μl of artificial perilymph is shown. At the start of injection the mean pressure increased to an ∼600 Pa higher level within seconds. At the end of injection, the pressure immediately decreased, returning to its initial value within seconds. This pattern was consistently observed in all individual experiments. The time constant for increase and subsequent decrease of inner ear pressure were 1.25 and 1.16 s, respectively (Fig. 1b, c).
The averaged f1 signal in the CM (Fig. 2a) showed a mirrored behavior in respect to the inner ear pressure. Directly after the start of injection, the f1 CM decreased by 30 μV. The corresponding time constant was 1.17 s (Fig. 2b), which is very close to the time constant for pressure increase. Conversely, at the end of injection the f1 amplitude returned to its initial value with τ = 3.49 s (Fig. 2c), which is slow with respect to the corresponding τ for pressure.
The averaged DPOAE 2f1 − f2-amplitude showed a similar typical course as the inner ear pressure (Fig. 3a). Almost directly after the start of injection the 2f1 − f2-amplitude increased to reach a maximum of 0.4 dB above its original value at the end of the injection period. When the injection was instantaneously terminated, the amplitude dropped and returned to slightly above its onset level. There was a small observable time lag with respect to the pressure curve. The time constant for increase and subsequent decrease of the 2f1 − f2-amplitude were 1.52 and 3.43 s, respectively (Fig. 3b,c).
Fig. 3a Averaged changes (n = 8) in 2f1 − f2 DPOAE amplitude (dB SPL) during and after injection of 0.5 μl of artificial perilymph into scala tympani. The gray area depicts the injection period. Standard error bars are shown at t = 5, 15, and 40 s. b The shape of the mean 2f1 − f2 DPOAE amplitude (μPa) curve at the start of injection (time axis corresponds with Fig. 3a). The recording (filled circles) is fitted (broken line) with an exponential function, yielding a time constant τ. c The same is done for 2f1 − f2 DPOAE amplitude recovery directly after injection
In Fig. 4a, the typical temporary decrease for the averaged 2f1 − f2 CMDP amplitude relative to the f1 CM is shown. The time constant for decrease in amplitude at the onset of injection was 1.12 s (Fig. 4b), which corresponds with τ for pressure. The recovery is somewhat slower (2.28 s, Fig. 4c), but still faster than the time constant for 2f1 − f2 DPOAE recovery.
Fig. 4a Averaged changes (n = 8) in 2f1 − f2 CMDP relative to f1 CM amplitude during and after injection of 0.5 μl of artificial perilymph into scala tympani. The gray area depicts the injection period. Standard error bars are shown at t = 5, 15, and 40 s. b The shape of the mean 2f1 − f2 CMDP (relative to f1 CM) amplitude curve at the start of injection (time axis corresponds with Fig. 4a). The recording (filled circles) is fitted (broken line) with an exponential function, yielding a time constant τ. c The same is done for 2f1 − f2 CMDP (relative to f1 CM) amplitude recovery directly after injection
Finally, in Fig. 5, the averaged results of the f2 − f1 CMDP amplitude relative to f1 CM depict an increase during injection of artificial perilymph. When the injection was terminated, the f2 − f1 CMDP amplitude returned to the pre-injection level. Unfortunately, due to a low signal-to-noise ratio for the f2 − f1 CMDP, the obtained curve could not be fitted properly and no time constants could be derived.
Fig. 5Averaged changes (n = 8) in f2 − f1 CMDP relative to f1 CM amplitude during and after injection of 0.5 μl of artificial perilymph into scala tympani. The gray area depicts the injection period
Discussion
The most striking result of this study is the minimal change in amplitude of the DPs in both OAE and CM during an acute increase of inner ear pressure. Besides, amplitudes changed to normal when inner ear pressure returned to its initial level. These results indicate a dynamically stable inner ear which is able to cope with a pressure increase of 600 Pa with only minimal and reversible cochlear function loss. According to Böhmer [2] this level of pressure increase would still be in the physiological range of the guinea pig ear varying between −100 and +700 Pa.
The inner ear is an elastic structure completely encompassed by bone, in which pressure changes depend on the compliance of the membranous structures [14]. In normal ears, hydrostatic pressure in the perilymph equals pressure in the endolymph, and pressure changes applied to one compartment are immediately transmitted to the other through the elastic walls. The perilymph is in a dynamic equilibrium with the cerebrospinal fluid via the cochlear aqueduct. An excess of perilymph will escape through the cochlear aqueduct restoring inner ear fluid level.
When increasing inner ear pressure, a cochlear disturbance with functional loss is expected. The observed decrease in CM f1 amplitude thus seems obvious. An increased cerebrospinal fluid pressure also suppresses the CM [16]. The increase in 2f1 − f2 DPOAE is counterintuitive, but in accordance with a previous publication in which we demonstrated the same change in cochlear function as measured by 2f1 − f2 DPOAE’s [11]. According to Büki et al. [3] a change in stiffness of the oval window might slightly improve middle ear transfer and thus explain the small increase in DPOAE amplitude. In contrast to the 2f1 − f2 DPOAE, the 2f1 − f2 CMDP showed a decrease in amplitude during injection. Unfortunately, in literature very little can be found about the relation between DPOAE and CMDP. Still, it were Kemp and Brown [6] who found close parallels in the behavior of the 2f1 − f2 DPOAE and CMDP during and after anoxia in the guinea pig. Group latencies and third-tone suppression were found to be very similar, linking both signals to have a common origin.
The DPOAE and CMDP are generated in response to the same acoustic stimuli. The acoustic primaries are transferred through the external and middle ear to eventually reach the sensory epithelia in the cochlea. The cochlea as a non-linear mechano-electrical transducer transforms sound pressure waves into electrical information. This forward transduction process is called the MET. The CM potential is a direct result of this process and is thought to be mainly produced by OHCs. On the other hand, there is also a reverse electromechanical transduction in which the OHC mechanically influences the vibration of the cochlear partition [7]. Sound is transmitted back to the external ear where the acoustic DPOAE can be measured; the backward transmission.
The rise and fall times of the DPs in response to changes in inner ear pressure were derived from the fitted curves (Figs. 2, 3, 4). As both CMDP and DPOAE were measured, this yielded information on forward and backward transmission. As anticipated, the time constants for the CMDP were smaller than the time constants for the DPOAE. Indeed, the latter would travel back and forth the middle ear which understandably costs more time. However, the difference has not the form of a simple longer delay. The explanation for this observation remains tentative. The averaged differences were 0.4 and 1.15 s for the 2f1 − f2 DP during increase and decrease of inner ear pressure, respectively.
At pressure increase all measured signals almost directly followed the change in inner ear pressure. Time constants varied from 1.17 to 1.52 s with respect to 1.25 s for the inner ear pressure curve. However, with pressure decrease there was a delay of 1–2 s for all measured signals. Measured differences cannot be accounted for by delay in measuring equipment, for this was <0.18 s. Assumed that the overall stiffness of the stapes-oval window-system governs the behavior of the DPOAE 2f1 − f2 amplitude, it is expected that it would follow stiffness changes directly and not 2 s later, as observed. It is conceivable that a delayed change in the endocochlear potential (EP) could be the cause of difference in τ values. However, in experiments by Salt and DeMott [9] the EP followed a temporary pressure change with a delay of <0.1 s. So, the mechanisms for the observed delays in inner ear pressure recovery remain to be elucidated.
Two-tone distortions are generated during non-linear mechanical amplification in the cochlea. Generation of the 2f1 − f2 depends on asymmetric component of a non-linear transfer function whereas the f2 − f1 depends on symmetric components. A change of the operating point is reflected in the level of the f2 − f1 distortion. In the gerbil low-frequency tones of 5 Hz were used to bias the position of the cochlear partition by Frank and Kossl [4, 5].
By low-frequency modulation of the 2f1 − f2 DPOAE, Bian et al. [1] described the forward MET by a sigmoid-shaped function, which related the displacement of the basilar membrane to the response of the hair cells.
By injecting artificial perilymph into scala tympani of the cochlea, the basilar membrane is temporarily displaced. Thus, by this displacement toward scala media, the operating state of the cochlear amplifier is also influenced. We measured a small decrease in the amplitude of the 2f1 − f2 distortion in CM (with respect to the amplitude of f1 in CM) and a small increase in the f2 − f1 in CM. Using the model by Bian et al. [1] this leads to the conclusion that the operating point for hair cell transduction is somewhat shifted from a symmetric position. This shift leads to an increase in even-order distortion and a decrease in odd-order distortion. Along these lines of reasoning we would also expect the 2f1 − f2 DPOAE to decrease. This might be the case at the generation sites [15], but this effect is then more than compensated for by an improved middle ear transfer.
Anyhow, the observed changes are small. Changes of overall inner ear fluid pressure have minor effects on cochlear function. It may be that creating an artificial endolymphatic hydrops, by injecting artificial endolymph into scala media instead of scala tympani, has a larger effect. This is currently tested in further research, which will hopefully give insight into diseases with pathologically changed inner ear fluid volumes, as in Ménière’s disease.
|
[
"cmdp",
"dpoae",
"inner ear pressure",
"guinea pig",
"injection",
"perilymph",
"2f1 − f2",
"f2 − f1"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Anal_Bioanal_Chem-3-1-1705481
|
Analytical methods for PCBs and organochlorine pesticides in environmental monitoring and surveillance: a critical appraisal
|
Analytical methods for the analysis of polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs) are widely available and are the result of a vast amount of environmental analytical method development and research on persistent organic pollutants (POPs) over the past 30–40 years. This review summarizes procedures and examines new approaches for extraction, isolation, identification and quantification of individual congeners/isomers of the PCBs and OCPs. Critical to the successful application of this methodology is the collection, preparation, and storage of samples, as well as specific quality control and reporting criteria, and therefore these are also discussed. With the signing of the Stockholm convention on POPs and the development of global monitoring programs, there is an increased need for laboratories in developing countries to determine PCBs and OCPs. Thus, while this review attempts to summarize the current best practices for analysis of PCBs and OCPs, a major focus is the need for low-cost methods that can be easily implemented in developing countries. A “performance based” process is described whereby individual laboratories can adapt methods best suited to their situations. Access to modern capillary gas chromatography (GC) equipment with either electron capture or low-resolution mass spectrometry (MS) detection to separate and quantify OCP/PCBs is essential. However, screening of samples, especially in areas of known use of OCPs or PCBs, could be accomplished with bioanalytical methods such as specific commercially available enzyme-linked immunoabsorbent assays and thus this topic is also reviewed. New analytical techniques such two-dimensional GC (2D-GC) and “fast GC” using GC–ECD may be well-suited for broader use in routine PCB/OCP analysis in the near future given their relatively low costs and ability to provide high-resolution separations of PCB/OCPs. Procedures with low environmental impact (SPME, microscale, low solvent use, etc.) are increasingly being used and may be particularly suited to developing countries.
Introduction
Scope of the review
Polychlorinated biphenyls (PCBs) and organochlorine pesticides (OCPs; aldrin, chlordane, DDT, dieldrin, endrin, heptachlor, mirex, toxaphene, hexachlorobenzene (HCB)) constitute ten of the twelve chemical substances/groups currently defined under the Stockholm Convention on Persistent Organic Pollutants (POPs). While bans and phase-outs of these chemicals occurred during the 1970s and 1980s in most developed countries, they were not in place in many developing countries. Article 7 of the Stockholm Convention requires National Implementation Plans (NIPs) to be developed by signatory countries [1]. For countries where PCBs and OCPs are still in use or loosely regulated, the NIPs will eventually lead to the phase-out of OCPs for agricultural use and reduced use for disease vector control. However, many countries will be faced with the legacy of past use of these persistent chemicals. Residues in soils, sediments and biota will pose a risk for wildlife and human consumers. The problems of destroying existing stockpiles, especially of OCPs and PCBs in electrical equipment, as well as preventing migration of contaminants from storage areas, will need to be solved. There is a need to monitor levels and trends now and into the future in order to assess compliance with the POPs Convention within each country as well as on a regional and global basis.
Methodology is available for PCBs and OCPs as a result of a vast amount of environmental analytical chemistry research and development over the past 30 to 40 years. However, the establishment of an analytical laboratory and the application of this methodology at currently acceptable international standards is a relatively expensive undertaking. Furthermore, the current trend to use isotope-labeled analytical standards and high-resolution mass spectrometry for routine POPs analysis is particularly expensive. These costs limit participation of scientists in developing countries and this is clear from the relative lack of publications and information on POPs from countries in Africa, south Asia and South/Central America. Thus, this review will summarize the best practices in developed countries and new advances in PCB/OCP analysis, while bearing in mind the need for low-cost methods easily implemented in developing countries.
This review will also mainly focus on methodology for the analysis of PCB/OCPs in biological samples (fish, aquatic and terrestrial mammals and birds) as well as soils, sediment and passive samplers such as semipermeable membrane devices (SPMDs). These matrices and environmental compartments have higher concentrations of PCBs and most OCPs than water or air, making them more suitable for routine monitoring and more relevant in the context of exposure of humans and wildlife. Milk and blood are important matrices for POPs monitoring in humans and, in the case of blood, are increasingly used in programs involving live-captured marine mammals and birds. Thus analytical methods for milk and blood are briefly considered here. The advantages and disadvantages of various bioindicator species are discussed in detail by Tanabe and Subramanian [2].
Substances to be analyzed
PCBs (except non-ortho-substituted congeners; no-PCB) and OCPs can be considered together because they are extracted and analyzed together in most cases, as discussed below. In practice, most laboratories determine about 30 or more individual PCB congeners, and 10–20 individual OCPs and their metabolites, regardless of the sample matrix. Ongoing POPs monitoring programs vary in their analyte lists. For example, the Integrated Atmospheric Deposition Network (IADN) in the Great Lakes region of North America includes over 100 PCB congeners [3], while the UNEP/World Bank/GEF project on Persistent Organic Pollutants, Food Security, and Indigenous Peoples in Arctic Russia included 15 PCB congeners [4]. The Arctic Monitoring and Assessment Program recommended 30 ortho-substituted PCBs [5].
The UNEP Global POPs monitoring workshop categorized individual PCB/OCPs as “essential” and “recommended,” taking into consideration the need for various levels of analytical capacity building in developing countries in order to conduct the monitoring (Electronic Supplementary Material, Table 1) [6]. For PCBs, a minimum of seven PCBs (28/31, 52, 101/90, 118, 138, 153 and 180) were regarded as essential for routine monitoring of fish and food products around the world. Using only seven congeners severely underestimates total PCB concentrations in some matrices—particularly in abiotic samples such as air, soil, and sediment. On the other hand, these seven are robustly determined, having been the subject of numerous interlab comparisons (e.g., QUASIMEME [7]). UNEP [6] recommended a suite of 30 ortho-substituted congeners for more comprehensive monitoring.
Toxaphene and chlordane are multicomponent mixtures and analytical standards are available for some, but not all, major components, which makes them the most difficult to analyze among the eight OCPs on the POPs list. Current programs in Europe are determining three toxaphene congeners, Parlar (P) 26, 50 and 62, and interlaboratory comparisons have shown reasonable agreement for these congeners among laboratories [8]. However, P26 is interfered with by coeluting chlorobornanes in some matrices, and P62 can vary widely in its instrumental response. Furthermore, in North America, most datasets for toxaphene are still based on quantification with technical toxaphene, which yields a single value and no congener-specific information. Capacity for determining toxaphene outside of the USA, Canada and western Europe is very limited judging from the relatively large number of measurements from these countries and the dearth of measurements elsewhere.
The situation for chlordane is better because analytical standards are available for the major components (cis- and trans-chlordane; cis- and trans-nonachlor, heptachlor). Nevertheless, other octachloro- and nonachlorochlordane components are important [9] and are often found in top predators [10, 11], and analysts need to be aware of possible interferences from these compounds in seabird and marine mammal extracts.
The chlorinated insecticides endosulfan, lindane and methoxychlor are also candidates for inclusion in analytical methods where impacts of OCPs are being investigated. Global use of these insecticides, particularly endosulfan, is very significant [12, 13] These current-use pesticides can generally be isolated and quantified along with other OCPs.
The no-PCBs need to be considered separately because their analytical methodology is normally quite different from that used for ortho-substituted PCBs and OCPs (see Reiner et al. [14] in this issue).
In summary, the minimum number of PCB/OCPs to be considered will vary with the objectives of a given study. A long list of analytes would be used for initial surveys where no information was available. Chemical structures and physical and chemical properties of selected PCBs and OCPs are listed in Electronic Supplementary Material Table 2 [143, 144].
Analytical standards
The availability of accurate analytical standards is a fundamental requirement of an analytical program designed to quantify PCB/OCPs. Standards are available from commercial chemical supply companies as well as from agencies involved in certification of reference materials, e.g., the National Institute of Standards and Technology (NIST) or from the Bureau Communautaire de Référence (BCR) via the Sigma-Aldrich company. Interlaboratory studies have shown that incorrect analytical standards and calibration errors can be a major source of error in POPs analysis [15, 16]. These authors recommended that multilevel calibration at 4–6 levels should be used to define the response profile, that calibration solutions should be checked against independent solutions, either from another reliable laboratory or by preparing duplicate solutions, and that calibrants should be ampouled (in sealed glass vials) for long-term storage. They noted that refrigerator storage can loosen the caps on screw-top vials. Thus, storage of calibrant solutions in this manner should only be temporary.
Quantification of PCBs can be carried out using two approaches; by Aroclor mixture or individual congener standard. The former method employs a standard that was available as a commercial product (e.g., Aroclor 1254 used in transformer oils) to quantitate the congeners contained in the Aroclor mixture against those matching in the sample. There are several mixtures available, with varying degrees of mean chlorination. Percent ratios of the congeners in various commercial mixtures have been reported [17]. The limitation of this method is the varying congener ratios within a given commercial mixture due to batch-to-batch production variability which, in turn, may lead to congener calculation error [18]. Also, a significant bias resulting in an overestimation of PCB measurements may occur if the PCB pattern in the sample does not match that of the standards. Sather et al. [19] suggested that accurate PCB determinations can be made if the chromatographic pattern of both the standards and the samples are similar. However, if the pattern differs, the study reported a bias doubling the values determined by the individual congener standard. The method of choice is congener determination using standard mixtures comprising composite individual congeners added at known quantities which were purchased individually. This provides an improved and known level of accuracy not possible with the commercial PCB mixture approach [19].
Commonly used analytical methods for PCB/OCP monitoring and surveillance
Overview
Numerous methods have been published over the past 30 years related to specific analytical techniques for the determination of PCBs and OCPs in food and environmental matrices. Laboratory standard operating procedures (SOPs) are available from the National Environmental Methods Index in the USA (which includes US EPA and US Geological Survey methods) (http://www.nemi.gov) and the Japan Environment Agency (http://www.env.go.jp/en/index.html), the National Oceanographic and Atmospheric Administration in the USA [20] and the US Food and Drug Administration [21]. Useful information may also be available from the International Committee for Exploration of the Sea [22], the Oslo–Paris Commission [23] (http://www.ospar.org), the Helsinki Commission (HELCOM) (http://www.helcom.fi), the International Organization for Standardization (http://www.iso.org), the Association of Official Analytical Chemists International (http://www.aoac.org), and Gosstandart of the Russian Federation (http://www.krohne.ru/russia_en/downloads/certificates/cis/russia/gosstandard/).
Not all of these sources provide analytical SOPs for all environmental media. Alternatively, numerous books are available that summarize methods. For example, Keith [24] summarized US EPA methods for PCBs and OCP in sediment and biological materials. Wells and Hess [25, 26] have reviewed and recommended methods for the separation, isolation and recoveries of OCPs and PCBs from soils, sediment and biological matrices. De Boer and Law [27] provide a useful overview of current analytical methodology for OCP/PCBs.
The UNEP POPs workshop on global monitoring (UNEP 2003) noted that, given the broad range of technical expertise for analysis of PCBs and OCPs, as evident from the extensive international participation in interlaboratory calibration projects for these compounds, no single, detailed, step-by-step analytical method can be recommended. Instead laboratories would use methods best-suited to their situation and take part in international interlaboratory comparisons to verify their work. This performance-based approach has also been adopted by the US EPA in an effort to introduce flexibility in conducting environmental monitoring. This approach also encourages analytical innovation by allowing the use of new and alternate methods provided that predetermined performance criteria are met [28].
Sampling
Acquisition of suitably representative samples is fundamental to any environmental measurement program and should not be overlooked. Contamination during field collection can occur, particularly from PCBs in electrical equipment and building products (e.g., marine paints, joint sealants) [29, 30], or due to the ongoing use of OCPs. “Clean” techniques need to be adopted, such as the use of special clothing and disposable gloves for sampling, sealed shipping containers, and field blanks. ICES [22], NOAA [20] and OSPAR [23] provide useful guidelines for sampling of fish/crustaceans, mussels, and sediment, respectively. Azcue and Mudroch [31] provide a detailed overview of sampling methods for lakes and river sediments.
Sample containers should be supplied by the laboratory and checked for contamination. Appropriate sample containers for biological tissue that is to be analyzed for POPs are glass jars or glass vials with screw caps. Screw caps should be lined with solvent-rinsed aluminum foil or Teflon cap inserts. Clear polyethylene bags, and polypropylene jars, are also appropriate for temporary storage but may not be suitable for long-term storage because of the possibility of migration of plasticizers (such as phthalates) into the tissue, especially for samples with high lipid contents. Freezing and storage of multiple small samples suitable for analysis, rather than larger masses, is recommended to avoid multiple freezing and thawing of tissue [25] and to reduce sample handling, which in turn reduces the potential for contamination. Samples can be wrapped in aluminum foil and then inserted into plastic bags.
Blood samples should be collected in ethylenediamine tetraacetic acid (EDTA) vials or vacutainers, centrifuged, and the plasma transferred to vials washed with hexane. Field blanks, consisting of sample containers taken to the lab and returned with other samples, should be included.
Air sampling for POPs is normally conducted using polyurethane foam (PUF) or XAD resin [3]. From 300 to 10,000 m3 of air are passed through the sampling medium [3, 32]. Sampling media are then transferred to clean, wide-mouthed glass jars, sealed and stored at 4 °C prior to extraction. Some sampling cartridges can be sealed and transported and the sampling media transferred to Soxhlet or pressurized liquid extraction systems in a clean laboratory environment.
Water sampling is particularly prone to contamination by PCBs due to ubiquitous contamination, especially on older ships and buildings, and to the very low sample concentrations likely to be encountered [33]. US EPA analytical methods for PCBs in water assume that drinking water is being analyzed [34, 35] and do not place much emphasis on sampling because the drinking water quality guideline is 0.5 ug/L and only small <1 L samples need to be collected to assess compliance. Solid-phase extraction (SPE) cartridges have been widely used to extract relatively small volumes (1–5 L) for OCP analysis [36]. For sampling of natural waters where PCB concentrations are in the low pg/L range, much larger samples are required. For example, Achman et al. [37] pumped lake water (∼100 L) directly into large-diameter filters (0.7 μm nominal pore size) and then through XAD-2 resin columns. Surrogate standards were added to the resin column prior to extraction. Similar solid-phase extraction approaches are used for sampling and extracting wet precipitation [38, 39]. Another large-volume application uses liquid:liquid extraction in a continuous flow approach [40]. This involves a one pass-through extraction unit where the sample is metered into a mixing chamber which contains a solvent (i.e., dichloromethane) and flows out to an overflow exit. The drawback to this method is the significant potential for contamination due to exposure to the surrounding air. Petrick et al. [41] describe an in situ sampler for the collection of very large volumes (∼500 L) of seawater on polyurethane foam for the determination of PCBs at low pg/L concentrations. The studies by Petrick et al. [41] and Sobek and Gustafsson [42] illustrate the great care that is needed to properly determine PCBs at the low pg/L level (femtograms/L of individual congeners) that are found in ocean waters.
SPMDs offer potential for relatively inexpensive sampling of water and air for POPs and potential for broad geographic coverage [43, 44]. For water sampling, lipid-based SPMDs (consisting of triolein inside polyethylene tubing) have generally been used to estimate bioavailable water concentrations. SPMD use in the aquatic environment has been validated in a series of studies [45]. The use of performance reference compounds (PRCs) added to the triolein prior to deployment has been shown to enhance the use of SPMDs by accounting for site-specific variables [45]. The PRCs are gradually lost from the sampler and thus permit an exposure adjustment factor or depuration rate to be calculated, which adjusts for changes in sampling rate.
Lipid-based passive air samplers have also been developed and deployed for air sampling of PCBs and OCPs [46]; however, more recently designs based on polyurethane foam (PUF) and XAD resin have become more widely applied [47, 48]. Care must be taken to prepare and transport the samplers under clean conditions. Following deployment, which is usually for several months for lipid-based and PUF samplers [49] for up to one year for XAD [44], samplers are placed in transport containers and returned to the lab. PRCs are also often added to the passive air samplers [47].
Sample storage and handling
A basic requirement for analytical laboratories involved in the measurement of PCBs and OCPs is the availability of freezer and refrigerator capacity for sample storage and archiving. Wells and Hess [25] and ICES [22] provide recommendations for storage of biological and sediment samples. Kiriluk et al. [50] found no significant differences in concentrations of PCBs in whole fish homogenates stored at −20 or −80 °C over four years. Norstrom and Won [51] observed dehydrochlorination of DDT in long-term-archived egg homogenates, which they attributed to reactions occurring during freeze-drying. De Boer and Smedes [52] found no significant changes in the concentrations of PCBs and DDT components in fish tissue stored at −20 to −70 °C in the dark for up to 24 months. However, they did find changes in lipid extractability caused by oxidation of triglycerides to free fatty acids. The authors pointed out that this could affect calculations of lipid-normalized concentrations. To minimize any effects, they recommended long-term (>2 year) storage at −70 °C.
Sample preparation
Wells et al. [53] noted that sample preparation affected the magnitude of the variance in the analysis of PCB congeners in mussel tissue. They found that freeze-drying samples reduced recoveries of OCPs and PCBs, probably due to tighter binding and occlusion of residues in the dried material. Volatilization losses might also occur during freeze-drying of volatile compounds such as HCB, and there is a greater potential for contamination in the lab [54]. Although freeze-drying has been recommended by some authors for sediments and soils [25], lower recoveries of PCBs from freeze-dried sediments have also been reported [55]. Karl [56] noted that loss of water led to an increase in OCP concentrations in fish products during processing. Thus, in general, maintaining environmental sample tissues and sediments in their original wet state is regarded as the most appropriate approach for preparing samples for PCB/OCP analysis. Furthermore, avoiding a drying step minimizes potential contamination from lab air, particularly in the case of PCBs that are ubiquitous contaminants of older (pre-1980s) buildings [30, 57]. Using wet material also avoids possible volatilization losses. Instead, homogenized samples should be mixed with a desiccant such as sodium sulfate, Celite or Hydromatrix to bind water. The desiccant must be certified to be free of analyates, e.g., by heating at high temperature in the case of sodium sulfate or pre-extraction (Celite; Hydromatrix).
Tables 1 and 2 provide general guidance for various preparation/extraction and isolation/QA steps in the analysis of PCBs and OCPs and sources of information. Starting with sample preparation, the basic approach is to assure that the sample is prepared for extraction in a room that is free from significant contamination. Ideally this would involve a well-ventilated lab with air prefiltered through HEPA and carbon filters, but any clean chemical laboratory facility should be adequate for most work on PCBs and OCPs in most matrices except water or soils and sediments from remote locations. Laboratory buildings built after about 1990 are likely to be free of PCB-containing materials (electrical equipment/paints/sealants) or OCPs used as insecticides (or termiticides in foundations) because of the effects of bans on the use of these substances promulgated during the 1970s and 1980s.
Table 1Extraction techniques used for solid environmental samples (from Lopez-Avila [64]; Wells and Hess [25])TechniqueOverviewMethod referenceConventional SoxhletSample + desiccant mixture in glass or paper thimble is leached with warm (condensed) solvent for 4–12 hrs. Solvents are, e.g., diethyl ether, DCM, hexaneEPA 3540 [65]Automated Soxhlet (e.g., “Soxtec”)Extraction thimble is immersed in boiling solvent (30–60 min) then raised for Soxhlet extraction. Solvent can also be evaporated.EPA 3541 [65]Supercritical fluid extraction (SFE)Sample (usually +desiccant) placed in high-pressure cartridge and carbon dioxide at 150–450 atm at temp of 40–150 °C passed through. After depressurization, analytes are collected in solvent trapEPA 3560–3562High-speed blendingUseful for high water content samples such as plant material. Homogenizes sample with acetone and NaCl.[21, 66]Column extractionSample (+desiccant) placed in large column with filter and stopcock. Eluted with large volume of extraction solvent, e.g., hexane:DCM; hexane[67]Sonication-assisted extractionSample in open or closed vessel immersed in solvent and heated with ultrasonic radiation using ultrasonic bath or probe.Method 3550C [65]Microwave-assisted extraction (MAE)Sample in open or closed vessel immersed in solvent and heated with microwave energy.EPA 3546 [65]Pressurized liquid extraction (PLE)Sample (usually +desiccant) placed in extraction cartridge and solvent (heated, pressurized) passed through then dispensed in extraction vial.EPA 3545 [61, 65]Table 2Guidance for various preparation, extraction and isolation steps in the analysis of PCBs and OCPsEnvironmental matrixAnalytical stepsGeneral proceduresEPA or other methodSoil and sedimentPreparationPrepare in a PCB- and pesticide-free room.[35, 86]Avoid air-drying. Wet sieve if necessary to remove large particles. Centrifuge sediment to remove excess water.Mix soils/sediments with desiccant such as Na2SO4.Separate determination of dry mass by oven drying. For sediments total organic carbon should be determined.QAOne blank, soil CRM every ten samples; spike all samples with recovery surrogate standards. Bake glassware overnight at 200 °C or higher.ExtractionSoxhlet, PLE, sonication, or MAE with acetone: hexane or DCMSolvent evaporation, transfer to hexane.Sulfur removal with activated copper turnings required for sediment.Isolation/cleanupAlumina, silica or Florisil elutions: non-polar (hexane) and polar (DCM:hexane or equivalent)VegetationPreparationHomogenize using food chopper or blender. Cryoblending using liquid nitrogen or dry ice is useful. Mix with dessicant. Separate determination of dry mass by oven-drying.[21, 35]QASame as soil. Use vegetation CRM if possibleExtractionSame as soil.Isolation/cleanupSame as soil.Aquatic biotaPreparationSelect muscle or liver depending on species. For mussels and crustaceans use soft tissue. Select tissue that has not been in contact with the sample container. Homogenize using food chopper or blender. Cryoblending is useful.[21, 35]Mix with drying agent. Separate determination of lipid content.QASame as soil. Use fish or mussel SRMs.ExtractionSoxhlet, pressurized liquid extraction, or column extraction. Use acetone:hexane or DCM.Isolation/cleanupRemove lipid using gel permeation chromatography if possible or by repeated washing of the extract with sulfuric acid. Follow with fractionation on silica or Florisil columns as described for soil.Marine mammal blubberPreparationSelect blubber that has not been in contact with the sample container. Blend or hand-mix with drying agent. Separate determination of lipid content.[10, 87]QASame as soil. Use fish oil or marine mammal SRMs.Isolation/cleanupSame as for fish extracts.Air (high volume)Extraction,QA and cleanupAssuming that air is collected on polyurethane foams or XAD resin, these would be extracted in a Soxhlet or pressurized liquid extractor. Other steps as for soil or sediments[32]Semi-permeable membrane devices (SPMD)PreparationSPMDs would be removed from their transport cases and rinsed with precleaned water to remove accumulated dust (air-borne samplers) or periphyton (water samplers).[46]QAUse PRCsExtraction, and cleanupAssuming that the SPMD is lipid-based, extraction of POPs by “dialysis” into hexane would be achieved in a large glass cylinder.Water (including melted snow, ice and wet precipitation)ExtractionLiquid–liquid, SPE (e.g., C18) extraction for small (<1 L) samples; XAD-2 or modified “Speedisk” for >1 L.US EPA 508.1 [34] and 525.2 [36, 37, 78, 88]QA and cleanupPre-spike XAD columns with surrogates.Blood plasmaExtractionExtract blood plasma with ammonium sulfate/ ethanol/hexane (1:1:3) or C18 SPE extraction.[72, 73]Determine lipid content.QASame as fish. Use NIST 1589a SRM.Isolation and cleanupSulfuric acid partitioning to remove lipids.Acid–base silica for additional lipid removal.MilkExtractionLiquid–liquid partitioning with acetone:hexane or C18 SPE extraction. Determine lipid content.[75, 77]QABCR SRM 284 & 533 milk powder.Isolation and cleanupAs with plasma.
Recommended extraction and isolation techniques for PCBs and OCP
Recovery surrogates/internal standards
A standard QA step in the analysis of PCB/OCPs is to include surrogate recovery standards in each sample. These are generally one or two PCB congeners (e.g., CB30 and CB204) and OCP-related, unlabeled compounds such as pentachloronitrobenzene, endrin ketone, 1,3,5-tribromobenzene, and other compounds that are readily separated from individual OCPs and not commonly found in the environment. If GC–MS is being used as the detection system then 13C-labeled or deuterated surrogates should be used. These standards are carefully added by pipetting the appropriate volume of a standard solution onto the sample matrix prior to extraction. Surrogates do not define absolute recoveries because they are not incorporated into the matrix; however, they do provide information on analyte recovery throughout the isolation process and can also provide sample-to-sample extraction efficiency comparability. If bioanalytical techniques are being used to screen extracts, it may not be possible to include recovery standards, particularly 13C-surrogates, since they may respond similarly to the native compounds.
Extraction techniques
The appropriately prepared sample can then be extracted by any one of a number of techniques (Table 1). The main points to consider are to allow adequate exposure time for the solvent system in the sample matrix and to limit sample handing steps, i.e., avoid filtration steps by using Soxhlet (sample in a glass thimble), extraction columns (sample matrix eluted after soaking in solvent) or semi-automated systems (e.g., pressurized liquid extractors, PLEs). Cross-contamination from residues left behind by high levels in other samples is a concern at this stage and equipment must be thoroughly cleaned and checked from batch to batch. Purity of extraction solvents is a major consideration here. Only high-purity glass-distilled solvents should be used, because later evaporation steps will concentrate any contaminants. PLE is increasingly being used to replace Soxhlet and column extraction methods [58–60]. PLE uses much less solvent than Soxhlet or column elution. However, it suffers the disadvantage of initial high cost and the need for a stable power supply to avoid premature instrument shutdown. Also, operation of an automated PLE requires regular scheduled maintenance by trained service personnel. Preparation of a homogeneous dry sample from wet tissue or sediment for PLE can be a challenge due to the limited size of PLE extraction vials, currently set at 100 ml. A US EPA method (3545) using PLE for solid waste extraction is available [61]. This method recommends acetone/hexane (1:1, v /v) or acetone/dichloromethane (DCM) for extraction of PCBs from solid waste, and these solvent systems appear to be the most commonly used [59]. There has been a recent trend toward not using chlorinated (potentially toxic) solvents such as chloroform and DCM, due to concerns over exposure to lab personnel. Binary mixtures such as hexane/acetone (1:1) are regarded as equally effective for the extraction of POPs from solid samples [62]. In general, extraction with a polar:apolar binary mixture has been found to be more efficient for recovering OCPs and PCBs from fish tissues of low lipid content than an apolar solvent [63].
The time- and solvent-consuming nature of Soxhlet extraction (or related techniques involving percolation of a solvent through the sample) is generally thought to be related to the slow diffusion and desorption of the analytes from the sample matrix. Semivolatile compounds such as naphthalenes can also be lost from Soxhlet apparatus via volatilization [68]. The use of microwave, sonication, supercritical fluids, or elevated temperatures and pressure (as in PLE) increases the rates of diffusion and desorption and thus speeds up extraction [58, 64, 69]. Pressurized hot water extraction has even been used to extract PCBs from sediment and soil. Under pressure, the dielectric constant of the water can be manipulated to facilitate the extraction of nonpolar analytes [70]. Sonication, microwave-assisted extraction (MAE) and supercritical fluid (SFE) extraction have all been successfully employed to recover PCBs and OCPs from solid samples (Table 1). Soxhlet is regarded as more reliable than sonication, SFE, or shake flask-type methods and equivalent to (but not as rapid as) MAE and PLE for a wide range of environmental matrices [25].
Blood plasma lipids can be rapidly extracted using conventional solvent partitioning using a mixture of ammonium sulfate/ethanol/hexane (1:1:3) [71, 72]. Solid-phase extraction using C18 cartridges is also a proven technique [73]. Needham et al. [74] have reviewed methods for the analysis of POPs in human milk. PCB/OCPs can be extracted from milk samples using conventional lipid extraction techniques involving liquid–liquid extraction with hexane/acetone [75]. Blending with acetonitrile and potassium oxalate [76] followed by C18-solid phase extraction has also been used in recent studies [77].
Water sampling devices such as solid-phase extraction (SPE) cartridges, XAD or SPMDs are extracted by elution or dialysis (with hexane for lipid-based SPMDs). The elution of reverse-phase or XAD resin water sampler cartridges generally involves the use of a water miscible solvent first to remove water followed by a solvent of intermediate polarity such as DCM, methyl t-butyl ether or ethyl acetate. Combined extracts are then partitioned into hexane. Usenko et al. [78] compared six different SPE devices for the recovery of PCBs and OCPs from natural waters. These included XAD-2 resin, C18-filter disks and various C18 - and divinyl benzene-based adsorbents. They found that C18-based products, particularly “Speedisk,” a modified divinylbenzene solid-phase extraction device, gave the best overall recoveries.
Another water sampling device, solid-phase microextraction (SPME), involves immersing a polydimethylsiloxane-coated syringe into 3–5 ml samples. Hydrophobic compounds are adsorbed onto the coated fiber and the syringe then placed into a hot injection liner which desorbs these compounds into the GC [79]. The benefit of this approach is that it requires no solvent or multistepped cleanup/concentration procedures. A drawback is that the adsorption efficiency can be affected by complex water matrices and may not be suitable for all water sample types [80]. Particulate phases collected by filtration on glass fiber filters are treated like sediments and should be extracted by Soxhlet or pressurized liquid extraction.
Determination of lipid content
Lipid contents of biological samples should be determined during the analysis of POPs. Most studies have determined total extractable lipid gravimetrically by drying a fraction of the sample extract to constant weight [25]. However, results can vary widely among laboratories due to different extraction efficiencies of various combinations of solvent and extraction apparatus [81]. The benchmark method for total lipid is that of Bligh and Dyer [82]. A modified version of this method has improved sample handling and method precision [83]. Smedes [84] demonstrated that propanol:cyclohexane:water (8:10:11) was an effective substitute for the Bligh and Dyer chloroform:methanol:water. The Smedes method gave more consistent results for extractable lipids in fish tissue with low lipid content (∼1% lipid). Samples with very low fractions of lipid, such as blood plasma, are best analyzed using a colorimetric method [85].
Isolation of analytes from coextractives
There are two aspects to this phase of POPs analysis. The analytes must be separated from nonvolatile materials which affect the performance of GC columns, such as pigments, inorganic sulfur, and triglycerides. Also, there is a need to separate the OCPs and PCBs from each other as much as possible prior to GC analysis in order to limit coelution problems. These isolation steps are also useful, and sometimes essential, for the application of bioanalytical methods to PCBs/OCPs and dioxin-like activity (See “Bioanalytical methods for the quantification of OCPs and PCBs”).
Adsorption “cleanup” columns
Separation of PCB/OCPs from coextractives can be relatively straightforward for low-lipid samples such as soils, sediments and vegetation (Table 2). Generally, small silica gel or Florisil columns (either prepared in the lab or pre-purchased) should suffice. The purpose of this step is to remove coextractive pigments and to separate nonpolar PCBs (plus p,p′-DDE) from more polar OCPs (HCH, most chlordanes, dieldrin/endrin). This is achieved by applying the extract in a small volume of apolar solvent to fractionation by eluting with hexane followed by one or two other elutions of increasing polarity. Polar compounds are retained on the column. Basic alumina and silica gel columns have been effective for the separation of OCPs from fish lipids [89], although there is the possibility of minor losses due to dehydrochlorination of some OCPs, e.g., p,p′-DDT, on the alumina. The effectiveness of these adsorption columns depends on the mass and the water content of the adsorbent together with the polarity of the solvent. In general 3–8 g of absorbent is used in a 0.5–1.0 cm-diameter column with silica gel or Florisil deactivated with a low percentage of water (0–5%) [51, 90]. Typically the sample is applied to the top of the column in hexane and then a first elution solely of hexane separates PCBs from many OCPs. Subsequent elutions with hexane–diethyl ether or hexane–DCM recover the OCPs (p,p′-DDT, dieldrin, toxaphene) [21, 76]. Alumina and Florisil have the capacity to retain about 100 mg lipid per 10 g of adsorbent [26].
Size-exclusion columns
For high-lipid samples (>10% lipid), such as some fish tissues (e.g., salmonid muscle, cod livers) and marine mammal blubber, a lipid removal step must be included. This can be achieved using size exclusion or gel permeation chromatography (GPC), either in automated systems, using HPLC columns or by gravity flow columns. A 60 g bed of Bio-Beads SX3, a neutral porous styrene divinylbenzene resin (BioRad Laboratories, Hercules, CA, USA) can accommodate up to 1 g of lipid and achieves the separation of fish lipids from most PCBs and OCPs with a size exclusion of ∼400 Da [67]. Equivalent HPLC size-exclusion columns can be used [91]. The advantages of GPC are that it is nondestructive and that columns can be reused. A disadvantage is a requirement for large volumes of purified solvent (low-pressure or gravity systems) or expensive columns (HPLC). Automated GPC units are commercially available and allow for unsupervised elution of 12–24 samples. Due to required rinses of the unit sample lines, a portion of the sample is lost and must be accounted for when calculating final analyte concentrations. GPC eluates generally require an adsorption column fractionation step on silica or Florisil to remove remaining low molecular weight lipids, waxes and pigments that are not completely separated from the PCB/OCPs.
Lipid destruction
Lipid removal using sulfuric acid washing or elution of the extract through sulfuric acid (50%) in silica or KOH-treated silica columns is also effective for most PCB/OCPs, but does result in loss of some analytes such as dieldrin [92, 93]. Ethanolic–KOH treatment of Soxhlet extracts is also suitable for lipid removal and works well for PCBs [87], but may degrade chlorinated aliphatics such as DDT, which is converted to DDE. The acid- or base-treated extracts containing the PCBs and OCPs are then subjected to an adsorption column fractionation step on silica or Florisil.
Sulfur removal
Sulfur is coextracted with PCBs and OCPs and presents a particular problem for GC–ECD analysis of sediment extracts because of its strong response in this detector. It is also cytotoxic and needs to be eliminated prior to use with in vitro bioassays [94]. Sulfur can be removed by GPC but can also be removed using activated copper turnings (washed with concentrated HCl and held in an apolar solvent) or using mercury. The latter method removes sulfur more efficiently but is not recommended due to potential for contamination of the lab and lab effluent. Shaking with tetrabutylammonium sulfide has also been used to remove sulfur [95].
Evaporation steps
Solvent evaporation is generally used several times within all analytical methods for PCB/OCPs and successful analyte recoveries critically depend on minimizing losses during this step. Surrogate standards may be added prior to evaporation to monitor compound losses [96]. The Pesticide Analytical Manual of the US FDA [21] includes a discussion of solvent evaporation and an overview of the most widely used equipment, rotary evaporators, Kuderna–Danish apparatus and TurboVap equipment. The US FDA [21] overview favors the rotary evaporator because of greater range of use, e.g., for temperature-sensitive and semivolatile analytes, but notes that it is applicable to only one sample while both the TurboVap and K–D apparatus can be set-up for multiple samples.
Preparation for GC analysis
Following fractionation on silica or Florisil, final extracts are prepared in GC vials for analysis. Addition of an internal standard to check solvent volume is recommended at this stage. Careful evaporation is required at this step, and only high-purity compressed gas (usually nitrogen) should be used. This can be done using a stream of regulated gas via a disposable glass pipet and heating block or via multineedle devices (e.g., “N-Evap”).
Quantification methods
Overview
Numerous analytical approaches are available for quantifying PCBs and OCPs. As with extraction/isolation steps, only general comments are provided here for GC analysis of ortho-substituted PCBs and OCPs. However, a major consideration is that the laboratory will have access to modern capillary GC equipment and either electron capture or mass spectrometry detection. Some general guidance on the application of gas chromatographic analysis of ortho-substituted PCBs and OCPs is provided in Table 4. Cochran and Frame [97], in their review of the GC analysis of PCBs, noted that separation requires attention to (1) proper injection to minimize analyte band-broadening; (2) choice of carrier gas; (3) optimized carrier gas velocity; (4) GC oven programming; (5) column dimensions such as length, inside diameter, film thickness, number of plates; and (6) the type of column stationary phase.
Bioanalytical methods based on immunoassays, or in vitro bioassays for dioxin-like activity, have become widely available over the past ten years for the screening of sample extracts for POPs. These methods and selected applications are briefly reviewed in this section.
GC injection ports
Optimization of injection conditions is critical to proper GC analysis [98, 99]. In the analysis of POPs and OCPs, problems often occur with nonvolatile coextractives such as triglycerides and pigments that, despite various isolation procedures, are still present in final extracts. Most GC applications for PCB and OCP analysis have employed split/splitless injection systems, although on-column injection has also been used. Although on-column injection avoids artifacts associated with heated split/splitless systems (i.e., degradation of labile compounds), it requires highly clean extracts to avoid matrix effects [99]. Retention gaps consisting of an empty capillary column have been used to reduce these effects and permit larger on-column injection volumes. Pressure programming allows improved control over carrier gas flow, resulting in improved separation of PCB and OCPs.
GC columns
Fused silica open tubular capillary columns, generally coated with nonpolar or medium-polarity chemically bonded liquid phases are almost universally used for GC separation of PCBs and OCPs [99]. The use of capillary columns revolutionized PCB analysis, allowing identification of the individual congeners. This improved the comparability of the analytical data from different sources and helped to establish a basis for the understanding of geochemical cycles and toxicological implications. The basic technology for separation of PCB congeners described by Mullins et al. [100] has not changed greatly over the years. These authors used a 5% phenyl methyl silicone phase and a long temperature program (100 min) to achieve high-resolution separation of PCB congeners. Improved routine separations of PCBs have been achieved using 60 m×0.25 mm i.d. columns with hydrogen carrier gas. PCBs within a homolog group elute according to their number of ortho chlorines: 4<3<2<1<0. Coelution of congeners remains a problem for routine analysis by GC–ECD. However, several modified polydimethylsiloxane phases with n-octyl or n-octyldecyl substituents, e.g., DB-XLB (J&W Scientific, Folsom, CA, USA) and HT 8 (SGE Inc., Austin, TX, USA) can resolve all but four pairs of significant congeners and five pairs of minor congeners [18, 101] using mass spectrometry detection. The coelution of PCB congeners on the more commonly used 30 m DB-5 columns is outlined in Table 3. Coelutions are thus important even for routinely monitored congeners such as CB 28, 31, 105 and 153. However, separation of these congeners can be routinely achieved on 60 m 5% phenyl methyl silicone phase capillary columns using H2 carrier gas and on n-octyl phases as well [18, 101]. Use of dual-column GC, in which the sample is injected simultaneously into nonpolar and medium-polar phase columns, is recommended if a suitable instrument is available. Alternatively, confirmation of peak identity in a subset of samples, using a second column of different polarity, can be done instead. Although other carrier gases such as He can be used, the best peak resolution and efficiency is achieved by using H2.
Table 3Significant PCB congener coelutions on 5% phenyl phases [97]ClassificationaCoeluting PCBsNumber of chlorinesA4, 101, 2A9, 72, 2A12, 132, 2A17, 152, 3 A27, 243, 3A32, 163, 3A28, 313, 3A33, 20, 533, 3, 4A43, 494, 4A47, 75, 484, 4, 4A44, 594, 4A37, 424, 4A71, 41, 644, 4, 4A66, 954, 5A56, 604, 4A84, 89, 101, 905, 5, 5, 5A117, 87, 1155, 5, 5A77, 1104, 5A135, 144, 1246, 6, 5A147, 1096, 5A123, 139, 149, 1185, 5, 6, 5A114, 1335, 6A131, 1226, 5A153, 132, 1056, 6, 5A176, 1307, 6A164, 163, 1386, 6, 6A158, 1296, 6A175, 1667, 6A173, 157, 2017, 6, 8A170, 1907, 7A198, 1998, 8A203, 1968, 8aCapital A indicates a major Aroclor congener
Coelution of major OCPs with each other or with PCBs is also a problem in GC–ECD analysis, despite preseparation by adsorption chromatography. For example, p,p′-DDD can coelute with cis-nonachlor, CB99 with a trans-nonachlor isomer [10], p,p′-DDE with dieldrin and CB85, and toxaphene congener Parlar 50 with CB128. Recently, the coelution of CB180 and brominated diphenyl ether congener 47 has also been noted [102].
Chiral GC separation of OCPs and PCBs
Most OCPs on the POPs list have optically active or chiral isomers (e.g., α-HCH, o,p′-DDT, the main constituents of technical chlordane, cis-/trans-chlordane, heptachlor, as well as chlorobornanes in toxaphene) [103]. There are also 19 PCB atropisomers which are also chiral as a result of restricted rotation at the central biphenyl bond. While PCBs and OCPs are racemic mixtures when manufactured, microbial degradation in soils and water as well as biotransformation reactions in biota can result in nonracemic patterns in environmental samples. Crucial for chiral analysis is the availability of chiral capillary GC columns such as those with various cyclodextrins chemically bonded to a polysiloxane. These phases are relatively heat-stable and have low bleed [104]. Current methods range from the simple use of 30 m chiral columns to a two-dimensional “heart-cutting” technique. This involves a redirection of a targeted portion of the sample after elution from a methyl/phenyl polysiloxane column into a chiral column via a computer-controlled pneumatic valve. This provides higher peak capacity and generally further separation of chiral compounds. While use of chiral GC separations is not part of routine OCP analysis, it is a well-developed technology that is relatively easy to implement in existing GC–ECD and GC–MS instruments. This technique has useful applications for tracking sources of OCPs and transformations of PCBs. For example, Jantunen et al. [105]) and Leone et al. [106] have shown that enantiomer fractions of chlordane isomers can be used to identify emissions of this pesticide from soils. Wong et al. [107] calculated minimum transformation rates for PCB atrophisomers CB95 and 136 in the Lake Superior food web using enantiomer ratios in predators and prey species.
GC–ECD
Since the 1960s, POPs have been determined by GC with electron capture detection (ECD), initially using packed columns. Capillary GC–ECD began to be routinely applied by the early 1980s. Use of GC–ECD is recommended for routine analysis of OCPs, except for toxaphene, as well as for ortho-PCBs (Table 4) but not for non-ortho-PCBs [14]. This instrumentation is widely available at relatively low cost from at least four instrument manufacturers. A substantial knowledge base exists on the use of this 40-year-old technology. GC–ECD is capable of determining PCB/OCPs at low ng/g levels or higher in environmental matrices. Although at one time tritium-based ECDs were once available, the 63Ni detector is now universally used. This detector is operated at high temperatures, 300–350 °C, which makes it relatively unaffected by column bleed. ECD suffers from the potential for false positives due to interferences such as those from sulfur, phthalate esters, and negative peaks generated by hydrocarbons. ECDs are normally operated with N2 or argon/methane gas, which combines with the flow from the GC column (He or H2 carrier gas). Gases used for GC–ECD must be ultrapure to protect both the GC column (which can be oxidized by trace oxygen or siloxanes hydrolyzed by trace water) and the ECD itself. Recent refinements in ECD technology include the use of microcells which have greater linear range than older detector cells [108] and can provide greater sensitivity.
Table 4General guidance on GC analysis and data reporting for PCBs and OCPsGC detectorAnalytesConfigurationAdvantages/disadvantagesDetection limitsaCapillary GC - with electron capture detectionAll ortho-subsituted PCBs & all OCPs on the POPs list except toxaphene30 or 60 m×0.25 mm id. Column with H2 carrier gas. Dual column nonpolar (DB-1) and intermediate polarity columns (DB-5).Relatively inexpensive and easy to operate. Similar response factors for most OCs.DDT/DDE ∼ 1 pgGood sensitivity for all POPs. Adequate for routine tasks. High potential for misidentification of some POPs due to coeluting peaks.HCB ∼0.5 pgQuadrupole mass spectrometry in electron ionization (EI) modeAll PCBs & all OCPs on the POPs list except toxaphene30 m×0.25 mm i.d. low-bleed columns with He carrier gas. Selected ion mode for target POPs. Moderately expensive and more complex to operate and maintain. Newer instruments (post 1997) have adequate sensitivity for routine POPs monitoring at low pg/μL concentrations. Much less potential for misidentification than with ECD.DDT/DDE ∼1–10 pgHCB ∼1–10 pgDieldrin ∼25 pgToxaphene ∼500 pg (as tech mixture)Quadrupole mass spectrometry in electron capture negative ionization (ECNIMS) modeToxaphene and other highly chlorinated OCPs and PCB with >4 chlorines30 m×0.25 mm i.d. low-bleed columns with He carrier gas. Selected ion mode for target POPs.Comparable sensitivity in ECNIMS mode to ECD in SIM mode for some POPs. Much less potential for misidentification than with ECD.DDT/DDE ∼0.1 pgHCB ∼0.1 pgDieldrin ∼1 pgToxaphene ∼10 pg (as tech mixture)Ion trap mass spectrometry using MS/MS modeAll PCBs, All OCPs on the POPs list30 m×0.25 mm i.d. low-bleed columns with He carrier gas. Same columns as quadrupole MS.Comparable sensitivity to ECD in MS/MS mode for some POPs. Much less potential for misidentification than with ECD.DDT/DDE ∼1 pgHCB ∼1 pgDieldrin ∼5 pgToxaphene ∼100 pg (as tech mixture)High-resolution magnetic sector mass spectrometry in electron ionization (EI) modeAll PCBs, all OCPs on the POPs list except toxaphene30 m×0.25 mm i.d. low-bleed columns with He carrier gas. Selected ion mode for target POPs at 10,000× resolution.Comparable sensitivity to ECD in SIM mode. Highly reliable identification at low pg/μL levels.DDT/DDE ∼0.05 pgHCB ∼0.05 pgDieldrin ∼0.1–0.5 pgToxaphene ∼10 pg (as tech mixture)aInstrumental detection limits at S/N of ∼10
GC–MS
Capillary GC coupled with mass spectrometry (GC–MS) revolutionized environmental organic analysis in the 1980s [109], particularly with the advent of bench-top instruments. Current GC–low-resolution (quadrupole) mass spectrometry (LRMS) instrumentation is capable of determining most PCB/OCPs at low pg concentrations using electron ionization (EI) in selected ion mode. Electron capture negative ion ionization (ECNI) is capable of detection of low femtogram amounts of highly chlorinated OCPs such as chlordane and toxaphene congeners. In this mode, a reagent gas, usually methane, is introduced into the ion source and negative ions are monitored. The ECNI mode also has a more limited linear range than the EI mode, is more difficult to operate routinely due to greater sensitivity to temperature variations, and needs more frequent cleaning due to the reaction of the reagent gas with surfaces in the ion source.
GC–ion trap MS (ITMS) in MS/MS mode offers an equivalent sensitivity to quadrupole ITMS in selected ion mode and improved specificity by examining product ions of major mass fragments. ITMS has been successfully applied to detect mono-ortho-PCBs at subpicogram levels in a range of environmental samples [110] and to determine pg levels of toxaphene congeners [90]. In general, both quadrupole and ion trap LRMS can suffer from false positives due to unit mass resolution. Thus experienced analysts are needed to interpret results using confirmatory information such as full-scan analyses, fragmentation patterns and ion ratios.
GC coupled to HRMS provides much higher specificity for individual PCB/OCPs due to its ability to provide millimass resolution and can also be used for the determination of all ortho-substituted PCBs (e.g., EPA Method 1668 [35]) and OCPs too, and provides a very high level of confidence in the results compared to GC–ECD and LRMS.
The use of GC–MS allows for the application of isotope-dilution techniques and these have increasingly come into routine use for PCB/OCP analysis in environmental samples due to the sensitivities of the latest generation of quadrupole and ion-trap MS systems [111, 112]. A wide range of PCBs and OCPs are available as 13C-labeled compounds from several chemical supply companies, which, when added at the extraction step, increase precision and reproducibility for the native analytes.
Bioanalytical methods for the quantification of OCPs and PCBs
Enzyme-linked immunoabsorbent assays (ELISA)
ELISAs have been used to quantify most DDT/DDE, HCH isomers, toxaphene and cyclodiene OCPs as well as PCBs in environmental samples [113, 114]. ELISAs are based on competitive binding in which the binder molecule, an excess amount of labeled analyte or coating antigen, and the target analyte are allow to approach equilibrium. The sample antigen competes with the coated antigen for binding sites on the labeled antibody; after a wash step, detection is performed by adding substrate and chromophore. Quantification is generally performed via spectrophotometric measurements and the amount of analyte in the sample is interpolated from a calibration curve. Although widely used to screen for current-use herbicides and insecticides as well as their polar metabolites [115], the development of competitive immunoassays for neutral hydrophobic OCPs has lagged [114]. This is in part due to the need for low detection limits. Also, ELISA analysis for some OCPs such as HCH (and lindane) has been challenging due to the small size of the HCH molecule, its structural symmetry, and possibly most importantly, its lack of aromatic structures or other atomic groups capable of supporting delocalized electron networks [116]. The development of a sensitive ELISA for DDT required synthesis of a ligand with Cl replaced by b-alanine carboxamide and lindane after degradation to trichlorobenzenes [117]. Commercial ELISA kits for detection of PCBs and most OCPs are available from Millipore Corp. (Billerica, MA, USA) and Strategic Diagnostics (Newark, DE, USA). Detection limits for PCBs and OCPs are typically in the low μg/L range in water or 0.1–1 μg/g range in soil and plant extracts. These kits are meant to be used with relatively little sample preparation and, although semiquantitative, are ideal for screening samples and complement more elaborate techniques involving GC analysis (Fig. 1).
Fig. 1Illustration of the basic components of an ELISA for detection of OCPs and PCBs in environmental samples or extracts. Sample antigen (analyte) competes with antigen for binding sites on coating protein; after a wash step, detection is performed by adding substrate and chromophore
Shivaramaiah et al. [118] used an ELISA specific for DDE to survey DDT residues in surface soils in three river valleys of northern New South Wales, Australia. The technique allowed the analysis of large numbers of samples, thus permitting the areas of contamination to be mapped. Galloway et al. [119] used an ELISA to detect PCBs in extracts of mussels from New Bedford Harbor (USA). The ELISA and GC–ECD results were highly correlated, although the latter were about 20% lower than quantitation by GC. Samples were extracted and lipid partially removed by chromatography on Florisil prior to exchanging the sample into a phosphate buffer/methanol solution for the immunoassay. The use of the ELISA saved additional isolation steps and GC analysis. In general, fatty coextractive materials must be removed prior to application of ELISAs for OCPs. Skerritt et al. [120] examined the application of ELISAs for DDT and cyclodiene insecticides, heptachlor and endosulfan in plant-derived foods. They found that cleanup was necessary for foods that yielded highly colored extracts such as coffee and spinach and for oily products such as cottonseed.
Quality assurance issues for PCBs and OCPs
Ancillary data
A significant amount of ancillary data is required in order for the PCB/OCPs measurements to be properly interpreted in terms of temporal and/or spatial trends and exposure of wildlife and humans. Tanabe and Subramanian [2] have extensively reviewed the factors to be considered for the selection of bioindicator species, including human tissues. Analytical lab reports need to include quality assurance data so that results and lab performance can be assessed independently [6], as outlined in Table 5. The objective here would be to have a record of the entire processing of the sample from preparation through to reporting concentrations that can be evaluated independently. Therefore, the individual labs should report concentrations for analytes, blanks and reference materials. Data reports should also include instrument calibration results and recoveries. This would enable method detection limits (MDLs) to be calculated independently of the lab if necessary.
Table 5Minimum reporting dataset for POPs analysisInformationDetailsSampling protocolsMethod, number, size and representativenessStorage temperature and locationSample tracking informationDate received, date analyzed, lab batch number or other unique identifiedPublished analytical methode.g., EPA methodLimit of detection/quantificationSee “QA procedures”BlanksReagents and also field blanks if possibleRecoveriesDuplicatesCalibrationSource of standards; date stocks preparedSurrogate and internal standardsQA of cofactorsSuch as lipid, organic carbon and moisture contentConfirmatory testse.g., Use of second GC column or other detection systemData manipulationsBlank subtraction, recovery correction
Field and lab blanks
Determination of PCBs and OCPs requires the analysis of blank samples because of the ubiquitous nature of these contaminants, as discussed previously. For air and water, the sampling media (i.e., pre-extracted water, or the XAD resin, C18 cartridges, or SPMDs) can serve as blanks. Ideally these media should be transported to the field and exposed briefly to the sampling equipment, thus serving as a field blank. Additional samples of the media should be archived in the preparation lab for use as lab blanks. For soils and fatty samples, US EPA Method 1668 for PCBs recommends sand and corn oil, respectively, as lab reference blanks [35]. Clean facilities, ideally with HEPA and carbon-filtered air are recommended, but are not essential, for the determination of PCBs, except for ultralow levels in water. Contamination of lab air can be checked by use of SPMDs or by exposing C18-disks in the lab for a specified period.
QA procedures
Key elements in QA/QC for POPs analysis are the use of reference materials, the use of quality charts, participation in interlaboratory studies, and the use of guidelines for sampling and analysis [121].
Recovery studies in which individual analytes are spiked into the sample matrix prior to extraction are useful for determining the performance of the analytical method. However, they do not provide much information on the true extractability of analytes from naturally contaminated matrices. This can only be accomplished by extracting certified reference materials (CRMs) or, if none are available for a particular type of sample, some other well-characterized material that has been analyzed by an experienced reference lab [25, 74]. Recovery studies can be also useful for determining method detection limits if conducted at concentrations within 3–5 times the estimated limit of detection determined from the instrumental response or from blanks [122].
As a routine measure, spiking surrogate recovery standards into each sample provides useful information on losses of analyte from the extraction step onwards. However, no single PCB or OCP can be representative of all of the organochlorines being determined, and thus recovery correction should be performed with caution. Isotopically labeled surrogates are ideal for analyses of PCBs and OCPs that are being performed by LRMS and HRMS, and isotope dilution techniques correct for the recoveries of these surrogates.
If blanks are significant (for example, averaging greater than 10% of the average level of total PCBs), then blank correction should be carried out. This should only be done with a robust method blank based on at least seven samples or a blank known to be representative of a batch of samples. Ferrario et al. [29] subtracted an MDL value based on the blank+2×σb on the rationale that this ensured a low probability (5%) of false positives compared to subtraction of the average blank only.
Detection limits
Detection limits for PCB/OCPs depend on the analytical method but also on the sample size and QA considerations, e.g., information available from blank or control samples and recovery studies. The selection of detection limits also depends on the goals of the program and how much emphasis is placed on the reliability of the results versus the need to achieve broad geographic coverage, to meet or exceed regulatory or risk assessment guidelines (Electronic Supplementary Material, Table 3 [145–152]). and to avoid reporting “less thans” for a high proportion of samples. A recent US EPA report has reviewed the approaches to calculating detection limits in the analysis of POPs and other environmental contaminants [122].
Detection limits are usually expressed as a concentration, i.e., they are based on the average weight of the sample analyzed, as summarized in Table 6. The recommended method for determining the MDL [122] is to analyze a series of samples, as free from the target analyte as possible (a minimum of seven replicates), that have a concentration of the analyte that is at least equal to, or in the same concentration range as (within a factor of five of), the estimated detection limit. All of the replicate aliquots are processed through the entire analytical method. The standard deviation (σ) of the replicate measurements is determined and the MDL is then determined by multiplying the standard deviation by the Student’s t-value at a 99% percentile for n−1 degrees of freedom. The MDL is thus an estimate of the measured concentration at which there is 99% confidence that a given analyte is present in a given sample matrix. The MDL is defined by the analytical laboratory and is likely to vary between labs because of different blanks and instrumental conditions.
Table 6Detection limits defined by various organizationsOrganizationTerminologyCalculationUS EPAMethod detection limit (MDL)Minimum level of quantitation (ML)American Chemical SocietyLimit of detection (LOD)Limit of quantitation (LOQ)International Organization for Standardization / International Union of Pure and Applied Chemistry (ISO/IUPAC)Critical value (CRV) minimum detectable value (MDV)Limit of quantitation. (LOQ ISO/IUPAC)
The MDL is analogous to the “CRV” described by Currie [123] and reported by ISO/IUPAC, which is tn−1, 95×σb of a sample blank. Note that the t-value is based on the 95th percentile. Thus the MDL and CRV differ in the multiplier used. There may be practical difficulties with measuring some analytes if there no blank value, however the σ of a low-concentration standard can be used. The MDV is similar to the CRV but measured at the detection limit.
Limit of detection (LOD) described by the American Chemical Society [124–126] is equal to the field blank b plus three times the standard deviation of the blank (σb). is numerically equivalent to the MDL as b approaches zero or when results are blank-subtracted. This definition has been widely used because of its publication in the peer-reviewed literature. It should be noted that the US EPA “MDL” and the ISO/IUPAC “CRV” implicitly assume that results are adjusted for interferences or blank values.
The US EPA [122] and other organizations have also defined various lowest levels of quantitation which give a higher degree of confidence in the results. The “minimum level of quantitation” is the lowest level at which the entire analytical system must give a recognizable signal and an acceptable calibration point for the analyte. It is “equivalent to the concentration of the lowest calibration standard, assuming that all method-specified sample weights, volumes, and clean up procedures have been employed. The ML is calculated by multiplying the MDL by 3.18 and rounding the results to the number nearest to (1, 2, or 5) ×10n, where n is an integer.” [122]. The ACS LOQ is defined as being b+10×σb above the gross blank. The ISO/IUPAC LOQ is defined as “the signal or analyte (true) value that will produce estimates having a specified relative standard deviation (RSD), commonly 10%.” [123]. It is calculated as: where σQ is the standard deviation of the blank, and where KQ=10.
Table 7 presents some approximate values for MDLs that should be achievable assuming typical responses of ECD and MS systems, as well as three types of bioanalytical methods, to various OCPs/PCBs and negligible blanks for the individual PCBs and OCPs in a 10 g sample.
Table 7Estimated MDLs for individual PCBs and OCPs (ng/g wet wt), after Liem [127]AnalyteELISA (soil/fly ash)ECDaMSb (low resolution; EI-SIM)MSc (high resolution; EI-SIM)PCB 28–0.050.10.01PCB 52–0.050.10.01PCB153–0.050.050.005PCB180–0.050.020.005p,p′-DDE500.050.050.01Toxaphene5000.050.020.005Lindane4000.010.040.01HCB–0.010.020.005Dieldrin100–5000.010.020.01cis-Chlordane100–5000.030.050.01Total PCB 100–5000.1–1.00.1–1.00.01–0.1aAssumes MDL=3×SD of low-level spike (US EPA 2003), assuming 10 g sample and a sample volume of 0.5 mLbAssumes use of isotope dilution (13C) surrogates; electron ionization, selected ion monitoring mode with 10 g sample, and a sample volume of 0.5 mLcAssumes use of isotope dilution (13C) surrogates; electron ionization, selected ion monitoring mode with 10 g sample, and a sample final volume of 0.05 mL
Reference materials
Certified reference materials (CRMs; also known as standard reference material or SRM) are essential for determining the accuracy of analytical methods for most PCBs and OCPs [128–130]. The CRMs currently available are shown in Table 8. While no single CRM is currently certified for all POPs, there are published uncertified values available for a broad array of individual PCBs, DDT isomers and toxaphene congeners, especially for NIST 1588a cod liver [131]. Sediment CRMs are freeze-dried and thus potentially not representative of sediment materials normally extracted. Some CRMs, notably SRM 1945a, cannot be shipped outside of the USA without CITES permits.
Table 8CRMs for PCBs/OCPs in biota and sediment (adapted from de Boer and McGovern [128])CRMSourceTissue/speciescis-CHLtrans-CHLDieldrinDDTHCBMirexPCB congenersSRM1974bNISTmusselXXXXSRM1976NISTlake troutXXXXSRM1588aNISTcod liverXXXXXSRM1945NISTwhale blXXXXXXSRM2974NISTmusselXXXSRM2977NISTmusselXXXXXSRM2978NISTmusselXXXX140/OCIAEAplantXXXBCR598BCRcod liverXXXXXXCARP-1NRCCcarpXBCR349BCRcod liverXBCR350BCRmackerelXBCR682BCRmusselXBCR618BCRherringXEDF 2525CILlake troutXEDF 2514CILsoilXSRM1944NISTsedimentXXXXSRM1939aNISTsedimentXXXIAEA383IAEAsedimentXIAEA408IAEAsedimentXXXHS-1NRCCsedimentXHS-2NRCCsedimentXBCR536BCRsedimentXDX-1BCRsedimentDX-1BCRsediment
Criteria for evaluating the desirability and efficacy of different analytical methods in environmental monitoring and surveillance
Overview
The data generated by chemical analytical and bioanalytical methods for PCBs/OCPs are often used to assess spatial and temporal trends of contamination in environmental media on a national, regional and international basis. The data will also be used by regulators and risk assessors to determine if concentrations exceed guidelines. Thus methods must give accurate results, which are comparable among labs, and at concentrations which are at or below action limits for food or guidelines on wildlife protection (see Electronic Supplementary Material Table 3 [145–152]).
Comparison of detection limits for widely used instrumentation for PCBs and OCPs in Table 7 with guideline limits (Electronic Supplementary Material Table 3) suggests that conventional chemical analytical methodology and bioanalytical methods can meet and exceed these limits in some cases by orders of magnitude. Sample size, lack of coextractive interferences, and final volume are important considerations.
The QA programs discussed above can assure laboratory personnel that accurate and reproducible results are being generated and can be validated by participating in regional, national or international interlaboratory comparisons using samples with assigned values [132]. In the context of implementation of the Stockholm POPs Convention, global comparability is the goal and thus successful participation in international programs should be a requirement of any participating lab. The challenge for future POPs monitoring, especially in developing countries, will be to assess the quality of data and to provide guidance for best laboratory practices.
Interlaboratory comparisons
Interlaboratory comparisons of POPs were conducted, along with the first environmental measurements, in the late 1960s, and became widely adopted in the 1990s [27]. The major international programs related to PCBs and OCPs are the QUASIMEME (Quality Assurance of Information for Marine Environmental Monitoring in Europe) program [53] and the determination of OCPs in food organized in 1993 and 1994 by the Global Environmental Monitoring Scheme (GEMS) of the World Health Organisation (WHO) [133]. Other important ongoing programs are the determination of OCPs/PCBs in human milk under the auspices of AMAP [134] and the intercomparisons operated by the National Institute of Standards and Technology and the Marine Mammal Health and the National Oceanic Atmospheric Administration, National Marine Fishery Service. These programs have been a major stimulus for improved analysis of POPs globally in the past ten years.
The results from interlab programs indicate that training and specific guidance on key aspects of the analytical methods, such as preparation of calibration solutions, are important [26].
PCB/OCP method accuracy
How accurate should analytical methods be? Obviously, they should be accurate enough to identify changes in temporal or spatial trends in contaminant concentrations [121]. Using the approach of Nicholson [135], accuracy is defined as being 95% certainty that the measured values will fall within upper and lower limits defined by:
where μ is the true concentration, σ is the precision, ±b is the bias, where 1.645=tn−1, 0.95, and where the dataset is normally distributed. Both σ and ∣b∣ can be calculated from multiple analyses of a CRM, where σ is the standard deviation of the analyte and ∣b∣ is the relative percent deviation from the certified value (μ). As a percentage of the true value, the accuracy is given as:
Between-laboratory accuracies of 15–20% have been achieved for PCB congeners among experienced labs participating in QUASIMEME [27].
The European Commission [136] has defined the acceptability of an analytical method based on repeated analysis of PCB/OCPs in a CRM as follows:
−50% to +20% for analytes in the range of <1 μg/kg−30% to +10% for analytes >1 μg/kg to 10 μg/kg−20% to +10% for analytes >10 μg/kg.
These percentages are also recommended acceptability guidelines for samples spiked with PCBs/OCPs where no CRM is available.
Emerging issues in analytical methods and future directions
Analytical methods for trace organic analysis are constantly evolving and improving, as can be seen from recent reviews of environmental analysis [58, 137, 138]. Although PCB/OCP analysis is a mature area within environmental analytical chemistry, with basic methods that have remained unchanged for the past 20 years, there have been many new technologies that have been gradually adopted and undoubtedly more will emerge in the future. Some emerging procedures with low environmental impact (SPME, microscale glassware, low solvent use, etc.) may be particularly suited for developing countries where analytical budgets are small and product delivery times are lengthy. Thus, strategies must be considered that will allow improved techniques to be adopted by such labs.
A major development in the analysis of POPs has been the introduction of semi-automated extraction instruments for PLE (Accelerated Solvent Extraction; Dionex Instruments, Sunnyvale, CA, USA). The PLE extraction can be combined with solid-phase adsorbents to extract and isolate analytes of interest, for example to remove lipids during the extraction of POPs from fatty samples [139]. In general, automated and semi-automated systems are available for cleanup and isolation of samples for PCB/OCP using disposable solid-phase cartridges, as well as high-performance/pressure LC columns.
Advances in GC that are potentially applicable to PCB/OCP analysis in developing countries are the commercial availability of multidimensional GC (2D-GC) and “fast GC.” In GCxGC, or comprehensive two-dimensional GC, all of the analyte mass is transferred to a second column, and thus resolving power is increased by an order of magnitude [140]. When applied to PCB separations using GCxGC ECD, three- to four-fold lower LODs were obtained compared to one-dimensional GC, and dioxin-like no- and mo-PCBs could be separated from other congeners [141]. Rapid GC separation of PCB congeners by “fast GC” were obtained on narrow-bore (0.1 mm) columns which reduced peak widths and shortened total run times for the full PCB congener mix to minutes [142]. Both techniques can be run with a micro-ECD and are thus relatively inexpensive to implement.
Conclusions
Regulators and managers require data on PCBs and OCPs concentrations in environmental media to assess water quality, food contamination, and to infer possible biological effects. Indeed, the global assessment of the success of the Stockholm POPs convention assumes a chemical measurement capability in all UNEP regions which will deliver data on concentrations in key media [6].
The analytical methodologies discussed here refer to an “analytical system” encompassing information on the collection and storage of samples, the procedures used to extract, isolate, concentrate, separate, identify, and quantify POPs residues in samples, as well as specific quality control and reporting criteria. All aspects of this system must be in operation for POPs to be analyzed and reported.
The chemical analytical methodology for the determination of PCBs and OCPs is a mature area within environmental analytical chemistry as a result of research and development over the past 30–40 years. Basic steps in the quantification of OCPs and PCB congeners have hardly changed in the past 20 years. Analytical methods for the determination of OCP/PCBs in foods, soils, sediments, fish, birds, mammals (including human milk and blood) are available and could be implemented at relatively low cost in developing countries. However, access to modern capillary GC equipment with either electron capture or mass spectrometry (MS) detection to separate and quantify PCBs/OCPs is required in order to conduct the analysis and to take part in regional and international intercomparsons.
In general, ELISAs are very useful tools for the rapid assessment of PCB/OCPs contamination, especially in areas of former heavy use. They are particularly well-suited to laboratories in developing countries which may have access to spectrophotometric equipment but not to GC instrumentation.
Existing analytical methods for PCB/OCPs can determine over 100 individual components at low ng/g concentrations in many environmental media using high-resolution capillary GC–ECD. However, the number of certified values for OCP/PCB congeners in certified reference materials is more limited (approximately 23 PCB congeners and 15 OCPs in NIST 1588a cod liver). At a minimum, the OCP/PCBs for which there are certified values in readily available CRMs should be determined (approximately 38). With this number of analytes, the information would be useful for both regulatory actions as well as for source identification using multivariate analysis or other “fingerprinting” methods.
Interlab comparisons of POPs analysis over the past ten years have shown that availability of accurate analytical standards is a fundamental requirement of an analytical program designed to quantify trace organic contaminants such as POPs. Agencies such as GEF and UNEP Chemicals should give top priority to ensuring that certified analytical standards are available to all labs on a continuing basis.
Quality assurance programs are critically important for demonstrating the performance of analytical methods for POPs within a lab and between labs. QA requirements for PCBs/OCP analysis are well known and include the use of certified reference materials, field and laboratory blanks, the use of quality control charts to monitor long-term lab performance, participation in interlaboratory studies, and the use of guidelines for sampling and analysis.
Determination of PCBs/OCPs requires the analysis of blank samples because of the ubiquitous nature of these contaminants. If blanks are significant (for example, averaging greater than 10% of the average level of total PCBs), then blank correction should be carried out.
As a routine measure, spiking surrogate recovery standards into each sample provides useful information on losses of analyte from the extraction step onwards. However, no single PCB or OCP can be representative of all the organochlorines being determined, and thus recovery correction should be performed with caution. Isotopically labeled surrogates are ideal for the quantification of PCBs and many OCPs via LRMS and HRMS; the application of isotope dilution techniques can correct for the recoveries of these surrogates.
Detection limits depend not only on the analytical method used but also on the sample size and QA considerations, e.g., on information available from blank or control samples and recovery studies. Detection limits should be calculated as described by US EPA or by IUPAC/ISO methodology. Comparison of detection limits for widely used instrumentation for POPs with action limits for POPs in food and tissue residue guidelines suggests that current GC–ECD and GC-MS analytical methodology for PCB/OCPs can meet and exceed these limits, in some cases by orders of magnitude.
Some emerging new analytical techniques, such 2D-GC and “fast GC” using GC–ECD, may be well-suited for use in developing countries in the near future given their relatively low cost and their ability to provide high-resolution separations of OCP/PCBs. Procedures with low environmental impacts (microscale, low solvent use, etc.) may be particularly well-suited to developing countries where analytical budgets are small and product delivery times are lengthy. Thus, strategies must be considered that will allow improved techniques to be adopted by such labs.
Electronic supplementary materials
Below is the link to the electronic supplementary material.
Table 1
(DOC 4 kb)
Table 2
(DOC 13 kb)
Table 3
(DOC 6 kb)
|
[
"pcbs",
"organochlorines",
"pesticides",
"analysis",
"pops",
"review",
"extraction",
"gc–ecd",
"gc–ms",
"sample preparation and qa/qc"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"P",
"R"
] |
Diabetologia-3-1-2100429
|
Vitamin B12 and folate concentrations during pregnancy and insulin resistance in the offspring: the Pune Maternal Nutrition Study
|
Aims/hypothesis Raised maternal plasma total homocysteine (tHcy) concentrations predict small size at birth, which is a risk factor for type 2 diabetes mellitus. We studied the association between maternal vitamin B12, folate and tHcy status during pregnancy, and offspring adiposity and insulin resistance at 6 years.
Introduction
The ‘thrifty phenotype’ hypothesis [1] introduced a new paradigm in the aetiology of chronic disease and stimulated research into the relationship between maternal nutrition, fetal growth and offspring risk of type 2 diabetes mellitus. India has the largest number of diabetic patients in the world [2] and this may be related to the fact that Indian babies are amongst the smallest in the world [3]. We were the first to demonstrate that low birthweight of Indian babies predicts insulin resistance and adiposity in childhood [4]. Subsequently we demonstrated that the apparently small Indian newborn babies are relatively adipose, hyperinsulinaemic and hyperleptinaemic compared with white babies [5, 6]. These babies grow into shorter, thinner (low BMI) but more adipose (higher body fat per cent and higher central fat) adults compared with whites, and have a many times higher risk of type 2 diabetes [7]. Thus, in addition to the role of putative genetic and modern day lifestyle factors, the intrauterine environment may also influence the risk of type 2 diabetes mellitus. The Pune Maternal Nutrition Study (PMNS) is the first study in India to investigate the relationship between maternal nutrition and offspring risk of type 2 diabetes and cardiovascular disease. Data were collected on the mother’s size, diet, micronutrient status and physical workload during pregnancy [6, 8, 9], and the newborns were measured in detail at birth and every 6 months thereafter. At 6 years of age, the offspring were investigated for risk of type 2 diabetes and cardiovascular disease.
We have reported that a higher frequency of maternal intake of green leafy vegetables (GLVs), milk and fruit and higher erythrocyte folate concentrations are associated with larger newborn size [8]. In a small sub-sample we found that higher maternal circulating total homocysteine (tHcy) predicted low offspring birthweight [10]. Other studies in Pune have shown that vitamin B12 deficiency is common and contributes to hyperhomocysteinaemia, while folate deficiency is rare [11, 12].
We therefore hypothesised that vitamin B12 and folate deficiency in the mother during pregnancy would predict greater adiposity and insulin resistance in the offspring.
Methods
The recruitment of mothers and follow-up of the children in the PMNS (Fig. 1) has been described [6, 8, 9]. The study started in 1993. Non-pregnant married women living in six villages near Pune had detailed anthropometric measurements every 3 months and menstrual dates were recorded monthly. Women who became pregnant were assessed at 18 ± 2 and 28 ± 2 weeks of gestation for anthropometry, physical workload, dietary intake (24 h recall and food frequency questionnaire [FFQ]) and measurement of circulating nutrients. According to the National Nutritional Anemia Control Programme, women were given 100 tablets of iron (60 mg per tablet) and folic acid (500 μg per tablet) from 18 weeks of gestation. Newborn babies were measured within 72 h of birth (1994–1996), and the children were followed up post-natally with repeat anthropometry every 6 months.
Fig. 1The PMNS: a flow diagram to describe data collection and exclusions from the study. LMP Last menstrual period
At 6 years of age (December 2000–February 2003) the children were investigated for body size, body composition and risk factors for type 2 diabetes mellitus and cardiovascular disease. Families were instructed to continue their usual diet and activity prior to the study, and were transported to the Research Unit the previous evening. They were provided with a standard dinner and then water only until the next morning. A fasting blood sample was collected, followed by a 1.75 g/kg body weight oral anhydrous glucose load and a further blood sample 120 min later.
The family’s socioeconomic status (SES) was assessed using the Standard of Living Index devised by the National Family Health Survey [13]. Weight and height were measured using standardised protocols. Whole-body dual-energy X-ray absorptiometry (DEXA) scans (Lunar DPX-IQ 240; Lunar Corporation, Madison, WI, USA; paediatric software) were carried out to measure total, truncal and leg fat mass, percentage body fat and lean mass. Standard quality assurance tests were performed every day. The DEXA machine was installed 3 months after the study started, and therefore 59 children had DEXA measurements later than the other investigations.
Plasma glucose was measured using an Hitachi 911 automated analyser (Hitachi, Tokyo, Japan) by the glucose oxidase peroxidase method. Plasma insulin was measured using a Delfia technique (Victor 2; Wallac, Turku, Finland) [14]. Insulin resistance was calculated using the fasting insulin and glucose concentrations (homeostatic model assessment of insulin resistance [HOMA-R]) [15]. As previously described, maternal erythrocyte folate concentrations were measured at the time of the original study (1993–1996), taking all necessary precautions in the collection, transport and storage of samples [8]. Based on the findings of a preliminary study [10], we measured plasma vitamin B12, tHcy and methylmalonic acid (MMA) concentrations in all the stored maternal fasting samples (−80°C) in February 2004, as described [11].
Ethical permission for the study was granted by the KEM Hospital Ethical Committee, and by the local village leaders. Parents gave informed written consent. None of these volunteers were paid for participating in the study.
Statistics Skewed variables were transformed to normality using the following transformations: log to the base e (ln) (plasma vitamin B12 [18 weeks], tHcy and MMA and erythrocyte folate concentrations, maternal fat intake, and the child’s trunk and leg fat mass), square root (maternal energy, protein, and carbohydrate intakes, and the child’s total fat mass) and reciprocal of square root (maternal vitamin B12 concentrations at 28 weeks). Maternal pre-pregnant fat mass was calculated from the sum of four skinfold thickness measurements [16]. Maternal intakes of specific foods, based on the FFQ, were analysed as categorical variables. Low vitamin B12 and erythrocyte folate concentrations were defined as <150 pmol/l and <283 nmol/l, respectively. Elevated tHcy and MMA concentrations were defined as >10 and >0.26 μmol/l [17]. Correlations between different nutritional measures in the mothers were tested using Pearson correlation coefficients. Relationships between maternal nutritional variables and outcomes in the children were analysed using multiple linear regression.
As a final stage of the analysis, we summarised the inter-relationships between maternal factors and outcomes in the children using principal components analysis (PCA) and conditional independence analysis. We first performed a PCA to condense the information contained in large groups of variables into a small number of ‘component’ variables. These components are independent of each other and therefore can be used in regression analysis without the problem of colinearity [18]. The groups of variables were: (1) maternal pre-pregnant size (height, skinfold thicknesses, head circumference, mid-upper-arm circumference [MUAC] and waist and hip circumferences); (2) maternal macronutrient intakes (energy, protein, carbohydrate and fat intakes at 18 and 28 weeks); (3) maternal micronutrient-rich foods in pregnancy (frequency of intake of dairy products, GLV, fruit and non-vegetarian items [meat, fish and eggs] at 18 and 28 weeks of gestation); (4) maternal micronutrient status (vitamin B12, folate, tHcy and MMA concentrations at 18 and 28 weeks); (5) newborn size (birthweight, length, skinfold thicknesses, MUAC, head and abdominal circumferences); and (6) 6 year body composition (height and DEXA measurements of lean mass, and total, truncal and limb fat). Other key variables (maternal SES and physical workload score [mean of 18 and 28 week scores], and HOMA-R in the child at 6 years) remained as single variables. These, and the relevant principal components were included simultaneously in a conditional independence analysis, which is a method of displaying ‘pathways’ of association between a pre-specified set of variables (in this case, the groups of variables described above). The partial correlation (that is, the correlation while holding the remaining variables constant in the set) of each pair of variables was calculated [19]. A ‘path’ diagram was then drawn, connecting pairs of variables that were significantly correlated (p < 0.01). Two-tailed significance was calculated at 5% level. Analyses were performed using SPSS 11.0 for windows (SPSS, Chicago, IL, USA) and STATA 7.0 (STATA, College Station, TX, USA).
Results
Many of the maternal size and nutritional measurements and their relationship to neonatal size have been described before [8]. The age (median [25th, 75th centiles]) of the mothers when they became pregnant was 21 (19, 23) years. They were short (152.0 [148.5, 155.4] cm) and had a low pre-pregnant BMI (17.8 [16.7, 19.1] kg/m2) but a relatively high percentage body fat (20% [18, 24]). Their dietary intakes and micronutrient status during pregnancy are shown in Table 1. The women’s energy and protein intakes were low compared with the recommended dietary allowances of 10,565 kJ (2,525 kcal) and 65 g/day, respectively [20]. At 18 and 28 weeks, their energy intakes were 1.57 and 1.43 times their calculated basal metabolic rate [21]. One-third of women were lacto-vegetarian, and only 15% of women ate non-vegetarian foods more frequently than once every alternate day. The portion sizes of non-vegetarian foods were small (<120 g/day for chicken, fish and meat dishes and ~60 g/day for eggs).
Table 1Maternal nutrition data during pregnancy (median and 25th and 75th centiles, unless otherwise stated) Number18 weeks of gestationNumber28 weeks of gestationMacronutrient intakes Energy (kJ)6927,293 (5,858, 8,774)6706,803 (5,523, 8,268) Energy (kcal)6921,743 (1,400, 2,090)6701,626 (1,320, 1,976) Carbohydrate (g)692313 (257, 376)670296 (236, 360) Protein (g)69244.9 (35.4, 54.8)67041.7 (33.8, 51.3) Fat (g)69232.7 (23.9, 42.9)67029.6 (22.4, 39.2)FFQ, n (%) GLVs Never69215 (2)67169 (10) <1 per week692139 (20)671197 (29) >1 per week692277 (40)671243 (36) >Alternate day692261 (38)671162 (24)Dairy products, n (%) Never692107 (15)67196 (14) <1 per week692141 (20)671137 (20) >1 per week692129 (19)671144 (21) >Alternate day692315 (46)671294 (44)Non-vegetarian foods, n (%) Never692228 (33)671254 (38) <1 per week692182 (26)671181 (27) >1 per week692182 (26)671149 (22) >Alternate day692100 (14)67187 (13)Circulating micronutrients Vitamin B12 (pmol/l)638135 (103, 175)594122 (94, 160) <150 pmol/l, n (%)380 (60)423 (71) Erythrocyte folate (nmol/l)618874 (687, 1,106)562961 (736, 1,269) <283 nmol/l, n (%)1 (0.2)1 (0.2) MMA (μmol/l)6360.80 (0.50, 1.34)5940.73 (0.44, 1.18) >0.26 μmol/l, n (%)586 (94)533 (90) tHcy (μmol/l)6398.1 (6.8, 10.3)5938.6 (6.7, 10.8) >10 μmol/l, n (%)177 (28)193 (33)Dairy products refers to whole milk plus milk products (milk in tea and other beverages, yoghurt, buttermilk, ghee, ice cream and other milk-based preparations). Non-vegetarian foods are meat, fish and eggs
Maternal vitamin B12 and folate status Women in whom vitamin B12, folate and related measurements were not available had similar pre-pregnant weight, SES and weight gain at 28 weeks of pregnancy compared with those who were studied. The majority of women (>60%) had low plasma vitamin B12 concentrations (Table 1), over 90% had elevated MMA concentrations, and one-third were hyperhomocysteinaemic. On the other hand only one woman had a low erythrocyte folate concentration. Between 18 and 28 weeks of gestation, plasma vitamin B12, MMA and tHcy concentrations remained similar (p ~ 0.11, for all), while erythrocyte folate concentrations increased (p < 0.001). Plasma tHcy and MMA concentrations were inversely related to plasma vitamin B12 (p < 0.001 for both at 18 and 28 weeks of gestation). Low vitamin B12 concentration contributed 41 and 24% to the population attributable risk of hyperhomocysteinaemia, and 40 and 12% to high MMA at 18 and 28 weeks of gestation, respectively; the contribution of low folate concentration could not be calculated because of small numbers. Higher frequency of intake of dairy products and non-vegetarian foods was associated with higher plasma vitamin B12 concentrations (p = 0.005 and 0.04, respectively) and lower tHcy (p = 0.1 and p = 0.04) and MMA concentrations (p = 0.01 and p = 0.003). Plasma vitamin B12 concentrations were also related to protein intakes (28 weeks, p = 0.03) but not to energy intake. Higher frequency of GLV intake predicted higher erythrocyte folate concentrations (p = 0.001). All these associations were independent of the energy intake and SES.
Maternal nutrition during pregnancy and newborn size Maternal vitamin B12 and MMA concentrations were unrelated to neonatal measurements. As previously described [8], lower maternal folate concentrations were associated with smaller newborn weight, MUAC and abdominal circumference (p = 0.003, 0.008, 0.008, respectively). Higher tHcy concentrations at 18 weeks were associated with smaller newborn size (MUAC, p = 0.02; abdominal circumference, p = 0.02; and subscapular and triceps skinfold thicknesses, p = 0.01 and p = 0.007).
Maternal nutrition during pregnancy and offspring size, body composition and HOMA-R at 6 years At 6 years, the children were light, short and had a low BMI compared with an international (UK) reference [22] (Table 2); none were overweight or obese as defined by International Obesity Task Force [23] criteria. However, skinfold thickness measurements showed that the children were relatively truncally adipose; the mean SD score for subscapular skinfold thickness was −0.42 compared with the UK growth standards [24], in contrast with −2.23 for weight and −1.86 for BMI. Higher fat mass and higher body fat per cent were associated with higher fasting insulin concentrations, higher HOMA-R and higher 120 min plasma glucose concentrations (p < 0.05 for all).
Table 2Characteristics of the children at 6 years (n = 653) Median (IQR)Mean SD scorea (SD)Age (years)6.1 (6.1, 6.2)Weight (kg)16.0 (14.8, 17.3)−2.23 (1.00)Height (cm)109.7 (106.7, 112.9)−1.39 (0.89)BMI (kg/m2)13.3 (12.8, 14.0)−1.86 (0.90)Subscapular skinfold thickness (mm)5.0 (4.4, 5.6)−0.42 (0.68)Triceps skinfold thickness (mm)6.2 (5.4, 7.1)−1.41 (0.79)Fat mass (kg)3.1 (2.4, 3.7)Fat (%)18.8 (15.5, 22.2)Leg fat mass (kg)1.3 (1.0, 1.5)Trunk fat mass (kg)0.9 (0.7, 1.2)Lean mass (kg)12.7 (11.7, 13.7)Insulin resistance (HOMA-R)0.70 (0.37, 1.06)120 min glucose (mmol/l)5.49 (4.82, 6.21)Vitamin B12 (pmol/l)224 (167, 311)aSD scores at 6 years were derived from the UK growth standards [26, 28]IQR Interquartile range
Higher frequency of maternal intake of GLVs and higher erythrocyte folate concentrations (28 weeks) were associated with higher fat mass and higher per cent body fat in the children (GLVs: p = 0.02, 0.08; and folate: p = 0.01, 0.002; Table 3). None of the maternal nutritional variables were related to lean mass in the children.
Table 3Outcomes in the children at 6 years (body composition and insulin resistance) according to maternal vitamin B12 concentrations at 18 weeks, erythrocyte folate at 28 weeks and GLV and dairy product intake at 28 weeks NumberHeight (cm)Fat mass (kg)Lean mass (kg)Insulin resistance (HOMA-R)Vitamin B12 at 18 weeks (pmol/l) <103146110.2 (4.3)3.3 (0.9)12.8 (1.6)0.78 (0.42, 1.34) 103–134150109.3 (4.6)3.0 (1.0)12.7 (1.5)0.68 (0.35, 0.99) 135–174145109.6 (4.9)3.1 (1.0)12.9 (1.7)0.69 (0.30, 0.95) ≥175151110.0 (4.9)3.2 (1.0)12.8 (1.7)0.61 (0.35, 1.00) p valuea0.80.50.30.03 p valueb0.60.40.30.04Erythrocyte folate at 28 weeks (nmol/l) <734129109.2 (4.6)3.0 (1.0)12.8 (1.7)0.52 (0.28, 0.81) 734–691136109.8 (4.5)3.1 (0.9)12.9 (1.6)0.65 (0.36, 0.92) 962–1,268131109.9 (4.7)3.2 (1.2)12.7 (1.8)0.71 (0.39, 1.12) ≥1,269127110.7 (4.8)3.4 (1.1)13.0 (1.6)0.85 (0.51, 1.27) p valuea0.020.0010.5<0.001 p valueb0.70.010.4<0.001GLVs at 28 weeks Never63109.1 (5.5)3.0 (0.9)12.5 (1.7)0.83 (0.25, 1.23) <Once per week183109.5 (4.7)3.0 (1.0)12.7 (1.7)0.73 (0.36, 1.05) >Once per week231109.8 (4.6)3.2 (1.0)12.8 (1.7)0.69 (0.34, 0.99) ≥Alternate days146110.5 (4.1)3.3 (1.2)13.0 (1.6)0.66 (0.41, 1.00) p valuea0.0020.0010.010.99 p valueb0.40.020.40.98Dairy products at 28 weeks Never87108.4 (4.9)2.9 (0.9)12.6 (1.8)0.60 (0.32, 0.95) <Once per week124110.2 (4.1)3.2 (1.0)12.8 (1.6)0.62 (0.37, 0.99) >Once per week134109.4 (4.6)3.2 (0.9)12.7 (1.7)0.73 (0.40, 1.06) ≥Alternate days278110.3 (4.6)3.2 (1.2)12.9 (1.7)0.74 (0.37, 1.11) p valuea0.0050.030.30.02 p valueb0.20.20.80.006Values are mean (SD) or median (interquartile range)ap value adjusted for the child’s age and sexbp value adjusted for the above, plus: SES, the mother’s pre-pregnant height and fat mass, the child’s birthweight, skinfold thicknesses and gestation at delivery and the mother’s protein intake at the time of measurement
Higher frequency of maternal intake of dairy products at 28 weeks, and lower plasma vitamin B12 (18 weeks) and higher erythrocyte folate concentrations (28 weeks) were associated with higher HOMA-R in the children (Table 3). The highest HOMA-R was in children whose mothers had the lowest vitamin B12 and highest folate concentrations (Fig. 2). There was no statistically significant interaction, however, between vitamin B12 and folate concentrations in relation to HOMA-R (Table 4). The associations remained significant after further adjustment for the child’s own fat mass and plasma vitamin B12 concentrations. None of the other maternal dietary or macronutrient variables were related to HOMA-R in the children.
Fig. 2Insulin resistance (HOMA-R) in the children at 6 years in relation to maternal vitamin B12 (18 weeks) and erythrocyte folate (28 weeks)Table 4Multiple linear regression models for child’s insulin resistance (HOMA-R) at 6 years as the dependent variableIndependent variableModel 1Model 2Model 3Model 4βp valueβp valueβp valueβp valuePlasma vitamin B12 at 18 weeks (pmol/l)−0.160.03−0.160.04Erythrocyte folate at 28 weeks (nmol/l)0.38<0.0010.38<0.001Interaction term (vitamin B12 × folate)−0.080.7The independent variable included maternal plasma vitamin B12 concentration at 18 weeks gestation, erythrocyte folate at 28 weeks gestation, child’s age, sex, fat mass and standard of living index. The models progressively included vitamin B12, erythrocyte folate, both, and an interaction term (models 1 to 4, respectively)
Maternal vitamin B12 concentrations were not related to the child’s insulinogenic index (a measure of beta cell function) but 28 week erythrocyte folate concentration was directly related (p < 0.01).
PCA and conditional independence analysis PCA and conditional independence analyses are summarised in Fig. 3 and the Electronic supplementary material (ESM) Table 1. Mothers of higher SES had higher intakes of the micronutrient-rich foods, except non-vegetarian foods (correlation A). Larger maternal body size (correlation B) and larger size of the offspring at birth (correlation C) were associated with larger 6 year size. None of the maternal nutritional variables were related to 6 year body composition in the children, but the contrast between folate and vitamin B12 concentrations (micronutrient status, factor 3: higher maternal folate and MMA and lower vitamin B12) was associated with higher HOMA-R in the child at 6 years (correlation D).
Fig. 3The groups of variables used in the PCAs are described in Statistical methods. The notation F1–F3 indicates the first to third principal components, or factors (F), derived from each group of variables. Lines connecting boxes indicate significant positive correlations (p < 0.01), and bold lines denote correlations significant at p < 0.001. Correlations labelled A–D are explained in the Results (PCA and conditional independence analysis). Veg Vegetables
Discussion
We have demonstrated for the first time in a purposeful, community-based prospective study an association between maternal nutritional measurements in pregnancy and two major risk factors for type 2 diabetes in the offspring. Among rural mothers in Pune, energy and protein intakes are lower than the recommended daily allowance, vitamin B12 status is poor but folate status is adequate. Maternal macronutrient intakes were unrelated to adiposity and insulin resistance in the offspring. However, higher maternal folate concentrations predicted greater adiposity (fat mass and body fat per cent) and higher insulin resistance, and lower vitamin B12 concentrations predicted higher insulin resistance. Children born to mothers with low vitamin B12 concentrations but high folate concentrations were the most insulin resistant. The 18 week vitamin B12 concentration was more strongly associated with insulin resistance than the 28 week value, while the reverse was true for erythrocyte folate. The lifespan of erythrocytes (4 months) means that 28 week erythrocyte folate reflects nutritional status earlier in pregnancy. Thus, early-mid pregnancy may be a critical period for the programming of adiposity and insulin resistance.
The PMNS children are short and thin but relatively adipose compared with white children, similar to the situation in Indian newborns [5, 6] and adults [25]. There were no excessively adipose or insulin-resistant children in this cohort, but adiposity measures were strongly related to metabolic risk factors (glycaemia and insulin resistance) in the normal range. Other studies have shown that higher levels of risk factors in the normal range in childhood predict an increased risk of disease in later life [26, 27].
We have demonstrated that the ‘thin-fat’ phenotype of Indians is associated with a higher risk of type 2 diabetes [7]. This phenotype reflects the simultaneous involvement of two major body compartments (less lean but more adipose), which contribute to the pathogenesis of type 2 diabetes [28]. Our study suggests that an intrauterine imbalance between two related micronutrients (vitamin B12 and folate) may be responsible. A related concept is that of sarcopenic obesity, which increases risk of metabolic and skeletal adverse outcomes [29].
We did not find an association between maternal vitamin B12 concentrations and size at birth, possibly because so many women had values in the deficient range. However, higher maternal tHcy concentrations (which were related to low vitamin B12 concentrations) predicted smaller newborn size. A study in Bangalore, South India, where mothers had higher vitamin B12 concentrations, found that low maternal vitamin B12 concentrations predicted fetal growth retardation [30]. These findings are consistent with the ‘thrifty phenotype hypothesis,’ which proposed that ‘poor’ maternal nutritional status increases the risk of type 2 diabetes in the offspring [1]. On the other hand, the positive associations of maternal folate and intakes of GLVs and dairy products with adiposity or insulin resistance in the children, are contrary to the hypothesis because they predicted larger offspring size at birth. Our results suggest a need for caution in designing nutritional strategies to improve fetal growth and future health based on relationships with birth size alone.
Previous studies linking maternal nutrition in pregnancy to cardiovascular risk factors in the offspring did not measure vitamin B12 or folate [31]. These two vitamins play a crucial role in nucleic acid synthesis and one-carbon metabolism. Low maternal folate status has been implicated in the aetiology of neural tube defects [32] and high tHcy levels have been associated with poor pregnancy outcomes [10, 33, 34]. This has led to a policy of supplementation with folic acid before and during pregnancy, and fortification of flour with folic acid in some countries. The Indian policy is to provide iron and folic acid (60 mg and 500 μg per day) to all pregnant mothers [35]. Despite the evidence of widespread deficiency [11] vitamin B12 supplementation is not a consideration in pregnant Indian women. Vegetarianism and low milk intakes contribute to low vitamin B12 status in Indians [36, 37]. Vegetarianism is multigenerational, and influenced by religious and socioeconomic factors. King Samrat Ashok (273 bc) first banned the killing of animals for food and this was institutionalised by three religions: Jainism, Hinduism and Buddhism. The role of socioeconomic factors is evident in the small portion sizes of non-vegetarian foods habitually eaten by our mothers. Adequate folate status in the PMNS indicates adequate dietary intake even before supplementation was started (18 weeks of gestation), which further increased the levels. In addition, it is our experience that many Indian obstetricians routinely prescribe high doses of folic acid (5 mg or more) in ‘early’ pregnancy with the intention of preventing neural tube defects, even though the majority of pregnant women approach the doctor after 12 weeks of gestation, when the neural tube is already closed. Thus religious and socioeconomic factors, and medical practices, contribute to create an imbalance in these two related vitamins. The association of higher maternal intake of dairy products with insulin resistance in the offspring requires further exploration. It is intriguing that higher milk consumption is associated with higher insulin resistance in European children [38].
We can only speculate about the possible mechanisms for our findings (Fig. 1). Vitamin B12 deficiency will trap folate as 5-methyltetrahydrofolate [39], prevent the generation of methionine from homocysteine and therefore reduce protein synthesis and lean tissue deposition. Elevated methylmalonyl-CoA could contribute to increased lipogenesis by inhibiting carnitine palmitoyltransferase [40] and thereby inhibit β-oxidation [41]. An analogous clinical situation is high-dose folic acid treatment of severely vitamin B12-deficient pernicious anaemia patients: anaemia improves but neurological damage worsens, possibly because of accumulation of odd-chain carbon fatty acids [42]. Epigenetic regulation, involving DNA methylation, may be another mechanism of nutritional programming, as demonstrated in animal models [43–45]. It would be interesting to study these aspects further.
The PMNS has many strengths including a population-based design and high participation and follow-up rates. The nutritional assessment methods were developed specifically for the population. The rural setting is representative of over 70% of the Indian population, in whom the prevalence of diabetes is rapidly rising [46]. The lacto-vegetarian food habits of our women are similar to ~40% of Indian households and ~10% of the world population [47]. Although vitamin B12 was measured in stored frozen samples, it is stable during long-term storage [48]. In the absence of accepted guidelines for interpreting vitamin B12 status in pregnancy [49], we cannot define the true extent of vitamin B12 deficiency, but high levels of MMA and tHcy suggested its presence in a majority of the women in our study [17]. Associations with folate and vitamin B12 concentrations but not with tHcy and MMA could be due to dependence of the latter two on non-nutritional factors that are altered in pregnancy (haemodilution and elevated glomerular filtration rate), although we did find an association between high maternal MMA and insulin resistance in the children in the conditional independence analysis. A limitation of the study is that it is observational, and therefore causality cannot be ascertained.
In conclusion, our data raise the important possibility that high folate intakes in vitamin B12-deficient mothers could increase the risk of type 2 diabetes in the offspring. This is the first report in humans to suggest that defects in one-carbon metabolism might be at the heart of intra-uterine programming of adult disease. There is a need for further studies to test our findings and to determine the correct management of low vitamin B12 concentrations in Indian mothers.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Figure S1 Suggested metabolic mechanisms for adiposity, insulin resistance and altered gene expression in a situation of dietary vitamin B12 deficiency combined with adequate folate status (PDF 35.7 kb)
Table S1
Unrotated factor loadings from the PCA (PDF 75.3 kb)
|
[
"vitamin b12",
"folate",
"pregnancy",
"insulin resistance",
"offspring",
"maternal nutrition",
"adiposity"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Pediatr_Cardiol-4-1-2267487
|
Polymorphisms of Human Leukocyte Antigen Genes in Korean Children with Kawasaki Disease
|
Background Kawasaki disease is a leading cause of acquired heart disease in children. The prevalence rate varies in different ethnic groups. Recently, with the clinical application of molecular genetic technology, human leukocyte antigen (HLA) polymorphisms associated with several diseases have been identified by DNA analysis. This study aimed to assess the association of HLA alleles with susceptibility and complications of Kawasaki disease in Korean children.
Kawasaki disease is a systemic vasculitis that currently is a leading cause of acquired heart disease in children. The etiology of Kawasaki disease continues to be unknown, although it was described 4 decades ago in 1967 by Kawasaki [15]. Diagnosis continues to depend primarily on clinical manifestations.
It appears that the pathogenesis of Kawasaki disease is related to infection of a susceptible host coupled with epidemiologic and ethnic factors. In a study examining the epidemiologic characteristics of Kawasaki disease in San Diego, CA, USA, Asians/Pacific Islanders were 2.7 times more likely and Hispanics one-third as likely to be hospitalized for Kawasaki disease compared with the children of other ethnic groups combined [4]. The sibling risk ratio is also known to be higher, than the risk of being affected with Kawasaki disease without sibling history, and cases of Kawasaki disease in parents and children have been reported [10, 21].
For these reasons, many reports have suggested involvement of the human leukocyte antigen (HLA) system. The results of such studies vary depending on the ethnic group studied. Several reports show the association of HLA genotypes with Kawasaki disease including HLA-Bw22 (now referred to as HLA-B54) in Japanese [14, 20], HLA-B51 in white populations [13, 16, 18, 19], and the major histocompatibility complex class I chain-related gene A (MICA) genes and others in southern Chinese population [6, 12].
The major histocompatibility complex (MHC) class II genes in Kawasaki disease show variable correlations [2, 8, 27]. To date, there have been no reports supporting a consistent relationship between Kawasaki disease and HLA genes in Korean patients. Therefore, the current study aimed to assess the association of HLA-A,-B,-C, and -DRB1 alleles with Kawasaki disease in Korean children.
Materials and Methods
Subjects
From March 2004 to December 2004, 74 patients with Kawasaki disease (44 boys and 30 girls) attending the pediatric outpatient clinic or admitted to St. Vincent’s Hospital in Korea were recruited for this study. The mean age of the patients at diagnosis was 2.7 ± 1.9 years. For comparison, genetically unrelated healthy Korean adults (n = 159) who had no history of Kawasaki disease were studied as a healthy control group. Blood samples were obtained after informed consent in compliance with the institutional review board of St. Vincent’s hospital, College of Medicine, The Catholic University of Korea.
The diagnosis of Kawasaki disease was established by clinical manifestations meeting the defined criteria according to the report of the Research Committee on Kawasaki Disease of the Japanese Ministry of Health and Welfare in addition to exclusion of any other possible illnesses [25]. Coronary complications included the following lesions detected by echocardiography: an internal diameter exceeding 3 mm in patients younger than 5 years and exceeding 4 mm in patients age 5 years or older, dilation of a segmental luminal diameter to at least 1.5 times that of an adjacent segment, or a definitely irregular internal lumen of the coronary arteries. The clinical characteristics of the patients are shown in Table 1.
Table 1Profiles of subjects with Kawasaki diseasen (%)Male/female44/30 (60/40)Typical KD/atypical KD65/9 (88/12)Family history of KD3 (4.1)Recurrence history of KD4 (5.4)Coronary complications21 (28)Other complications of KDGallbladder hydrops14 (19)Pyuria25 (34)Hepatopathy38 (51)Arthritis10 (14)KD, Kawasaki disease
Genomic DNA was extracted by standard methods using the AccuPrep DNA extraction kit (Bioneer, Daejeon, Korea) from peripheral blood collected (4 ml) with ethylenediaminetetraacetic acid (EDTA) and kept at −20°C.
HLA-A, -B, and -C Genotyping
The genotyping of HLA-A, -B, and -C was performed by the amplification refractory mutation system (ARMS)–polymerase chain reaction (PCR) method. Each reaction contained a primer mix consisting of the allele- or group-specific primer pairs as well as internal control primers matching nonallelic sequences. Specific amplifications of the HLA-A, -B, and -C genes were performed using forward and reverse primers (44 for HLA-A, 47 for HLA-B, and 33 for HLA-C) designed according to the published nucleotide sequences [5, 17, 26].
The PCR procedure was carried out in a reaction (13 μl), containing 100 to 200 ng genomic DNA, 0.8 X buffer (40 mmol/l KCl, 1.2 mmol/l MgCl2, 8 mmol/l Tris-HCl pH 8.8, 0.08% Triton X-100), 5% dimethylsulphoxide (DMSO), 200 μmol/l of each dNTP, 0.25 U Taq DNA polymerase (Boehringer, Mannheim, Germany), 1 μmol/l of each sequence-specific primer, and 0.2 μmol/l of internal control primers.
The amplifications were performed in a My Cycler thermocycler (Bio-Rad, Hercules, USA). For amplification, 30 cycles were performed using the following steps: heating to 96°C for 1 min to denature the DNA; denaturation at 96°C for 25 s and at 70°C for 45 s; annealing and extension at 72°C for 30 s (for the first 5 cycles), 96°C for 25 s, 65°C 45 s, 72°C for 30 s (for the next 21 cycles); 96°C for 25 s, 55°C for 60 s, 72°C for 120 s (for the last 4 cycles); and a final 1 min extension at 72°C. The presence or absence of PCR products was determined after separation of the samples on a 1.5% agarose gel containing 0.5 μg/ml of ethidium bromide.
HLA-DRB1 Genotyping
The PCR sequence-specific primer (SSP) method was used for HLA-DRB1 genotyping. Each reaction contained a primer mix consisting of the allele- or group-specific primer pairs as well as the internal control primers that matched the nonallelic sequences. Specific amplification of the HLA-DRB1 gene was performed using 47 forward and reverse primers for HLA-DRB1 designed on the basis of the published nucleotide sequences [23].
The PCR procedure was carried out in a reaction (10 μl) containing 100 ng genomic DNA, 1 X buffer (60 mmol/l Tris-HCl pH 9.0, 15 mmol/l ammonium sulphate, 2.5 mmol/l MgCl2), 200 μmol/l of each dNTP, 0.25 U Taq DNA polymerase (Boehringer Mannheim, Germany), 1 μmol/l of each sequence-specific primer, and 0.2 μmol/l of the internal control primers.
The amplifications performed in a My Cycler thermocycler (Bio-Rad, Hercules, USA). For amplification, 27 cycles were performed using the following steps: heating to 96°C for 3 min to denature the DNA; denaturation at 96°C for 20 s; annealing and extension at 66°C for 60 s (for the first 10 cycles), 62°C for 80 s (for the next 10 cycles), and 61°C for 120 s (for the last 7 cycles); and a final 10 min extension at 72°C. The presence or absence of PCR products was determined after separation of samples on a 1.5% agarose gel containing 0.5 μg/ml of ethidium bromide.
Statistical Analysis
Statistical differences between Kawasaki disease patients and healthy control subjects and between subgroups of Kawasaki disease patients were tested with chi-square and Fisher’s exact test. The quantification of the relationship of the frequencies studied, in cases with Kawasaki disease, was performed by calculating the relative risks using the method of Woolf. Haldane’s modification was applied for cases in which the variables included zero. A p value less than 0.05 was considered to be significant.
In this study, the data were not adjusted for multiple comparisons because the sample size was not large enough for multiple-comparison analysis correction.
Results
The Genotype Distribution and Allele Frequencies of HLA-A for the Patients With Kawasaki Disease and the Healthy Control Group
In the analysis of the polymorphisms of HLA-A alleles, no statistical difference was found in the frequency of alleles between the patients with Kawasaki disease and the control group. However, the frequency of HLA-A26 alleles was significantly decreased in the Kawasaki disease patients without coronary complications (CC) compared with the healthy control group (p < 0.04; relative risk (RR) = 0.2). In a comparison between subgroups, with and without CC, no significant difference was identified (Table 2).
Table 2The allele frequencies of human leukocyte antigen-A (HLA-A) in the patients with Kawasaki disease and the control subjectsHLA allelesHealthy controls (n =159) n (%)KD patients (n = 74) n (%)Coronary complicationsPresent (n = 21) n (%)Absent (n = 53 n (%)HLA-A015 (3.1)5 (6.8)2 (9.5)3 (5.7)0285 (53.5)43 (58.1)12 (57.1)31 (58.5)035 (3.1)3 (4.1)0 (0.0)3 (5.7)1131 (19.5)21 (28.4)6 (28.6)15 (28.3)2457 (35.8)33 (44.6)12 (57.1)21 (39.6)2625 (15.7)6 (8.1)4 (19.0)2 (3.8)a3020 (12.6)4 (5.4)1 (4.8)3 (5.7)3114 (8.8)6 (8.1)0 (0.0)6 (11.3)323 (1.9)0 (0.0)0 (0.0)0 (0.0)3338 (23.9)15 (20.3)2 (9.5)13 (24.5)681 (0.6)0 (0.0)0 (0.0)0 (0.0)KD, Kawasaki disease; RR, relative riskap < 0.04; RR = 0.2 (95% confidence interval [CI], 0.05–0.90) vs healthy control subjects
The Genotype Distribution and Allele Frequencies of HLA-B for the Patients With Kawasaki Disease and the Healthy Control Group
In analysis of the polymorphisms of HLA-B alleles, there was a significant increase in the frequency of HLA-B35 and -B75 alleles in the patients with Kawasaki disease compared with the control group (p < 0.006; RR = 3.1 vs p < 0.02; RR = 8.2). When the patients with Kawasaki disease were divided into two subgroups, with or without CC, the Kawasaki disease patients without CC showed a significantly increased frequency of HLA-B35 and -B75 alleles compared with the healthy control group (p < 0.02; RR = 3.1 vs p < 0.003; RR = 11) (Table 3). Although the frequency of the HLA-B35 alleles were similar in the groups with and without CC (19% vs 18.9%), the HLA-B75 allele was found only in the Kawasaki disease patients without CC (0.0% vs. 13.2%). Therefore, the HLA-B35 allele was implicated in susceptibility to Kawasaki disease and the HLA-B75 allele in susceptibility to Kawasaki disease without CC. In a comparison between the subgroups, with and without CC, there was no overall significant difference identified between the two groups.
Table 3The allele frequencies of human leukocyte antigen-B (HLA-B) in the patients with Kawasaki disease and the control subjectsHLA allelesHealthy controls (n = 159) n (%)KD patients (n = 74) n (%)Coronary complicationsPresent (n = 21) n (%)Absent (n = 53) n (%)HLA-B0719 (11.9)9 (12.2)3 (14.3)6 (11.3)082 (1.3)0 (0.0)0 (0.0)0 (0.0)1320 (12.6)5 (6.8)2 (9.5)3 (5.7)146 (3.8)0 (0.0)0 (0.0)0 (0.0)2711 (6.9)3 (4.1)1 (4.8)2 (3.8)3511 (6.9)14 (18.9)a4 (19.0)10 (18.9)b375 (3.1)4 (5.4)2 (9.5)2 (3.8)387 (4.4)3 (4.1)0 (0.0)3 (5.7)392 (1.3)3 (4.1)2 (9.5)1 (1.9)4429 (18.2)10 (13.5)3 (14.3)7 (13.2)4618 (11.3)7 (9.5)2 (9.5)5 (9.4)470 (0.0)1 (1.4)0 (0.0)1 (1.9)4813 (8.2)2 (2.7)0 (0.0)2 (3.8)5129 (18.2)13 (17.6)5 (23.8)8 (15.1)529 (5.7)3 (4.1)2 (9.5)1 (1.9)5423 (14.5)10 (13.5)3 (14.3)7 (13.2)553 (1.9)4 (5.4)1 (4.8)3 (5.7)562 (1.3)0 (0.0)0 (0.0)0 (0.0)572 (1.3)1 (1.4)0 (0.0)1 (1.9)5815 (9.4)5 (6.8)0 (0.0)5 (9.4)594 (2.5)3 (4.1)0 (0.0)3 (5.7)6011 (6.9)7 (9.5)2 (9.5)5 (9.4)6123 (14.5)13 (17.6)6 (28.6)7 (13.2)6229 (18.2)14 (18.9)1 (4.8)13 (24.5)674 (2.5)2 (2.7)1 (4.8)1 (1.9)717 (4.4)1 (1.4)0 (0.0)1 (1.9)752 (1.3)7 (9.5)c0 (0.0)7 (13.2)dKD, Kawasaki disease; RR, relative riskap < 0.006; RR = 3.1 (95% confidence interval [CI], 1.29–7.62) vs healthy control subjectsbp < 0.02; RR = 3.1 (95% CI, 1.16–8.43) vs healthy control subjectscp < 0.02; RR = 8.2 (95% CI, 1.74–38.68) vs healthy control subjectsdp < 0.003; RR = 11 (95% CI, 2.79–51.13) vs healthy control subjects
The Genotype Distribution and Allele Frequency of HLA-C for the Patients With Kawasaki Disease and the Healthy Control Group
In the analysis of the polymorphisms of HLA-C alleles, there was a significant increase in the frequency of the HLA-Cw09 allele in the patients with Kawasaki disease compared with the healthy control group (p < 0.04; RR = 2.0). When the Kawasaki disease patients were divided into two subgroups, with or without CC, the Kawasaki disease patients without CC showed a significantly increased frequency of the HLA-Cw09 allele compared with the healthy control group (p < 0.05; RR = 2.1). However, no significant difference was found between the subgroups with and without CC (Table 4).
Table 4The allele frequencies of human leukocyte antigen-C (HLA-C) in the patients with Kawasaki disease and the control subjectsHLA allelesHealthy controls (n = 159) n (%)KD patients (n = 74) n (%)Coronary complicationsPresent (n = 21) n (%)Absent (n = 53) n (%)HLA-Cw0159 (37.1)22 (29.7)6 (28.6)16 (30.2)020 (0.0)1 (1.4)0 (0.0)1 (1.9)0421 (13.2)11 (14.9)2 (9.5)9 (17.0)056 (3.8)1 (1.4)0 (0.0)1 (1.9)0625 (15.7)10 (13.5)5 (23.8)5 (9.4)0744 (27.7)20 (27.0)6 (28.6)14 (26.4)0829 (18.2)10 (13.5)3 (14.3)7 (13.2)0923 (14.5)19 (25.7)a5 (23.8)14 (26.4)b1039 (24.5)24 (32.4)5 (23.8)19 (35.8)1211 (6.9)4 (5.4)2 (9.5)2 (3.8)1434 (21.4)18 (24.3)7 (33.3)11 (20.8)158 (5.0)1 (1.4)1 (4.8)0 (0.0)KD, Kawasaki disease; RR, relative riskap < 0.04; RR = 2.0 (95% confidence interval [CI], 0.97–4.29) vs healthy control subjectsbp < 0.05; RR = 2.1 (95% CI, 0.92–4.87) vs healthy control subjects
The Genotype Distribution and the Allele Frequencies of HLA-DRB1 for the Patients With Kawasaki Disease and the Healthy Control Group
Among the HLA-DRB1 alleles, there was no increase in the frequency of alleles in the Kawasaki disease patients compared with the healthy control group. However, after subgrouping of the Kawasaki disease patients, with or without CC, the frequency of HLA-DRB1*11 was significantly increased in the Kawasaki disease patients with CC compared with the healthy control group (p < 0.04; RR = 4.7). In a comparison between the subgroups, with and without CC, the frequency of the HLA-DRB1*09 allele was increased in the Kawasaki disease patients with CC compared with the Kawasaki disease patients without CC (33.3% vs 9.4%; p < 0.04). However, there was no significant difference in the frequency of these alleles when each was compared with the healthy control group. The frequency of the HLA-DRB1*04 allele was slightly increased in the Kawasaki disease patients without CC compared with the healthy control group (p < 0.05; RR = 1.9) (Table 5).
Table 5The allele frequencies of human leukocyte antigen-DRB1 (HLA-DRB1) in the patients with Kawasaki disease and the control subjectsHLA allelesHealthy controls (n = 159) n (%)KD patients (n = 74) n (%)Coronary complicationsPresent (n = 21) n (%)Absent (n = 53) n (%)HLA-DRB10129 (18.2)8 (10.8)1 (4.8)7 (13.2)037 (4.4)1 (1.4)0 (0.0)1 (1.9)0448 (30.2)30 (40.5)6 (28.6)24 (45.3)a0726 (16.4)8 (10.8)1 (4.8)7 (13.2)0832 (20.1)14 (18.9)3 (14.3)11 (20.8)0928 (17.6)12 (16.2)7 (33.3)b5 (9.4)104 (2.5)4 (5.4)2 (9.5)2 (3.8)1110 (6.3)8 (10.8)5 (23.8)c3 (5.7)1218 (11.3)11 (14.9)2 (9.5)9 (17.0)1332 (20.1)10 (13.5)3 (14.3)7 (13.2)1421 (13.2)10 (13.5)3 (14.3)7 (13.2)1535 (22.0)20 (27.0)5 (23.8)15(28.3)164 (2.5)2 (2.7)1 (4.8)1 (1.9)KD, Kawasaki disease; RR, relative riskap < 0.05; RR = 1.9 (95% confidence interval [CI], 0.96–3.82) vs healthy control subjectsbp < 0.04 between KD patients with and without coronary complicationcp < 0.04; RR = 4.7 (95% CI, 1.26–17.17) vs healthy control subjects
Discussion
The prevalence of CC among the patients with Kawasaki disease in this study was higher than previously reported (28.4% vs 5%) [22]. This is likely because the patients included had a previous diagnosis of Kawasaki disease and were attending our clinic for this reason. Patients with more severe Kawasaki disease, including those with CC, present for medical care more frequently than patients without CC, and this may explain the higher proportion of CC in our study. Three patients (4%) had a positive family history, and four patients (5.4%) had a history of Kawasaki disease recurrence. Prior studies reported the familial occurrence and recurrence rates for Kawasaki disease to be 1% and 3%, respectively [11, 24, 28].
Our data were confined to Korean children and focused on patients who had Kawasaki disease with CC. Therefore, further research on the effect of the family history and recurrence of Kawasaki disease in association with HLA polymorphisms is needed.
Studies have associated HLA alleles with many different human diseases, such as HLA-B27 with ankylosing spondylitis and Reiter syndrome, HLA-B35 with subacute thyroiditis, and HLA-DR8 with juvenile rheumatoid arthritis. Most of the early studies on the genetics of Kawasaki disease were conducted using serologic testing that measured HLA antigens. The association of the HLA-Bw22 allele with Kawasaki disease in Japanese studies was evaluated with the microcytotoxicity test used for tissue typing of 205 patients with Kawasaki disease and 500 control samples in the late 1970s [14]. On basis of these studies, a gene was thought to control susceptibility to Kawasaki disease and was linked to the Japanese-specific HLA antigen.
Since the clinical application of genetic technology became available in the 1980s, polymorphisms at many HLA loci have been identified. In addition, there were changes in nomenclature, and the alleles were reclassified. For example, the previously labeled HLA-Bw22 allele was changed to the HLA B54 allele [1, 3].
Our study on the polymorphisms of HLA genes was conducted using the PCR-ARMS method for the HLA class I genes and the PCR-SSP method for the HLA class II genes. There was a significant increase in the frequencies of HLA-B35, -B75, and -Cw09 alleles in the patients with Kawasaki disease compared with the healthy control group. As previously mentioned, this study showed that the HLA-B35 allele was related to susceptibility to Kawasaki disease and the HLA-B75 allele to Kawasaki disease without CC in Korean children. These results suggest that polymorphisms (in some of the B and C loci) in HLA class I genes are associated with Kawasaki disease in Korean children; consistent with prior reports on other ethnic groups [12, 14, 16, 19].
Review of the medical literature on Kawasaki disease and HLA shows that there is a trend toward an association between the HLA-B loci and Kawasaki disease. However, to date, there is no confirmed relationship to a particular locus. It is likely that the HLA-B locus is not the only locus associated with Kawasaki disease. This can be the case because of its functional variation or the potential effects from other related genes around the HLA-B locus. The HLA-B locus and other associated genes might be useful genetic markers for Kawasaki disease.
The HLA-B35 gene is known to play a major role in a variety of infections. For example, in cases with HIV infection, the patients with an increased frequency of HLA-B35 show rapid progression of HIV [9]. Infection has been implicated in the etiology of Kawasaki disease. The finding in our study that the frequency of the HLA-B35 allele was increased in Korean children with Kawasaki disease suggests a role for infection in Kawasaki disease.
The HLA-DRB1 alleles showed no increased frequency in the Kawasaki disease patients compared with the healthy control group. However, when the patients were compared by subgroups, with and without CC, the frequency of HLA-DRB1*11 was significantly increased in the Kawasaki disease patients with CC (p < 0.04; RR = 4.7), and HLA-DRB1*04 was increased in the Kawasaki disease patients without CC (p < 0.05; RR =1.9) compared with the healthy control group. Comparison of the subgroups with and without CC showed that the frequency of the HLA-DRB1*09 allele was increased in Kawasaki disease patients with CC compared with the Kawasaki disease patients who had no CC (33.3% vs 9.4%; p < 0.04), and that there was no significant difference in comparison with the healthy control group. For the HLA-DRB1*11 allele, the finding of a higher allele frequency compared with the healthy control group was amplified by the subgrouping of Kawasaki disease patients with and without CC. Future studies with a larger sample size are needed to confirm these findings.
Several reports describe the presence of HLA-DRB1 and the progression of infections [7]. However, most of the reports on the association of HLA-DRB1 with Kawasaki disease concluded that despite a regional association, their data failed to support a consistent role for HLA class II alleles in Kawasaki disease. If there was any role, it was predicted to be minor [2, 8].
Interpopulation discrepancies of the frequencies of HLA alleles make generalization of results difficult. Matched population profiles are needed for disease association studies because ethnic differences can provide altered disease associations.
In this study, we compared HLA alleles of Kawasaki disease patients with known polymorphic loci of HLA genes in healthy Korean adults. The results we report on the polymorphisms of HLA genes in Kawasaki disease patients may be limited to the Korean population. However, this information may be helpful in future studies on HLA genetic polymorphisms in Kawasaki disease. Further studies with a larger sample size are needed for confirmation of our findings.
|
[
"polymorphism",
"gene",
"kawasaki disease",
"hla"
] |
[
"P",
"P",
"P",
"P"
] |
Immunogenetics-3-1-1914291
|
MHC class I A region diversity and polymorphism in macaque species
|
The HLA-A locus represents a single copy gene that displays abundant allelic polymorphism in the human population, whereas, in contrast, a nonhuman primate species such as the rhesus macaque (Macaca mulatta) possesses multiple HLA-A-like (Mamu-A) genes, which parade varying degrees of polymorphism. The number and combination of transcribed Mamu-A genes present per chromosome display diversity in a population of Indian animals. At present, it is not clearly understood whether these different A region configurations are evolutionarily stable entities. To shed light on this issue, rhesus macaques from a Chinese population and a panel of cynomolgus monkeys (Macaca fascicularis) were screened for various A region-linked variations. Comparisons demonstrated that most A region configurations are old entities predating macaque speciation, whereas most allelic variation (>95%) is of more recent origin. The latter situation contrasts the observations of the major histocompatibility complex class II genes in rhesus and cynomolgus macaques, which share a high number of identical alleles (>30%) as defined by exon 2 sequencing.
Introduction
Major histocompatibility complex (MHC) represents a multigene family that plays a crucial role in the generation of adaptive immune responses in vertebrate species. A key feature of the system is that some of its genes display abundant polymorphism at the population level. In addition, the number of Mhc class I or II genes may differ significantly between species, as well as between individuals of a species (Kelley et al. 2005). MHC polymorphisms may have a profound impact on several features such as disease susceptibility, organ transplantation, and reproductive success (Lagaaij et al. 1989; Goulder and Watkins 2004; Bontrop and Watkins 2005; Ziegler et al. 2005; Smith et al. 2006). The MHC systems of various primate species, including humans, have been studied extensively (Watkins 1995; Antunes et al. 1998; Bontrop et al. 1999; Adams and Parham 2001; de Groot et al. 2002; Lafont et al. 2004; Middleton et al. 2004; Marsh et al. 2005; Penedo et al. 2005; Abbott et al. 2006; Huchard et al. 2006). For example, the MHC of the rhesus macaque (MhcMamu), an Old World primate species, has been shown to share many similarities with the human leukocyte antigen (HLA) system in humans. The evolutionary orthologs of the HLA-DP, -DQ, and -DR genes are also present in rhesus monkeys, and as in humans, these loci are polymorphic (Bontrop et al. 1999). The number of Mamu-DRB region configurations appears to be expanded in comparison with humans, and some of these regions seem to harbor an extended number of genes (Doxiadis et al. 2000). Subsequent cDNA studies have illustrated, however, that in humans and rhesus macaques, comparable numbers of DRB genes are transcribed (de Groot et al. 2004).
Apart from Mamu-E (Knapp et al. 1998), -F (Otting and Bontrop 1993), -G (Boyson et al. 1996a; Castro et al. 1996), and -AG (Boyson et al. 1997), rhesus macaques also possess a B-like sequence, designated Mamu-I, which appears to be present in each haplotype (Urvater et al. 2000). As observed in humans, these nonclassical genes display low levels of polymorphism. An ortholog of the HLA-C gene appears to be absent in the rhesus macaque, whereas evolutionary equivalents of the classical HLA-A and -B genes have been described (Boyson et al. 1996b). The Mamu-A and -B genes have been subjected to several rounds of duplication as was shown by genomic sequencing (Daza-Vamenta et al. 2004; Kulski et al. 2004). Analysis of a panel of rhesus macaques, mainly originating from the Indian subcontinent, illustrated that the number and combination of Mamu-A and -B genes that are expressed per haplotype may differ at the population level (Otting et al. 2005). In the same study, marked differences in expression levels were shown. At this stage it is not known whether rhesus macaques originating from other geographic areas have unique Mhc class I alleles and/or region configurations. Comparative studies have illustrated that many Mhc loci and lineages predate speciation events. The sharing of Mhc alleles between two primate species seems to be a rare event, and only a few cases have been documented (Cooper et al. 1998; Evans et al. 1998). An exception is provided by rhesus and cynomolgus macaques, which seem to share a high number of Mhc class II alleles as was defined by exon 2 sequencing (Blancher et al. 2006; Doxiadis et al. 2006). Whether this sharing is because of introgression or purifying selection remains to be elucidated. Thus far, about 50 Mafa-A sequences have been published for the cynomolgus macaque (Uda et al. 2004; Krebs et al. 2005). The absence of pedigreed material did not allow us to define Mafa-A loci or region configurations. For that reason, a panel of pedigreed cynomolgus macaques was incorporated in the present study. Comparison with rhesus macaque Mhc class I sequences obtained from different populations enabled us to draw conclusions on the evolutionary stability of A region variations.
Materials and methods
Animals and cell lines
The Biomedical Primate Research Centre houses a self-sustaining colony of approximately 1,000 rhesus macaques, mainly of Indian origin, that have been pedigreed based on the segregation of serologically defined MHC allotypes (Bontrop et al. 1999; Penedo et al. 2005; Doxiadis et al. 2006). Furthermore, a large collection of DNA samples as well as B-cell lines is available. In this study, B-cells derived from rhesus macaques of Chinese origin and from cynomolgus monkeys were used to isolate RNA. Most cynomolgus monkey samples originate from a pedigreed group housed at the campus of the University of Utrecht.
cDNA cloning and sequencing
RNA was isolated from B cells (Rneasy kit, Qiagen) and subjected to One-step reverse transcriptase polymerase chain reaction (RT-PCR), as recommended by the supplier (Qiagen). The primers (5′MAS) AATTCATGGCGCCCCGAACCCTCCTCCTGG, and (3′MAS) CTAGACCACACAAGGCGGCTGTCTCAC were used, which are specific for class I A transcripts in macaques. The final elongation step was extended to 30 min to generate a 3′dA overhang. The RT-PCR products were cloned using the InsT/Aclone kit (Fermentas). After transformation, 16 to 32 colonies were picked for plasmid isolations. The separate Mamu-A loci show differences in expression levels, as indicated by the numbers of clones, picked from each animal. The alleles with high expression levels (majors) may be associated with serotypes and thus involved in classical antigen presentation, whereas those with low expression levels (minors) are considered nonclassicals and may exhibit more specialized types of functions (Otting et al. 2005).
Sequencing reactions were performed using the BigDye terminator cycle sequencing kit, and samples were run on an automated capillary sequencing system (ABI Genetic Analyzer 3100) as has been described previously (Otting et al. 2005).
Locus-specific PCR reactions
To test the presence of Mamu- and Mafa-A4*14 alleles, 1 μl of the RT-PCR samples was taken for a single-specific-primer PCR (SSP-PCR) reaction in advance of the cloning step. The relevant primers, (5′A*14) GGGACCCGACGGGCGCCTCCAA and (3′A*14) GGCCCTCCAGGTAGACTCTGTC have annealing sites in exon 3. Amplifications were carried out starting with 2 min at 94°C, followed by 25 cycles at 94°C, 65°C, and 72°C for 1 min each. The PCR products were subjected to direct sequencing, and the reactions were performed as described above.
Phylogenetic analysis and nomenclature
The sequences were analyzed with the Sequence Navigator Software version 1.0.1 (Applied Biosystems), and alleles are based on at least three clones with identical full-length sequences. To define loci and lineages, alignments of the sequences were made using the MacVector™ version 8.1.1 (Oxford Molecular Group), followed by manual adjustments. Phylogenetic analyses on the full-length (1,065 bp) sequences were also performed with the MacVector software. Neighbor-joining trees were constructed with the Kimura 2 parameter method. Bootstrap analyses were performed based on 1,000 replications.
In total, 130 unreported Mamu-A and Mafa-A sequences were submitted to the European Bioinformatics Institute and European Molecular Biology Laboratory (EBI-EMBL) database. Relevant information such as accession number and a reference cell-line are provided (Table 1). Moreover, all novel Mamu-A and Mafa-A sequences are named in accordance with a generally accepted nomenclature proposal (Klein et al. 1990; Robinson et al. 2003; Ellis et al. 2006). For example, Mamu-A1*0101 defines a Mhc allele in the rhesus macaque, which is encoded by one of the class I loci: namely, A1. The first two digits after the asterisk define the lineage, whereas the third and fourth digits define the allele number. These allele numbers are arbitrary, as they reflect the order in which the alleles were discovered. A fifth and sixth digit are used to mark a synonymous basepair difference between two sequences.
Table 1Summary of Mamu-A and Mafa-A alleles detected in this studyAlleleAccessionAnimalsAlleleAccessionAnimalsMamu-A1Mafa-A1A1*0102AM295885EKKA1*0101AM295828YabaaA1*0103AM295886Ri056A1*0301AM295829HoebaA1*030102AM295887Ri185A1*0302AM29583077A1*0302AM295888r300A1*1002AM295831k71A1*0303AM2958898822A1*1003AM295832TrufoA1*0304AM295890Ri301A1*1702AM295833BufoA1*0305AM295891Ri037A1*1901AM29583477A1*040102AM295892Ri165A1*2202AM295835k94A1*0402AM295893Ri366A1*3101AM295836GayoA1*1001AM295894Ri002A1*3102AM295823BilboaA1*1002AM295895Ri142A1*3201AM29582480A1*1102AM295896Ri136A1*3801AM295825k66A1*1701AM295897Ri145A1*4002AM295826PedroA1*1801AM295898Ri011A1*4102AM295827k65A1*1802AM295899Ri018A1*4302AM295837k65A1*1803AM295900Ri260A1*5201AM29583888A1*1902AM295901Ri081A1*5301AM295839k390A1*1903AM295902BB10A1*5501AM295840k127A1*1904AM295903Ri197A1*5601AM295841k766A1*2201AM295904Ri009A1*5701AM295842k2,A1*2502AM295905Ri088A1*5801AM295843RastafaA1*2602AM2959068765A1*5901AM295844RozaA1*2802AM295907Ri082A1*6001AM295845VivaaA1*3201AM295908Ri094A1*6101AM295846k135A1*3301AM295909Ri253A1*6201AM29584781A1*4001AM295910Ri165A1*6301AM295848HippoA1*4101AM295911Ri284A1*6401AM295849BufoA1*4701AM295912Ri189A1*6402AM295850k95A1*4801AM2959138822A1*650101AM29585181A1*4901AM295914Ri088A1*650102AM295852k390A1*5001AM2959154053A1*6601AM295853k73A1*5101AM295916Ri281A1*6602AM295854MilvaA1*5201AM295917Ri136A1*6701AM29585583A1*5301AM295918Ri056A1*6801AM295856JuanitaA1*5401AM295919Ri260A1*6901AM295857ZazaaA1*5402AM295920Ri011A1*7001AM295858FrikoA1*5601AM295921Ri146A1*7101AM295859ZolaA1*5602AM295922Ri228A1*7201AM295860CanadaA1*5701AM295923Ri006A1*5801AM295924Ri026Mamu-A2Mafa-A2A2*050301AM295925Ri056A2*0501AM295861ClintA2*050302AM295930Ri226A2*0502AM29586284A2*050402AM295926Ri009A2*0503AM295863k95A2*0513AM295927Ri159A2*0504AM295864ClintA2*0514AM295928Ri095A2*0505AM295865BufoA2*0515AM295929Ri009A2*0506AM295866TrufoA2*0517AM295931Ri284A2*0507AM29586780,A2*0518AM295932Ri094A2*0508AM295868ZebraA2*0519AM295933Ri137A2*0509AM29586981A2*0520AM295934Ri197A2*0510AM295870MilvaA2*0521AM295935Ri126A2*0511AM295871HippoA2*0522AM295936Ri026A2*051201AM295872k73A2*0523AM295937Ri028A2*051202AM295876DojoA2*0524AM295938Ri301A2*0513AM29587381A2*0525AM295939Ri137A2*0514AM295874k95A2*0526AM295940Ri002A2*0515AM295875JuanitaA2*2402AM295941BB67A2*0516AM295878JawaA2*2403AM295942Ri366A2*0517AM295877LabaMamu-A3Mafa-A3A3*1307AM295943Ri136A3*1301AM295879k2,Mamu-A4Mafa-A4A4*1404AM295944Ri137A4*1401AM29588077A4*1402AM295881BufoMamu-A5Mafa-A5A5*300101AM295945BB67A5*3001AM295882GayoA5*300102AM295946Ri081A5*3002AM295883k766A5*3002AM295948Ri088A5*3003AM295947Ri366Mamu-A6Mafa-A6A6*0101AM295949Ri145A6*0101AM295884k73Mamu-A7A7*0101AM295950Ri082A7*0102AM295951Ri315A7*0103AM295952Ri009The Chinese rhesus macaques are indicated by Ri numbers. Nine alleles found in rhesus macaques of mixed breeding are also reported.
Results and discussion
Polymorphism and diversity of the Mamu-A region: comparison of Chinese and Indian rhesus macaques
In a previous communication, five different Mamu-A region configurations were defined in a population of Indian rhesus macaques (Otting et al. 2005). These configurations display diversity with regard to the number and combination of distinct Mamu-A genes present per chromosome. The loci have been designated Mamu-A1, -A2, -A3, and -A4, respectively (Fig. 1). As can be seen, each region configuration comprises a Mamu-A1 gene characterized by high transcription levels (major) combined with one or two other Mamu-A genes characterized by lower transcription levels (minors). The Mamu-A1 gene is probably responsible for executing the classical antigen presentation function (Evans et al. 1999; Sidney et al. 2000; Sette et al. 2005), whereas the others can be considered as nonclassicals and may interact, for instance, with the KIR gene family present on NK cells. Two complete MHC haplotypes in the rhesus macaque are sequenced (Daza-Vamenta et al. 2004; Kulski et al. 2004), and they represent two-region configurations. One haplotype contains the -A1/-A2 combination, whereas in the other, the -A1/-A4 pair is observed. Both region configurations are present in the animals of Indian origin (Fig. 1).
Fig. 1Schematic representation of different Mamu-A region configurations observed in Indian rhesus macaques. The exact order and physical distances of the loci on the genome are still unknown. The relative levels of polymorphism and transcription of the loci are indicated. Two Mamu-A region configurations (1 and 5) are confirmed by the sequencing of the complete rhesus MHC region (Daza-Vamenta et al. 2004; Kulski et al. 2004)
Within the present panel of 42 Chinese animals, 59 different full-length Mamu-A cDNAs were detected that can be grouped into various loci and lineages (Fig. 2). A complete listing of the alleles detected in the Chinese macaques is provided as electronic supplementary material (Table 5). Because of codominant expression, rhesus macaques can be heterozygous for the Mamu-A1 gene and thus can express up to two allotypes. Nine animals were found to express three different Mamu-A1-like sequences. These animals shared one of these sequences in separated clades in phylogenetic tree (Fig. 2; -A5, -A6, and -A7). The triplets of the three animals are indicated by asterisks. The clustering of the three alleles, obtained from one animal, in clades distinct from -A2, -A3, and -A4 indicates that one allele belongs to a separate locus. The newly detected loci are named Mamu-A5, -A6, and-A7, according to the generally accepted nomenclature system (Klein et al. 1990; Robinson et al. 2003; Ellis et al. 2006). Representatives of two loci, A5 and A6, are also observed in cynomolgus macaques that express more than two A1-like sequences.
Fig. 2Phylogenetic tree of Mamu-A and Mafa-A gene/alleles detected in this study. The -A2 locus is represented by four alleles only. The tree is based on full-length sequences, although analyses on only exon 2 and 3 showed no significant differences. The nomenclature of the new alleles is based on a more extended tree (not shown) containing all the Mamu-A and Mafa-A alleles now available. Asterisk, three alleles of animal Ri081, of which one represents the locus Mamu-A5. Double asterisks, three alleles of animal Ri145, of which one represents the locus Mamu-A6. Triple asterisks, three alleles of animal Ri078, of which one represents the locus Mamu-A7
Thirty-three new Mamu-A1 alleles were detected in the Chinese rhesus macaques. Although some of the alleles in this population group into lineages that are also present in Indian animals, sharing of alleles for the Mamu-A1 locus is rare and was observed only once. The Mamu-A1*26 allele is detected in both populations, whereas the Mamu-A1*03 and for -A1*04, alleles are highly similar and differ for one synonymous basepair substitution. Hence, it is concluded that most allelic polymorphism observed for the Mamu-A1 locus was probably generated after the rhesus macaque populations were separated. Sequence comparisons illustrated that most of the variations map at the contact residues of the peptide-binding site (data not shown). Thus, polymorphism at the highly divergent Mamu-A1 locus, characterized by high expression levels, must have resulted from positive Darwinian selection (Hughes and Nei 1988; Borghans et al. 2004).
The Mamu-A2 gene displays differential haplotype distribution in the Indian population of rhesus macaques (Fig. 1). This also appears to be the case for Chinese animals, as Mamu-A2 cDNAs were detected in 39 out of 42 of them. In this panel, 16 alleles that differ from the ones earlier detected in Indian animals were defined. In the Indian population, low expression levels characterize the Mamu-A2 gene, which appears also to be the case in Chinese animals.
Five Chinese macaques possess the Mamu-A3*1307 sequence, which is always observed in combination with Mamu-A1*1102. In concordance with the Indian animals, the number of Mamu-A3 clones detected also reflects low expression levels. A full-length Mamu-A4*1404 cDNA was detected in only three animals. In contrast to the Mamu-A4 allele in the Indian macaque, the Chinese allele has a stopcodon at the last triplet of exon 5 encoding the transmembrane part of the class I protein. This suggests that the corresponding gene product may settle in the membrane, but the signal transduction is impaired. As was observed earlier in Indian animals, this locus is characterized by extremely low expression levels (Fig. 1). For that reason, the presence of Mamu-A4 cDNAs in the other animals was tested with SSP on the RT-PCR samples, and 17 out of 42 Chinese animals appeared to be positive for this locus. The Mamu-A5, -A6, and -A7 genes are also characterized by low expression levels and are considered to represent minors.
In contrast to the previously studied Indian population, the Chinese animals were selected randomly from a large population. Because of the lack of pedigree data and segregation profiles, it was not possible to firmly establish the segregation of different Mamu-A genes present per chromosome. Haplotypes can only be deduced based on sharing of sequences between animals, and such haplotypes are listed in Table 2. At this stage, it is impossible to determine in which region configurations the Mamu-A5,-A6, and -A7 genes are present.
Table 2Chinese Mamu-A haplotypes deduced by sharing of alleles grouped into two-region configurations Mamu-A1Mamu-A2Mamu-A3Number1*1801*15163*2201*05156*5201*051332*1102*13075
Mafa-A region polymorphism and diversity in cynomolgus monkeys
Sixty-two alleles were detected in the 96 cynomolgus macaques analyzed. As observed in the rhesus macaques, most animals had two polymorphic Mafa-A1 sequences, in combination with orthologs of the Mamu-A2, -A3, -A4, -A5, and/or -A6 genes. In total, 38 Mafa-A1 alleles were detected, illustrating the high level of polymorphism of this locus. These alleles were compared with sequences reported earlier by two other research groups (Uda et al. 2004; Krebs et al. 2005) that have used different nomenclature systems. Only four alleles -A1*3101, -A1*3201, -A1*3801, and -A5*3001 detected in our panel are identical to the earlier described sequences A*310101, A*320101, A*380101, and A*300101.
A total of 79 out of 96 animals have at least one Mafa-A2 sequence, and 18 alleles could be distinguished (Table 1). The oligomorphic A2 locus is also present in the pigtailed macaque (Macaca nemestrina) as was recently described (Pratt et al. 2006; Lafont et al. 2007). In these studies, 90% of the animals possess the gene, and lower transcription levels were also observed in comparison to the other Mane-A alleles. The Mafa-A3 gene is observed in only two animals, whereas the Mafa-A4 locus is detected in 41 individuals. Two Mafa-A4 alleles were found that differ in only one basepair (nonsynonymous) substitution (Fig. 2). As in the Chinese rhesus macaques, the Mafa-A4 alleles have a stopcodon at the end of exon 5. The Mafa-A5 and -A6 alleles were detected in, respectively, eight and four animals and display low levels of polymorphism. Phylogenetic analyses illustrate that apart from loci rhesus and cynomolgus monkeys also share lineages (Fig. 2). As found in rhesus macaques, the Mafa-A1 alleles are characterized by high transcription levels, whereas the other loci display moderate or low expression levels.
Because most of the cynomolgus macaques are pedigreed, the combination of sequences that are inherited on one chromosome could be defined. As an example, the kinship tree of one breeding group is provided (Fig. 3). The haplotypes are grouped based on the combination of loci (Table 3). Seven combinations of loci or region configurations are recognized (Fig. 4), and four of them (two to five) are also observed in the Indian rhesus macaques (Fig. 1), which indicates that they predate speciation. Mafa-A5*3001 segregates in combination with Mafa-A1*6001 and -A1*7201 but not exclusively. The exact region configurations of Mafa-A6 containing haplotypes are not yet known.
Fig. 3Pedigree of a cynomolgus macaque family showing segregation of Mafa-A alleles. The animals analyzed are indicated by shading. A question mark indicates that the sire has not been identifiedTable 3Mafa-A haplotypes as defined by segregation grouped into seven-region configurations Mafa-A1Mafa-A2Mafa-A3Mafa-A4Mafa-A5Number1A1*40026A1*58014A1*700132A1*0101A2*05144A1*1003A2*050610A1*1901A2*05133A1*3101A2*05044A1*5901A2*050812A1*6301A2*05017A1*6301A2*05113A1*6401A2*05059A1*6801A2*05153A1*6601A2*0512013A1*6602A2*051010A1*6801A2*051533A1*5701A3*130124A1*1702A4*140211A1*7101A4*1465A1*650102A2*05A4*1436A2*051202A4*1497A1*6001A5*30012A1*7201A5*30013One region configuration appears to lack the Mafa-A1 gene. At this stage, it is unknown whether the locus is really absent or was missed because of PCR failure, for example, because of a mutation in a primer site.Fig. 4Schematic representation of different Mafa-A region configurations in cynomolgus macaques. Four configurations (2–5) are shared with the Indian rhesus macaques (Fig. 1). One region configuration lacks the Mafa-A1 gene. It is not yet known whether the locus is really absent or missed because of PCR failure. Should a Mafa-A1 gene be found in extended studies than this region, configuration has to be deleted from the list
Sharing of -A1 and -A2 sequences between different macaque species
A comparison of all available Mamu-A to Mafa-A sequences showed that six full-length cDNA transcripts are shared between both species of macaque. This is the case for five -A1 pairs and one -A2 pair (Table 4). One of these cynomolgus monkey alleles is identical to an Indian rhesus allele, whereas the other five are shared with animals of Chinese origin. This observation was unexpected as the two rhesus macaque populations investigated share only one allele. The fact that cynomolgus macaques share more alleles with Chinese rhesus macaques than with Indian individuals may be explained by the fact that there is an overlap in the geographic areas inhabited by both species in Indochina (eastern Asia). It is known that rhesus macaques and cynomolgus monkeys can interbreed and produce offspring (Tosi et al. 2002). Sharing of alleles was not observed for the minors controlled by the A3–A7 loci.
Table 4Identical Mhc class I sequences detected in the two species of macaqueMacaca mulattaMacaca fascicularisMamu-A1*0103Mafa-A*260101aMamu-A1*4001Mafa-A*400101aMamu-A1*1001Mafa-A1*1002Mamu-A*21bMafa-A1*0302Mamu-A1*5301Mafa-A1*5301Mamu-A2*0519Mafa-A2*051202aPublished by Krebs et al. 2005.bPublished by Otting et al. 2005.
Different modes of selection operating on the Mhc class I and II genes in macaques
Although Mamu-A and Mafa-A sequences are interspersed in the phylogenetic tree, the vast majority of the alleles are species unique, and for rhesus macaque, most of them appear to be population specific. This is in sharp contrast to the situation observed for the macaque class II region, where sharing for different exon 2 sequences at the Mhc-DR, but especially the -DQ and -DP genes, is far more common (Blancher et al. 2006; Doxiadis et al. 2006). About half of the Mafa-DPB1, -DQA1, and -DQB1 and one third of the Mafa-DRB sequences are identical to rhesus orthologs. As animals of the same populations have been used to study Mhc class I and II sequences, the present results exclude the possibility that this high sharing of Mhc class II sequences is because of introgression. A more likely explanation is that in macaques, the exon 2 sequences of the Mhc class II genes, which encode the peptide-binding site, have been subjected to purifying selection. As a consequence, many Mhc class II alleles in these two macaque species predate speciation processes. A databank search indicated that the phenomenon also extends to other macaque species (Robinson et al. 2003).
The Mhc class I alleles in macaques are largely unique, illustrating that selection has favored diversity. Mhc class I proteins are involved in the presentation of intracellular pathogens such as viruses and parasites that are known to evolve at high mutation rates. Recognition of Mhc class I molecules combined with a foreign peptide may result in the lysis of an infected cell. The Mhc class II molecules select peptides, originating from extracellular pathogens such as bacteria and fungi, for binding. Mhc class II-mediated activation may result in antibody production and/or providing help to cytotoxic T cells. The present results illustrate that, probably because of coevolution with intracellular pathogens, macaques have generated a highly complex and divergent Mhc class I repertoire. The fact that Mhc class I alleles evolve quickly, even within a species, has been documented in humans for instance (Belich et al. 1992; Watkins et al. 1992). The high degree of sharing of Mhc class II sequences seems to be unique for macaques and has not been observed in any other group of vertebrate species.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Table 5
Overview of Mamu-A genes/alleles in the panel of Chinese rhesus macaques. Sequences represented by two ciphers (for example A*05) are not confirmed and are not submitted to the to the EBI-EMBL database (PDF 127 KB).
|
[
"nonhuman primates",
"rhesus macaque",
"cynomolgus macaque",
"evolution"
] |
[
"P",
"P",
"P",
"U"
] |
Ann_Surg_Oncol-3-1-1914273
|
Isolated Limb Perfusion with Tumor Necrosis Factor Alpha and Melphalan for Locally Advanced Soft Tissue Sarcoma: Three Time Periods at Risk for Amputation
|
Background The aim of this study was to investigate the long-term limb salvage rate and overall survival after isolated limb perfusion (ILP) with tumor necrosis factor alpha and melphalan for locally advanced soft tissue sarcoma (STS).
Limb salvage in patients with locally advanced extremity soft tissue sarcoma (STS) continues to be a challenge. Survival in these patients is determined by the development of distant metastases and is not improved with the amputation of the affected limb.1,2 Besides amputation, an extensive surgical procedure followed by radiotherapy is a treatment option.3 Rosenberg et al.1 showed the same disease-free and overall survival as amputation in the early 1980s with this treatment regimen. Preoperative therapies to improve limb salvage rates have been proposed. Suit et al.4 reported in 1981 on the use of preoperative radiotherapy. Eilber et al.2,5 combined preoperative (intra-arterial or systemic) chemotherapy and radiotherapy to improve resectability rates. In a randomized trial, O’Sullivan et al.6 reported a greater risk of wound complications in the preoperative radiotherapy group compared with the postoperative radiotherapy group. The use of brachytherapy may also improve local control and avoid amputation.7 The current treatment strategy of high-grade limb sarcomas is wide local resection, with the goal of achieving a R0 resection with a 2-cm margin. If the margin is < 2 cm or if a R1 resection (microscopically involved margin) is performed, adjuvant radiotherapy with 50–70 Gy is indicated to reduce the risk of local failure.3 The question whether radiotherapy should be given before or after surgery is still unanswered.6
Another strategy for limb salvage in locally advanced extremity STS is to perform an isolated limb perfusion (ILP) with cytostatic agents. This procedure, which was originally developed for the treatment of melanoma of the limb in 1957, was also applied to the treatment of STS of the limb. In their first experience, Krementz et al.8 showed an early response rate of 83% with melphalan alone; however, complete regression of the tumor was rarely seen. Other perfusion agents in the treatment of limb STS were therefore investigated. Rossi et al.9 claimed that doxorubicin was efficacious; another study10 showed that doxorubicin alone was ineffective and that combined with melphalan, it was too toxic. Cisplatin also proved to be less effective than melphalan in the limb perfusion setting of sarcomas, and carboplatin was too neurotoxic.11–13
With the addition of tumor necrosis factor alpha (TNF-α) to the perfusion circuit, Lienard et al.14 made a step forward in the treatment of locally advanced extremity STS. A large European multicenter study proved the ILP concept in the limb salvage procedures for locally advanced STS with TNF-α and melphalan. The objective response rate was 75%, and a limb salvage rate of 82% was achieved with minimal treatment-related morbidity.15 Since 1991, patients with locally advanced STS of the limbs have been treated at the University Medical Center Groningen by ILP with TNF-α and melphalan with or without interferon gamma as perfusion agents, followed by delayed surgical excision and postoperative radiotherapy if a marginal resection or nonradical resection was performed. Recently we encountered long-term local morbidity, and therefore the aim of the present study was to analyze the limb salvage rate and survival in patients with locally advanced STS of the extremities that were treated in our center and to report the late effects of this treatment modality.
PATIENTS AND METHODS
Patient Characteristics
During 1991–2003, a total of 73 patients with STS of the extremity underwent 77 perfusions with a combination of TNF-α and melphalan, with (n = 19) or without (n = 58) interferon gamma. Thirty-six men and 37 women with a median age of 54 (range 14–80) years were treated. Tumors were considered unresectable because of size, their multicentricity in the limb, or fixation to the neurovascular bundle and/or bone, and therefore amputation was the only treatment option. Perfusion was performed at the iliac level in 32 cases (42%), at the popliteal level in 23 cases (30%), and at the femoral and axillary level in 11 cases each (14%). There were 60 primary (82%) and 13 recurrent (18%) sarcomas. Sixty-two sarcomas were located in the leg (85%) and 11 were located in the arm (15%). All patients were treated after informed consent was obtained according to institutional guidelines. Nineteen different histological types of STS were distinguished. The pathological grade of the tumor was scored following the criteria of Coindre et al.,17 and the stage of the tumor was scored according to the American Joint Committee on Cancer (AJCC) criteria16 (Table 1). Median tumor size was 16.2 (range 8.3–23) cm. In case of multifocal disease, the largest diameter was used.
TABLE 1.Histological grade and stage of tumorsaCharacteristicn%Grade I1014 II2332 III4054Stage I1014 II11 III5069 IV1216a Grade according to Coindre et al.,17 and stage according to American Joint Committee on Cancer.16
Perfusion Technique
The perfusion technique used at the University Medical Center Groningen is based on the technique developed by Creech et al.18 and described elsewhere.19 The major modifications during the last 30 years were the use of modern thermal blankets, improvement in leakage monitoring, and the introduction of a membrane oxygenator and heat exchanger to ensure optimal perfusion at 39–40°C. Because extensive washing with 6 L of saline is used, systemic inflammatory response syndrome (SIRS) is rarely seen.20 After surgery, patients can be monitored on the recovery ward instead of the intensive care unit.
Assessment of Tumor Response, Tumor Remnant, and Follow-up
Responses were assessed by standardized World Health Organization criteria and on the basis of physical examination and/or imaging investigations (magnetic resonance imaging and computed tomographic scans).21 Complete response was defined as the disappearance of all measurable disease in the limb for > 4 weeks, partial response as regression of the tumor size by > 50% of the largest diameter for > 4 weeks, and no change as regression of < 50% of the tumor in the limb or progression of < 25% for longer than 4 weeks. Resection of the tumor remnants was performed 2–15 weeks (median, 8 weeks) after perfusion. After resection, response was also assessed by pathological examination. The tumor remnants were measured in three dimensions and the percentage of necrosis estimated in relation to the complete tumor volume. Representative tumor sections were taken, encompassing macroscopically different tumor areas, including necrosis. As a general rule, one section per centimeter largest diameter with a minimum of three was taken. On the basis of an integration of gross and microscopic findings, a final estimate of the percentages of viable and necrotic or regressive tumor was made.
Excision margins were also evaluated on pathological examination and classified as radical when the resection margins were free of tumor cells (complete resection, R0), as R1 when resection margins were microscopically involved, or as R2 when resection margins were macroscopically positive involved. Postoperative radiotherapy (60–70 Gy) was considered indicated in case of < 95% necrosis on pathological examination of the tumor or with marginal or microscopically positive resection margins. All patients were followed after perfusion treatment in a standardized protocol. Median follow-up was 27 (range 2–138) months.
Statistical Analysis
Survival and limb salvage curves were calculated according to the Kaplan-Meier method and log rank test.22 Values of P < 0.05 were considered to be statistically significant. GraphPad Prism version 2.0 for Windows statistical software was used.
RESULTS
Tumor Response
A clinical complete response was observed after 19 ILPs (25%), a partial response after 53 ILPs (69%), and no change after 5 ILPs (6%); local progression was never observed. Resection of the remnant tumor was performed in 68 patients (93%). The pathological response is illustrated in Fig. 1. After 17 ILPs (23%), no viable tumor cells were found on pathological examination. In 29 ILPs (37%), ≥ 90% necrosis was found on pathological examination. When both groups were added together, a good response to ILP was found in 60% of the patients. In 17 ILPs (22%), an intermediate response was found on pathological examination (necrosis 50%–80%). After eight ILPs (10%), < 20% of necrosis or no necrosis was found on pathological examination. In five patients (7%), tumor response was not assessed because of progression of distant metastases in four patients and a local recurrence in one patient that necessitated a second perfusion, resulting in 90% necrosis of the tumor. No correlation could be demonstrated between grade and percentage of tumor necrosis after perfusion (Pearson’s correlation). Postoperative radiotherapy (total dose 60–70 Gy, 25 × 2 Gy daily and 10–20 Gy boost) was given in 37 patients with microscopically involved or marginally free resection margins. Radiotherapy was initiated within 5–6 weeks after tumor resection. Radiation treatment was delivered through a multiple-field technique with computed tomographic treatment planning on a linear accelerator, 6–15 MV.
FIG. 1.Percentage of necrosis estimated at pathological examination of resected tumor remnant in relation to number of patients.
Amputations and Limb Salvage
A total of 21 amputations (28%) were performed. Table 2 presents the time interval between ILP and amputation and the rationale for amputation. Overall 1, 5, and 10 years’ limb salvage was 80.1% ± 4.8%, 68.2% ± 6.5%, and 60.6% ± 9.2%, respectively (Fig. 2). When we analyzed the limb salvage curve, we found that the risk for amputation could be divided into three time episodes. The first period occurs within the first year or year and a half after perfusion (n = 17 patients), with amputation performed to treat massive perfusion-induced necrosis of the tumor and overlying skin, which resulted in a soft tissue deficit (n = 6); tumor recurrence after perfusion (n = 5); wound complications after ILP followed by radiotherapy (n = 2); a microscopically involved resection margin with the rejection of the patient for adjuvant radiotherapy of the foot (n = 2); insufficient clinical response (one patient); and arterial thrombosis with no vascular reconstruction possibilities 2 months after resection of a local recurrence in the groin (n = 1).
TABLE 2.Amputations performed in 21 patients according to interval durationPatient no.DiagnosisAge (y)Interval (mo)ResectionRTRationale for amputationCurrent status1PUS600R0NoPostperfusion necrosis38NED 120 months2Angiosarcoma741R2NoLocal recurrenceDOD 11 months3Fibrosarcoma761R0NoPostperfusion necrosisNED 2 months4PUS672R0NoPostperfusion necrosisDOD 9 months5Epithelioid Sarcoma212R0NoPostperfusion necrosisDOD 54 months6Leiomyosarcoma172R0NoInsufficient clinical responseDOD 7 months7Liposarcoma602R1NoR1 resection, RT not possibleAWD 10 months8PNET623R0NoLocal recurrenceDOD 17 months9Synovial sarcoma393R0NoPostperfusion necrosisDOD 50 months10PUS633R1NoPostperfusion necrosisNED 72 months11Angiosarcoma804R1NoLocal recurrenceDOD 10 months12Synovial sarcoma654R1NoR1 resection, RT not possibleNED 6 months13Epithelioid Sarcoma226R2NoLocal recurrenceaDOD 39 months14Hemangioma pericytoma508R0YesWound complications after ILP with radiotherapyaAWD 65 months15PUS7112R0YesWound complications after ILP with radiotherapyaNED 14 months16PUS6115R0YesArterial occlusionAWD 17 months17Synoviosarcoma4218R1YesLocal recurrenceNED 20 months18Liposarcoma5337R0YesLocal recurrenceDOD 110 months19Liposarcoma3958R0NoLocal recurrenceDOD 120 months20PNET56110R0YesCritical leg ischemiaNED 118 months21Chondrosarcoma18125R0YesCritical leg ischemiaNED 134 monthsPUS, pleomorphic undifferentiated sarcoma; RT, radiotherapy; NED, no evidence of disease; AWD, alive with disease; DOD, death of disease; PNET, malignant peripheral nerve sheath tumor; R1 resection, microscopically involved resection margin; ILP, isolated limb perfusion.a Patients treated for second ILP with tumor necrosis factor alpha and melphalan.FIG. 2.Limb salvage curve in patients with locally advanced soft tissue sarcoma treated with tumor necrosis factor alpha, melphalan, and isolated limb perfusion.
The second time period was within 5 years after ILP, with two amputations performed for late local recurrent disease (37 and 58 months after perfusion). The third episode occurred approximately 10 years after perfusion. Amputation was performed for critical leg ischemia with neuropathy due to treatment-induced atherosclerosis of the remaining tibial artery that was not suitable for arterial reconstruction (110 and 125 months after perfusion). An example of the clinical appearance of patient 21 is shown in Fig. 3. In this patient, a popliteal ILP was performed at the age of 18 years for a chondrosarcoma. After marginal resection, this patient received 66 Gy adjuvant radiotherapy. Ten years after ILP, an amputation was performed because of critical leg ischemia. No recurrent disease was found on pathological examination of the amputated specimen. Another two patients developed a pathological fracture of the femur due to radiation-induced osteonecrosis (78 and 129 months after perfusion). These four patients with late post-ILP complications received high-dose postperfusion radiotherapy (60–70 Gy).FIG. 3.Clinical appearance of lower leg of patient 21 (Table 3) before amputation for critical leg ischemia.
Systemic Metastases and Survival
Twelve patients presented with distant metastases at time of ILP (16% stage IV AJCC); half of these patients had lung metastases, and the other half had lymph node metastases. Eleven of these patients died of disease after a median period of 9 (range 2–54) months; one patient is alive with no evidence of disease after 11 months. During follow-up, 25 patients (36%) developed distant metastases at a median interval of 9 (range 2–100) months. A statistically significant difference (P < .001) was observed between patients with no distant metastases at the time of ILP compared with patients with metastases at the time of ILP (Fig. 4). Overall 1, 5, and 10 years’ survival for all patients was 82.9% ± 9.2%, 58.7% ± 13.1%, and 42.5% ± 18.2%, respectively (Fig. 4).
FIG. 4.Overall survival in patients with locally advanced soft tissue sarcoma treated with isolated limb perfusion (ILP), tumor necrosis factor alpha, and melphalan. Statistically significant difference was observed between patients with no distant metastases at time of ILP (mets −) compared with patients with metastases at time of ILP (mets +).
DISCUSSION
The results of a European multicenter trial performed in the 1990s led to the approval of the use of TNF-α for ILP in patients with locally advanced extremity sarcomas by the European Medicine Evaluation Agency.15 ILP with TNF-α is currently available in more than 30 centers, and in 2002, a total of 350 so-called TNF-α perfusions were performed. As one of the first centers that participated in the TNF-α ILP experience, and with more than a decade’s experience, we recently encountered long-term treatment-related morbidity necessitating amputation of the perfused limb 10 years after treatment. For this reason, we analyzed our results of ILP with TNF-α and melphalan and describe our results in the present study.
We observed an overall response rate of 82%, which is in the range of the 63%–91% response rates reported in the literature.15,23,25 Although a suggestion has been made for a relation between the grade of sarcoma and the response to TNF-α ILP, we could not demonstrate a correlation between grade and the percentage of necrosis after ILP with TNF-α. This is in concordance with the results of the Amsterdam group.23 Various reports have shown that a limb salvage rate of 81%–86% can be achieved in patients with locally advanced limb sarcoma.15,23,25 An independent review committee reconsidered the unresectability criteria of all patients enrolled onto the European study. Eighty percent of the patients in this study met the criteria for unresectability, and survival curves based on a match-control study with cases of the Scandinavian Soft Tissue Sarcoma Databank showed that TNF-α ILP had no negative effect on survival.26
We used the Kaplan-Meier method to calculate limb salvage because this method adjusts for censored observations, i.e., patients who were alive and well at the time of last contact or patients who died of distant metastases but who retained preserved limb function. By use of this method, we calculated a 1-year limb salvage rate of 80%. Amputations were performed mostly because of postperfusion-related complications or early local recurrence in the first year after ILP. A second curve in limb salvage was observed within 5 years after TNF-α ILP in two patients with late local recurrences. A third bend in the limb salvage curve was observed approximately 10 years after ILP. This was a new observation in two patients who had critical leg ischemia with ulceration and continuous pain. Besides ILP with TNF-α and melphalan, both patients received adjuvant radiotherapy (66 and 70 Gy) after marginal tumor resection.
What seems to be the cause of this late morbidity? Analysis of the functional and long-term morbidity in 97 patients with stage I melanoma treated with ILP with melphalan as the sole perfusion agent in our center showed after a median follow-up of 36 (range 12–76) months, no patients had critical leg ischemia.27 The Rotterdam and Amsterdam perfusion group reported long-term morbidity consisting of muscle atrophy or fibrosis in 11% of the patients after ILP with melphalan; however, cases of critical leg ischemia are not described.28 The fact that in our series no muscle atrophy or fibrosis was found might be explained by the fact that we always perform a lateral fasciotomy after ILP to prevent a compartment syndrome. When we performed a literature search for late morbidity after ILP with TNF-α and melphalan, no studies could be retrieved.
The clinical importance of late morbidity after radiotherapy has evolved since Eifel et al.29 retrospectively reviewed the medical records of 1784 FIGO stage IB patients receiving primary radiotherapy at the M. D. Anderson Cancer Center between 1960 and 1989. She showed that after 5 years, there was a small but continuous risk of experiencing major complications of radiotherapy (i.e., urinary, rectal, and small bowel complications) at up to 20 years of follow-up. Johansson et al.30 described a high occurrence of severe neuropathy closely linked to the development of fibrosis around the nerve trunks after aggressive postoperative telecobalt therapy received in 1963–1965 in a group of 71 patients with breast cancer who were initially treated with modified radical mastectomy. Radiotherapy damage to the vascular system was demonstrated by Hopewell31 in an experimental setting; arteries of hamster cheek pouch showed localized constrictions after irradiation. These constrictions were caused by clones of endothelial cells and may be the predominant factor influencing the degeneration of the capillary bed after radiotherapy.32 Evidence of this occlusive effect of vessels by proliferating endothelial cells after radiation have also been reported by other investigators.33
Another argument to explain the observed late morbidity, at least in part, to radiotherapy is the fact that another two patients who received adjuvant radiotherapy after ILP with TNF-α and melphalan developed a pathological fracture of the femur (78 and 129 months after ILP with TNF-α and melphalan). Radiotherapy-induced osteonecrosis is a well-known phenomenon. Lin et al.34 described 12 fractures of the femur after surgery and irradiation for STS of the thigh. Treatment of these fractures was difficult and demanding, with only four bony unions after a mean follow-up of 37 months. When we add up the evidence of developing fibrosis after ILP with melphalan and the development of fibrosis after radiotherapy, the combination of the two regimens could explain the observed late morbidity rate in the present series.
Overall survival for all patients showed a steadily decline, with a 10-year percentage of 42%. Even after 110 and 120 months, patients die of distant metastases. Sixteen percent of the patients had metastases at the time of ILP. A marked difference in survival was observed when comparing patients with pulmonary or lymph node metastases with patients lacking metastases at time of ILP. Five-year overall survival of 59% in this series is higher than the reported 5-year survival of 48% in the Amsterdam experience23 and the reported 32% of Lejeune et al.24 This is an unexpected observation because selection criteria for ILP with TNF-α between the institutes are comparable. High-grade tumors and tumors > 5 cm in diameter have a great tendency to metastasize. Patients with such tumors could theoretically benefit from neoadjuvant chemotherapy. A quantitative meta-analysis of data from 14 trials of doxorubicin-based adjuvant chemotherapy showed a benefit from systemic adjuvant chemotherapy of 6% for local relapse-free interval, but there was no overall survival benefit at 10 years.35 Delaney et al.36 developed a regimen of preoperative chemotherapy consisting of mesna, Adriamycin, ifosfamide, and dacarbazine (MAID) interspersed with radiotherapy, followed by resection and postoperative chemotherapy with or without radiotherapy to improve outcome in patients with high-grade extremity STS. Compared with a historical group of control patients, outcome in the MAID group was better. In an update of 64 patients, 5 required amputation because of disease, 3 had unresectable disease, and 1 patient refused surgery. Estimated 3-year survival and local-regional control were 75.1% and 79.3%, respectively. These results are comparable with the results of the present study.37 However, systemic therapy is associated with systemic toxic effects, in contrast with the mild systemic side effects observed after ILP with TNF-α and melphalan.
Since 1992, we have not changed the indication for TNF-α perfusion. Patients who were candidates for amputation of the involved limb, as assessed by preoperative magnetic resonance imaging, were offered an ILP with TNF-α and melphalan, with the goal of preserving the limb with a locally advanced STS. After ILP patients received a delayed surgical resection, adjuvant radiotherapy was provided to patients with marginal or microscopically positive resection margins. This treatment resulted in a high limb salvage rate in patients with locally advanced STS, although late morbidity can occur, especially when adjuvant postoperative radiotherapy is applied. Therefore, continuous follow-up of these patients is warranted.
|
[
"isolated limb perfusion",
"tumor necrosis factor alpha",
"melphalan",
"sarcoma",
"amputation",
"radiotherapy",
"late morbidity"
] |
[
"P",
"P",
"P",
"P",
"P",
"P",
"P"
] |
Neuroradiology-2-2-1592467
|
Cerebral blood volume, genotype and chemosensitivity in oligodendroglial tumours
|
Introduction The biological factors responsible for differential chemoresponsiveness in oligodendroglial tumours with or without the −1p/−19q genotype are unknown, but tumour vascularity may contribute. We aimed to determine whether dynamic susceptibility contrast (DSC) magnetic resonance imaging (MRI) could distinguish molecular subtypes of oligodendroglial tumour, and examined the relationship between relative cerebral blood volume (rCBV) and outcome following procarbazine, lomustine and vincristine (PCV) chemotherapy.
Introduction
Oligodendroglial tumours with combined loss of chromosomes 1p and 19q have a more indolent clinical evolution, respond more favourably to chemotherapy and show prolonged survival compared to those with intact 1p/19q, but the basis of these clinical differences is not well understood [1–5]. In addition, up to 30% of oligodendroglial neoplasms with intact 1p/19q may respond to chemotherapy [3, 6, 7], but the genetic and biological factors associated with response are unknown and not all patients with oligodendroglial tumours derive durable benefit from therapy. Assessment of the clinical utility of biological markers and non-invasive imaging techniques is essential to advance the clinical management of these patients.
Recently dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) has been used to calculate perfusion parameters in gliomas [8, 9], guide biopsy [10], provide prognostic information [11] and demonstrate differences in the vascularity of low-grade astrocytomas compared with oligodendrogliomas [8]. It has been suggested that increased vascularity and concomitant improved drug delivery may contribute to the chemoresponsiveness of oligodendroglial tumours, especially as lipophilic drugs (procarbazine, lomustine and temozolomide), which easily cross the blood–brain barrier, are used in the treatment of oligodendrogliomas [12–14]. Currently DSC-MRI is used in the radiological evaluation of gliomas, but its role in oligodendroglial neoplasms classified by genotype and the relationship with response and outcome following chemotherapy has not so far been investigated.
The aim of this study was to determine whether DSC-MRI obtained through routine pre-therapy MR diagnostic examinations may be used to distinguish molecular subtypes of oligodendroglial tumours, and to investigate the influence of tumour blood volume on outcome following procarbazine, lomustine and vincristine (PCV) chemotherapy compared to molecular genetics.
Methods
Patient selection
Patients for investigation were selected from those participating in a prospective study of adult oligodendroglial tumours eligible for chemotherapy at the Walton Centre for Neurology and Neurosurgery/Clatterbridge Centre for Oncology between May 2000 and June 2003 [7, 15–18]. The study had full ethical approval. For inclusion in this study, patients had to have (1) given research consent, (2) a histological diagnosis of oligodendroglioma or oligoastrocytoma based on the current World Health Organisation (WHO) classification [19], (3) a known 1p/19q status, and (4) undergone DSC-MRI before or after biopsy (2 image-guided; 35 serial stereotactic), but prior to commencing therapy. All patients (with primary and recurrent disease) underwent biopsy prior to PCV. Patients with post-resection imaging were excluded.
DCS-MRI
Imaging studies were obtained using a 1.5-T Signa MR scanner (General Electric Medical Systems). The MR imaging protocol consisted of T1-weighted spin-echo (SE) sagittal sections (400/14 TR/TE), axial fluid attenuated inversion recovery (FLAIR 8802/142/2200 TR/TE/TI), axial fast spin-echo (FSE) T2-weighted imaging (4360/102 TR/TE), 2D dynamic susceptibility-weighted, SE, echoplanar imaging (EPI 1900/80 ms, bandwidth 62.5) and postcontrast T1-weighted SE (600/14 TR/TE) or FLAIR (2280/11/750 TR/TE/TI) axial imaging. The T2-weighted FLAIR, FSE and postcontrast T1-weighted images were used to define regions of T2 hyperintensity and T1 enhancement. For the dynamic EPI series, the tumour size and location, and the position of the margins were determined from the T2-weighted FLAIR or FSE images to equally define low-grade and high-grade tumours. Nine slices were selected to cover the majority of the tumour volume. Slice thickness was 7 mm, with no gap. A standard dose of 0.1 mmol/kg body weight of dimeglumine gadopentetate (Gd DTPA) was injected intravenously manually at an approximate rate of 3 ml/s. A series of T2*-weighted multislice image sets (field of view 30×30, matrix 192×128) were acquired every 2 s during the first pass of the contrast agent. Image postprocessing was performed using commercial software (Functool2, version 6.0; Sun Microsystems). Negative enhancement integral (NEI) maps were generated for each patient relative to the signal intensity in the superior sagittal sinus (Fig. 1). The axial slice with the greatest cerebral blood volume and tumour area was chosen for analysis.
Fig. 1Calculation of rCBV using DSC-MR. CBV is derived from the drop in the T2* signal after injection of a bolus of paramagnetic compound [8, 9]. Using Functool2, the radiological tumour margin was delineated on the T2* sequence axial image and a single circular ROI placed over normal brain mirrored in the contralateral hemisphere (a, b). Negative enhancement integral (NEI) colour maps were generated (c, d) relative to the signal intensity in the superior sagittal sinus using image data between dynamic scans immediately before and after the contrast transient. CBV is proportional to the area under the contrast agent concentration-time curve and was calculated for each ROI by Functool2 using algorithms to integrate ΔR2i values (where ΔR2 is the change in the reciprocal of T2*). rCBV for each tumour was calculated from signal intensity time curves for individual ROIs placed within the delineated tumour region or in the large ROI in the contralateral hemisphere (as illustrated in e and f, arrows scans adjacent to the contrast transient) as described in the Materials and methods. a, c, e Low rCBV (0.73±0.17, mean±SD) in a grade II oligoastrocytoma with intact chromosomes 1p and 19q and; b, d, f high rCBV (4.15±0.87) in a grade II oligoastrocytoma with the −1p/−19q genotype
The radiological tumour margin was delineated on the T2* sequence axial image using a nongeometric region of interest (ROI), and a single circular region was placed over normal brain mirrored in the contralateral hemisphere. Depending on tumour location, the mirrored area could include both white and grey matter. ROIs were placed according to published methods giving the lowest inter- and intraobserver variability [20] within the delineated tumour region. Briefly, five ROIs (area range 12–27 mm2) were placed over the tumour areas with the highest blood volume. Cystic and hemorrhagic regions were avoided and care was taken not to place ROIs over normal blood vessels or a biopsy tract where seen. Three independent analyses were performed by a neuroradiologist (T.S.S.) and a neurosurgeon (M.D.J.) blinded to histology and genotype, to assess inter- and intraobserver variability. For each observation, data were expressed relative to the internal reference (four ROIs placed within the delineated mirrored region of normal brain without deliberate bias towards either grey or white matter) to generate relative cerebral blood volume (rCBV) values (mean ROItumour/mean ROInormal brain) and accommodate physiological differences between patients [21]. For comparison with clinical data, the mean rCBV from these three data sets was used.
Molecular genetics
For each patient regions of tumour histology in pretherapy biopsy specimens (formalin-fixed, paraffin-embedded tissues or intraoperative diagnostic smear preparations) representative of the most aggressive tissue available and the overall pathology diagnosis were selected for laser capture microdissection to enrich the tumour component in the samples for analysis. Allelic imbalance was investigated using paired normal (microdissected uninvolved brain or DNA from blood) and tumour tissue samples and multiple simultaneous PCR amplification of a panel of microsatellite markers, capillary electrophoresis and data analysis as described previously [16–18]. The microsatellite panel included: chromosome 1 – D1S2667, D1S508, D1S214 (1p36); chromosome 19 – D19S412, D19S112, D19S596 (19q13); chromosome 17 – D17S796, D17S1176, D17S1353 (17p13); chromosome 10p – D10S89, D10S179 (10p12-10p15); chromosome 10q – D10S1687, D10s2491, D10S583, D10S587, D10S212 (10q22-10q26). Additional markers, D1S468, D1S2795 (1p36) and D19S217, D19S572 (19q13) were used in cases that lacked informativity for given loci. p53 mutations (exons 5–8) were detected by direct sequence analysis of PCR products.
PCV chemotherapy, response and clinical outcome
All patients receiving oncology therapy had evidence of advancing disease (either radiographic or clinical progression) prior to therapy. Patients with mass effect or evidence of malignancy were treated at presentation; others were followed clinically until progression. Oncology therapy was based on individual patient’s clinical status/choice. Patients fit for PCV were offered PCV chemotherapy, or radiotherapy if fertility was an issue. PCV chemotherapy was administered according to standard clinical protocols [22]: lomustine (CCNU) (110 mg/m2), procarbazine (60 mg/m2) and vincristine (1.4 mg/m2, maximum 2 mg) were given on days 1, 8–21, and 8 and 29. respectively. Cycles were repeated every 6 weeks for a maximum of six cycles. As described previously [18], a consultant neuroradiologist reviewed all available MR or CT images taken before, during and after chemotherapy, and at follow-up. The largest perpendicular diameters of the tumour were measured on axial sections of T2-weighted MR images or contrast-enhanced regions on T1-weighted MR or CT images. Response was assessed using the criteria of Macdonald et al. [23] and postcontrast T1-weighted MRI in 25enhancing tumours. In tumours showing no enhancement (n=6) or in those in which contrast enhancement was not assessable (n=1), response was based on T2-weighted images.
Information regarding current management, follow-up and outcome was collected prospectively. Progression-free survival (PFS) and overall survival (OS) were calculated from the start of PCV to progression, last seen, or death as appropriate.
Statistical analysis
Inter- and intraobserver variability was determined using Bland and Altman plots [24]. Differences in rCBV between genetic and clinical factors were assessed using Student’s t-test for numerical data and the Chi-squared or Fisher’s exact test for categorical data. Receiver operator characteristics (ROC) analysis was performed to define the optimal rCBV cut-off for predicting 1p/19q status. Kaplan-Meier survival curves were obtained and differences in OS or PFS were tested for statistical significance using the Log Rank test. Cox regression multivariate analysis of factors shown to be significant in univariate analysis was by forward stepwise entry of parameters at a significance of 0.05 for entry and 0.01 for removal. Probability (P) values less than 0.05 were considered significant.
Results
The study included 30 patients with primary tumours without previous therapy and 7 with recurrent or persistent disease following radiotherapy. Median time between previous radiotherapy and MR imaging was 6.5 years (range 1.1–14.2 years). Median age was 44 years (range 28–71 years).
Interobserver and intraobserver variability
Tumour rCBV was obtained from DSC-MRI (Fig. 1) by two independent observers. Bland and Altman plots revealed low inter- and intraobserver variation (Fig. 2). The closest observed limits of agreement were between the first observations of the neuroradiologist and neurosurgeon (mean −0.17, range −1.27 to 0.93; Fig. 2a). The widest observed limits of agreement were between the repeat observations of the neurosurgeon (mean 0.08, range −1.59 to 1.95; Fig. 2c). The greatest variation was observed in tumours with higher rCBV values.
Fig. 2Bland and Altman plots illustrating interobserver (a, b) and intraobserver (c) variation of rCBV measurements. The greatest variation is seen at the right of the plot for tumours with the higher rCBV values
Histopathology, genotype and rCBV
Oligodendroglial tumours of different histopathology subtypes and grade did not have significantly different rCBV measurements (Figure 3a,b). Loss of both 1p36 and 19q13 was seen in 6/9 grade II oligodendrogliomas, 6/14 grade II oligoastrocytomas, 4/4 grade III oligodendrogliomas, and 3/10 grade III oligoastrocytomas. Tumours with the −1p/−19q genotype had significantly higher rCBV than those with intact 1p and 19q (Fig. 3c). Similar findings were observed when primary tumours only were analysed (data not shown). ROC analysis was used to determine the specificity and sensitivity of rCBV as a potential non-invasive tool to distinguish oligodendroglial neoplasms with or without the −1p/−19q genotype. The area under the ROC curve was 0.853 (95% confidence interval: 0.778–0.927). A cut-off value of 1.59 for rCBV (sensitivity 92%, specificity 76%) best discriminated oligodendroglial tumours with and without the −1p/−19q genotype (Fig. 3c).
Fig. 3Box plots of rCBV against (a) histopathology subtype (P=0.279*), (b) histopathology grade (P=0.442*), and (c) genotype (P=0.001*) in oligodendroglial tumours. The optimal rCBV cut-off that distinguishes tumour genotype is shown by the horizontal line. *Student’s t-test
To facilitate comparison of rCBV with genotype and clinicopathological characteristics (Table 1), tumours were considered to have high rCBV if this was greater than 1.59 as determined by the ROC analysis (Fig. 4). High rCBV values were significantly more common in tumours with loss of 1p36 and 19q13 and less common in tumours with 17p13 loss and p53 mutation. Three tumours in the series had loss of chromosome 10q22–26, of which two also had loss of 10p11–15; all had high rCBV. One tumour with high rCBV had loss of 1p36 without loss of 19q13. As for numerical data, there was no significant association between high rCBV and histology subtype or grade. Contrast enhancement was weakly associated with high rCBV in the series but not in primary tumours. No associations were seen between tumour rCBV and patient age <50 years versus age ≥50 years, gender or temporal versus non-temporal location. Similar associations between rCBV, genotype and clinicopathological characteristics were obtained if only primary tumours were analysed (Table 1).
Table 1rCBV, molecular genetics and clinical characteristics in the series and primary tumours All tumoursPrimary tumoursnNumber of tumours with high rCBV P valueanNumber of tumours with high rCBV P valueaLoss of 1p36 and 19q13Yes1917 (89%)0.000*1614 (88%)0.000No184 (22%)142 (14%)Loss of 17p13Yes112 (18%)0.003*80 (0%)0.001No2619 (73%)2216 (73%)Loss of 10q 22–26 ±10p11–15Yes33 (100%)0.23811 (100%)1.0no3317 (52%)2915 (52%)p53 mutation(exon 5–8)Yes112 (18%)0.004*80 (0%)0.001No2518 (72%)2115 (71%)Patient age (years)<502513 (52%)0.491209 (45%)0.260>50128 (67%)107 (70%)Histology subtypeOligodendroglioma1310 (77%)0.091118 (73%)0.142Oligoastrocytoma2411 (46%)198 (42%)WHO histology gradeII2311 (48%)0.19199 (43%)0.118III1410 (71%)217 (78%)Contrast enhancementAbsent82 (25%)0.04682 (25%)0.092Present2819 (68%)2114 (68%)Tumour locationTemporal147 (50%)0.733136 (46%)0.713Nontemporal2314 (61%)1710 (59%)Tumour statusPrimary3016 (53%)0.674Recurrent75 (71%)Response to PCV All tumoursResponders2114 (67%)0.1421711 (65%)0.097Nonresponders114 (36%)82 (25%) Tumours with loss of 1p/19qResponders1614 (88%)1311 (85%)Nonresponders00 Tumours with intact 1p/19qResponders50 (0%)0.24540 (0%)0.515Nonresponders114 (36%)82 (25%)aProbability calculated by Fisher’s Exact test. Responders showed complete, partial or minor response and non-responders had stable or progressive disease.Fig. 4rCBV, molecular genetics and clinical characteristics. rCBV for each tumour is given as the mean±standard deviation of three independent observations. The arrow indicates tumours with rCBV above (high rCBV) and below the cut-off value of 1.59 (low rCBV). CR complete response (disappearance of all tumour, off steroids and neurologically stable or improved), PR partial response (50% or greater reduction in cross-sectional area, steroids stable or reduced, and neurologically stable or improved); MR minor response (>25 to <50% reduction in cross-sectional area, steroids stable or reduced, and neurologically stable or improved); PD progressive disease (25% or greater increase in cross-sectional area or any new tumour on CT/MR images and/or neurologically worse with steroids stable or increased); SD stable disease (all other situations); P primary tumour, R recurrent tumour
rCBV and outcome following PCV chemotherapy
Of the 37 patients, 33 had PCV chemotherapy following DSC-MRI (median 0.7 months, range 0.1–15.2 months), of whom 26 with primary tumours received PCV chemotherapy as the first oncology therapy, 6 with recurrent or persistent disease treated initially with radiotherapy received PCV as second oncology therapy, and 1 had PCV following radiotherapy and temozolomide. One patient with primary tumour had therapy switched to radiotherapy at the end of the first cycle because of reaction to procarbazine. The remainder received a median of four cycles of PCV (range one to six). Response to therapy was assessed in 32 patients (Fig. 4). Measured radiological changes and clinical factors permitted response assessment using standard criteria of Macdonald et al. [23] in all patients showing enhancement. However, in tumours showing no enhancement assessed using T2-weighted MR images, three showed radiological change in cross-sectional area of >25% to <50%, accompanied by clinical benefit; these were included in the minor response (MR) category. None of those showing enhancement had measured radiological change in this range. Response (i.e. complete, partial or minor; see Fig. 4 legend for definition) was seen in tumours with and without high rCBV. Similarly, non-responders (i.e. patients with stable or progressive disease) had either high or low tumour rCBV (Table 1, Fig. 4). Response was strongly associated with genotype; all 16 tumours with loss of 1p36 and 19q13 responded compared with 5 of 17 tumours with intact 1p36/19q13 (Chi2 P<0.001). Of the 16 responders in the 1p36/19q13 loss group, 14 had high rCBV. The five responders in the intact 1p36/19q13 group all had low rCBV and the four tumours with high rCBV failed to respond.
The relationship between rCBV, genotype, histopathology and survival following PCV in the series and in those with primary tumour is given in Table 2. The −1p/−19q genotype and low-grade pathology were associated with longer PFS and OS. Histopathology subtype was not significantly associated with outcome. Increased rCBV was only associated with short PFS and OS when adjusted for the −1p/−19q genotype. Tumours with intact 1p/19q and high rCBV (two grade III primary oligoastrocytoma, one grade II recurrent oligodendroglioma, one grade III recurrent oligoastrocytoma) were associated with shorter PFS and OS than tumours with intact 1p/19q and low rCBV or those with lp/19q loss and high rCBV (Fig. 5a,b). Similar findings were observed if primary tumour only were analysed (Fig. 5c,d). Compared with significant clinicopathological parameters, high rCBV was an independent prognostic factor for PFS when adjusted for 1p/19q status (Cox regression: PFS: hazard ratio 13.4, 95% CI 1.3–133; P=0.008) and for PFS and OS in patients with intact 1p/19q (Cox regression: PFS: hazard ratio 16.5, 95% CI 1.8–152.2; P=0.004; OS: hazard ratio 7.4, 95% CI 1.6–33.3; P=0.013).
Table 2rCBV and survival in the series and in those with primary tumours. PFS and OS were calculated from start date of PCV (n number of cases, nc not calculated, nr median survival not reached). Contrast enhancement (present vs absent), age (<50 vs ≥50 years), ECOG performance status (0–1 vs ≥2) and gender (male vs female) were not significantly related to PFS or OS nProgression-free survival (months)Overall survival (months)Median (95%CI)P valueaP valuebMedian (95%CI)P valueaP valuebAll tumoursLoss of 1p36 and 19q13Yes16>56 (nr)0.005>56 (nr)0.020No178 (0–21)>49 (nr)Histopathology subtypeOligodendroglioma11>56 (nr)0.0550.187>56 (nr)0.1130.288Oligoastrocytoma218 (9–41)50 (9–91)Histopathology gradeII19>56 (nr)0.0040.002>56 (nr)0.0070.012III144 (8–26)21 (8–34)rCBVLow14>46 (nr)0.7660.003>49 (nr)0.5850.002High1946 (nc)>56 (nr)Primary tumoursLoss of 1p36 and 19q13Yes14>56 (nr)0.037>56 (nr)0.056No12>39 (nr)49Histopathology subtypeOligodendroglioma10>56 (nr)0.0550.137All censored0.0920.209Oligoastrocytoma1646 (6–86)50 (6–101)Histopathology gradeII17>56 (nr)0.0470.009>56 (nr)0.0480.041III96 (13–37)50 (nc)rCBVLow12>46 (nr)0.6810.007>49 (nr)0.6110.003High14>56 (nr)>56 (nr)aProbability calculated by the Log-Rank test.bProbability calculated by the Log-Rank test after adjustment for 1p/19q genotype.Fig. 5rCBV and outcome following PCV chemotherapy: a, b all patients; c, d patients with primary tumour only; a, c progression-free survival; b, d overall survival.① Patients with tumours with low rCBV (<1.59) and loss of 1p36 and 19q13 (all n=2, primary n=2).② Patients with tumours with high rCBV (>1.59) and loss of 1p36 and 19q13 (all n=15, primary n=12).③ Patients with tumours with low rCBV (<1.59) with intact 1p36 and 19q13 (all n=12, primary n=10).④ Patients with tumours with high rCBV (>1.59) and intact 1p36 and 19q13 (all n=4, primary n=2)
Discussion
To our knowledge this is the first study investigating the relationship between rCBV, genotype and outcome in oligodendroglial tumours treated with PCV chemotherapy. The cohort was drawn from a larger series in an ongoing prospective study of oligodendroglial neoplasms [15–18] and reflects the range of patients given PCV chemotherapy in routine clinical practice. Tumours with rCBV >1.59 were more likely to have 1p/19q loss. PCV chemosensitivity was predicted by 1p/19q loss, but high rCBV was not necessary for response. Following chemotherapy, rCBV was significantly associated with PFS and OS only after adjustment for genotype.
DSC-MRI can be used to determine cerebral perfusion parameters such as rCBV, relative cerebral blood flow and permeability, but with the instrumentation available to us was restricted to rCBV. As in other studies [8, 20, 25, 26], placement of ROI in regions within tumour of greatest signal intensity yielded data with good inter- and intraobserver variability. The greatest variation was seen in tumours with high rCBV, reflecting heterogeneity of blood volume within these tumours. In this series, tumours with the −1p/−19q genotype were more likely to have high rCBV than those with intact 1p/19q. The ROC analysis determined an optimal cut-off (rCBV >1.59) for predicting the −1p/−19q genotype with 92% sensitivity and 76% specificity. These data suggest that in conjunction with histopathology, rCBV may be used to distinguish tumours of the oligodendroglial lineage.
MRI studies of tumour vasculature in gliomas have demonstrated that rCBV may be useful in the non-invasive determination of tumour grade and vascularity [21, 25, 27, 28]. Compared to their low-grade counterparts, high-grade tumours have elevated rCBV, associated with increased mitotic activity [29]. rCBV correlates with conventional angiographic assessment of tumour vascular density and histological measures of angiogenesis such as microvascular density and VEGF expression [26, 30–32], which reflects the histopathological finding of increasing neovascularization with grade [19]. However, most studies have investigated astrocytic gliomas or included a variety of glioma pathologies and large series of oligodendroglial neoplasms are uncommon. Recent reports suggest that rCBV values in oligodendrogliomas render DSC-MRI-based tumour grading inaccurate [28, 33], and rCBV may be similar in low- and high-grade oligodendrogliomas [33]. Indeed low-grade oligodendrogliomas have significantly elevated rCBV compared to low-grade astrocytomas [8]. In our series, rCBV did not distinguish oligodendrogliomas from oligoastrocytomas, nor was rCBV significantly different between grade II and III tumours. This lack of correlation between rCBV and grade is consistent with the increased microvessel density seen in both low- and high-grade oligodendroglial tumours [34, 35].
It has been postulated that the tumour vasculature in oligodendrogliomas contributes to their chemosensitivity [12–14]. In an anaplastic oligodendroglioma treated with PCV chemotherapy, post-therapy volume changes and normalization of rCBV were more evident in regions of the pretherapy tumour that showed high rCBV [13]. In the present study, tumours with and without the −1p/−19q genotype, and high or low rCBV showed response to chemotherapy when assessed using conventional criteria [1, 23]. These data indicate that rCBV measured through DSC-MRI is not a key factor in determining PCV chemosensitivity. These findings may reflect the limitations of DSC-MRI associated with contrast recirculation and disruption of the blood–brain barrier in tumours [8, 9]. Alternatively, chemosensitivity may be influenced more by blood flow than by volume. Further studies to measure additional parameters associated with vascular perfusion, including blood flow and permeability are essential to establish whether these factors contribute to chemosensitivity.
The −1p/−19q genotype is known to be associated with longer PFS and OS [5] but the impact of tumour vasculature on prognosis in oligodendrogliomas is confined to a few histopathological [34, 36, 37] and imaging studies [12]. Although our cohort is small, we have presented preliminary data that the prognostic significance of rCBV may differ in tumours with or without the −1p/−19q genotype. 1p/19q loss and low-grade pathology were associated with favourable PFS and OS following PCV. In other studies of astrocytic gliomas [28] or mixed low-grade gliomas [11], high rCBV predicted poor prognosis. In our series, rCBV alone was an unreliable indicator of outcome, showing prognostic significance only after stratification for genotype, suggesting different effects for the two genetic subgroups at the cut-off used to classify tumours with high rCBV. With this cut-off, tumours with high rCBV and 1p/19q loss were associated with longer survival than those with high rCBV and intact 1p/19q. Unlike astrocytic gliomas, high rCBV values do not necessarily indicate aggressive biology associated with poor outcome in oligodendroglial neoplasms with 1p/19q loss, suggestive of differences in the baseline biology of these two genetic subtypes. These preliminary data therefore indicate that further investigation of the prognostic significance of rCBV in molecular subtypes of glioma in a much larger series is warranted.
Conclusion
In conclusion, we have shown that high rCBV is more likely in oligodendroglial tumours with the −1p/−19q genotype and may be a useful aid to diagnosis; however, rCBV does not predict chemosensitivity. Lack of correlation with histopathology grade and the data suggesting that the prognostic significance of rCBV may differ in patients with or without the −1p/−19q genotype, indicates that different interpretation of rCBV data may be necessary for oligodendroglial compared with astrocytic gliomas. Further study is essential to fully evaluate the role of DSC-MRI in the non-invasive diagnosis and prediction of prognosis in gliomas with an oligodendroglial phenotype.
|
[
"chemosensitivity",
"oligodendroglial tumour",
"rcbv",
"outcome"
] |
[
"P",
"P",
"P",
"P"
] |
Pharm_World_Sci-3-1-1805038
|
Determinants of potential drug–drug interaction associated dispensing in community pharmacies in the Netherlands
|
Objective: There are many drug–drug interactions (D–DI) of which some may cause severe adverse patient outcomes. Dispensing interacting drug combinations should be avoided when the risks are higher than the benefits. The objective of this study was to identify determinants of dispensing undesirable interacting drug combinations by community pharmacies in the Netherlands.
Statements on the impact of the article on practice
Process and structure characteristics may influence the dispensing of undesirable interacting drug combinations in community pharmacies but probably to a minor degree.
Medication surveillance in Dutch pharmacies seems to be effective.
Introduction
Drug–drug interactions (D–DIs) are responsible for many adverse patient outcomes. Different studies suggest that D–DIs may cause up to 3% of all hospital admissions [1–4]. A D–DI is defined as a pharmacokinetic or pharmacodynamic influence of drugs on each other, which may result in desired effects, in reduced efficacy and effectiveness or in increased toxicity [5]. Although many D–DIs exist, only a small part of these D–DIs is clinically relevant [6–8]. The potential benefits of drug combinations should be weighed against the seriousness of the D–DI, taking into account the availability of alternatives. Only in cases that the risks associated with the D–DI are higher than the benefits or if a better alternative is available, the D–DI should be avoided.
In the Netherlands, one of the tasks of the pharmacist is to intervene in case of D–DIs, which involve a high risk for the patient. Hereto, the pharmacist uses patient characteristics and the medication history. All prescriptions, which are submitted to the pharmacy, are screened on potential interactions with the help of medication surveillance software. These D–DIs are evaluated by the pharmacist who intervenes if necessary. This task is important but cumbersome, and requires great attention from the pharmacist. The organisational aspects, such as the tuning of the medication surveillance software and instructions of technicians, should be managed by the pharmacist in such a way that in case of D–DIs with a high risk the pharmacist intervenes. This is important for the prevention of adverse patient outcomes [9].
The objective of this study was to assess process and structure characteristics associated with the dispensing of interacting drug combinations, which carry a high risk of adverse patient outcomes.
Methods
Setting
The data for this study were retrieved from the Drug Information Project, a division of the Health Care Insurance Board. This is a database containing the reimbursement data from eight health care insurance companies in the Netherlands. The reimbursement data from January 1st, 2001 until October 31st, 2002 were analysed. Eleven potential D–DIs, that contained a high risk and could be substituted, because a good alternative was available, were selected and counted for each pharmacy in the database. These undesirable potential D–DIs were mostly interactions between chronically used drugs which cannot be interrupted and short-term use of antibiotics or antimycotics, and were selected from the Dutch guidelines for the management of D–DIs (Table 1) [10, 11]. A D–DI was counted as such, when the chronically used drug was dispensed both in the period 150 days preceding and in the period 150 days after the dispensing of antibiotics or antimycotics for short-term use in the same pharmacy. Pharmacies with less than 5,000 dispensings in the database were excluded.
Table 1Number of dispensings in the database of the individual drugs involved, the eleven potential D–DIs and the calculated ratioDrug–drug interactionNumber of dispensings drug A × 1,000 (range)Number of dispensings drug B × 1,000 (range)Number of D–DIs counted (range)Average ratio (range)aDrug ADrug B1Erythromycin, clarithromycin, azithromycin, roxithromycinDigoxin440.8(0–2754)487.0(0–3064)3,993(0–41)1.39(0–18.52)2ItraconazoleDigoxin88.7(0–349)487.0(0–3064)245(0–7)0.45(0–21.69)3CiprofloxacinTheophylline105.4(0–769)100.9(0–756)944(0–14)6.39(0–534.38)4Miconazole oral gelAcenocoumarol, fenprocoumon44.6(0–233)608.2(5–3156)154(0–3)0.38(0–21.30)5ErythromycinCarbamazepine49.7(0–531)193.6(0–871)35(0–4)0.24(0–40.92)6Erythromycin, clarithromycin, azithromycinDisopyramide426.6(0–2754)9.4(0–151)61(0–4)–7Erythromycin, clarithromycinPimozide274.4(0–2004)57.4(0–394)70(0–15)0.46(0–46.12)8Propranolol, oxprenolol, pindololbeta2-mimetics, inhalation corticosteroids250.6(1–1075)2,546.9(27–10504)5,127(0–94)0.54(0–12.98)9Erythromycin, clarithromycinCisapride274.4(0–2004)127.5(0–821)586(0–11)1.16(0–40.45)10Itraconazole, fluconazole, ketoconazoleCisapride199.9(0–727)127.5(0–821)347(0–12)0.95(0–57.10)11Acenocoumarol, fenprocoumonAzapropazon608.2(5–3156)8.4(0–164)32(0–19)–aCalculated with Formula 1
Procedure
For each pharmacy, we calculated the dispensing-ratios for the eleven potential D–DIs with Formula 1. This formula was used because the risk of dispensing a D–DI between drug A and drug B is dependent on the number of times each drug is dispensed. The more drug A or drug B are dispensed, the higher the risk that these drugs are combined on the basis of chance alone. In case the dispensing of drug A is independent from the dispensing of drug B and the D–DI is never intervened, the ratio will on average be one. The number of times this ratio was above one was calculated, because a ratio above one might indicate that medication surveillance fails. In this calculation, there were 342 D–DIs between norfloxacin and theophylline, which were excluded in the analysis because the guidelines concerning the management of this D–DI were inconsistent. Two groups of pharmacies were selected, and the pharmacists were asked in August 2003 by the Inspectorate for Health Care (IHC) to fill in a questionnaire. The first group included pharmacies with a high risk of dispensing these D–DIs, while the second group consisted of a random sample from the remaining pharmacies. These groups were equal in size. The selection criteria are described in Fig. 1. The selection criterion for receiving a questionnaire (≥4 times a ratio >1) was chosen on pragmatic reasons to have enough power for statistical analysis with manageable numbers. A concept questionnaire was composed on basis of a literature search and interviews with experts [9]. The questions concerned process and structure characteristics of several quality aspects and those questions were selected that could discriminate between high and low quality pharmacies. Mostly questions with objective answers were included, for example about written instructions for technicians, filing of data, tuning of the software (which signals were shown and which not) and personnel. The concept questionnaire was tested in three pharmacies and some questions were amended on the basis of their comments. The final questionnaire contained 183 questions, divided into 12 subjects (Table 2). The questionnaire was accessible via the Internet. Pharmacies who had no access to the Internet received the questionnaire by post. Pharmacies who failed to fill in the questionnaire received reminders at regular intervals. A sample from both groups was visited by the IHC (Fig. 1). Also here, the selection criterion (≥5 times a ratio >1) was chosen to have enough power with manageable numbers. Thirty-seven questions from the questionnaire were selected and during the visits these questions were verified. The pharmacies were informed in advance that a selection of the pharmacies would be visited. The selected pharmacies were acquainted after completing the questionnaire. The visiting inspectors were blinded to the number of interacting drug combinations.
Fig. 1The selection of the pharmacies receiving a questionnaire and IHC visitTable 2The subjects and number of questions in the questionnaireChapterSubject (number of questions)General pharmacy dataOwnership of the pharmacy (1), cooperation with other pharmacies (1), cooperation with general practitioners (1), electronic submission of prescriptions (4)FacilitiesAlterations (2)Quality policySetting up and implementing a quality system (4), certification (2), attitude towards quality management (12)Quality measurementMeasurement of errors (2), complaints (1), patient satisfaction (2), interventions (3), and participation in mystery guest investigations (2)Receipt procedureNumber of personnel involved in dispensing a receipt (2), checks in dispensing a receipt (3)Medication surveillance—tuning softwareMedication surveillance system used (1), tuning of the system e.g. which signals are showed and which are regarded as irrelevant (55a), surveillance of pharmacy preparations (2)Medication surveillance—organisationThe way technicians are instructed to manage medication surveillance signals (5), the way this is supervised (2), number of interventions (1), use of resources (2), participation in courses (4), management of the D–DI between carbamazepine and erythromycin (5) and between Sulfamethoxazole/trimethoprim and Acenocoumarol (7)Medication surveillance—recording managementThe way the management of signals is recorded (4)Pharmacy preparationsThe way instructions for pharmacy preparations are recorded (1), the way pharmacy preparations are supervised (3), the number of pharmacy preparations (2), the policy regarding analysing pharmacy preparations (3)Personnel and workloadSubjective workload (3), absence through illness (1), number of receipts dispensed per technician (2), personnel and experience of personnel (18)Patient careInformation given to patients (6), information exchange with hospitals (4), participation in health care projects (4)Farmacotherapeutic consultation groupsParticipation in farmacotherapeutic consultation groups (3), agreements made (3)aAs the questions for the four systems (Pharmacom, Aposys, Euroned, others) differed, pharmacists had to fill in only a quarter of these questions
Formula 1:
Formula 2:
whereki,ab: number of dispensings of interacting drug combination AB in pharmacy i;ki,a: number of dispensings of drug A in pharmacy i;ki,b: number of dispensings of drug B in pharmacy i;Ni: total number of dispensed drugs known in the database in pharmacy i;Ntot: total number of dispensed drugs known in the database in all pharmacies
Statistical analysis
For each pharmacy, dispensing-ratios for the D–DIs, comparable to the standardised mortality ratio, were calculated using Formula 2. With this formula, we standardise for the total number of dispensings per pharmacy in the database. In case all pharmacies dispense the D–DIs in equal numbers, the ratio will be one for all pharmacies, and therefore the ratios have a better comparability. Pharmacies which have only a small number of dispensings in the database, will have extremely high numbers in case they dispense one or a small number of D–DIs. Therefore, the results were equalized with Bayesian statistics to prevent extreme ups and downs due to low numbers of dispensings [12]. The pharmacies were divided into two sets. One set was used for the analyses and contained two-third of the pharmacies, the other set was used for the validation of the results obtained in the analyses. In the univariate analysis, correlations were searched between the answers in the questionnaire and these ratios. Correlations are only given if in both sets a correlation was found (p < 0.01). In the multivariate analysis, models were composed using the analysis set, predicting the dispensing of the interacting drug combinations. The models were validated using the validation set. The number of questions was too large for the multivariate analysis, and only a limited number of questions were selected. From every chapter, those questions were selected that correlated with the other questions and that could discriminate between pharmacies.
Results
The database contained a total of 100,295,311 dispensings in the selected study period. One thousand one hundred and forty-two pharmacies were recorded in the database with 5,000 or more dispensings. The number of dispensings per pharmacy varied from 5,019 to 264,631. Because pharmacies receive reimbursements from several health care insurance companies and because not all health care insurance companies were included in the database, these numbers do not correspond with the total number of dispensings per pharmacy. The eleven potential D–DIs were dispensed 11,594 times. In 5%, more than one pharmacy was involved. As these D–DIs could not be assigned to a single pharmacy, they were excluded from further analyses. The number of dispensings and D–DIs are shown in Table 1. Disopyramide (D–DI number 6) and azapropazon (D–DI number 11) were not dispensed by 44% and 46% of the pharmacies, respectively. Therefore, a ratio could not be calculated for these pharmacies and these D–DIs were excluded from the analyses.
The number of times a ratio above one was found was calculated (Table 1) and pharmacies were selected as shown in Fig. 1. Two hundred and sixty-eight pharmacies were selected to receive a questionnaire and 74 pharmacies were selected for a visit by the IHC. For several reasons, such as recent visitations and duplications in the database, 12 pharmacies were excluded. Eventually, 256 pharmacies received a questionnaire and 62 pharmacies were selected for a visit. Two hundred and forty-six questionnaires were filled in (response rate 96.1%) and 58 (93.5%) pharmacies were visited after the questionnaire was completed. The judgements during the visits by the IHC were compared with the answers by the pharmacists. In 33 of the 37 verified questions, the IHCs judgement matched in more than 90% the answer of the pharmacist. Except four questions, the judgement by the IHC was equally more positive and more negative than the answers by the pharmacist.
In the univariate analysis, all combinations between the questions and D–DIs were searched for significant correlations. Two correlations were found with D–DI number 1 between macrolide antibiotics and digoxin (Table 3). Pharmacies, which are part of a health care centre dispensed this interacting drug combination more often than other pharmacies. A correlation with the type of medication surveillance system was also found. Pharmacies using the Euroned system dispensed this interacting drug combination more often, while pharmacies using the Pharmacom system dispensed this interacting drug combination less often.
Table 3Significant univariate correlations between the questionnaire and the number of dispensings of the D–DIs between macrolide antibiotics and digoxin (number 1)QuestionCorrelationSignificanceIs the pharmacy part of a health care centre? (1 yes, 2 no) (yes n = 18, no n = 228)−0.1650.009Which medication surveillance system is used in the pharmacy?Pharmacom (1 yes, 0 other) (n = 81)−0.2610.000Aposys (1 yes, 0 other) (n = 62)0.0880.170Euroned (1 yes, 0 other) (n = 89)0.1970.002
For the multivariate analysis, 32 variables were selected, representative of the whole range of questions. These variables were used in the analysis-set to compose models. The adjusted explained variance ranged from 2.6% to 28.9% (Table 4). The model explaining the D–DI between macrolide antibiotics and digoxin had by far the highest adjusted explained variance. The models were validated in the validation-set, calculating the unexplained variance (Table 4). The six variables in this model explaining the D–DI between macrolide antibiotics and digoxin are shown in Table 5.
Table 4Predictability of the models composed in the multivariate analysisD–DIAdjusted explained variance (R2) (analysis set)Unexplained variancea (validation set)128.90.61212.8−0.22317.331.547.0−0.18514.4−0.4176.56.4816.10.68914.0−0.43102.60.90aAn unexplained variance of zero means that the predictability found in the validation set equals the predictability in the analysis-set. The higher the unexplained variance, the worse the predictability in the validation-set compared to the analysis-setTable 5The questions in the multivariate model predicting the dispensing of the D–DI between macrolide antibiotics and digoxin (number 1)Variable:Answer (coding)Direction coefficientConstant3.3679Is the pharmacy part of a health care centre? (yes n = 18, no n = 228)Yes (0) versus no (1)−2.2749Co-trimoxazole—acenocoumarol: no appointments were made with the GPs. The drug will be dispensed.Option 1 ‘with all GPs’ (1) versus other option (0) (n = 11)ReferenceEight options of choice option 1 ‘with all GPs’ and optionOption 2 (1) versus other option (0) (n = 10)1.0308Eight ‘with no GPs’Option 3 (1) versus other option (0) (n = 4)0.3788Option 4 (1) versus other option (0) (n = 4)−0.4542Option 5 (1) versus other option (0) (n = 3)0.9026Option 6 (1) versus other option (0) (n = 2)−0.5100Option 7 (1) versus other option (0) (n = 4)−0.1912Option 8 ‘with no GPs’ (1) versus other option (1) (n = 202)0.0886Are separate signal texts in the medication surveillance program adjusted to the situation in the pharmacy? (yes n = 72, no n = 165)Yes (0) versus no (1)0.1793Is the management of signals traceably recorded on the receipt? (yes n = 211, no n = 35)Yes (0) on the receipt, no not on the receipt (1)0.2691The supervision on management of signals takes place on the basis of signal lists (yes n = 158, no n = 86)Yes (0) on the basis of signal lists, no (1) not on the basis of signal lists0.0723How many receipts are dispensed per year divided by the number of fte technicians<10–4
Discussion
In this study, we investigated determinants for the dispensing of 11 undesirable interacting drug combinations. In general, our results are in line with the expectation that the medication surveillance system plays an important role in medication surveillance. Although the 11 potential D–DIs were counted 11,594 times which suggests that a considerable number of patients is exposed to potential and avoidable adverse patient outcomes, these results should be judged against a background of approximately 100 million dispensings. It is possible that in these cases due to particular circumstances any other option, such as substituting or not dispensing one of the drugs, is a less favourable choice than dispensing the D–DI. In 5% of the total number of D–DIs more than one pharmacy was involved, indicating the importance of communication. For the D–DI between macrolide antibiotics and digoxin, two determinants were found. Although the type of medication surveillance system was a determinant, this does not mean that the differences are determined by the quality of the system itself because they may also correlate with the attitude of the pharmacists using the systems. The three medications surveillance systems differ in the extent to which communication with other healthcare providers is possible and developments were made in recent years. The Pharmacom system has the most advanced communication possibilities and compared to the other systems, new developments to the Euroned system were modest. Unexpectedly, pharmacies part of a health care centre dispensed this D–DI more often than other pharmacies. In health care centres, the communication lines between pharmacists and general practitioners are much shorter, suggesting that intervening undesirable D–DIs will be easier. Possibly, pharmacies which are part of a health care centre oppose the opinions from the general practitioners less often, to avoid harming the cooperation within the health care centre but, of course, there may be several other reasons.
For the other eight assessed D–DIs no determinants were found in the univariate analysis, neither did the models in the multivariate analysis have a good predictability. A possible explanation is that the quality of medication surveillance in community pharmacies in the Netherlands is high. Therefore, the number of pharmacies dispensing high-risk D–DIs seems to be small.
Our study has some potential limitations. First, because we used strict inclusion criteria to prevent false-positive results, it is likely that the number of dispensings of undesirable interacting drug combinations in this study is an underestimation and it is possible that important determinants were not recognized or difficult to assess. In the univariate analyses, only those questions are given which had a significant (p < 0.01) correlation in two independent sets. Although we included 183 questions and nine D–DIs in the univariate analysis, the possibility of including a significant correlation by chance was small (on average 0.16 question). Second, the reimbursement data from eight health care insurance companies were used. In the Netherlands, these companies work mostly regionally. It is nevertheless not to be expected that the determinants of dispensing interacting drugs differ per region or that pharmacies differ in their management of D–DIs between patients of different health care insurance companies. Third, from all potential D–DIs, only 11 (but highly clinically relevant ones) were selected for this study. According to the Dutch guidelines, for all 11 combinations the dispensing of an alternative was strongly advised as a good alternative was available. Nevertheless, it is possible that these dispensings were not an error because any other option was not possible. For example, when a patient is hypersensitive to the alternative drug recommended in the guidelines or when the alternative drug is not effective. In these cases, the benefit of both drug therapies should be weighed against the potential risks of the D–DI. The potential risks can partly be avoided by taking appropriate measures such as monitoring of drug levels. In this study, we could not retrieve why the pharmacist had dispensed the interacting drug combination, and whether the dispensing was erroneous or not.
Fourth, the questionnaire was composed on the basis of a literature search and interviews with experts. It is possible that not all characteristics correlating with the dispensing of undesirable interacting drug combinations were disclosed, such as differences in population characteristics between pharmacies. For example, pharmacies with an elderly population using more drugs simultaneously have a higher risk of dispensing interacting drug combinations than pharmacies with a younger population. Also, it is possible that in areas with many general practitioners who use a medication surveillance system for prescribing, the background chance of a D–DI is much smaller. Fifth, it is possible that the differences between pharmacies were too small compared with the power of this study to distinguish determinants.
All associations found in this study were directly related to the management of signals. In our questionnaire, we also included other topics, such as pharmacy preparations and patient care. Future research should focus on the management of a larger variety of signals than the ones in our study and on how D–DI associated dispensing could be further reduced.
Conclusion
In conclusion, both medication surveillance systems and being part of a health care centre may play an important role in the management of D–DIs and the avoidance of adverse patient outcomes. Pharmacies in a healthcare centre dispensed D–DIs more often. For most D–DIs, no determinants were found possibly indicating that the quality of medication surveillance in the Netherlands is high.
|
[
"drug–drug interaction",
"pharmacy",
"netherlands",
"drug interaction",
"practice management"
] |
[
"P",
"P",
"P",
"R",
"R"
] |
Antonie_Van_Leeuwenhoek-4-1-2268727
|
Control and possible applications of a novel carrot-spoilage basidiomycete, Fibulorhizoctonia psychrophila
|
A novel cold-tolerant fungus, Fibulorhizoctonia psychrophila, was isolated from a refrigerated carrot storage facility and identified as an anamorph of Athelia, often classified in Rhizoctonia s.l. Growth of this fungus was observed between 0 and 20°C with an optimum at 9–12°C, while incubation of mycelium grown at 15–32°C resulted in absence of growth even after the fungus was transferred back to 15°C. Growth was inhibited in the presence of the antifungals sorbic acid or natamycin, in particular when the fungus was incubated at 18°C. F. psychrophila produces polysaccharide degrading enzymes that, when compared to enzymes from the ascomycete fungus Aspergillus niger, retain a larger proportion of their activity at lower temperatures. This indicates that F. psychrophila could be used as a source for novel industrial enzymes that are active at 4–15°C.
Introduction
Fungi occupy every natural environment on earth as well as man-made indoor environments. In fact, some fungi have ecological requirements and amplitudes that are ideally suited to some artificial (human made) environments. Some of these fungi create large problems with respect to human health (e.g. Aspergillus, Candida) or to food quality (e.g. Penicillium). Among standard methodologies to prevent food spoilage are the use of preservatives and refrigeration of food. The latter method, however, is not effective when cold-tolerant spoilage fungi are present. Cold-tolerant fungi are commonly found in nature and comprise species of different branches of the fungal kingdom, such as Geomyces, Leucosporidium (Panikov and Sivoza 2007), Cystofilobasidium and Mrakia (Nagakawa et al. 2004).
Cell walls form the majority of the plant biomass and consist mainly of polysaccharides. As fungi are not able to import polymeric compounds into their cells, they rely on extracellular enzyme systems for the degradation of polysaccharides into monosaccharides and short oligosaccharides. Hydrolytic enzymes acting on polysaccharides are commonly produced by fungi. These enzymes have many applications in the food and feed, paper and pulp, textile, and pharmaceutical industries and have therefore been the subject of many studies (de Vries and Visser 2001; de Vries 2003). Most of these studies and applications involve enzymes of saprophytic ascomycete fungi (mainly species from Aspergillus and Trichoderma) that have optimal growth temperatures between 30 and 37°C. Consequently, most of these enzymes have optimum activities between 30 and 50°C and their activity rapidly decrease at lower temperatures (de Vries and Visser 2001). Efficient use of these enzymes in industrial applications is therefore often only possible by incubation above 30°C. This not only increases production costs due to the energy required to reach this temperature, but can also cause spoilage problems, especially with respect to food and feed applications.
One of the most commonly known cold-tolerant spoilage fungi is the basidiomycete Rhizoctonia carotae (Adams and Kropp 1996; Jensen 1969; Jones and Aldwinckle 1990; Punja 1987), which is found in storage facilities of several vegetables such as carrot. Although the optimum growth-temperature of this fungus is between 15 and 20°C (Punja 1987), this fungus causes crater rot characterized by sunken lesions and abundant mycelial growth on carrots stored at temperatures between 1 and 4°C. The anamorph genus Fibulorhizoctonia (Adams and Kropp 1996) (as Fibularhizoctonia) was proposed to accommodate Rhizoctonia carotae and R. centrifuga Lév (Rader 1948). The teleomorphs of these species belong to Athelia Pers., while those of the Rhizoctonia solani-complex belong to Thanatephorus Donk, currently placed in a different order (Hibbett et al. 2007). The teleomorph of F. carotae was identified as Athelia arachnoidea (Berk.) Jülich (Adams and Kropp 1996). However, there is some controversion considering the scope of this species, as the name has been used for litter decomposing specimens and lichen parasites. Arvidsson (1976) concluded that the use of the name A. arachnoidea should be confined to the lichen parasite, while the material of Adams and Kropp (1996) contained only litter decomposers.
Other basidiomycete species reported from cold storage are: Corticium centrifugum (Bielenin 1986; Stalpers and Loerakker 1984; Weresub and Illman 1980), from stored apples and pears, causing fisheye rot. Although the name C. centrifugum is connected with Athelia, the species concerned differs by having constant clamps at the septa, narrower hyphae and no sclerotia, and the production of a basidiome in culture, which is now known as Butlerelfia eustacei (Weresub and Illman 1980). This species is known from Europe and North America.
In this paper we describe a new carrot-spoilage fungus, Fibulorhizoctonia psychrophila, which is a predominant species detected in refrigerated storage facilities for carrots in The Netherlands. We have analysed the cold-tolerance of this fungus and its sensitivity to commonly used fungicides. In addition, we have analysed F. psychrophila for production of plant polysaccharide degrading enzymes to assess its potential as a source of cold-active enzymes.
Materials and methods
Strains and growth conditions
F. psychrophila CBS 109695 (IMI 395943) was isolated as a mycelial sample by H.A.B. Wösten from a wooden crate containing Daucus carota in a refrigerated storage facility (4°C) in Bant, The Netherlands in 2002. Athelia arachnoidea (Berkeley) Jülich CBS 418.72 was isolated from fallen leaf-litter of Populus sp. in The Netherlands. The type strain of Fibulorhizoctonia carotae (Rader) G.C. Adams & Kropp CBS 464.48 and Aspergillus niger N402 were described previously (Adams and Kropp 1996; Bos et al. 1988).
F. psychrophila CBS 109695 and A. niger N402 were routinely propagated on malt extract agar and minimal medium (de Vries et al. 2004), respectively. For growth of F. psychrophila on carrot, potato and onion, these vegetables were ground using a coffee grinder. The vegetable pulp was used at a concentration of 10% in 1.5% agar plates in water. Liquid cultures of F. psychrophila and A. niger were performed in Schizophyllum commune minimal medium (Dons et al. 1979) and Aspergillus minimal medium (de Vries et al. 2004), respectively, using 1% of a crude arabinoxylan preparation obtained from wheat after extraction of starch and proteins as the substrate. Plate cultures to determine the influence of several fungicides on growth of F. psychrophila were performed on malt extract agar plates. Comparison of the optimal growth temperature for F. psychrophila, Athelia arachnoidea and F. carotae was performed on malt extract agar and cherry decoction agar at temperatures of 0–27°C with intervals of 3°C.
Powdered wood was obtained by grinding wood shavings of a storage crate in a coffee grinder until a fine powder was obtained. Of this powder, 1 g was added to 1.5 g agar and 100 ml water and autoclaved to prepare solid media. The crude arabinoxylan preparation was a gift from Latenstein (Nijmegen, The Netherlands) and is in fact a waste stream of a protein and starch extraction process from wheat.
Molecular biology methods
Genomic DNA of CBS 109695 was extracted using the FastDNA kit (Bio 101 Systems, Q-Biogene). The 5.8S gene and flanking ITS1 and ITS2 were amplified using the primers ITS1 and ITS4 (White et al. 1990) and the sequence was deposited at genbank (Acc. Nr. EF492880). The sequence was compared with ITS sequences from F. carotae (U85789) and A. arachnoidea (U85791).
Enzyme assays
F. psychrophila and Aspergillus niger were grown at 15 and 30°C, respectively in liquid medium containing a crude arabinoxylan preparation. Culture filtrate was harvested over a nylon filter after 2 days of growth and analysed for enzyme activities at 4, 15 and 30°C. Enzyme activity was determined using p-nitrophenyl-β-d-galactopyranoside, p-nitrophenyl-β-d-glucopyranoside, p-nitrophenyl-α-d-galactopyranoside, and p-nitrophenyl-β-d-xylopyranoside (Sigma) as a substrate. A mixture was made consisting of 10 μl culture filtrate, 50 μl 50 mM sodium acetate (pH 5.0) and 30 μl sterile MiliQ water. The reactions were performed in triplicate and were started by the addition of 10 μl of 0.1% stock of the p-nitrophenyl-linked substrate and incubated for 2 h at 25°C, unless stated otherwise. The reactions were stopped by the addition of 100 μl 0.25 M Na2CO3 and measured at 405 nm in a micro plate reader (model 550, Bio-RAD). The amount of free p-nitrophenol was calculated using a calibration curve.
Results and discussion
Taxonomy
Fibulorhizoctonia psychrophila Stalpers & de Vries spec. nov. (Fig. 1).
Fig. 1Colony morphology of Fibulorhizoctonia carotae (464.84), Athelia arachnoidea (418.72) and Fibulorhizoctonia psychrophila (109659) after 2 weeks. (a) 0°C; (b) 6°C; (c) 18°C
Mycelium ad 9°C cotoneum, albidum, ad 18°C cremeum, avellaneum vel brunneum, velutinum vel crustosum. Liquor exsudatus brunneus. Sclerotia presentia vel absentia, irregularia, in statu maturitate subfusca a liquore exsiccata. Hyphae hyalinae vel subfuscae, (2.5–)3–6 μm. Septa fibulatae vel afibulatae. Cellae sclerotiorum doliformes, tenui-tunicatae vel crasse-tunicatae, hyalinae vel fuscae. Fungus psychrophilus, ad temperaturas −3° ad 21°C crescens.
Typus: CBS 109695
Colonies at −3 to 9°C are cottony, rather high and reaching the lid of the Petri dish, white. Margin raised, rather dense. At increasing temperatures the mycelium grows less high and becomes nearly velvety at 18°C and nearly crustose at 21°C. At 12° the centre of the colony is cream-coloured, at 15°C becoming Light Cinnamon Drab to Avellaneous (Munsell 7.5YR6/2, 10YR6/2, 10 YR6/3). At these temperatures exudate droplets are not present. At 18°C the colony is low, nearly velvety, Ochraceous Tawny (7.5YR5/8) caused by exudate drops and at 21°C there is hardly any growth; the mycelium is crustose and Cinnamon Brown to Prouts Brown (5YR4/3, 5YR4/4), and finally Mummy Brown. Sclerotia are generally not produced within 6 weeks, but have occurred in a tube; they start as white semiglobose pustules, generally up to 1 mm diameter, aggregated, often fusing, and producing a brown exudate, which dries in, leaving a dark brown sclerotium, up to 5 mm diam, finally Mummy Brown.
Marginal hyphae at −3 to 3°C are irregular, rather long-celled, with swellings and granular contents, thin-walled, (2.5–)3–6 μm wide, swellings up to 12 μm wide. Clamps present, not abundant. Branching at various angles, sometimes a constriction is present at the base of the side branch. At 6°C and above the hyphae are regular, thin- to firm-walled, 2.5–6(–7.5) μm wide, contents hyaline to pale brownish, with abundant clamps, especially at the wider hyphae, but the narrower hyphae (2.5–3.4–4 μm) generally have clampless septa. Cell wall hyaline, but irregular brownish encrustations present at some hyphae, probably dried exudate.
Aerial mycelium regular, thin- to firm-walled, (2.5–)3.2–6.5 μm wide, with hyaline to brownish contents. Most septa with clamps; clamps regular, but sometimes of the medallion type.
Sclerotia rarely produced, uniform, not forming a distinct cortex, consisting of thin- to thick-walled hyphae, usually consisting of slightly elongated swollen cells with granular contents, 4–10 μm wide, not unlike the (thin-walled) advancing hyphae at lower temperatures.Cardinal temperatures: minimum below −3°C, optimum range between 9 and 12°C, maximum 20°C.Holotype: CBS 109695 (IMI 395943), also preserved dried in herb. CBS.Mycobank number: MB501325.The sterile, sclerotium producing mycelia are traditionally classified in the artificial genus Rhizoctonia, which is currently restricted to basidiomycetous fungi and comprises anamorphs of various, not closely related genera such as Tulasnella, the Ceratobasidium–Thanatephorus complex and Athelia (Stalpers and Anderson 1996). These groups have been elevated to the genus level as Epulorhiza, Rhizoctonia s.str. and Fibulorhizoctonia. Although teleomorphs have not been observed in F. carotae and F. psychrophila, morphological, physiological and molecular characters indicate that F. psychrophila, F. carotae and Athelia arachnoidea are closely related and belong to Athelia, a corticioid genus, currently classified in the Atheliales (Rader 1948).
F. carotae has been described from cold stored carrots. It differs from F. psychrophila in growing between 18 and 27°C (Fig. 2), abundant production of sclerotia, and colour of the mat above 15°C. A. arachnoidea has been described from both lichens and leaf litter. It differs from F. psychrophila in its growth at higher temperatures (Fig. 2), less abundant aerial mycelium, and pale colour at higher temperatures. ITS sequencing of F. psychrophila revealed 95% identity to the ITS sequence of A. arachnoidea and A. carotae (data not shown), indicating that F. psychrophila is very closely related to these two species.
Fig. 2Growth comparison at different temperatures of Fibulorhizoctonia psychrophila, Fibulorhizoctonia carotae and Athelia arachnoidea. The means of duplicate values are represented
Both morphological and molecular characters indicate without doubt, that F. psychrophila, F. carotae and Athelia arachnoidea are closely related.
Temperature-tolerance and growth of F. psychrophila
F. psychrophila was inoculated on malt extract agar plates and incubated between 0 and 27°C. Growth was monitored by measuring the colony diameter. Optimal growth occurred at 9–12°C (Fig. 2) and growth was absent above 20°C. When plates incubated for 3 days at 20 and 25°C were placed back at 15°C the fungus re-initiated growth, but incubation for 3 days at 32°C prevented growth of F. psychrophila after the plates were transferred to 15°C (data not shown). Optimal growth of F. carotae and A. arachnoidea was at between 18 and 21°C (Fig. 2). The strong decrease in growth of F. psychrophila when the incubation temperature is raised above 15°C, suggests an even stronger adaptation of this fungus to cold biotopes.
Growth of F. psychrophila was tested on three vegetable crops that are commonly stored under refrigerated conditions (carrot, potato and onion) in comparison to agar plates without carbon source. Water agar alone already permits growth of F. psychrophila, but growth is significantly improved in the presence of carrot (data not shown). The presence of potato only resulted in a small increase in growth compared to water agar, but onion reduced growth of F. psychrophila. These data demonstrate a preference of F. psychrophila for carrot as a substrate. The reduced growth on onion is likely caused by the high levels of phenolic compounds in onions.
Inhibition of growth of F. psychrophila using antifungals
Sorbic acid and natamycin are compounds commonly used for the inhibition of growth of spoilage fungi and the effectiveness of these compounds was tested on F. psychrophila on malt- and water agar plates with powdered wood. F. psychrophila is believed to survive on the crates used for the storage of carrots and therefore causes repeated spoilage when these crates are re-used. The presence of 250 mM sorbic acid resulted in a significant reduction in growth on malt extract agar, but not on powdered wood at 4°C (Fig. 3a). However, higher levels of sorbic acid (625 and 1,250 mM) prevented growth of F. psychrophila on both substrates. In light of the high concentrations of sorbic acid required to inhibit growth, the expected effectiveness of this compound in preventing spoilage by F. psychrophila is limited.
Fig. 3Influence of antifungals on the growth of Fibulorhizoctonia psychrophila grown on malt extract agar (MA) or powdered wood from storages crates (HW). (a) sorbic acid, 18°C; (b) natamycin, 4°C; (c) natamycin, 18°C
Natamycin was more effective in preventing growth of F. psychrophila, resulting in a significant decrease in growth at 2 g/l at 4°C. Increasing levels of natamycin reduced growth of the fungus at 4°C similarly on both malt extract agar and powdered wood, although a significant difference was observed at 50 and 100 g/l (Fig. 3b). However, at 100 g/l natamycin, some residual growth could still be observed for powdered wood. A stronger inhibition with natamycin was observed when the incubations occurred at 18°C instead of 4°C, resulting in nearly complete inhibition of growth on both media at 0.5 g/l natamycin (Fig. 3c). The results described here indicate that spraying wooden crates with natamycin and incubating them at 18°C or higher likely reduces spoilage F. psychrophila.
F. psychrophila produces cold-active hydrolytic enzymes during growth on polysaccharides
Both carrots and powdered wood consist largely of polysaccharides. As these are the main carbon sources for F. psychrophila during growth in the storage facilities, it can be expected that the fungus produces polysaccharide degrading enzymes that are active at low temperatures. To study whether these enzymes have in fact a higher relative activity at low temperatures than those currently used in industrial applications, we compared enzyme activities from F. psychrophila to those from Aspergillus niger. A. niger is one of the most commonly used fungi for the production of industrial enzyme preparations and produces a wide range of polysaccharide degrading enzymes (de Vries and Visser 2001). However, these enzymes are mainly active at higher temperatures.
For A. niger, enzyme activities dropped to 30–50% when incubation at 15°C was compared to 30°C (Fig. 4a). At 4°C, only 10–20% of the 30°C-activity was observed. For F. psychrophila only a 5–35% drop was detected when comparing 30°C to 15°C and 30–60% of the activity at 30°C was still observed at 4°C (Fig. 4b). The strongest difference was observed for α-galactosidase and β-glucosidase, where more than 90% of the activity was still detected at 15°C for F. psychrophila, while for A. niger only 35–45% of the activity was detected at this temperature.
Fig. 4Relative hydrolase activities in culture filtrate of Fibulorhizoctonia psychrophila and Aspergillus at different temperatures. (a) A. niger culture filtrate; (b) F. psychrophila culture filtrate. AGL = α-galactosidase, LAC = β-galactosidase, BGL = β-glucosidase, BXL = β-xylosidase. Activity at 30° is set at 100%
F. psychrophila and other cold-tolerant fungi are potential sources for enzymes with high activity at low temperatures. This idea is supported by studies with cold-tolerant yeast and fungi, which were shown to produce cold-active pectinases (Nagakawa et al. 2004, 2005a, b), although in these studies a direct comparison with industrially used enzymes was not made. Cold-active enzymes can be important for many applications, for example in detergents or for the removal of lactose from milk for lactose-intolerant people.
|
[
"carrot-spoilage",
"fibulorhizoctonia psychrophila",
"cold-tolerant fungus",
"cold-active enzymes"
] |
[
"P",
"P",
"P",
"P"
] |
Clin_Rheumatol-3-1-1820749
|
Central sensitization: a biopsychosocial explanation for chronic widespread pain in patients with fibromyalgia and chronic fatigue syndrome
|
In addition to the debilitating fatigue, the majority of patients with chronic fatigue syndrome (CFS) experience chronic widespread pain. These pain complaints show the greatest overlap between CFS and fibromyalgia (FM). Although the literature provides evidence for central sensitization as cause for the musculoskeletal pain in FM, in CFS this evidence is currently lacking, despite the observed similarities in both diseases. The knowledge concerning the physiological mechanism of central sensitization, the pathophysiology and the pain processing in FM, and the knowledge on the pathophysiology of CFS lead to the hypothesis that central sensitization is also responsible for the sustaining pain complaints in CFS. This hypothesis is based on the hyperalgesia and allodynia reported in CFS, on the elevated concentrations of nitric oxide presented in the blood of CFS patients, on the typical personality styles seen in CFS and on the brain abnormalities shown on brain images. To examine the present hypothesis more research is required. Further investigations could use similar protocols to those already used in studies on pain in FM like, for example, studies on temporal summation, spatial summation, the role of psychosocial aspects in chronic pain, etc.
Introduction
Chronic fatigue syndrome (CFS), as defined by the Centers for Disease Control and Prevention (CDCP), is a complex illness characterized by prolonged debilitating fatigue and multiple non-specific symptoms including headaches, recurrent sore throats, fever, muscle and joint pain, and neurocognitive complaints [1, 2]. In addition to the chronic fatigue, widespread and persistent pain is common in individuals with CFS [3–5]. A population-based study revealed that 94% of the persons diagnosed with CFS report muscle aches and pain and 84% report joint pain [6]. Nishikai et al. [7] reported muscle pain in 85 CFS patients of 114 patients (74.6%). Seventy-four patients (64.9%) complained of arthralgia. In another study, 24 of 44 patients suffered from chronic widespread pain [8]. Chronic fatigue accompanied by chronic musculoskeletal impairments such as myalgias and arthralgias could be considered an important subclass of CFS [9]. Evidence supportive of the clinical importance of widespread pain in CFS has been provided [10]: chronic pain accounts for up to 34% of the CFS patients’ self-reported activity limitations and participation restrictions. Chronic pain is more disabling than chronic fatigue [10].
Given these facts, it may be surprising that the etiology of these pain complaints has not been studied extensively in patients with CFS. The systematic literature review by Meeus et al. [11] shows that only little progress has been made in understanding chronic widespread pain in patients with CFS. A few hypotheses have been proposed, but they have not been studied in depth or relatively little work has been performed to test these hypotheses. In contrast, a large body of scientific literature regarding the etiology of chronic pain complaints in fibromyalgia (FM) is currently available. The diagnosis of FM is based on the 1990 American College of Rheumatology criteria. Following these criteria, FM patients present with 11 of 18 positive tender points and with widespread pain [12]. Validity of both the definition for CFS and FM has been shown [12, 13].
Especially, investigations focusing on the phenomenon “central sensitization” are presented in force in FM. In CFS, the theory of central sensitization has only been suggested, to our knowledge [14]. Given the great overlap between CFS and FM [15] and given the dearth of studies focusing on the explanation for the chronic widespread pain in patients with CFS, it would be interesting to propose a theoretical model for the chronic pain in CFS based on the current knowledge of CFS and on the evidence for central sensitization in FM, giving rise to further research on that matter. Besides the knowledge on chronic pain in FM, it is necessary to gather knowledge on musculoskeletal pain in CFS.
The syndromes may overlap, but despite the similarities between the two syndromes, there are also differences. For example, immunological dysregulations such as the abnormal 2–5A synthetase/RNase L pathway [16] have been revealed in CFS but have never been detected in FM patients. Furthermore, there is not yet any good evidence for similar pain mechanisms in CFS and FM. Some authors already found evidence suggesting differences in pain processing. For example, patterns of functional brain activity in patients with FM are quite different from those in patients with CFS.
Patients with CFS, relative to controls, showed significantly lower blood perfusion in the brain stem [17, 18]. Patients with FM exhibited significantly lower rCBF levels, during rest, in the thalamus and the caudate nucleus [19]. Furthermore, Substance P has been found to be elevated in CSF of FM patients [20] and not in patients with CFS [21]. Therefore, the knowledge on pain in FM cannot be applied on CFS patients without further study. Based on the similarities and differences between the two syndromes, further research on pain in CFS is advised to get an image of pain processing in the two diseases.
The goal of this article is to provide a rational basis for future investigations. First, the concept of central sensitisation as a cause of chronic pain will be explained. This theoretical background will then be applied to FM and an overview of the evidence for central sensitization in FM will follow. Finally, based on the theoretical background and the findings in FM, the hypothesis concerning central sensitization in CFS will be unfolded, supported with the present knowledge on CFS.
Central sensitization
Introduction Pain is a complex perception that is influenced by prior experience and by the context within which the noxious stimulus occurs; “nociception” is the physiologic response to tissue damage or prior tissue damage [22]. The definition of pain is endorsed by the International Association for the Study of Pain: “Pain is an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage” [23]. There are a host of physiologic mechanisms by which injuries lead to nociceptive responses and ultimately to pain [22]. However, not all nociceptive signals are perceived as pain and vice versa, not every pain sensation originates from nociception. Nevertheless, acute pain almost always originates from nociceptors in somatic or visceral tissue. Mainly two types of pain receptors are activated by nociceptive input. These include low-threshold nociceptors that are connected to fast conducting A-delta pain fibers, and high-threshold nociceptors that conduct impulses in slow (unmyelinated) C fibers. Within the dorsal horn of the spinal cord, these pain fibers synapse with spinal neurons via synaptic transmission. Many neurotransmitters (i.e., glutamate, substance P, etc.) are able to modulate the postsynaptic responses with further transmission to supraspinal sites (thalamus, anterior cingulated cortex, insular cortex, and somatosensory cortex) via the ascending pathways [22, 24, 25].The simplest form of plasticity in nervous systems is that repeated noxious stimulation may lead to habituation (decreased response) or sensitization (increased response) [26]. Prolonged or strong activity of dorsal horn neurons caused by repeated or sustained noxious stimulation may subsequently lead to increased neuronal responsiveness or central sensitization [25, 27]. Neuroplasticity and subsequent CNS sensitization include altered function of chemical, electrophysiological, and pharmacological systems [22, 28, 29]. These changes cause exaggerated perception of painful stimuli (hyperalgesia), a perception of innocuous stimuli as painful (allodynia) and may be involved in the generation of referred pain and hyperalgesia across multiple spinal segments [25, 30–33].While the exact mechanism by which the spinal cord becomes sensitized or in “hyperexcitable” state currently remains somewhat unknown, some contributing factors have been proposed.
Temporal summation or wind-up “Wind up” refers to a central spinal mechanism in which repetitive noxious stimulation results in a slow temporal summation that is experienced in humans as increased pain [34]. In 1965, animal experiments showed for the first time that repetitive C-fiber stimulation could result in a progressive increase of electrical discharges from the second-order neuron in the spinal cord [35]. This mechanism of pain amplification in the spinal cord is related to temporal summation of second pain or wind-up. Second pain, which is more dull and strongly related to chronic pain states, is transmitted through unmyelinated C fibers to dorsal horn nociceptive neurons. During the C-fibres transmitted stimuli, N-methyl-d-aspartate (NMDA) receptors of second-order neurons become activated.It is well-known that NMDA activation induces calcium entry into the dorsal horn neurons [36]. Calcium entry into sensory neurons in the dorsal horn induces activation of nitric oxide (NO) synthase, leading to the synthesis of NO [37]. NO can affect the nociceptor terminals and enhance the release of sensory neuropeptides (in particular, substance P) from presynaptic neurons, therefore contributing to the development of hyperalgesia and maintenance of central sensitization [38]. Substance P (SP) is an important nociceptive neurotransmitter. It lowers the threshold of synaptic excitability, resulting in the unmasking of normally silent interspinal synapses and the sensitization of second-order spinal neurons [39].Furthermore, SP can extend for long distances in the spinal cord and sensitize dorsal horn neurons at a distance from the initial input locus. This results in an expansion of receptive fields and the activation of wide dynamic neurons by non-nociceptive afferent impulses [24].Wind-up can be elicited in human patients if identical nociceptive stimuli are applied to the skin or muscles more often than once every 3 s. The resulting progressive increase of pain sensations represents wind-up and has been demonstrated to result from a central rather than a peripheral nervous system mechanism, because the input from C nociceptors has been shown to decline or stay the same with stimulus repetition [40].
Endogenous pain modulatory systems The presence of several pain inhibitory and facilitatory centers in the brainstem is well recognized. The dorsolateral funiculus appears to be a preferred pathway for descending pain inhibitory systems [41]. Experimental evidence for the existence of descending inhibitory pathways, and their connection with central sensitization, include the observations that bilateral lesions of the dorsolateral funiculus in the rat led to a significant decrease in latency for paw withdrawal to noxious stimulus [42]. Similarly, temporary spinal cord block (lidocaine) caused dorsal horn nociceptive neurons to expand their receptive fields and their responsiveness to afferent input [43]. In addition, selective chemical lesion of serotonergic inhibitory neurons in experimentally “inflamed” animals resulted in demonstrable behavioral “pain” hypersensitivity [42].The foregoing investigations suggest that disruption of one or more of the elements of the inhibitory system can result in, among other things, the equivalent of central sensitization [44]. One function of the descending inhibitory pathway is to “focus” the excitation of the dorsal horn neurons. The effect is to generate a more urgent, localized, and rapid pain signal by suppressing surrounding neuronal activity [45]. This role is attributed to the “diffuse noxious inhibitory controls” (DNIC) phenomenon [46]. According to this model, descending pathways effectively enhance the biologically valuable pain signal by reducing the level of irrelevant “noise” in the system.Facilitatory pathways leading from the brainstem have also been identified. There is now behavioral evidence that forebrain centers are capable of exerting powerful clinically significant influences on various nuclei of the brainstem, including the nuclei identified as the origin of the descending facilitatory pathway [44]. The activity in descending pathways is not constant but can be modulated, for example, by the level of vigilance or attention and by stress [47]. Brosschot [48] refers to it as cognitive emotional sensitization. Forebrain products such as cognitions, emotions, attention, and motivation have influence on the clinical pain experience [44].Dubner and Ren [41] rewarded subjects for responding to a randomly delivered transient tissue threatening peripheral stimulus. It was found that sensitization of second-order pain pathway neurons was directly related to the strength of attention. The evidence suggests that selective attention to relevant stimuli activated descending pain modulatory systems, turning the balance in favor of facilitation. The dominance of descending facilitation then led to sensitization of second-order neurons [44]. Behavioral variables such as attention to a potentially threatening stimulus result in sensitization of dorsal horns spinal cord neurons. Moreover, behavioral modulation associated with selective attention to a perceived threat utilizes the same forebrain and brainstem structures and mechanisms as are involved in the development, amplification, and maintenance of persistent pain after actual tissue damage and inflammation [41].Certain cognitive styles and personality traits have been associated with amplification of pain and its extension in the absence of tissue damage. These include somatization, catastrophizing, and hypervigilance [49–52]. Thus, via descending pathways behavioral and cognitive therapies might also effect synaptic transmission in the spinal cord and thereby have the capacity to prevent or reverse long-term changes of synaptic strength in pain pathways [47].
Evidence in FM
Exaggerated pain is common in whiplash and FM patients. Theoretically, peripheral mechanisms can account for the pain hypersensitivity. In FM, however, there is no evidence for peripheral sensitization as the cause of hyperalgesia, given the absence of real tissue damage. Peripheral sensitization is defined as a reduction in the threshold of nociceptive afferent receptors caused by a local change in the sensitivity of sensory fibers initiated by tissue damage [53]. Peripheral sensitization almost always depends on local inflammation, which may lead to decreased nociceptor thresholds. Despite extensive investigations, no tissue pathology, structural abnormalities, or evidence for a source of chronic stimulation of pain afferents have been detected in fibromyalgic patients [54].
Furthermore, FM pain lacks a distinct spatial localization [55]. This suggests that pathophysiological central mechanisms contribute to or are responsible for FM pain [56]. This has led to the hypothesis that the central nervous system is hyperexcitable in these patients. Central hypersensitivity could explain exaggerated pain in the presence of minimal and undetectable tissue damage, in that the nociceptive signal is amplified by the hyperexcitable neurons [57].
Pain measurements Studies in patients with chronic pain after whiplash injury and with FM have demonstrated exaggerated pain responses after sensory stimulation of healthy tissues [55, 58–63]. For example, at same level of thermal stimulation, FM patients perceived pain as 49 and 52% more intense than healthy controls and patients with low-back pain [61]. After-sensations at 15 s after heat taps were regarded as painful on 83% of FM patients, compared to 37% pain reports of healthy controls. The late after-sensations (2 min after heat stimulation) were rated as painful in 55 and 5% of FM and control subjects, respectively [55]. It was not mentioned in the results if all FM patients were subject to increased sensitivity, but there were always significant differences between the FM patients and the healthy controls.Although most comparisons were made to pain-free subjects, some researchers, like Julien et al. [61], even compared the pain responses of FM patients to other pain patients, such as patients with chronic low-back pain. They could also report significant differences. Despite the lack of research into the contribution of psychological factors in wind-up, these results suggest that input to central nociceptive pathways is abnormally processed in patients with FM. Intramuscular electrical stimulation has been used to assess the efficacy of temporal summation of painful muscle stimuli. Temporal stimulation was found to be more pronounced and to cause stronger pain and larger referred areas in FM patients compared to controls [60].The increased efficacy of temporal summation in FM has even so been reproduced with cutaneous heat stimulation [55] and with cold and heat taps [59]. Facilitated temporal summation in patients with pain suggests that the efficacy of central processing is increased (central sensitization) in these patients [64]. In addition, after-sensation was greater in magnitude, lasted longer, and was more frequently painful in patients with FM [55]. The prolonged decay together with the augmentation provides evidence for the presence of central sensitization [24, 25]. Immersion of the arm in circulating noxious cold water resulted in a 49% more intense pain in FM patients compared to healthy controls [61]. Further evidence of central sensitization in FM is seen by enlarged referred pain areas. Sörensen et al. [60] found that fibromyalgic patients experienced stronger pain and larger referred areas after intramuscular injection of hypertonic saline.Moreover, spatial summation effect during increase of the stimulation area was found only in patients with FM and not in healthy controls or in patients with chronic low-back pain [61]. This indicates that pain inhibitory systems are not optimally recruited in patients with FM. FM patients perceived pain at the same intensities and unpleasantness during the ascending and the descending sessions (fingertip to shoulder and shoulder to fingertip immersion). In healthy controls and patients with low-back pain, the noxious stimulation of a large surface area results in an activation of a large population of nociceptive afferents that induce endogenous inhibitory responses, resulting in a decreased response in the dorsal horn neurons, and subsequently leads to lower pain intensities afterward.Given that FM patients experience similar pain intensities in the descending session after noxious stimulation of the whole arm, this study clearly demonstrated that FM patients present with a lack of activation of endogenous inhibitory systems [61]. Secondly, Lautenbacher and Rollman [65] showed that tonic painful and non-painful thermal stimulation of the foot increased the pain thresholds to electrical stimulation applied to a non-tender point (inner forearm) in healthy controls but had no effect on patients with FM. Kosek and Hansson [66] found that tourniquet ischemic pain in an arm increased the pressure pain threshold in healthy controls but not in FM patients, suggesting once more a deficiency in the latter of a pain-inhibitory phenomenon termed “diffuse noxious inhibitory control” (DNIC) [66, 67]. Similarly, aerobic exercise has been shown to decrease wind-up pain in normal subjects but increased it in FM patients, suggesting the possibility of reduced endogenous analgesic mechanisms [68]. Also, isometric exercise (i.e., hand-grip exercise) resulted in increased thermal pain ratings and decreased pain thresholds, both ipsilateral and contralateral to the exercised extremity [69]. These investigations support a general hypothesis that FM reflects a disorder affecting modulation of pain sensitivity [66, 67]. In one such scenario, tonic DNIC is present in the normal situation, and its pathological absence results in the spontaneous pain and evoked pain sensitivity associated with FM [34].
Measurements of excitability The results of the above-mentioned studies, however, are based on pain reports of the patients and thus subjective in nature, and it was not clear whether this hypersensitivity was the result of central mechanisms or whether the hypersensitivity was the cause of hypervigilance. Banic et al. [57] could provide objective evidence by quantifying the minimal intensity of transcutaneous electrical stimulation of the sural nerve required to evoke flexion reflex in the biceps femoris. This study clearly demonstrates that spinal cord neurons are sensitized in chronic pain after whiplash and in patients with FM; because the stimuli were delivered at random time intervals and the latency of EMG response was measured, voluntary symptom amplification could be ruled out. Moreover, the electrical stimulation bypasses peripheral receptors.
Cognitive emotional sensitization Patients with FM or other pain disorders often receive the message that it is “all in their head”. One construct that has been hypothesized to explain the pain amplification in FM is that of hypervigilance [34]. This hypothesis of hypervigilance has been argued by McDermid et al. [70]. FM showed an increased aversiveness to non-painful stimuli such as loud noise. Also, Crombez et al. [71] reported significant correlations between hypervigilance and pain intensity in FM patients. Furthermore, exposure to stressful situations, including noise, lights, and weather, is known to exacerbate symptoms of FM, including pain [72].In addition, FM patients with catastrophic thoughts report increased pain intensities [73–75]. Hassett et al. [74] found catastrophizing (27% of the variance) and depression (30% of the variance) to be significant predictors of pain. Finally, kinesiophobia and fear of pain are related to pain severity in patients with FM [76].The foregoing relations between emotions or cognitions and reported pain severity support the hypothesis of cognitive emotional sensitization in FM.
Central abnormalities in FM FM patients differ from healthy persons in regional cerebral blood flow (rCBF) distribution in several brain structures involved in pain processing and pain modulation both at rest and during experimental pain induction. Patients with FM exhibited significantly lower rCBF levels, during rest, in the thalamus and the caudate nucleus [17]. Dysregulation of thalamic activity and hypoperfusion of the caudate nucleus may contribute to the abnormal pain modulation, given the results of previous investigations [77–79]. In addition to nociceptive transmission, the thalamus also plays an important role in pain modulation. Animal studies proved that thalamic stimulation induces analgesia and lesions of the thalamus cause hyperalgesia [77, 78].Similarly, stimulating the caudate nucleus decreases pain behavior [79]. During pain induction in patients with FM, the absence of significant thalamic activation and a bilateral activation of the somatosensory cortices and the right anterior cingulated cortex was seen [80]. The patterns of cortical activation may be characteristic of patients with allodynia, and pain-induced activation of the right anterior cingulated cortex is associated with the use of maladaptive coping strategies [81]. Patients recruited by Bradley et al. [80] indeed reported significantly greater use of emotion-focused coping strategies (e.g., praying and hoping) during pressure stimulation. In addition, increased right frontal brain activity seemed related to increased pain sensitivity [82].Furthermore, FM patients are characterized by relatively high cerebrospinal fluid levels of substance P [20]. This finding indicates that abnormal brain activity in persons with FM is associated with abnormal CSF levels of a neuropeptide involved in pain transmission [80]. In addition, it is shown that a subgroup of FM patients present with mycoplasma infections [83]. Infection triggers the release of the pro-inflammatory cytokine interleukin-1β, which is known to play a major role in inducing cyclooxygenase-2 (COX-2) and prostaglandin E2 expression in the central nervous system [84].Peripheral nerve terminals can be sensitized by elevated COX-2 amounts and prostaglandin E2. Peripheral infections are even able to activate spinal cord glia, leading to the release of NO and proinflammatory cytokines, enhancing the pain response [85]. Physiological symptoms, such as pain, can be explained by these mechanisms (“sickness response”).
Central sensitization in CFS?
Direct evidence supporting the central sensitization hypothesis in CFS patients is currently lacking. But the present knowledge concerning CFS is suggestive of a central process similar to that seen in FM, given the great overlap between the two diseases and the observed similarities.
First, lower pain thresholds at different sites (hyperalgesia) are reported in patients with CFS, compared to controls [86, 87]. Similar to FM, the lack of peripheral tissue damage and the lack of a distinct localization of the pain complaints are suggestive of a central abnormality responsible for the chronic widespread pain. To our knowledge, there are no investigations that focused on abnormal wind-up, temporal summation, or spatial summation in CFS, to collect evidence for central sensitization in CFS. Yet, evidence of a dysfunctional central anti-nociceptive mechanism in CFS has been proved by Whiteside et al. [88]. They reported a decrease of pain threshold in patients with CFS after graded exercise, while healthy controls present an increased pain threshold. These findings are similar to those of Vierck et al. [68] in FM patients.
Secondly, the frequently reported opportunistic infections [89, 90] may lead to the “sickness response” and complies with the central sensitization hypothesis, as explained above. Indeed, elevated NO levels have been documented in CFS patients [91]. As earlier mentioned, NO plays an important role in the history of central sensitization and, therefore, central sensitization caused by NO would be likely in patients with CFS. The release of excessive amounts of Substance P in the CSF, however, could not be documented in patients with CFS, in contrast to FM patients [21]. The CFS patients in this study, however, did not experience chronic widespread pain, and it is not clear if they fulfilled the 1994 CDC criteria [2]. Given that 70% of the CFS patients do fulfill the ACR criteria for widespread pain [12], it is striking that the patients included in the investigation of Evengard et al. [21] (focused on the source of pain) did not report these pain complaints. Further research on that matter should analyze the CSF of CFS patients (fulfilling the Fukuda criteria) suffering from chronic widespread pain.
A third important argument in the central sensitization theory for CFS concerns the cognitive, psychological, and behavioral changes in patients diagnosed with CFS. CFS patients often present with depression [6, 92], catastrophizing [93, 94], somatization [95, 96], and kinesiophobia or fear avoidance [97–99]. In CFS, it is known that these psychosocial aspects are important factors in maintaining the complaints of CFS. These cognitions and emotions are able to influence pain perception via modulation of the descending pathways [44]. Furthermore, these “cognitive styles” and “personality traits” have earlier been mentioned to be associated with the amplification and the extension of pain [49–52].
Finally, brain imaging already provided evidence for altered brain activity in CFS. Patterns of functional brain activity in patients with FM are quite different from those in patients with CFS. Patients with CFS, relative to controls, showed significantly lower blood perfusion in the brain stem [17, 18]. Patients with FM exhibited significantly lower rCBF levels, during rest, in the thalamus and the caudate nucleus [19]. However, the areas are different in FM and in CFS; both the affected areas could be related to central pain processing. Low brain stem rCBF levels may contribute to abnormal function of the locus ceruleus in patients with CFS. The locus ceruleus is involved in controlling descending anti-nociceptive pathways from the brain to the spinal dorsal horn [56]. In consequence, pain experiences of patients with CFS may be related to low resting state levels of functional activity in the brain stem [80].
Conclusion
Chronic widespread pain can be the consequence of central sensitization. Central sensitization is known as an increased central neuronal responsiveness and causes hyperalgesia, allodynia, and referred pain and hyperalgesia across multiple spinal segments, leading to chronic widespread pain. Possible triggers for sensitization of the spinal cord have extensively been discussed, such as wind-up or temporal summation, dysregulated descending inhibitory pathways, and upregulated facilitatory modulation. Wind-up or temporal summation is the result of repetitive noxious stimuli, leading to an increase in electrical discharges in the dorsal horn. Inhibitory modulation can be impaired by abnormalities in the central nervous system and the facilitatory pain pathways can be stimulated by certain behavioral and cognitive factors.
This theoretical background can be applied to FM. In FM, studies already provided evidence for central sensitization as the cause of chronic pain. Temporal summation was found to be more facilitated, and the inhibitory pain modulation seemed impaired in FM patients. These findings can explain the chronic spontaneous pain in FM. Furthermore, some central abnormalities could be examined/objectified in FM: 1) hyperexcitability of the spinal cord, 2) decreased perfusion of pain-related brain structures, and 3) high levels of substance P in CSF. In addition, FM patients often present with pain hypervigilance, maladaptive coping strategies, and catastrophic thoughts, leading to cognitive central sensitization.
Based on the knowledge on central sensitization, on FM and on CFS, it is suggested that chronic widespread pain in CFS is the consequence of central sensitization. There are arguments and probable mechanisms that could explain this phenomenon in CFS. Also, in other chronic pain populations, central sensitization may play a key role. In fact, there are many similarities between CFS patients and other chronic pain populations such as patients with chronic low-back pain, whiplash, FM, etc. The psychosocial factors, for example, have been proved to contribute to pain perception in these different pain populations. But the specific nature of CFS such as the immunological abnormalities, elevated NO amounts, preceding infections etc., invites further research, in particular, on the possible contributory role of these abnormalities to pain processing in CFS.
In FM, many researches have been conducted to prove the theory of central sensitization. In CFS, however, it sticks to “supposing.” To give a scientific basis to the theory, the protocols applied in FM investigations could be used for patients with CFS. It would, for example, be interesting to test the efficacy of temporal summation in CFS. The inhibitory control of pain could be another point of interest. The influence of exercise on pain tolerance has already been studied in CFS [88], however, on a relatively small sample. On the contrary, spatial summation has, to our knowledge, never been investigated in CFS. Furthermore, the role of depression, hypervigilance, kinesiophobia, catastrophising, etc. on chronic pain in CFS requires further research. To obtain more direct and objective information on central sensitization, the protocol described by Banic et al. [57] could be used to test the sensitivity of the central nervous system. Clearly, there are many possible research areas to test the hypothesis, but there is still a long way to go to elucidate the nature of the chronic pain in CFS.
|
[
"central sensitization",
"fibromyalgia",
"chronic fatigue syndrome",
"chronic pain"
] |
[
"P",
"P",
"P",
"P"
] |
Eur_J_Health_Econ-_-_-1388083
|
Benefits and entitlements in the Hungarian health care system
|
This contribution considers entitlements and benefits in the Hungarian health care system. After a brief introduction to the organizational structure of the system the decision-making processes are discussed in detail, including the most important actors, types and pieces of legislation, formal structures, decision-making criteria, and outputs in terms of benefit catalogues. Within the two main public financing systems (social insurance and tax-funded services) there are four types of regulatory regimes: (a) traditional political decision making, (b) price negotiations, (c) updating of classification systems for payment purposes, and (d) the procedure for the inclusion of registered medicines in the scope of the social health insurance system. As an example we discuss the benefit regulations and benefit catalogues in the category of services of curative care (HC.1) of the OECD classification of health services.
With the crisis and the collapse of the communist dictatorship Hungary witnessed the beginning of a thorough health sector reform. Replacing the tax-based financing of the state-socialist system, Hungary reverted to the earlier Bismarckian model of compulsory social insurance in 1990. New performance-based provider payment methods were introduced (capitation in primary care, fee-for-service for outpatient specialist services, and diagnosis-related groups (DRGs) in the acute inpatient care sector) together with cost-containment mechanisms to ensure that the preset budgets were not exceeded. Ownership of the majority of hospitals and other health care facilities was transferred to local governments. The vast majority of medical doctors and other health workers remained salaried public employees, with the only exception of primary care, where the bulk of family physicians work as contracted private entrepreneurs. The minimum salary for public employee physicians is determined by law, whereas the average salary in the health sector has remained among the lowest in the economy [1].
Entitlement to health care is based mainly on the participation in the social insurance scheme (with compulsory membership, opting out is not permitted), and to a few services on citizenship. The national Health Insurance Fund (HIF) provides a virtually universal population coverage with an almost comprehensive benefit package, which applies to the whole country (i.e., there are no variations by region or by payer). Nevertheless, the HIF covers only the recurrent costs of services. The owners of health care facilities, mainly local governments, are obliged to cover the capital costs of services, which usually come from general and local taxation. Tax revenues are also used for covering the deficit of the HIF [1997 Act LXXX on Those Entitled for the Services of Social Insurance and Private Pensions and the Funding of these Services, Sect. 3(2)], certain special services (e.g., public health, catastrophe medicine, experimental medical technologies, family planning and maternal care), which are financed entirely from the central government budget [1997 Act LXXXIII on the Services of Compulsory Health Insurance, Sect. 18(5)a–d, h; 1997 Act CLIV on Health, Sects. 141(2)b, 142(2)], and the copayment for certain medicines and therapeutic devices for socially disadvantaged (1993 Act III of on Social Services). In addition to informal payments to health workers, copayments for medicines and therapeutic devices constitute the most important private source of health care financing, which are almost exclusively out-of-pocket, as private health insurance is still insignificant in Hungary [1].
Decision-making processes and actors
Entitlement to publicly financed health services is regulated by various types of regulatory instruments (Table 1). These represent the decisions of various decision-making actors, with varying decision-making rules/processes and power. There is a hierarchy of regulations with the constitution on top, followed by governmental and then ministerial decrees. In the case of conflicting provisions the higher level regulation precedes the lower one, but the provisions of lower regulations are usually more detailed.Table 1 Main types of regulations in Hungary [1949 Act XX on the Constitution of the Republic of Hungary; 1987 Act XI on Codification; Resolution No. 1088/1994 (IX.21) on the Decision-Making Procedures of the Government; 1990 Act LXV on Local Governments]Type of regulationDecision makerMethod of decision makingConsultationLegal regulations, laws ConstitutionNational AssemblyTwo-thirds majority voteRequired Fundamental actsNational AssemblyTwo-thirds majority voteRequired ActsNational AssemblySimple majority voteRequired Governmental decreesCentral governmentSimple majority vote of the government or exceptionally the prime minister (one person)Required Ministerial decreesMinistersOne personRequired Local government decreesLocal government assembliesSimple majority voteRequiredRulings of the Constitutional CourtConstitutional CourtSimple majority vote–Other means of state control ResolutionNational Assembly, central government, governmental committees, local governmentsAccording to decision maker– OrderMinisters, heads of organizations with national scope of authorityOne person– PolicyNational Assembly, central government, ministers, heads of national organizationsAccording to decision makerRequired Statement of interpretation of legislationsNational Assembly, central governmentAccording to decision maker? Announcement, communicationMinisters, heads of organizations with national scope of authorityOne person–
The National Assembly (Parliament) and the central government (and Ministry of Health, MOH) are the key actors in national-level decision making. Parliament determines, for instance, the scope of publicly funded services, the benefit package, and the budget of the HIF. While most decisions of Parliament require a simple majority, the constitution and other fundamental acts (e.g., on local governments) can be changed only with a two-third majority vote. Passed bills are promulgated as acts, on the basis of which governmental and ministerial decrees are issued, which regulate the implementation of acts in detail [1].
In addition to acts, governmental, ministerial, and local government decrees, which represent generally valid and obligatory behavioral norms and are called “legal regulations” or “laws,” there are lower level regulations, which are categorized as “other means of state control,” such as resolutions, orders, policies, statements, and announcements (1987 Act XI on Codification, Sects. 46–56) as well as organizational operational rules and procedures of various organizations and decision-making bodies. Resolutions, for instance, can be issued by the Parliament, the government, governmental committees, local governments, and their organs to regulate the tasks of organizations controlled by them, the rules of their own operation and plans within their scope of authority, while orders can be issued by ministers and the heads of organizations with national scope of authority to regulate the activities of the organizations under their control [1987 Act XI on Codification, Sect. 46(1)].
The process of decision making and the content of the benefit basket are usually regulated at least in ministerial decrees with only a few cases, when the decision on benefits has been decentralized to other organizations (Table 2). This does not mean that the decision making has no input from a wide range of actors in the health care arena. The most important actors are as follows: (a) National Health Insurance Fund Administration (NHIFA), which administers the HIF, in particular its Department of Payment Informatics (formerly the Information Center for Health Care of MOH, (GYÓGYINFOK)) which is responsible for the provider payment and performance measurement; (b) various advisory bodies and organizations of the MOH, including the national institutes of health and the professional colleges, which provide an expert input concerning a particular medical specialty; (c) professional organizations (e.g., Hungarian Medical Chamber, HMC; Hungarian Chamber of Pharmacists, HCP), unions, and provider and patient associations [1].Table 2 Regulations of entitlements and benefits in HungaryRegulation (type according to Table 1)What is regulated?C/PAct XX of 1949 (1)Right to healthCAct CLIV of 1997 (3)Right to health services; scope and broad content of health servicesCAct LXXX of 1997 (3)Participation in and contribution to the social insurance scheme (who is covered)–Act LXXXIII of 1997 (3)Framework for benefits of social insuranceC, PGovernment Decree No. No. 217/1997. (XII. 1 Korm (4)Executive order of Act LXXIII of 1997P, CDecree No. 46/1997. (XII. 17) NM of the Minister of Welfare (5)Services excluded from public financingCGovernment Decree No. 284/1997 (XII. 23) Korm (4)Copayments; exclusions (full fee)CAct XCIII of 1993 (3)Occupational health servicesCGovernment Decree No. 89/1995. (VII. 14) Korm (4)Occupational health servicesCDecree No. 27/1995. (VII. 25) NM of the Minister of Welfare (5)Occupational health servicesCDecree No. 5/2004. (XI. 9) EüM of the Minister of Health (5)Benefit catalogue of balneotherapyCDecree No. 20/1995. (VI. 17) NM of the Minister of Welfare (5)Benefit catalogue of treatment in sanatoriaCDecree No. 48/1997. (XII. 17) NM of the Minister of Welfare (5)Benefit catalogue of dental careCDecree No. 49/1997. (XII. 17) NM of the Minister of Welfare (5)Benefit catalogue of infertility treatmentsCDecree No. 47/1997. (XII. 17) NM of the Minister of Welfare (5)Eligibility for free breast milkCDecree No. 50/1997. (XII. 17) NM of the Minister of Welfare (5)Eligibility for patient transportationCDecree No. 51/1997. (XII. 18) NM of Minister of Welfare (5)Benefit catalogue of screeningCDecree No. 56/2003. (IX. 19) ESzCsM of the Minister of Health, Social and Family Affairs (5)Benefit catalogue of balneotherapyCDecree No. 19/2003. (IV. 29) ESzCsM of the Minister of Health, Social and Family Affairs (5)Benefit catalogue of medical aids and prostheses (therapeutic devices)CGovernment Decree No. 43/1999. (III. 3) Korm (Annex 8) (4)Benefit catalogue of chronic careCDecree No. 9/1993. (IV. 2) NM of the Minister of Welfare (5)Benefit catalogue of outpatient specialist services; acute inpatient care; dental care; day cases of curative care; dialysis; chronic outpatient care;CDecree No. 1/2003. (I. 21) ESzCsM of the Minister of Health, Social and Family Affairs (valid until 1 July 2005 then NHIFA announcement) (5, 12)Benefit catalogue of medicinesCDecree No. 20/1996. (VII. 26) NM of the Minister of Welfare (5)Benefit catalogue of home careCDecree No. 49/2004. (V. 21) ESzCsM of the Minister of Health, Social and Family Affairs (5)Tasks of MCH nursesCDecree No. 26/1997. (IX. 3) NM of the Minister of Welfare (5)Tasks of school health services (physician, MCH, dentist, assistant)CGovernment decree No. 168/2004. (V. 25) Korm (4)Regulatory regime (price negotiations for special medicines)PDecree No. 6/1998. (III. 11) NM of the Minister of Welfare (5)Regulatory regime (payment)PGovernment Decree No. 112/2000. (VI. 29) Korm (4)Regulatory regime (price negotiation)PDecree No. 32/2003. (IV. 26) ESzCsM of the Minister of Health, Social and Family Affairs (5)Regulatory regime (medicines)POrder No. 6/2005. (Eb. K. 3) OEP of the Chief Executive Officer of the National Health Insurance Fund Administration (9)Regulatory regime (medicines)PC content regulation,
P Process regulation
The 1987 Act XI on Codification explicitly requires relevant nongovernmental and interest representation organizations to be consulted in the phase of preparation of laws (Sects. 27–32).
Given that the existing benefit catalogues are almost exclusively incorporated into acts, governmental decrees and ministerial decrees, the general features of the decision-making processes (regulatory regimes) on benefits are common, as described above. While in these cases it is straightforward who makes the decision on the basis of what method, the real issue is who is consulted in what form in the preparatory phase of the decision making process, for which the acts discussed so far only provide a very general and vague guidance.
In many cases the decision support mechanisms have not yet been formalized in lower level regulations, and therefore that the process of preparation and codification is based on tradition. For instance, the need for the creation of a ministerial decree can originate from the provision of an act or governmental decree—in this case the MOH Legal Department initiates the codification process at the relevant professional department—or it can be initiated internally or by an external actor/stakeholder. Although each department must work on the basis of specific regulations on issue handling, in what cases, who is to be consulted, and how are usually not specified in these regulations but passed from one civil servant on the other (Zs. Kovácsy, personal communication, 2005).
This traditional regulatory regime is the most commonly used decision-making mechanism regarding entitlements and benefits (Fig. 1). The two key acts in the center of the definition of the benefit package are 1997 Act CLIV on Health and 1997 Act LXXXIII on the Services of Compulsory Health Insurance. These acts define only a general framework in which both exclusions and inclusions are usually stipulated only at the level of broad functional categories. According to the 1997 Act CLIV, the right to health services is unconditional only for emergency life-saving services, services, which prevent serious or permanent health damage, and for the reduction in pain and suffering (Sect. 6). Patients have a right to other health services only within the limits set by another legislation (Sect. 7). The Act states that the state is responsible for the operation of the social insurance scheme to enable the individuals to exercise their right to health, and then lists the services, which must be financed from the central government’s budget (Sects. 141, 142). The 1997 Act LXXXIII defines health services which are free of charge (Sects. 10–17), covered with copayment (Sects. 23–25), or excluded [Sect. 18(5–6)]. The starting point of the Act is that all health services are fully covered and exclusions are stipulated. In the frame of social insurance all professionally justified treatments can be used [Sect. 18(4)], but diagnostic and treatment protocols issued by the MOH can further specify the actual services for which the patients are entitled to [Sect. 19(1)]. Physicians are allowed not to adhere to the protocols if the deviation is justified by the status of the patient and by therapeutic considerations. It must be noted, however, that broad functional areas are also listed (i.e., there is a scope for implicit exclusions) in the Act in three main categories: (a) services for the prevention and early detection of diseases (Sect. 10), (b) curative services, including family physician services, dental care, outpatient specialist, and inpatient care (Sects. 11–14), and (c) other services including deliveries, medical rehabilitation, patient transport and emergency ambulance services (Sects. 15–17, 22).Fig. 1 Traditional regulatory regime for entitlements and benefits in Hungary
These laws were passed by Parliament with a simple majority vote but are not updated or modified on a regular basis. Their modification can be initiated by the government (Minister of Health), the President of the Republic, Members of Parliament, Parliamentary Committees, or other stakeholders through these actors [Resolution No. 46/1994 (IX. 30) OGY of the National Assembly].
The above acts contain a large number of clauses, which call the government, the Minister of Health, or both of them to further specify certain functional areas. There are two main types of ministerial decrees. The 1997 Act CLIV calls on the Minister of Health to regulate the professional requirements, including minimum standards and procedures of certain service categories, and these decrees may include provisions related to the benefit package. For instance, these types of ministerial decrees can specify the tasks which must be fulfilled in the framework of a particular service (e.g., school health services Decree No. 26/1997 (IX. 3) NM of the Minister of Welfare on School Health Services) and what service can be ordered for what patients’ conditions (e.g., rehabilitative treatments in sanatoria Decree No. 20/1995 (VI. 17) NM of the Minister of Welfare on the Treatment in Sanatoria in the Frame of Medical Rehabilitation) and define who is allowed to provide specific services [e.g., home care; Decree No. 20/1996 (VII. 26) NM of the Minister Welfare on Home Care)].
On the other hand, the 1997 Act LXXXIII authorizes ministerial (and governmental) decrees to specify entitlements and benefits within a particular service category [e.g., dental care; Decree No. 48/1997 (XII. 17) NM of the Minister of Dental Services which Can Be Utilized in the Frame of the Compulsory Health Insurance]. In certain cases it is not the service concerned specified but the criteria of eligibility [e.g., the supply of breast milk; Decree No. 47/1997 (XII. 17) NM of the Minister of Welfare on the Supply of Breast Milk in the Frame of the Compulsory Health Insurance]. The updating of these decrees is made on an ad hoc basis as the need arises. Benefits are uniform throughout the country and explicitly defined, although services are usually not detailed.
Nevertheless, there are certain areas, such as pharmaceuticals and therapeutic devices, where various factors such as the presence of powerful suppliers (manufacturers of medical goods) have led to more formalized decision making. These developments received a new impetus in Hungary as a result of the country’s integration into the European Union, as the requirements of EU directives, for instance, in the case of pharmaceuticals, have had to be incorporated into Hungarian regulations. The decision-making process for the inclusion of a particular medicine in the benefit package of the social insurance scheme has been regulated down to the details of decision-making criteria.
Furthermore, the introduction of new payment mechanisms created indirect means of defining benefits through the classification of cases and services for the purpose of provider reimbursement (Fig. 2). In the case of outpatient care there is a list of services with the World Health Organization International Classification of Procedures in Medicine (ICPM) codes and point values, while acute inpatient cases are categorized into one of the 786 DRGs on the basis of the diagnosis (coded on the basis of the 10th edition of the International Classification of Diseases) and the procedures performed [Decree No. 9/1993 (IV. 2) NM of the Minister of Welfare of the Social Insurance Financing of Specialist Services]. While the 1997 Act LXXXIII of declares that all professionally justified curative services are included in compulsory social health insurance financing, there is no incentive for the providers to provide services which are not reimbursed. Therefore these classification systems can be regarded as indirect benefit catalogues, with the updating procedure being their indirect modification. Since the introduction of the new payment mechanisms they have been regularly updated and the process of instituting changes has been formalized since 1998 [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care]. The updating process includes the modification and extension of the two fundamental classification systems as well as the various payment catalogues. The extension of the ICPM has special relevance for the benefit package, since those health care interventions, which are not listed in the Hungarian version of the ICPM have been excluded from public financing (1997 Act LXXXIII on the Services of Compulsory Health Insurance).Fig. 2 Updated Classification Systems for Payment Purposes. [Decree No. 6/1998. (III. 11) NM of the Minister of Welfare; Procedure of the Payment Codes Updating Committee]
The key actors in the process are the GYÓGYINFOK and the so-called Payment Codes Updating Committee (PCUC), which is an advisory body of the minister of health [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care]. The former is responsible for preparing decision support documents, including the collection and the analysis of the necessary financing data, while the PCUC makes the decisions formulated as proposals, guided by the criteria of public health impact and of the efficient allocation of resources [Procedure of the Payment Codes Updating Committee (Working Committee), December 2003, Sect. 2.5]. The ultimate decisions rest with the MOH. The list of outpatient specialist services, DRGs, day cases of curative care, chronic care services, and the various forms of dialysis (as well as their modifications) are always issued as ministerial decrees. Various announcements, communications, and guidelines are also published to promote the lawful use of payment catalogues [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care Sects. 2(2–4), 7(1)].
The PCUC has 15 permanent members, delegated by the MOH (2), CEO of the NHIFA (2), GYÓGYINFOK (2), Hungarian Hospital Association (2), and HMC (1), while the Minister of Health appoints six medical experts and the head of the PCUC [Procedure of the Payment Codes Updating Committee (Working Committee), December 2003, Sect. 3]. The committee prepares its procedure, a yearly workplan and a methodological document as the basis of the updating process [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care Sect. 5(4)].
The updating process can be initiated in two ways. Regular updates are planned in the workplan of PCUC, while all the relevant stakeholders are allowed to ask for unscheduled updates which must then be evaluated and answered within 30 days [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care, Sect. 1(2)]; Procedure of the Payment Codes Updating Committee (Working Committee), December 2003, Sect. 2.4]. First, the latter proposals must be submitted to the head of the relevant national institute or directly to the Committee if there is no national institute concerned. The relevant professional college(s) then provides an expert opinion, and the GYÓGYINFOK then prepares a cost estimation and budget impact analysis usually on the basis of data provided by a sample of health care institutions. The PCUC has at least one meeting per month, whose proceedings are to be submitted to the MOH in the form of a proposal for changes at least once in a year [Decree No. 6/1998 (III. 11) NM of the Minister of Welfare on the Regulation of Updating Professional Classification Systems and Financing Parameters Used in Health Care, Sect. 5(6–7); Procedure of the Payment Codes Updating Committee (Working Committee), December 2003, Sect. 4.4].
In summary, there are four main regulatory regimes for the definition and modification of entitlements and benefits: (a) the traditional decision-making process, guided by the general rules of codification with less formalized preparatory phase, for services such as primary care and home care, (b) formalized as “price negotiations” for therapeutic devices and balneotherapy, (c) formalized as the “procedure for the inclusion of registered medicines in the scope of the social health insurance system” in line with the provisions of Council Directive 89/105/EEC (21 December 1988), and (d) formalized as the “procedure of updating professional classification systems and payment parameters” for outpatient specialist and inpatient care services.
As a result of the various regulatory regimes, universal benefit catalogues exist for almost all service areas.
Entitlements and benefits: services of curative care
Table 3 summarizes entitlements and benefits in the category of services of curative care (HC.1) of the International Classification for Health Accounts (ICHA) taxonomy [2].Table 3 Entitlements and benefits: services of curative care (HC.1)HCFunctional categoryRegulatory regimeaBenefit catalogue (regulation no. from Table 2, method of classification, date passed, last updateda), taxonomy1.1Inpatient curative careUpdating of classification systems for payment purposes (4)Acute inpatient care–somatic, intensive care, mental special area: transplantation21, Annex 3 (and Annex 8 for high cost high tech interventions; Annex 10 for course-type treatments); itemized by case (linked to diagnosis and therapy); 2 April 1993, 6 August 2004. 26 main groups (major diagnostic categories), altogether with 786 items (DRGs)1.2Day cases of curative careSpecialist outpatient care—day surgery21, Annex 9; itemized by service; 2 April 1993, 6 August 2004.227 items in two lists: (a) 194 elective surgical procedures (Sect. I), (b) 33 other (“clinical”) interventions (Sect. III)Special area: hemodialysis21, Annex 11; itemized by service; 2 April 1993, 2 March 2001.6 items: (a) hemodialysis, hemofiltration, hemodiafiltration; (b) same for patients under 18 years old; (c) peritoneal dialysis, (d) hemoperfusion, (e) hemodialysis with reusable dialysator, (f) mobile treatment1.3Outpatient care1.3.1Basic medical and diagnostic services(Curative) primary careTraditional (2, 4)–1.3.2Outpatient dental careDental prophylaxis, care, surgeryTraditional (2, 4)13; itemized by service; 17 December 1997, 1 November 2001.3 main groups: (a) dental screening, with 7 items in three subgroups; (b) dental emergency services with 10 items; (c) dental primary and secondary care, defined as the payment list, detailed in a separate decree (no. 21)Updating of classification systems for payment purposes (4)21, Annex 12; itemized by service; 2 April 1993, 11 March 2003.125 items in 10 groups: (a) examinations, documentations, 9; (b) prevention, 5; (c) radiography, 5; (d) anesthesia and drug prescription, 5; (e) tooth preserving services and endodontia, 12; (f) oral diseases and parodontology, 11; (g) dental surgery, 24; (h) dental prosthetics, 26; (i) child, school, and youth dental services, 13; (j) orthodontia, 151.3.3All other specialized health careSpecialist outpatient care–somaticUpdating of classification systems for payment purposes (4)21, Annex 2; itemized by service; 2 April 1993, 27 August 2004.List contains 3,166 (and together with the dispensary specific services 3,204) items (based on the 1978 ICPM of WHO)Specialist outpatient care–mental; special area: psychoanalysis, -therapy21, Annex 15; itemized by service; 11 May 2004.Altogether 120 items (82 items overlap with Annex 2) in five main groups: (a) skin and STD, 29 items; (b) oncology, 26 items; (c) pulmonology, 25 items; (d) psychiatry, 19 items; (e) addictology, 21 items1.3.9All other outpatient curative careOutpatient care by other professionalsUpdating of payment classification systems (4)As with 1.3.3, which are included (e.g., logopedia, physiotherapy, dietetics, optometry, acupuncture); the rest are implicitly excluded (e.g., chiropody, bioenergy treatments, iris diagnostics, hydrotherapy)Alternative and complementary medicineSpecial area: balneotherapy and physiotherapyTraditional (4)11 (definition of entitlements, regulation of prescription, utilization, and professional requirements); itemized by service; indications and contraindications; 9 November 2004.Annex 1, balneotherapy services, 10 items; Annex 4, physiotherapy services, 13 items; Annexes 5 and 6, indications and contraindicationsPrice negotiations (4)18; itemized by service; 19 September 2003.3 groups with the same 10 items; copayment is different for spas, which have national, regional, and local qualification1.4Services of curative home careTraditional (4)23, Annex 1 (specifies professional requirements); itemized by service; 26 July 1996, 9 June 1999.13 service categories, 22 items (includes hospice at home)a Benefit catalogues with the traditional regulatory regime are updated on an ad hoc basis, while the updating of payment lists is regular. All benefit catalogues are issued as ministerial decrees.
The general framework for these services is set by the 1997 Acts CLIV and LXXXIII, as discussed above, but all health services in this category are considered within the frame of the social insurance scheme. Although the 1997 Act LXXXIII states that all professionally justified services are included, in most cases the detailed payment catalogues imply implicit exclusions, such as most services of alternative medicine. With the exception of family physician services (primary care), there are benefit catalogues for each service category, either as a list of cases/services for payment purposes (1.1, 1.2, partly 1.3.2, 1.3.3, and partly 1.3.9) or for further specification of broad functional categories enumerated in the Act (partly 1.3.2, partly balneotherapy) or for the specification of professional requirements (1.4, partly balneotherapy).
These nationally valid benefit catalogues are the output of either the traditional decision-making process or the process of updating classification systems for payment purposes. They are issued as a ministerial decree in both cases, with the MOH being the final decision maker. Updating is ad hoc in the case of the former and regular in the case of the latter, where the preparatory phase is much more formalized and even a few decision-making criteria are defined (budget and public health impact), as discussed above.
Undoubtedly the two most detailed benefit catalogues are the lists of DRGs for inpatient curative care (1.1), and the so-called WHO point, or German point system for outpatient specialist care (1.3.3, partly 1.3.9). The essence of the DRG based hospital payment is that it pays a standard fee for discharged acute hospital cases and not for any individual service items such as laboratory tests, hospital days, drugs, and procedures. The DRG system classifies cases into a manageable number of categories, which are more or less medically meaningful and in which resource use is the same or at least similar (homogeneous resource use) [3]. The DRG system was introduced countrywide in Hungary in 1993 after a pilot project that began in 1987. The DRG system in the United States was adapted to the local situation using the cost data collected [4]. The current version of the Hungarian adaptation of DRGs (homogeneous disease groups, HDGs) is 5.0, which contains 786 groups in 26 main diagnostic categories [Decree No. 9/1993 (IV. 2) NM of the Minister of Welfare of the Social Insurance Financing of Specialist Services]. The first step of classifying patients is to determine the major diagnostic category, in which the main diagnosis and/or the interventions can be a principal classification criterion, which is to be further modified by comorbidity and age. For instance there are three groups of classifying factors in the major diagnostic category 14 (pregnancy, birth and childbed): the type of pregnancy (without complications, pathological pregnancy), the type of delivery (cesarean, vaginal, vaginal with operation, vaginal with epidural anesthesia) and the conditions of delivery (high risk, other comorbidity; Fig. 3). While in principle inpatient cases are sorted into HDGs primarily on the basis of diagnosis, which leaves hospital physicians to select treatment options freely, several treatment modalities are costly enough to be a principal classification criterion.Fig. 3 Examples of the homogeneous disease groups in Hungary. [Decree No. 9/1993. (IV. 2) NM of the Minister of Welfare on the Social Insurance Financing of Specialist Services]
The WHO point system, however, is entirely itemized by service. It is based on the 1978 ICPM, but the original list has been modified frequently ever since it has been introduced as the basis of payment in the outpatient specialist setting. Along with the chronic outpatient care items, the current catalogue contains 3,204 service items and has been in effect since 27 August 2004. Each service has a five-digit identification code and the services are listed in numerical order, from item 110011 (first aid) to 97550 (ambulatory developmental-neurology follow-up care of infants with spina bifida) and the list closes with 12 items of “complementary” points for special transfusion (two items) for noncompliant patients who threaten or attack attending medical staff (one item), for supervision of patients after treatment (five items), for drug loading (one item), and for patients of young age (three items). The list in the ministerial decree is not structured in groups or subgroups, but it is generally service (e.g., ultrasound examinations) and organ-oriented [e.g., eye-examinations; Decree No. 9/1993 (IV. 2) NM of the Minister of Welfare of the Social Insurance Financing of Specialist Services].
Discussion
The provision of universal and comprehensive coverage was the founding principle of the previous, state-socialist health care system. The tension between the changing and increasing needs and the available resources created shortages which have not been acknowledged and explicitly dealt with. Rationing effectively occurred through queuing, implicit waiting lists, the dilution of services, and informal payments [1].
The social insurance and payment reforms have brought about little change in this respect. Successive governments have faced the chronic deficit of the HIF, but measures to balance the budget have almost exclusively targeted the revenue side of the system, and only minor modifications have been implemented in the almost comprehensive benefit package [1]. Although entitlements are in principle linked to paying the contribution, the coverage is universal in practice since entitlement is not checked by the providers [5]. Despite payment catalogues the benefit package is defined rather negatively. While services financed by the HIF should be provided according to treatment protocols issued by the MOH, no such protocols have yet come into effect.
The general opinion in Hungary is that more explicit priority setting with more exclusions or more significant copayments would not be accepted by the majority of the population, and therefore politicians are reluctant to touch the issue of priority setting in a more systematic manner. On the other hand, the financial pressure on the system is high, and indirect, implicit rationing does occur, for instance, through informal payments. It is a question of how long this schism can be uphold, especially in the light of the challenges by joining the EU.
The plan of the current government is to revise entitlements to health care by expanding the scope of services to all emergency care for which all citizens are eligible [6]. The rest of the health services will be provided on the basis of participation in the social insurance scheme, but it will be checked whether the patient is in fact entitled to services. Furthermore, the government plans to revitalize the system of treatment protocols, which is obviously a means to exclude certain interventions. It is not yet known whether this is only a first step towards a more explicit priority setting, or whether the benefit package remains essentially unchanged.
|
[
"hungary",
"health services",
"health benefit plans",
"health priorities",
"national health programs"
] |
[
"P",
"P",
"R",
"R",
"M"
] |
Neuroimage-2-1-2225446
|
Interpreting scan data acquired from multiple scanners: A study with Alzheimer's disease
|
Large, multi-site studies utilizing MRI-derived measures from multiple scanners present an opportunity to advance research by pooling data. On the other hand, it remains unclear whether or not the potential confound introduced by different scanners and upgrades will devalue the integrity of any results. Although there are studies of scanner differences for the purpose of calibration and quality control, the current literature is devoid of studies that describe the analysis of multi-scanner data with regard to the interaction of scanner(s) with effects of interest. We investigated a data-set of 136 subjects, 62 patients with mild to moderate Alzheimer's disease and 74 cognitively normal elderly controls, with MRI scans from one center that were acquired over 10 years with 6 different scanners and multiple upgrades over time. We used a whole-brain voxel-wise analysis to evaluate the effect of scanner, effect of disease, and the interaction of scanner and disease for the 6 different scanners. The effect of disease in patients showed the expected significant reduction of grey matter in the medial temporal lobe. Scanner differences were substantially less than the group differences and only significant in the thalamus. There was no significant interaction of scanner with disease group. We describe the rationale for concluding that our results were not confounded by scanner differences. Similar analyses in other multi-scanner data-sets could be used to justify the pooling of data when needed, such as in studies of rare disorders or in multi-center designs.
Introduction
Techniques utilizing in vivo MRI-derived measures of brain tissue morphometrics show increasing promise for aiding clinicians in assessment and diagnosis of disease. Studies that use longitudinal and/or multi-site data-sets, such as the Alzheimer's Disease Neuroimaging Initiative (Mueller et al., 2005), have the potential to provide a wealth of information. The large numbers of subjects resulting from pooling multi-scanner data-sets has numerous advantages. It increases sensitivity thus allowing detection of subtle effects and enables the analysis of subgroups within a cohort. Additionally, pooling offers increased reliability and confidence about the size of effect by averaging out unforeseen confounds and hence a method for carrying out meta-analyses or analyses of rare subjects with orphan-diseases scanned at home rather than in distant centers. Any one study may have unforeseen bias which is lessened by pooling. However, one important confound of combining images gathered from different scanners is the potential for scanner effects to introduce systematic error, thus making the interpretation of results difficult. Partial volume effects and image intensity inhomogeneity can introduce error into automatic segmentation with any given scanner (Li et al., 2005). Noise of the electronics of the MRI system, subject-specific physiological noise and imaging gradient non-linearities also contribute to image intensity variability (Jovicich et al., 2006; Littmann et al., 2006). Furthermore, differences in subject positioning between sites add to the variability in distortion fields (Jovicich et al., 2006) and images can vary as a function of protocol differences between a baseline and a later scan or with drifts in instrument signal to noise over time (Preboske et al., 2006). The interaction of scanner differences with segmentation remains a particular concern and potential cause of varied measures of regional tissue volume in the brain (Ashburner and Friston, 2000). It is therefore vital to confirm that there is no important interaction between scanner and effect of interest and/or account for the effects of different scanners in a principled manner before pooling from different scanners can be recommended as a routine.
The use of phantoms to calibrate different scanners and account for scanner drift or upgrades is well established (Tofts, 1998; Van Haren et al., 2003; Schnack et al., 2004; Jovicich et al., 2006) and a useful way to guarantee quality. Studies using a small number of healthy subjects repeatedly scanned on different scanners or after upgrades are also recommended to test reliability (Tofts, 1998; Han et al., 2006; Jovicich et al., 2006). However, it is not clear that the extra calibration efforts can actually correct for the majority of these sources of variability (Jovicich et al., 2006), or are even necessary. Also, it has yet to be established how confidently the results from multi-scanner data-sets can be interpreted irrespective of up-front calibration.
Multi-scanner studies, though not common, are increasing. A longitudinal aging study by Raz and colleagues, one of the few which discussed the analysis of images derived from multiple scanners, concluded that using different scanners in 4 subjects did not affect measured intracranial volume with a manual tracing method (Raz et al., 2005). Likewise, manual hippocampal measurements performed on both 1.5 T and 3.0 T scanners in 8 healthy controls were not affected by field strength (Breillmann et al., 2001). We are not aware of any study that has described an automated analysis of a large, multi-scanner data-set with the aim of assessing whether scanner-associated biases are significant.
We set out to investigate scanner effects in an automatically preprocessed data-set of subjects with mild to moderate AD and cognitively normal elderly people whose MRI scans had been collected over a period of 10 years with multiple scanners and upgrades on a consistent platform and repeated calibrations. Even though the concern has been raised that segmentation errors are a particular problem in atrophic brains (Good et al., 2002), we did not find this to be an issue in our sample, possibly because the disease was early. We hypothesized that the changes due to AD pathology are large compared to scanner induced distortions and that there would be no interaction of scanner with case and/or control groups. In this paper, we report the results in our data-set and outline a method by which multi-scanner data-sets might be pooled with confidence.
Methods and results
Our data-set included a total of 6 scanners and 136 subjects scanned over a 10-year period. None of the subjects were scanned more than once. All scans were done on the same platform, General Electric Signa 1.5 T scanners (slice thickness 1.6, matrix dimensions 256 × 192). There were minor variations in the TR, TE, and flip angle (see Table 1). Scanners also underwent upgrades over time. Importantly, all scanners were monitored with daily phantom quality checks which calibrated the gradients to within ± 1 mm over a 200-mm volume centered at iso-center, monitored signal to noise and radio frequency (RF) transmit gain. The major hardware elements (body resonance module gradient coil and birdcage head transmit–receive volume coil) were unchanged throughout time and across scanners, except that for the oldest scanner (scanner 2), the wiring in the resistive shim set was not cooled to super conducting temperatures, whereas for the other 5 scanners, the shim coils were inside the dewar and were cooled to the superconducting range.
We used voxel-based morphometry (VBM) to evaluate the interaction of scanner and grey matter segmented modulated images for the 6 different scanners. VBM has the advantage of assessing the whole brain and not being biased to one particular region or structure. It entails a voxel-wise comparison of local volume of grey matter between groups after the images are spatially normalized into the same space, segmented, modulated, and smoothed. Voxel-wise statistical parametric maps result from statistically thresholded contrasts after corrections for multiple comparisons (Ashburner and Friston, 2000) using false discovery rate (FDR) (Benjamini and Hochberg, 1995).
Images were visually inspected for artifacts or structural abnormalities unrelated to AD. They were firstly segmented into white (WM) and grey matter (GM) using SPM5 (Wellcome Trust Centre for Neuroimaging, Institute of Neurology, UCL, London UK – http://www.fil.ion.ucl.ac.uk/spm). Then, WM and GM segments were further normalized to a population template generated from the complete image set using a diffeomorphic registration algorithm. This non-linear warping technique minimizes structural variation between subjects (Ashburner, 2007). For comparison, we also repeated the analysis using the more widely used standard SPM5 segmentation code (Ashburner and Friston, 2005) instead of the diffeomorphic registration algorithm. Resolution before normalization was − .9, .9, 1.6 and after normalization was − 1.5, 1.5, 1.5 for the diffeomorphic registration algorithm and − 2.0, 2.0, 2.0 with the standard SPM5 procedure. A separate ‘modulation’ step (Ashburner and Friston, 2000) was used to ensure that the overall amount of each tissue class was not altered by the spatial normalization procedure. Modulation was performed by multiplying the warped tissue probability maps by the Jacobian determinant of the warp on a voxel-by-voxel basis, which represents the relative volume ratio before and after warping, thus allowing voxel intensities in the segmented grey matter map, together with the size of the voxels, to reflect regional volume and preserve total grey matter volume from before the warp. Modulated grey matter scans were smoothed using a 6-mm full-width at half-maximum Gaussian kernel.
The smoothed grey matter images were analyzed in a factorial design, with the 6 different scanners as one factor (SCANNER) with 6 levels and the presence of AD (GROUP) as the second factor with two levels (present and absent). Age, gender, and total intracranial volume were entered as covariates (Fig. 1). We performed F-tests correcting for multiple comparisons across the brain (FDR correction). The degrees of freedom was 121 for all comparisons. The effect of group revealed strong effects (p < 0.001) in the left medial temporal lobe (− 29, − 24, − 9 [x, y, z]; F = 100.34) and right medial temporal lobe (30, − 27, − 9 [x, y, z]; F = 91.44) (Fig. 2). The effect of scanner showed significant differences (p < 0.05) in the right (9, − 30, 0 [x, y, z]; F = 11.93) and left (− 9, − 11, 8 [x, y, z]; F = 9.69) thalami (Fig. 3), but this effect was less than the effect of group. T-tests contrasting each scanner against the others revealed the scanner effect in the thalamus was mostly due to the oldest scanner (scanner 2). One newer scanner (scanner 6) revealed an effect that did not survive FDR correction in the right thalamus (8, − 5, − 3 [x, y, z]; T = 4.03), whereas the oldest scanner contrasted with the others revealed a significant effect (p < 0.05) in the left thalamus (− 9, − 11, 6 [x, y, z]; T = 5.94). Despite the effect of scanner, there was no significant interaction of scanner with group, the highest Z-score being 3.82 with a corrected p value of 0.942. We performed F-tests for each possible combination of scanners and found no significant interaction with any of these groupings. When using the standard VBM procedure implemented in SPM5 for normalization and segmentation, the results of the various contrasts followed the same patterns as when using the diffeomorphic normalization/segmentation procedure. We failed to find an interaction of scanner with group; the highest Z-score was 3.78 with an FDR corrected p value of 0.954. The effect of scanner showed significant differences (p < 0.05) in the left thalamus (− 10, − 10, 6 [x, y, z]; F = 9.49) and the right thalamus (8, − 28, − 4 [x, y, z]; F = 9.41). The effect of group also showed strong effects (p < 0.001) in the left medial temporal lobe (− 26, − 14, − 20 [x, y, z]; F = 101.46) and right medial temporal lobe (26, − 8, − 18 [x, y, z]; F = 77.98).
Importantly, we found no significant scanner effects in the medial temporal lobe cluster (− 29,− 24, − 9 [x, y, z]; F = 0.01) and the disease effect size in the thalamus was minimal compared to the effect size in the medial temporal lobe (9, − 30, 0 [x, y, z]; F = 1.06), suggesting minimal scanner effects in the areas that are most affected by AD and minimal disease effects in the areas showing scanner differences. Confidence intervals, which are reflective of the standard deviations, for the contrast estimates are shown in Fig. 4. At the voxel of greatest effect of group, the confidence interval is small relative to the effect size for the main effect of group. The opposite is true for the effect of scanner at the area of greatest disease, which is further evidence of lack of effect. For the main effect of scanner, confidence intervals are similar between the different scanner contrasts (see Fig. 4A), indirectly suggestive of relatively little variance across scanners.
Because there were 4 software upgrades, we also analyzed the interaction of software version and disease. Using the same basic design matrix as described for the interaction with scanners, this time the contrasts included cases and controls from each software version, covaried with age, gender and intracranial volume. As with the effect of scanner, there was no significant interaction of software version with group, the highest Z-score being 3.65 with a corrected p value of 0.880.
Discussion
In our data-set, we found the effect of disease to be substantially larger than the effect of scanner and failed to find a significant interaction of disease with scanner or software upgrades. In general, the effect of disease in AD is liable to be larger than the effect of scanner, which is supported by our result of no important interaction between scanner and effect of interest. Ideally, further studies with an even larger data set to better calculate the effect sizes and quantify the distance between the scanner effect cluster and group effect cluster could be done to validate our findings. However, comparison of the magnitude of the scanner effect versus disease effect in the medial temporal lobe cluster in our data-set demonstrates that scanner differences had minimal effects in the areas that are important in the study of AD. Furthermore, even though there are likely to be some differences among data from different scanners, our experiments were explicitly designed to detect scanner-related differences. By modeling appropriate confounds in the design matrix, it appears possible to remove these small effects. We were able to model this interaction easily because relatively homogeneous cases and controls were scanned in each machine. We are unable to say whether lesser or more subtle and distributed changes would be as resistant to scanner effects.
Though we did not detect a significant interaction of scanner with disease, we cannot be absolutely certain it is indeed due to the absence of such effect. The lack of significance may be a reflection of the lack of statistical power, e.g., insufficient number of scans. Other causes such as a high average residual variance or residual variance inhomogeneities could also under-power the detection of the effects. However, the variance inhomogeneity was considered in our analysis by assuming unequal variance for the different levels of each of the two factors in our full factorial design (the two factors are the scanner and group). In a post-hoc manner, we explored the residual variance across scanners to assess whether that explained the lack of sensitivity in the results. In the area of greatest disease effect, the variance was low and did not reflect significant inhomogeneity, but in the area of greatest scanner effect there was lower average residual variance and more variability. Though our tests for variance inhomogeneity across the 6 scanners was not voxel-by-voxel over the whole brain volume, our findings seem to support that the variance inhomogeneity is location-dependent and should be accounted for when analyzing data acquired from different scanners as we attempted to do in our analysis.
The greatest effect of scanner was in the thalamus. The effect of the thalamus was largely driven by the scanner with the resistive shim set that was not cooled to superconducting temperatures, which suggests an impact of such hardware differences on thalamic segmentation. The composition of the thalamus is an issue of debate as it is not completely grey matter receiving numerous white matter tracts from other parts of the brain. Additionally, the grey matter intensity value of the thalamus is different from that of cortical grey matter. The intrinsically poor intensity contrast in the thalamus renders it susceptible to small differences in image contrast due to scanner differences. There is also less variability in this part of the brain, so tests will be more sensitive to such differences. These factors may contribute to the difficulty of accurate segmentation of the thalamus in addition to scanner effects.
The relatively small effect of scanner is potentially attributable to quality control measures during data acquisition and/or the robustness of the segmentation method. The SPM5 algorithm, by using spatial information together with intensity information, should be more robust to such differences than a segmentation algorithm that is purely intensity-based. However, this study did not directly investigate which forms of preprocessing are most affected by scanner parameters. It is important to note, though, that the effect was not due to the diffeomorphic normalization procedure, since we obtained similar patterns to those with the standard SPM5 normalization. Presumably, the use of daily scanner calibrations allows for a relatively constant intensity contrast between grey and white matter voxels between scans. Routine procedures designed to minimize image intensity differences over time are useful, particularly since it has been demonstrated that several common scan functions can potentially introduce measurement errors as high as 100% (i.e., much greater than the disease effect of AD), without appropriate quality control measures in place (Preboske et al., 2006). Even though it is common experience that current generation scanners are remarkably stable over time, they do still drift in a way that is correctable with judicious calibration. Though we collected no data to determine the frequency and level of sophistication necessary, basic calibration akin to the current standard in most major centers is likely sufficient. Furthermore, while it can be useful to collect data on inter-subject variability by scanning the same subjects in different scanners, such efforts would not directly answer the question of whether data from different subjects on different scanners can be pooled.
The unified segmentation method of SPM5 (Ashburner and Friston, 2005) produces an estimate of the tissue class intensities from a fitting of spatial priors to the image, which allows for differences in image intensities between scans. Additionally, any image intensity differences due to subject to subject variation in coupling of the RF coil to the head that may be introduced by different head shapes and sizes are likely to be well-accounted for by bias corrections in the unified segmentation algorithm.
As long as provision for different scanners and/or upgrades is made within an analysis, the effect of scanner regardless of magnitude is not likely to devalue the integrity of results. However, if there is a true physical interaction of the biological effect of interest with the method of measurement, perfect calibration or even using a single scanner would not prevent bias. On the other hand, any unusually large effect from one scanner would be attenuated by the totality of scans from different scanners that make up the template, which averages the different scanner effects for normalization.
To date, methodological differences between individual studies have prevented comprehensive pooling of data in meta-analyses of AD MRI studies (Zakzanis et al., 2003; Whitwell and Jack, 2005; Wahlund et al., 2005). To the extent that data pooling was possible, the meta-analysis compared regional volume between disease groups (Zakzanis et al., 2003). Although there may be concern about error introduced by the variability inherent in preprocessing many individuals across studies, automated VBM methods successfully overcome the problems to combine information from multiple scans and studies. VBM meta-analyses of pooled scans permit the usual range of analyses beyond simple categorical comparisons (e.g., regression, age interaction, subgroup, nonlinear, etc.).
Our results suggest that for Alzheimer's disease, particularly if the imaging platform remains constant, variations attributable to individual scanners and upgrades may have negligible effects on segmented grey matter images. We expect, though do not yet have the data, that the approach is robust enough to accommodate using both 1.5 T and 3 T scans in the same meta-analysis. A principled approach to test the validity of a meta-analysis is to carry out an a priori interaction analysis of scanner by biological effect to be studied to exclude scanner-specific compromising effects. Additional comfort can be gained by finding that the biological effect is of greater magnitude than any effect of scanner. An alternative approach is to model each scanner's effects separately, thereby permitting a principled meta-analysis on pooled data. That data can be pooled from different scanners without corroding the integrity of results is reassuring for large multi-site studies.
|
[
"alzheimer's disease",
"multi-scanner",
"magnetic resonance imaging (mri)",
"voxel based morphometry"
] |
[
"P",
"P",
"M",
"M"
] |
J_Urban_Health-2-2-1705474
|
An Analysis of Respondent Driven Sampling with Injection Drug Users (IDU) in Albania and the Russian Federation
|
Injection drug users in Tirana, Albania and St. Petersburg, Russia were recruited into a study assessing HIV-related behaviors and HIV serostatus using Respondent Driven Sampling (RDS), a peer-driven recruitment sampling strategy that results in a probability sample. (Salganik M, Heckathorn DD. Sampling and estimation in hidden populations using respondent-driven sampling. Sociol Method. 2004;34:193–239). This paper presents a comparison of RDS implementation, findings on network and recruitment characteristics, and lessons learned. Initiated with 13 to 15 seeds, approximately 200 IDUs were recruited within 8 weeks. Information resulting from RDS indicates that social network patterns from the two studies differ greatly. Female IDUs in Tirana had smaller network sizes than male IDUs, unlike in St. Petersburg where female IDUs had larger network sizes than male IDUs. Recruitment patterns in each country also differed by demographic categories. Recruitment analyses indicate that IDUs form socially distinct groups by sex in Tirana, whereas there was a greater degree of gender mixing patterns in St. Petersburg. RDS proved to be an effective means of surveying these hard-to-reach populations.
Introduction
Accessing hard-to-reach populations for HIV prevention and research activities has historically been challenging. Certain high-risk populations are fundamental in the proliferation of HIV, particularly in the early stages of a country’s epidemic. Owing to the stigmatized and often criminalized behaviors of injection drug users, for example, these populations avoid open contact with the rest of civil society. Respondent Driven Sampling (RDS) offers a solution to that dilemma, allowing researchers to access hard-to-reach populations through their social networks.2,3 A social network person includes other persons with whom he/she has a social relationship and who tend to engage in the same behaviors as the index person. This paper presents approaches used to implement RDS in Tirana, Albania and St. Petersburg, Russia and lessons learned in these two cities.
Materials and Methods
The studies in St. Petersburg and Tirana utilized RDS to recruit IDUs into an HIV-related behavioral survey. While the methodology behind RDS has been described elsewhere,3 in brief it is a referral-based system whereby participants refer a limited number of friends or acquaintances (who are also members of the target group) to participate. Study participants receive an incentive both for participating and successfully recruiting others. Data are analyzed using the recruiting linkages and yield a probability sample of the population in the target area.
For inclusion in either study, participants had to be at least 15 years old and report having injected drugs recreationally in the past 6 months. Data collection for both surveys included face-to-face interviews and encompassed a 7–8 week period at three sites in each city. In Tirana, data collection took place in offices of two non-governmental organizations (NGOs), conducting outreach to IDU and in one treatment hospital. Data collection in St. Petersburg was conducted in three sites in various parts of the city. One was a fixed site in the center of the city. The other two sites were located just outside the city center and open on weekend afternoons, operating from mobile vans at stable locations. One site was selected due to its proximity to an illegal drug market.
The data collection team at each site and shift consisted of two interviewers and one person serving as coupon manager, supervised by a study coordinator. The coupon manager recorded data for each participant on a form that included each individual’s self-reported network size, the coupon number brought to the interview, and the coupon numbers each was given for distribution. The Tirana sites also had a physician, laboratory technician and counselor as data collection included biological samples for HIV and other STIs.
Questionnaire topics included sociodemographic indicators, sexual and drug injection risk behaviors, knowledge of HIV and STIs, and previous HIV testing. Survey duration in Tirana was approximately 1.5 hour for both the behavioral and biological portions of the study, while in St. Petersburg data collection lasted approximately 45 minute. In accordance with RDS methodology, the initial participants of the study, known as “seeds”,4 were members of the target IDU population in their respective cities, and were selected to begin recruiting their peers into the study. In Tirana, 15 seeds launched the study, while in Russia recruitment was initiated with 11 seeds; 2 seeds were added later to generate more recruiting. The seeds selected in both cities were referred from local NGOs conducting outreach with the IDU population, as well as personal IDU contacts of study staff. Each participant in St. Petersburg received three coupons with an eight-week expiration date to recruit their peers while two coupons per participant were used in Tirana (no expiration dates). In St. Petersburg, three coupons continued to be distributed until approximately 200 participants were attained, after which participants were warned that the study would be ending within a few days. In Tirana, coupons were distributed until the desired sample size of 200 was reached; however, participants were warned from the beginning that once this was reached, no more coupons would be honored.
In Tirana, participants received cash incentives of 10 euros for participating and an additional 5 euros for each successful recruit. In St. Petersburg, use of monetary incentives was discouraged by all local stakeholders; therefore, participants were given a gift package consisting of a box of chocolates and personal hygiene products for completing the survey, and a box of chocolates or a can of coffee for each successful recruit. The gifts were equivalent to U.S. $10 for participating and U.S. $5 for recruiting.
Two hundred ten IDUs (not including seeds) were surveyed between July and August 2005 in Tirana, and 200 IDUs (not including seeds) were surveyed in St. Petersburg during the same time period.
Biometric measures were taken in Tirana to discourage study members from participating in multiple sites. Measurements consisting of wrist circumference multiplied by forearm length were taken and recorded. Participants were told that this was a unique measurement that could be used to determine if they attempted to return or access other data collection sites. Project staff also rotated between sites to decrease the likelihood of participants visiting multiple sites. In St. Petersburg, biometric measurements were not taken since local researchers thought it would induce suspicion among IDUs and consequently discourage participation. Recognition of potential duplicate participants depended upon study staff, who were outreach workers familiar with the target population and operated out of the same site for the entire 8 weeks of data collection.
The data were analyzed using RDS Analysis Tool version 5.4 (RDSAT). RDS population estimates and confidence intervals as calculated by RDSAT are reported. The homophily index describes the extent of in-group ties. A homophily index of H = 1.0 reflects perfect homophily, indicating that ties are formed with other members of the same group. In contrast, H = −1.0 reflects perfect heterophily, indicating that ties are formed completely outside of the group.3 Both studies were approved by the FHI Protection of Human Subjects Committee and the Tirana study was also approved by the Medical Ethical Committee of Albania.
In both cities, information about whether anyone refused to accept coupons was elicited from those who returned to collect their compensation for successful recruiting. Approximately 20% (Tirana) and 5% (St. Petersburg) of those approached to participate refused to accept the coupon from participants. Of a total of 450 coupons distributed in Tirana, 210 were returned. In St. Petersburg, 200 out of a total of 629 coupons distributed were returned. Furthermore, in Tirana, only two potential participants were found to be ineligible for participation and once consent had been given, no participants withdrew from the process. This information is not yet available for St. Petersburg.
Results
Demographic Characteristics
Table 1 describes the characteristics of the study samples in Tirana and St. Petersburg. The sample in Tirana consists predominately of male IDUs, with only 14 female IDUs recruited for the survey. This contrasts substantially from the sample in St. Petersburg, where more females were represented compared to Tirana, with 36% being female and 64% being male.
Table 1Demographic information AlbaniaRussiaSex Males91% (86, 96)64% (55, 73) Females9% (4, 14)36% (27, 45)Education completed Did not attend school9% (5, 16)0% Primary17% (11, 26)0% Secondary/vocational67% (58, 76)84% (77, 92) University7% (3, 8)16% (8, 23)Age 15–22 years38% (33, 48)11% (7, 15) 23–27 years36% (27, 41)38% (29, 47) 28 years or more26% (19, 32)51% (42, 62)Current co-habitation situation Married living with spouse12% (6, 18)9% (4, 15) Married living with other sex partner0%4% (2, 9) Married not living with any partner3% (1, 6)4% (1, 7) Not married, living with sex partner20% (14, 26)33% (25, 42) Not married, not living with sex partner65% (58, 72)49% (39, 60)The numbers in parentheses represent the 95% confidence intervals as calculated by RDS.
In Tirana, the majority had completed at least middle school (eight classes), with a third of the sample completing middle school and a third completing high school (12 classes). Only 7% in Tirana had completed university. In St. Petersburg, the majority (84%) had completed either secondary or vocational education, and 16% had completed university. In Tirana, the sample was comprised of young IDUs with 38% between 15 and 22 years of age and 74% below 27 years of age. The St. Petersburg study included an older IDU population, with over half of participants 28 years of age or older and only 11% between 15 and 22 years of age. In St. Petersburg, males were slightly older than females; a much higher proportion of males were 28 years of age or older (66 vs 26% of females; data not shown).
The majorities of the samples are unmarried (85% in Tirana and 82% in St. Petersburg). In Tirana, the majority also did not live with a sex partner (65%), while another 20% were unmarried but living with a sexual partner. A higher proportion of IDUs in Russia reported being unmarried but living with a sex partner (33%).
Network Sizes
As a part of RDS methodology, respondents were asked about the size of their personal networks. Specifically, participants were asked how many people they know personally who used injecting drugs (i.e., “you know who they are, they know who you are, you have seen them in the last six months”).
In Tirana, male IDUs reported a larger average personal network size than female IDUs. In contrast, in St. Petersburg, females had a larger average personal network size than males (20 vs 17, respectively).
The pattern of network size by age is again very different between IDUs in Tirana and St. Petersburg. In Tirana, network size increases with age (from 16 among the youngest age group to 21 in the oldest age group), whereas in St. Petersburg, network size decreases with age (from 25 in the youngest age group to 15 in the oldest age group). The same trend and differences between the two cities is observed with regard to education, which is likely due to the collinearity between age and education.
Similarly, in Tirana, there was a trend of increasing network size with increasing duration of injection drug use. However, in St. Petersburg, the network size was larger for recent initiates and long-term injectors compared to those who injected for 1–5 years (Table 2).
Table 2History of injection drug useDuration of injection drug useAlbaniaRussia<1 year21% (11, 24)6% (3, 9)1–5 years72% (69, 83)43% (31, 51)>5 years6% (3, 11)52% (43, 63)
Recruiting
Table 3 represents the recruitment patterns of IDUs in Tirana and St. Petersburg with respect to sex, age and length of injection. RDS provides the opportunity to learn about the degree to which people tend to affiliate with others sharing their same characteristics. In Tirana, the gender affiliation patterns demonstrate a strong heterophily in women, whereby women recruited only men (H = −1.00), while men recruited women 7% of the time (H = 0.16; this is probably due to the small number of women in the Tirana sample, with only 14 women participating). From a possible 28 coupons given to these women, 16 coupons were returned (57%). In contrast, the gender affiliation patterns of IDUs in St. Petersburg reflects neither strong homophily nor heterophily. Both males and females recruited males about the same percentage of times (63 and 57%, respectively). The heterophily index in males was −0.02 and homophily index was 0.12 for females.
Table 3.Recruitment patterns by age, sex, and duration of drug injection in Tirana and St. PetersburgRecruitersRecruiteesTiranaSt. PetersburgYoungest (15–22 years)Middle (23–27 years)Oldest (≥ 28 years)Youngest (15–22 years)Middle (23–27 years)Oldest (≥ 28 years)Age Youngest (15–22 years)42%35%23%10%54%36% Middle (23–27 years)33%38%29%19%36%45% Oldest (≥28 years)29%39%32%12%44%44%SexMaleFemaleMaleFemale Male93%7%63%37% Female100%0%57%43%Duration of drug injectionNew (<1 year)Mid-term (1–5 years)Long-term (>5 years)New (<1 year)Mid-term (1–5 years)Long-term (>5 years) New (<1 year)20%64%16%0%43%57% Mid-term (1–5 years)8%57%35%7%40%54% Long-term(>5 years)4%65%32%6%40%54%
The affiliation patterns of age recruiting relations in IDUs in Tirana reflects neither homophily nor heterophily. The homophily indices were 0.02 for the younger (15–22 years), 0.05 for the medium age (23–27 years) categories, and 0.09 for the older age group (≥28 years). As can be seen, both youngest and the oldest age categories recruited the medium age group approximately one-third of the time. Further, both the youngest and the medium age groups equally recruited the oldest age group (approximately 23–29% of the time). Lastly, the medium and the oldest age groups recruited the youngest group with equal vigor at about one-third of the time.
Among IDUs in St. Petersburg, the affiliation patterns of age relations reflect neither homophily nor heterophily. The homophily indices were −0.03 for both the younger (15–22 years) and the medium age (23–27 years) categories, and −0.15 for the older age group (≥28 years).
In Tirana, as with the relations by age, the affiliation patterns of relations based on duration of injection drug use reflects neither homophily nor heterophily. The homophily indices were 0.09 for new injectors (i.e., <1 year of injection experience), −0.10 for the mid-term users (1–5 years), and 0.09 for the longer-term injectors (>5 years). New users were recruited the least by the mid- and longer-term injectors (8 and 4%, respectively), and both the new and longer term injectors recruited mid-term injectors (64 and 65%, respectively).
In St. Petersburg, however, the affiliation patterns of relationships by duration of injection drug use differed between new injectors and longer-term injectors. The group of new injectors was completely heterophilous (H = −1.0), reflecting a lack of in-group ties. However, both mid-term and longer-term injectors exhibited neither heterophily nor homophily (H = −0.07 and H = 0.05, respectively).
Discussion
These two studies are among the first to report on the level of sexual and drug injection risk behaviors and knowledge about HIV prevention in a probability sample of IDUs in Tirana and St. Petersburg using RDS. The use of a systematic peer recruitment method afforded the possibility of capturing a wider range of IDUs than studies recruiting with outreach workers or from specific venues, such as STI clinics or needle exchange programs. Additionally, these two studies shed light on the social context by identifying social network relations among the IDUs.
The recruitment analysis suggests that IDUs in Tirana form socially distinct groups by sex. Female IDUs do not seem to interact with each other; no female IDU referred another woman. This suggests that intravenous drug use amongst females may be a highly stigmatized activity, or it may reflect their initiation into drug use by men.
In contrast, IDUs in St. Petersburg do not form socially distinct groups with regard to sex. The lack of strong ties within sex groups may be a reflection of injection behaviors in that they often inject in group settings which are likely to be of mixed sex and share needles with members of the opposite sex.
In Tirana, the lack of strong ties both within and between age groups most likely indicates that younger IDUs are probably not being initiated into injection drug use by injectors who are older in age. However, in St. Petersburg, there were no intra-group ties among younger females or among newer IDUs. All new IDUs appeared to be recruited only by long- and mid-term IDUs, and all younger female IDUs were primarily recruited by older male and female IDUs. This suggests that new IDUs, particularly female IDUs, have been initiated into injection drug use by older IDUs. The high prevalence of HIV and STIs5–7 in the IDU population in Russia places these new injectors at extremely high risk for HIV acquisition. The differences in average network sizes of male and female IDUs between Tirana and St. Petersburg may have implications for prevention programs. The larger network size of male IDUs compared to female IDUs in Tirana is not surprising given that the IDU population in Albania appears to be predominately male and clandestine for women. Therefore, it is apparently rare for female IDUs to know other female IDUs. In contrast, in St. Petersburg, female IDUs on average had a larger network size compared to male IDUs, which may be a reflection of the overlap of this population and the sex worker population. A sizable proportion are engaged in sex work so they likely know other sex workers from working in the same areas.
The younger age of IDUs in the Albanian sample indicates a potential for longer term drug use, allowing for greater transmission. Additionally, continued migration between borders with neighboring countries where prevalence amongst high-risk groups is higher (i.e., to Greece, Italy or Eastern European countries) may create a transmission–migration effect.8
The male-to-female ratio of almost 2:1 in St. Petersburg is consistent with the decreasing trend in the male-to-female ratio of IDUs previously reported, from 7:1 in 1991 to 5:1 in 2001.9 However, it is also possible that previous studies may have underestimated the proportion of female IDUs. RDS may reach female IDUs better than other sampling methodologies. This could be evidence that RDS can capture a more diverse and possibly a more representative sample of IDUs in St. Petersburg.
The Tirana sample is proportionately younger than the St. Petersburg sample, perhaps describing an epidemic that is newer or reflecting the population structure in Albania where 40% is under 25 years of age.10 Records indicate an increase in substance abuse treatment since 2000 and the first HIV case was detected in Albania in 1993—lending credence to the possibility of a younger epidemic.11
The co-habitation situation in Tirana differs from that in St. Petersburg, with a far greater proportion in St. Petersburg unmarried but living with a sex partner. It is possible that the more conservative nature of Albanian society and a holdover from the Albanian communist regime policies are reflected in this indicator.12 The finding in St. Petersburg that all new injectors were recruited only by long- and mid-term injectors is of concern. These social relations combined with the high prevalence of HIV in the IDU population in Russia place these new injectors at extremely high risk for HIV acquisition. This scenario is likely exacerbated by their large average network size (approximately 18).
The research conducted in Albania with IDUs is of vital importance. For the first time, policy makers and those involved with drug treatment and prevention of HIV and STIs have information to better plan and target interventions.
The pace of recruitment in both countries (approximately 8 weeks) seems to be fairly efficient. Differing numbers of coupons (two in Tirana and three in St. Petersburg) did not seem to affect overall recruitment with both studies reaching the desired sample size in approximately the same amount of time. In St. Petersburg, slow recruitment particularly at the fixed site at the city center was boosted by the addition of two seeds at week 6; participation of only 50 IDUs at week 6 increased to 200 by the eighth week. Over 200 IDUs were reached by the study in 7–8 weeks with 13–15 seeds in each city.
The body of literature regarding the use of RDS amongst hard-to-reach populations in developing countries is still in its nascence. Little is known about the effect of incentives on populations in developing countries or if random recruiting of peers (a fundamental assumption underlying the RDS methodology) is valid in countries where a large proportion of the population is poor. A greater body of research employing these methods in developing countries is needed.
Lessons learned from RDS methodology
Data collection during the summer was ideal in St. Petersburg given that some sites were mobile vans and participants had to wait outside. The season was also beneficial in Tirana, with data collection taking place before summer vacation. Given the large size of both cities, having three study sites in different parts of each city facilitated recruitment. Evening and Saturday hours were important to ensure that workers could participate. In St. Petersburg, upon recommendation from local NGOs, letters of support obtained from city police officials helped ensure that IDUs would not be harrassed at the study sites. In Tirana, biometrics were useful in creating a disincentive to participate multiple times. Involving NGO staff to recruit seeds and conduct data collection in the NGO offices enabled participants to feel comfortable about the process.
Limitations
Possible limitations to the study could have affected the results. In Tirana, the small numbers of women participating in the surveillance may indicate a strong desire to remain hidden, their limited numbers, or a reflection of poor recruiting. Because few women have been arrested or attended treatment facilities, there are no data regarding intravenous drug use amongst women in Tirana, or anywhere in Albania.11 In both cities, it is possible that RDS may have selectively attracted those IDUs desiring the incentives, which may have biased our sample. Formative research was conducted to assess if the amount would be coercive, so this is unlikely.
The major limitation of RDS is that it cannot control for non-response bias. Barriers to participation due to unknown characteristics of certain sub-groups or to factors beyond the control of the project could occur and are not measurable. Both studies attempted to determine the rate of non-response by asking participants returning to claim recruiting incentives about the number of attempts made before their coupons were accepted and the demographic characteristics of those who refused. They were also asked characteristics of those accepting the coupons. It is hoped that forthcoming analysis can assess if there is a difference between those who refuse, those who accept but do not show up, and those who accept and participate.
The coupon return rate in St. Petersburg was considerably lower than that of Tirana most likely due to the greater number of coupons given out by each participant in St. Petersburg (three coupons) than in Tirana (two coupons).
Conclusion
Because IDU participants were recruited through RDS, it is likely that the study sampled from a wide range of IDUs thereby increasing the sample representativeness in these two cities. Our study methodology likely drew a more diverse sample of IDUs compared to traditional recruitment by outreach workers or through needle exchange programs.
The studies in Tirana and St. Petersburg greatly enhance our understanding of IDUs and their network characteristics. They also provide evidence for the effectiveness of using RDS in recruiting hard-to-reach populations in environments where stigmatization and criminalization against their behaviors are commonplace.
|
[
"injection drug users",
"albania",
"respondent driven sampling."
] |
[
"P",
"P",
"R"
] |
Pediatr_Nephrol-3-1-1805049
|
Transient severe metastatic calcification in acute renal failure
|
Metastatic calcification, a known complication of prolonged end-stage renal disease, is herein described for the first time in a 10-month-old boy with acute renal failure, manifesting as a painful and swollen arm. Imaging revealed diffuse calcification and technetium-99 methylene diphosphonate (99Tc-MDP) uptake around the humerus and axilla. Calcium and vitamin D restriction, followed by intravenous administration of sodium thiosulfate caused a full symptomatic, radio- and scintigraphic improvement.
Introduction
Vascular (“metastatic”) calcification (MC) (or calcific uremic arteriolopathy) is the most common type of extra-osseous calcification in end stage renal disease, manifesting as both medial and intimal calcification of large arteries [1]. It usually results from a deposition of calcified products in otherwise normal tissues as a result of hyperphosphatemia with or without hypercalcemia. It may affect the visceral organs, such as kidneys, lungs, and gastric mucosa, as well as joints, eyes, and skin [2]. The predisposing conditions of hyperphosphatemia and high (calcium×phosphorus) concentration products need to be persistent over weeks and months, since MC has been not been described in the setup of acute renal failure (even though a similar biochemical profile is seen there too).
We describe a case of MC in a 10-month-old child presenting with hemolytic uremic syndrome (HUS) and acute renal failure (ARF).
Case report
A 10-month-old male infant was recovering from his first episode of atypical HUS. He belonged to a Bedouin tribe previously diagnosed by us [3] with recessive factor H deficiency. His clinical course was remarkable for microangiopathic hemolytic anemia, oliguric renal failure and hypertension with a depressed myocardial function, associated with resistant hypocalcemia (total calcium as low as 5.4 mg/dl) and hyperphosphatemia (as high as 13 mg/dl). The latter was eventually controlled with phosphor binders (aluminum hydroxide at first for several days and, later, calcium gluconate), intravenous administration of calcium (mainly due to an episode of cardiac arrest without hyperkalemia, as previously described for HUS [4]), low phosphorus diet and alphacalcidol, 0.25 μg q.d. (the latter only after serum phosphorus had been normalized). Medications were given through a right femoral vein catheter and peripheral veins in the right arm. Peritoneal dialysis was started, using regular calcium (3.5 meq/l) dialysate. Fresh frozen plasma (20 ml/kg) was given daily until there was evidence of HUS remission, and was then switched to 20 ml/kg twice a week, as previously done in homozygous factor H-deficient children [5].
Ten days after hospitalization, he was found to have a painful and mildly swollen left arm, mainly in the elbow area, with flexion limitation. There was no intravenous device in this arm. There were no high fever, local redness or swollen joints. The plain X-ray did not reveal any lesions or fractures. A week later the left shoulder and humeral area were more swollen and tender, with no specific nodule or bone deformity. On physical examination, the swelling was deeper than the dermis (mainly in the humeral area) and painful. The superficial skin did not show signs of ischemia, necrosis, ulceration or livedo reticularis. This time the plain X ray revealed diffuse calcification, involving mostly the arm muscles and the axilla (Fig. 1a). Calcium and phosphorus levels were 9.9 mg/dl and 6.6 mg/dl, respectively. Renal function remained stable but impaired (serum creatinine and urea concentrations were as low as 1.05 mg/dl and 90 mg/dl, respectively). Parathyroid hormone (PTH) levels were normal (25.7 pg/ml). Renal sonogram showed no evidence of nephrocalcinosis. A bone scan revealed significant technetium-99 methylene diphosphonate (99Tc-MDP) uptake along the areas that had seemed to be affected on the X-rays and physical examination (Fig. 1b). Treatment with peritoneal dialysis, alphacalcidol and calcium salts was stopped. Ketoconazole 1 mg/kg per day was given for 3 days, in addition to prednisolone 2 mg/kg per day for a week. Renal function improved, but the (Pi×Ca) products remained above 60 mg2/dl2. No improvement in the arm lesion was seen. On day 25 of hospitalization, intravenously administered sodium thiosulfate, 100 mg/kg q.d., was started for 12 days. A gradual symptomatic improvement in arm range of movement and resolution of local pain was seen. At follow-up 8 weeks afterwards, X- ray and bone scan revealed significant resolution of MC (Fig. 2). At long-term (>12 months) follow-up, no recurrence of MC was seen, in spite of recurrent events of HUS, eventually needing periods of peritoneal and hemodialysis.
Fig. 1Chest X-ray on day 18 (a) black arrow calcification on the axilla, arrowheads calcification lines along the humerus. Tc-99m MDP scintigraphy (b) shows increased abnormal uptake in the soft tissue of the left shoulder, humerus and left scapula. Kidneys are enlarged, with increased diffuse abnormal radiotracer activityFig. 2Repeated X-ray (a) and bone scan (b) after 1 month show no abnormal findings
Discussion
The unique MC features in our patient were its appearance early after the diagnosis of renal failure and its transient nature. Evidence for such early appearance of this complication comes only from an animal model of ARF (adenine-induced renal failure in rats), where microscopic metastatic calcification in aorta, coronary artery and other soft tissues were found after 4 weeks of uremia [6]. This was associated with hyperparathyroidism, hypocalcemia and hyperphosphatemia. In addition, a high index of suspicion may detect the MC early in the renal failure course, before it becomes ulcerated [7].
Current sophisticated imaging techniques suggest an incidence of subclinical or asymptomatic MC in 20–100% of patient with end-stage renal disease (ESRD) [8]. MC with soft tissue involvement in acute renal failure, to the best of our knowledge, has not been reported in children: using the PubMed and MESH search systems, and the terms “calcinosis” and “kidney failure, acute” (the term “calciphylaxis” is not an official indexing term), we found only three case reports of adults with such combination. There was a combination of risk factors for MC in our patient, including high dose parenteral calcium administration via both intravenous and peritoneal dialysis; high serum phosphorus levels due to renal failure and the hemolytic state; and the use of alphacalcidol [9]. These may have led to MC so early in the course of his disease. The possible correlation between renal failure due to HUS, a state where the basic damage is to the endothelium in small vessels, and MC is not supported by the literature search. Uremic calcification was thought to be a largely physico-chemical process facilitated by elevated Ca×P (i.e., “metastatic” calcification). Evidence now shows that vascular smooth muscle cells actively take up phosphate to form bio-apatite. This process is associated with a phenotypic transformation of vascular smooth muscle cells during which they express osteoblast markers [10].
Soft tissue accumulation of Tc99m MDP may be due to passive localization of tracer in slow fluid spaces (ascites, pleural effusion) when the blood concentration of tracer is high and these spaces do not clear as rapidly as the blood pool. Tracer may also bind in necrotic tissues (myositis, myonecrosis). Other causes of metastatic calcification include hypercalcemia from widespread destruction of bone (from metastatic cancer), parathyroid neoplasm and hypervitaminosis D. When the solubility product of calcium and phosphate is exceeded, there is precipitation of calcium in the extracellular space. Soft tissue calcification is, thus, another cause of soft tissue uptake of Tc99m MDP [11].
Reported therapeutic strategies for treating and preventing MC include: increasing dialysis dose, lowering serum calcium phosphorous and Ca×P solubility products, and avoiding calcium-based Pi binders and vitamin D analogs. Unfortunately, these methods have been ineffective in some patients with MC. Intravenously administered sodium thiosulfate (ST) increases the solubility of calcium deposits [12]. ST was shown to be successful in the treatment of both nephrolithiasis and tumoral calcinosis. ST has antioxidant effects on endothelial cells, but its exact mechanism of effect is unclear. It undergoes mainly renal clearance. Its half-life is increased from 15 min in the normal glomerular filtration rate (GFR) to 478 min in dialysis-dependent patients. The recommended adult dose is 12.5–25 g at the end of dialysis. The duration of this treatment is unclear. In our patient, a short ST course plus the withdrawal of enteral calcium and vitamin D were sufficient to resolve MC, which did not recur in the following events of HUS.
|
[
"sodium thiosulfate",
"calcific uremic arteriolopathy",
"hemolytic uremic syndrome",
"peritoneal dialysis",
"radionuclide imaging",
"alfacalcidol"
] |
[
"P",
"P",
"P",
"P",
"M",
"U"
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.