text
string | predicted_class
string | confidence
float16 |
|---|---|---|
We hypothesized that acquired deficiencies in ATM and/or MRE11 would lead to a dysfunctional repair mechanism and an inability to restore genomic stability, resulting in increased tumor cell death following radiotherapy. At least in the relatively small, but well characterized subgroup of microsatellite unstable CRCs with deficiencies in DNA mismatch repair (MMR) proteins, the accumulation of errors during replication and recombination presents a mechanism for inactivation of ATM and MRE11 through function-ablating mutations. As a corollary, we proposed that ATM and MRE11 might be used as a predictive marker of radiosensitivity in rectal cancer patients, which would correlate with improved patient outcomes.
|
study
| 100.0 |
Thus, we investigated the value of expressions of MRE11, ATM and combined MRE11/ATM as a predictive marker of response to radiotherapy in rectal cancer and as a prognostic marker in general, and performed univariate and multivariate analyses on various clinicopathologic factors according to their expression levels.
|
study
| 100.0 |
This study received ethics approval from the South Western Sydney Local Health District (SWSLHD) Human Research Ethics Committee (reference number HREC/12/LPOOL/102). Specifically, the research was conducted under the Protocol numbers X01-0138 and X03-0291 for the IHC assessment in paraffin embedded rectal cancer tissues and retrospective survival analyses. The ethics committee waived the need for direct consent for patients as samples and clinical data were all de-identified before access.
|
other
| 99.94 |
Specimens were obtained from 262 rectal cancer patients who underwent surgery in the SWLHD from 2000–2011. Clinicohistopathological data were collated and are summarized in Table 1. Standard radiotherapy treatments were at a dose of 25 Gy in 5 treatment fractions when performed alone or at 50.4 Gy in 28 treatment fractions when combined with 5 fluoro-uracil based chemotherapy. Short-term response to radiotherapy was measured by tumor regression grade (TRG) per the 7th edition of the American Joint Committee on Cancer (AJCC) manual on a scale of 0 to 3: 0, complete response with no viable malignant cells; 1, moderate response with single or small groups of malignant cells; 2, minimal response with residual malignancy outgrown by fibrosis; 3, poor response with extensive residual malignancy. TRG 0, 1 and 2 were categorized as responders and TRG 3 as non-responders. Long-term response to radiotherapy was measured by survival outcomes including DFS and OS. Specimens obtained from primary surgery for rectal or rectosigmoid cancers were obtained from the South-Western Area Pathology database, Australia. Surgery consisted of total mesorectal excision, with anterior or abdominoperineal resection. Variables of interest included age, gender, pathological TNM stage, grade, vascular invasion, perineural invasion, tumor-infiltrating lymphocytes, and treatment. Outcomes of interest were DFS, OS and histologic TRG in the resected bowel for cases treated with neoadjuvant chemoradiation. DFS was defined as the time from diagnosis to first recurrence, if any. OS was defined as the time from diagnosis to last follow-up or death. Follow-up consisted of regular clinic visits, colonoscopy, blood tests, and imaging at the discretion of the treating specialist.
|
study
| 99.94 |
For each patient, donor blocks of archived paraffin-embedded tissue were retrieved and two 1mm diameter cores were obtained from each sampling site: tumor center (TC), tumor periphery at invasive edge (TP), normal mucosa close/adjacent to tumor, normal mucosa well away from tumor, usually at the resection end margin, and from involved lymph nodes (LN). These were transferred into pre-drilled wells in a tissue microarray block and mounted on slides for immunohistochemistry.
|
study
| 99.94 |
Slides were deparaffinized in xylene and rehydrated in graded ethanol. Antigen retrieval was performed with Envision™ FLEX Target Retrieval Solution, pH 9.0, in a 98°C water bath for 10 minutes for ATM and 45 minutes for MRE11. This was followed by incubation for 5 minutes at room temperature with Envision™ FLEX Peroxidase-Blocking Reagent to block endogenous peroxidases. Slides were incubated with monoclonal anti-ATM primary antibody [2C1 (1A1)] (1:800 dilution, Abcam, Cambridge, UK) for 30 minutes at room temperature and with anti-MRE11 primary monoclonal mouse antibody (1:600 dilution, ab214, Abcam). After washing with TBST buffer, slides were incubated for 15 minutes with Dako mouse linker, rinsed, and incubated for 30 minutes with anti-mouse secondary antibody. The peroxidase substrate used was a mixture of Envision™ FLEX DAB+Chromogen DM827 and Envision™ FLEX Substrate Buffer DM823 (DAKO). Slides were counterstained with hematoxylin, washed with cold water, and then dipped 10 times in Scott Bluing solution. Slides were immediately rinsed with cold water before dehydration and mounting.
|
other
| 99.8 |
ATM expression was initially scored as the product of percentage and intensity of staining based on Angele et al . The percentage of positive staining was scored as: 1, <25%; 2, 25%–50%; 3, 50%–75%; or 4,>75%. The intensity of staining was graded as: null, 0; low, 1; moderate, 3; or high, 5. The two scores were multiplied to create a composite score between 0 and 20 and analyzed against clinicohistopathological and clinical outcome data. The percentage of MRE11-positive cells and staining intensity were also scored using the methods published previously . Intensity was graded as: negative, 0; weak, 1; moderate, 2; or strong, 3. The percentage of positive cells was graded as: 0, <5%; 1, 5%–25%; 2, 26%–50%; 3, 51%–75%; and 4, >75%. These two measures were multiplied to give weighted scores ranging from 0–12. To combine the two proteins as a biomarker panel, all cases were categorized into either a low expression group (score range: 0–5) and a high expression group (score range: 6–32).
|
study
| 100.0 |
Statistical analysis was performed with SPSS Statistics for Windows 20.0 (Chicago, IL, USA). ATM expression was compared with clinicohistopathological data using Pearson’s χ2 test, and the association of MRE11 expression with clinicohistopathological variables was assessed by Fisher's exact test. ATM and MRE11 expression levels were compared and combined by binary logistic regression as described previously . Survival analyses were performed in the overall cohort and separately in patients who received preoperative radiotherapy. Univariate and multivariate analyses were performed using Kaplan Meir curves and Cox’s proportional hazards survival modeling for the combined two-marker expression levels from cancer core and periphery samples. Covariates were sex, age, TNM stage, grade, vascular invasion, perineural invasion, treatment with chemotherapy and radiotherapy, and TRG. Univariate analysis by the Mann–Whitney U test was also used to assess associations between the single and combined two-marker expression levels in rectal tumor tissue with TRG, which was further characterized with receiver operating characteristic—area under curve (ROC-AUC) analysis. P<0.05 was considered statistically significant.
|
study
| 100.0 |
Patient characteristics are detailed in Table 1. The patients’ median age was 71 years (range: 35–100 years). Among 262 patients, 174 (66.4%) were male and 88 (33.6%) were female. Seventy six of 245 (31.0%) patients were treated with radiotherapy, of which 54 of 76 (71.1%) received preoperative therapy. Disease-free and OS data were available for 211 and 248 patients, respectively. Patients were followed for a median period of 3.2 years (range: 0−12.6 years). Local recurrence of disease occurred in 82/211 (38.9%) patients. The median time to recurrence was 2.12 years. At the time of the study, 141/248 (56.9%) patients were alive. The median time to death was 2.5 years following surgery (range: 0−11.1 years). Among patients who received preoperative radiotherapy, local recurrence occurred in 6/54 (11.1%) patients and median time to recurrence was 2.61 years (range: 0.75−4.29 years). All six patients (100%) had died by the end of this study. The median time to death following recurrence was 3.81 years (range: 0.6−10.9 years).
|
study
| 99.94 |
ATM and MRE11 protein expression levels in the TC were tested in a forward and reverse binary logistic regression analysis using a data set of immunohistochemical scoring derived from 257 tumor samples and 255 normal tissues. The final biomarker model gave an average receiver operator characteristic area under the curve (ROC-AUC) value of 0.849 for the combination of the two proteins. Similarly, ATM and MRE11 protein expression levels in the TP (tumor, n = 258; normal, n = 255) were also tested, and the model gave an average ROC-AUC value of 0.837. The sensitivity and specificity of the two-protein combined panel were 80.9% and 70.3% for TC and 61.6% and 48.8% for TP, respectively, and overall accuracy was 75.6% (TC) and 55.2% (TP) (S1 Table).
|
study
| 100.0 |
We examined the association of ATM and MRE11 expression levels with clinicopathological variables independently. ATM protein expression was analyzed in 259 central and 260 peripheral tumor cores. Samples from TC displayed high expression in 84% and low expression in 16%. Samples from TP showed high expression in 79% and low expression in 21%. There was a higher level of ATM expression in TC (mean (M) = 6.84) compared to TP (M = 5.28). This difference was found to be significant (P<0.001).
|
study
| 100.0 |
ATM expression was seen in 199/228 (87%) of adjacent normal mucosa and 204/250 (82%) of distal normal mucosa. ATM expression was higher in normal mucosa taken near the tumor (M = 6.43) compared to samples taken away from the tumor (M = 4.25). Again, this finding was statistically significant (P<0.001). ATM expression in both TC and TP were significantly higher than in distal normal mucosa (P<0.001 and P = 0.005 respectively), but not significantly different to that in the adjacent normal mucosa.
|
study
| 100.0 |
Low ATM expression in the TP was associated with older age (P = 0.013) and higher histologic grade (P = 0.044).(S2 Table). There was no correlation with sex, TNM category, vascular invasion, or perineural invasion. Expression of ATM in TC was not associated with any clinicohistopathological variables.
|
study
| 100.0 |
MRE11 expression in the TC was high in 45% of cases and low in 52%. Expression in the TP was high in 56% and low in 44%. Mean weighted scores were significantly different between TC and TP (5.5 versus 5.8, P<0.05 by paired t-test). The mean MRE11 score was 4.2 for both adjacent and distal normal tissue, which was significantly different from the score for both the TC and TP (P<0.001). There were no significant associations between MRE11 score in the TC or TP and clinicopathologic characteristics (S3 Table).
|
study
| 100.0 |
Next, we examined a possible association of clinicopathological characteristics and survival outcomes with the ATM/MRE11 two-protein combined panel. Fig 1A shows representative immunohistochemical staining of high and low/absent ATM and MRE11 expression in rectal cancer tissues. Results from the combined marker analysis differed from our initial single biomarker studies of ATM and MRE11 alone, in that a high combined expression of ATM and MRE11 was significantly associated with a number of clinicopathological variables. These included neoadjuvant status (P = 0.001 for TC), TRG (P = 0.03 for TC, P = 0.011 for TP), age (P = 0.02 for TP), and nodal stage (P = 0.042 for TP) (Table 2).
|
study
| 100.0 |
(A) Representative immunohistochemical staining of ATM and MRE11 in rectal cancer tissues. Bar = 200 μm. Staining for each protein was scored as high or low, as described in the Materials and Methods. (B) Overall survival according to combined protein expression levels of ATM and MRE11 was determined by Kaplan−Meier survival analyses and compared using the log-rank test. Patients with high combined protein expression levels of ATM and MRE11 in the TC (green line) showed significantly worse OS than those with low expression (blue line; P = 0.003). (C) Similarly, patients with high expression levels in the TC (green line) exhibited worse disease-free survival than those with low expression (blue line; P = 0.035). (D and E) When measured in the TP, no significant survival difference was seen between the high- and low-expression groups for overall survival (P = 0.208) or disease-free survival (P = 0.748).
|
study
| 100.0 |
In Kaplan Meier survival analysis, a high score for the two-protein combined expression in the TC was significantly associated with worse OS (P = 0.003) and DFS (P = 0.035) (Fig 1B and 1C). However, no significant survival difference was seen between high and low groups for expression in the TP (OS, P = 0.208 and DFS, P = 0.748; Fig 1D and 1E, respectively). In univariate Cox regression analysis, a high two-protein combined status in the TC (combined TC high versus low: HR = 1.944, 95% CI 0.037–3.645, P = 0.038) was significantly associated with reduced DFS (Table 3). In multivariate Cox analysis (adjusted for combined expression of ATM and MRE11, and perineural invasion), the two-protein combination panel (HR = 2.178, 95% CI 1.115–4.256, P = 0.023) as well as perineural invasion (HR = 2.183, 95% CI 1.222–3.899, P = 0.008) remained significantly associated with DFS (Table 3).
|
study
| 100.0 |
As noted above, DFS of rectal cancer patients with overexpression of the two-protein combined panel was significantly worse than that of patients with lower expression. When patients were grouped according to LN involvement, high expression of the two-protein combined panel was associated with decreased DFS in patients with LN-positive tumors (P = 0.029; Fig 2B) but not in those with LN-negative tumors (P = 0.480; Fig 2A). By multivariate Cox analysis, expression of the two-protein combined panel in TC in the LN positive subgroup significantly correlated with DFS (HR = 3.474, 95% CI 1.054–11.451, P = 0.041) (Table 3).
|
study
| 100.0 |
(A) and (B) respectively show survival curves of high (green line) and low (blue line) ATM/MRE11 two-protein expression groups in lymph node (LN) negative and LN positive rectal cancers. These show the effect of LN status on the association between expression levels and disease-free survival (DFS). (C) and (D) respectively show overall survival (OS) and DFS survival curves of 54 patients who received preoperative radiotherapy in terms of high (green line) and low (blue line) two protein expression groups. Worse OS (P = 0.024) and DFS (P = 0.028) were seen with high expression.
|
study
| 100.0 |
The OS estimates in the subgroup with neo-adjuvant radiotherapy are shown in Fig 2C, demonstrating that higher combined expression level was significantly associated with worse OS (P = 0.024). Similarly, DFS of radiotherapied patients with overexpression of two-protein combined panel was significantly worse than that of patients with lower expression (Fig 2D, P = 0.028), implicating that the MRE11/ATM two-protein panel has specific potential as a predictive marker of tumour response to radiotherapy.
|
study
| 100.0 |
To investigate the relationship between TRG and tissue levels of ATM, MRE11, and the two-protein combined panel, univariate analysis was carried out using the Mann−Whiney U test (SPSS). A significant association was seen between increasing TRG and TC expression of the MRE11 protein (P = 0.015, Fig 3B) and of the combined two-protein panel (P = 0.011, Fig 3C). The discriminatory power of each protein biomarker and their combination in TC was characterized using a ROC-AUC analysis of good response (TRG 0–2) versus poor response (TRG3) groups. The average ROC-AUC was 0.745 for the combined panel of two proteins, compared with 0.618 for ATM and 0.711 for MRE11 proteins (Fig 3D). These results suggest that the combined two protein biomarkers provide good discrimination between good and poor tumor responses after radiotherapy.
|
study
| 100.0 |
(A−C) Box plots show levels of ATM, MRE11, and their combined expression in the TC categorized by TRG as 0–2 (good response) or 3 (poor response). The association between protein expression with TRG was examined by Mann−Whitney U test (ATM, P = 0.27; MRE11, P = 0.015; and combined proteins, P = 0.011). (D) Receiver operating characteristic (ROC) curve analysis comparing the performance of ATM and MRE11 alone with the combined 2-protein panel.
|
study
| 100.0 |
Despite intensive research into biomarkers and prognostic factors for CRC, little is known about the factors influencing the survival of patients with rectal cancer. Although there is overlap with the treatment of colon cancer in the metastatic setting, rectal cancers in the locoregional setting are managed differently from colon cancers . The 1997 Swedish Cancer Trial of patients with resectable rectal cancer showed that preoperative radiotherapy improved both short-term and long-term local recurrence rates and OS compared with surgery alone , although the authors noted that the use of radiotherapy was associated with side effects . Since then, the benefits of preoperative radiotherapy or chemoradiotherapy have been corroborated in several meta-analyses .
|
review
| 99.9 |
At the individual level however, there can be significant variation in the degree of initial tumour response to radiotherapy. In this context, the TRG score provides an indicator of the regressive response of a tumor to cytotoxic treatment such as radiotherapy. The AJCC Staging Manual (7th ed.) recommends consideration of the TRG independent of the TNM system. The TRG provides important prognostic information because complete or subtotal tumor regression is widely recognized to be associated with better patient outcomes . Unfortunately, the majority of patients show no response or only a partial response to neoadjuvant therapy. For example, none of the patients in our study showed a complete TRG response; 82.4% showed a poor response and 17.6% showed a minimal to moderate response. However, a meta-analysis performed by Lee et al indicated that even partial tumor regression after preoperative chemoradiotherapy improves DFS, and should be considered as a favorable prognostic factor .
|
study
| 99.94 |
Given the dismal prognosis of patients with rectal cancer and the variation in response to radiotherapy, a reliable biomarker of intrinsic tumor radiosensitivity would improve assessment of prognosis and aid in making proper treatment decisions. It would help to avoid adverse side effects of neoadjuvant radiotherapy in patients who are unlikely to benefit, and at the same time identify those who may subsequently benefit from intensification of therapy. Although several biomarkers have been studied as predictors of radiotherapy response , the most extensively studied are MRE11 and the radiosensitivity index (RSI), a multigene expression model of tumor radiosensitivity, which was developed to predict treatment outcomes independent of disease site . However, to date, these biomarkers have not been validated in randomized trials, and their clinical value remains unclear. For instance, Sheridan and colleagues found no correlation between MRE11 expression and survival or radiosensitivity in patients with CRC, albeit with the limitations of a small sample size and short follow-up period. In our study, MRE11 expression alone was not significantly associated with any clinicopathologic characteristics, while in terms of ATM only low expression in TP was associated with older age and higher histologic grade, whereas ATM expression in TC was not associated with any clinicopathologic characteristics.
|
study
| 99.94 |
Based on the hypothesis that both ATM and MRE11 are integral to the detection of DNA damage and subsequent intracellular signalling following radiotherapy, and hence their deficiency would equate to increased radiosensitivity, we established a two-marker panel of ATM and MRE11 expression by binary regression analysis of tumor samples and normal tissues. Subsequent analyses showed greater association with general clinicopathologic parameters than either marker alone, as well as yielding a ROC-AUC value of 0.745 in predicting poor histological tumour regression (i.e. TRG 3) following radiotherapy. The latter was superior to using ATM (0.618) or MRE11 (0.711) alone. Furthermore, in both the neoadjuvant radiotherapy sub-cohort and the overall cohort, the combined ATM/MRE11 expression levels correlated with clinical survival outcomes. Thus these markers, when used together, related to both early (histological tumour regression) and late (clinical survival) response to neoadjuvant therapy in rectal cancers. These findings support our hypothesis and are also consistent with the fact that inhibitors of ATM and the MRN complex have been shown to have potential as radiosensitizing agents .
|
study
| 99.94 |
It is evident in our results that there are differences in the expression levels and their association with clinicopathological variables and survival outcomes when comparing the results from the centre versus periphery of the tumour mass. Effects of the sampling site on associations with tumor protein expression have been noted previously. For example, in an evaluation of biomarkers in ovarian cancer, Permuth-Wey et al reported sampling variability in protein expression analysis using tissue microarrays . The authors concluded that, for ovarian cancer at least, more reliable data were obtained from the TP than from the TC. This was attributed to optimal exposure to fixatives at the periphery of the processed tissues, although the effect of tumor heterogeneity on sampling variation should also be considered. For instance, TCs in general are more liable to hypoxia and necrosis compared to TP, and ischaemic injury may be another potential source for the biological differences seen within the same tumour.
|
study
| 99.94 |
On a final note, the role of LN status in predicting outcomes is intriguing, and hints at the complexity of using biomarkers to predict patient outcomes. Among patients with LN-negative tumors, there was no significant difference in survival between patients with high vs. low two-protein combined panel score in TC. In contrast, in patients with LN-positive tumors, high two-protein combined panel scores in TC were significantly associated with worse survival. This suggests that the combined expression of ATM and MRE11 may be associated with LN involvement in relation to patient survival.
|
study
| 100.0 |
Overall, our findings suggest that the optimal management of rectal cancer requires tailored treatment based on biomarker expression such as the two-protein MRE11/ATM panel described here, as well as clinicopathologic characteristics. Further elucidation of biomarkers that predict the response to radiotherapy and/or chemoradiotherapy will constitute important progress toward the individualized treatment of patients with rectal cancer. In this regard, we are continuing studies on ATM and MRE11 in independent datasets focusing on pre-treatment specimens in comparison to post-treatment specimens.
|
study
| 99.94 |
The number of cancer cases, which are diagnosed each year continues to rise, primarily due to an aging population. According to a recent report by Weir et al. (2015), this increase, in the United States, in all races and all sites, will be of 24,1% among men and of 20.6% among women within 2020. Cancer, along with cardiovascular disease (CVD), remains the most common cause of death. Most of the cancer-related deaths are due to the metastatic process, which is regulated by different mechanisms including the interaction of cancer cells with other cellular components present either in the tumor microenvironment or in the bloodstream (De Palma et al., 2017). Emerging evidence has recognized a central role for platelets in both tumor progression and metastasis (Contursi et al., 2017) and paraneoplastic thrombocytosis is observed in more than 30% of subjects diagnosed with different types of solid tumors, where it is associated to poor prognosis (Haemmerle et al., 2017).
|
review
| 99.9 |
Interestingly, it has been pointed out that low-dose aspirin (75–100 mg), which mainly targets platelets, reduces the incidence and mortality of colorectal cancer (CRC) and other types of solid tumors (Rothwell et al., 2010, 2011, 2012; Patrignani and Patrono, 2016).
|
review
| 99.8 |
In updating its recommendations, the US Preventive Services Task Force stated that low-dose aspirin should be used for the primary prevention of CVD and CRC in adults aged 50–59 years “who have a 10% or greater 10-year CVD risk, are not at increased risk for bleeding, have a life expectancy of at least 10 years, and are willing to take low-dose aspirin daily for at least 10 years”(Bibbins-Domingo and U.S. Preventive Services Task Force, 2016).
|
review
| 99.6 |
The antiplatelet effect of low-dose aspirin is dependent on the preferential inhibition of platelet cyclooxygenase (COX)-1 which translates into a virtually complete inhibition of thromboxane (TX) A2, i.e., a potent lipid mediator which acts as an amplifier of the response to primary platelet agonists, such as thrombin and collagen.
|
other
| 69.44 |
Another secondary agonist for platelets is adenosine diphosphate (ADP) which is released from platelet dense granules in response to the primary agonists (McNicol and Israels, 1999) (Figure 1). Adenine nucleotides and nucleosides [adenosine triphosphate (ATP), ADP and adenosine] modulate platelet aggregation, shape change and the release of alpha granules (Enjyoji et al., 1999; Burnstock, 2017). The role of ATP, ADP, and adenosine in platelet function is controlled by an organized enzymatic chain, including ecto-nucleoside triphosphate diphosphohydrolase (CD39) and ecto-5′-nucleotidase (CD73), which catalyzes the hydrolysis of released ATP into ADP, adenosine monophosphate (AMP) and finally to adenosine (Bakker et al., 1994; Robson et al., 2006; Burnstock, 2017). Extracellular nucleotides act on platelets through distinct receptors belonging to the P2 ATP family: the P2Y1 and P2Y12 metabotropic, G protein-coupled receptors, involved in transient platelet shape change and platelet aggregation (Burnstock, 1972) and the P2X1 ionotropic receptor (Burnstock, 1972; Cattaneo et al., 2002; Oury et al., 2004; Gachet, 2012) (Figure 1).
|
review
| 99.44 |
Schematic overview of the contribution of the three P2 purinergic receptor subtypes, P2Y1, P2Y12, and P2X1 in platelet activation. P2Y12 receptor plays a central role in ADP-mediated platelet activation. Platelet exposure to primary platelet agonists including thrombin, collagen and Von Willebrand factor (vWF) results in their adhesion followed by the release of ATP and ADP from dense granules. Extracellular ATP is rapidly metabolized into ADP, AMP and finally to adenosine by the ecto-enzymes CD39 and CD73. Secreted ATP and ADP, as well as ADP deriving from ATP degradation, activate P2Y1, P2Y12, and P2X1 receptors (as depicted). P2Y12 receptor stimulation modulates the growth and stability of thrombus by potentiating platelet dense granule release, platelet aggregation and procoagulant activity (Kunapuli et al., 2003).
|
review
| 99.9 |
The expression and function of P2Y12 in other cell types is still poorly investigated. P2Y12 congenital deficiency results in bleeding disorders characterized by a platelet impaired response to ADP (Cattaneo et al., 1992; Nurden et al., 1995; Cattaneo, 2011), but, in these patients, information is lacking on potential modifications induced in other tissues and organs.
|
review
| 85.9 |
In this minireview, the expression pattern in both normal and malignant cells and the signaling pathways of the P2Y12 receptor (P2Y12R) will be overviewed. The P2Y12R involvement in cancer development, progression, and metastasis, as well as the role of P2Y12R antagonists in these pathological processes, will be also discussed.
|
review
| 99.9 |
A full platelet aggregation in response to ADP occurs by the stimulation of P2Y1 receptor, followed by P2Y12 activation (Figure 1). P2Y1 is a Gq-coupled receptor that initiates ADP-induced platelet aggregation through the stimulation of phospholipase C and phosphatidylinositol-signaling pathway. P2Y12R is a seven transmembrane domain receptor. It mediates the inhibition of adenylate cyclase and, in turn, cyclic AMP (cAMP) production via the coupling to Gαi leading to impaired protein kinase A (PKA) activation and a subsequent inhibition of vasodilator-stimulated phosphoprotein (VASP), which restrains either secretory or adhesive events in platelets (Figure 1).
|
study
| 99.94 |
Vasodilator-stimulated phosphoprotein phosphorylation flow cytometry assay is used to monitor platelet responsiveness to P2Y12 targeted antiplatelet therapy (particularly in tailoring the treatment with the oral P2Y12 inhibitor clopidogrel) (Gachet, 2012; Fitzgerald and FitzGerald, 2013; Siller-Matula et al., 2013; Danese et al., 2016).
|
review
| 91.8 |
P2Y12R activation also recruits Gβγ subunits, causing phosphoinositide-3-kinase-(PI3K) dependent Akt phosphorylation and Rap1b activation, a key positive regulator pathway for the integrin GPIIb/IIIa. In this way, the sustained activation of P2Y12R contributes to thrombus stabilization. PI3Kβ isoform has been reported to be essential for ADP-induced TXA2 generation and platelet aggregation (Garcia et al., 2010) and to cooperate with PI3Kγ isoform in sustaining integrin activation (Cosemans et al., 2006; Schoenwaelder et al., 2007).
|
study
| 99.94 |
P2Y12R-Gi signaling leads positive regulation of other intracellular pathways including extracellular-signal-regulated kinase (ERK), myosin light chain kinase and Src family kinases as well as to membrane lipid shifts toward a pro-coagulant state such as phosphatidylserine and P-selectin exposure (Leon et al., 2003; Gachet, 2012). By acting on P2Y12R, ADP also contributes to the release of several agonists such as TXA2 (Cattaneo, 2015).
|
study
| 99.94 |
P2Y12R was originally found to be expressed only by platelets (Hollopeter et al., 2001), however further studies reported that it is functionally present in microglial cells, the resident immune cells of the brain, where it can play a role in their activation (Haynes et al., 2006). In vivo experiments confirm a role for P2Y12 in microglia. P2Y12-deficient mice showed a diminished early response to focal injury and microglia from these animals was much less responsive to purine nucleotides in terms of cell migration (Haynes et al., 2006). Recently, it was shown that ADP stimulation of microglia P2Y12R induced ERK1/2 and paxillin Ser83 phosphorylation, which play a role in the regulation of focal adhesions and actin cytoskeleton rearrangement (Lee et al., 2012). Moreover, in hippocampal slices, the receptor has been shown to stimulate process extension through the activation of integrin- extracellular matrix interaction (Ohsawa et al., 2010; Swiatkowski et al., 2016).
|
study
| 99.75 |
P2Y12R has also been shown to regulate migration of vascular smooth muscle cells (VSMCs). In these cells, ADP, through P2Y12-Gαi activation, inhibited cAMP/PKA signaling pathway resulting in cofilin dephosphorylation, actin disassembly and, as a consequence, an increase in VSMCs motility and migration (Niu et al., 2017).
|
study
| 99.94 |
A role for P2Y12R in inflammation and immune modulation has been recently reported (Wang et al., 2004; Diehl et al., 2010; Burnstock and Boeynaems, 2014; Cattaneo, 2015; Hechler and Gachet, 2015). Interestingly, it has been shown that platelets negatively affect the adoptive T cell therapy (ACT) in cancer by producing high levels of active TGFβ. Moreover, platelets are the only cell type known so far to constitutively express the TGFβ-docking receptor glycoprotein A repetitions predominant (GARP) which allows them to capture TGFβ from both other cells and the extracellular matrix. This platelet-specific TGFβ-GARP-axis seems to play a critical role by constraining the antitumor activity of T cell immunity (Rachidi et al., 2017). In B16-F1 melanoma-C57BL/6 mice, clopidogrel, a P2Y12R antagonist, in combination with aspirin made the ACT therapy highly effective compared to the control group, which received water. Indeed, most mice survived without relapse for more than 3 months (Rachidi et al., 2017).
|
review
| 98.1 |
Expression of P2Y12R in cancer cells has been poorly investigated. The receptor protein has been found in glioma and astrocytoma cells (Jin et al., 2001; Czajkowski et al., 2002; Burnstock and Di Virgilio, 2013) where it has been reported to increase cancer cell proliferation. In basal condition, C6 glioma cells expressed predominantly P2Y1 mRNA with lower levels of P2Y12 mRNA, but, when the cells were cultured in serum-free medium, the expression of P2Y1 mRNA decreased, whereas that of P2Y12 significantly increased (Czajkowski et al., 2004). In these conditions, ADP enhanced ERK1/2 phosphorylation and PI3K signaling by activating the P2Y12R (Czajkowski et al., 2004).
|
study
| 99.94 |
More recently, P2Y12 expression has been also described in breast cancer cell lines (Sarangi et al., 2013). The baseline expression of the receptor protein was low in both normal breast epithelium (MCF 10A cells) and in human breast cancer cell lines, namely MCF7 and MDA-MB-231 (Sarangi et al., 2013). Interestingly, as for the glioma cells, the P2Y12 protein levels were enhanced by serum starvation. Also, cell treatment with cisplatin, a well-known chemotherapeutic agent, enhanced P2Y12 expression in breast cancer cells (Sarangi et al., 2013; Dasari and Tchounwou, 2014). The inhibition of P2Y12 reduced cisplatin-mediated increase of hypoxia-inducible factor 1-alpha, a factor involved in the resistance to cytotoxic therapy (Ai et al., 2016; Zhao et al., 2016), in angiogenesis and in metastatic processes (Choi et al., 2016; Wang et al., 2016).
|
review
| 99.0 |
In platelet rich plasma from healthy subjects the P2Y12R antagonist cangrelor reduced the production of ADP-stimulated vascular endothelial growth factor (VEGF) (Bambace et al., 2010) a key protein in angiogenesis. Moreover, platelet secretion of other proangiogenic molecules, including IL-1α, IL-1β, GM- CSF, MMP-1 and uPAR, can be controlled by P2Y12R inhibitors in non-small cell lung cancer cell-stimulated platelets (Wu et al., 2015).
|
study
| 99.94 |
These evidence, together with the increasing interest in the anticancer properties of metal-based compounds (Chen et al., 2017), have recently lead to the development of innovative gold (III) complexes of prasugrel, a newer oral P2Y12R inhibitor, with promising chemotherapeutic activities (Benkli, 2016).
|
review
| 99.75 |
Finally, a role for P2Y12R has been postulated in neuropathic pain, which is often caused by chronic diseases such as cancer. It has been shown that peripheral nerve injury was able to increase the expression of different ATP P2Y receptors, including P2Y12, in microglial cells present in the spinal dorsal horn (Kobayashi et al., 2008, 2012). More recently, in a rat model of tongue cancer pain, obtained by inoculation of squamous cell carcinoma cells, a marked activation of microglia through P2Y12R was found in the trigeminal spinal subnucleus caudalis. This resulted to be associated with increased excitability of nociceptive neurons and consequent allodynia after mechanical stimulation. The administration of the P2Y12 antagonist MRS2395 strongly reduced the “nocifensive” behavior and microglial activation in these animals (Tamagawa et al., 2016).
|
review
| 95.2 |
At present, there are two groups of P2Y12 antagonists (Figure 2) which, after aspirin, are the most widely prescribed antiplatelet agents in CVD. Thienopyridines, including ticlopidine, clopidogrel, and prasugrel, irreversibly inhibit P2Y12R upon metabolic conversion into active metabolites by the hepatic cytochrome P-450 system (Cattaneo, 2010; Schrör et al., 2012). On the other hand, ticagrelor, cangrelor, and elinogrel reversibly and directly bind the receptor without any need for bioactivation (Cattaneo, 2010; Schrör et al., 2012) (Figure 2).
|
review
| 99.8 |
The potential contribution of P2Y12 receptor in cancer and metastasis. P2Y12 receptor (R) antagonists include thienopyridines, such as ticlopidine, clopidogrel and prasugrel, that irreversibly inhibit P2Y12R upon metabolic conversion into active metabolites by the hepatic cytochrome P-450 system (Cattaneo, 2010; Schrör et al., 2012) and ticagrelor, cangrelor and elinogrel, which reversibly and directly bind the receptor without any need for bioactivation (Cattaneo, 2010; Schrör et al., 2012). By targeting P2Y12R these drugs may contribute to constrain metastasis development through the prevention of platelet activation by ADP. Once activated, platelet-derived PGE2 and a direct platelet-tumor cell interaction synergize to promote epithelial-mesenchymal transition (EMT), migration and metastasis formation (Guillem-Llobat et al., 2016). Moreover, like platelets, cancer cells have been shown to release a significant amount of ATP (Beigi et al., 1999; Pellegatti et al., 2008; Burnstock and Di Virgilio, 2013), which is hydrolyzed into ADP through the activity of the ecto-enzymes CD39 (Bakker et al., 1994; Robson et al., 2006; Burnstock, 2017). Extracellular ADP, deriving from both activated platelets and cancer cells, in turn could activate P2Y12R expressed both in platelets and in malignant cells, thus triggering a vicious circle, which could contribute to cancer progression and dissemination.
|
review
| 96.9 |
Kohga et al. (1981) firstly demonstrated that ticlopidine treatment suppressed the formation of pulmonary nodules in rodents injected with a B16 melanoma or AH130 rat ascites hepatoma cells. In a rodent model of spontaneous pulmonary metastasis, ticlopidine was able to inhibit lung metastasis when administered per os (Kohga et al., 1981).
|
study
| 99.4 |
More recently, in a mouse model of either spontaneous or experimentally induced lung metastasis, obtained by injection of Lewis lung carcinoma (LLC) cells and B16 melanoma cells respectively, P2Y12 deficiency reduced the weight of lung metastasis without affecting the primary tumors (Wang et al., 2013). This suggests a role for the ADP receptor in promoting the metastatic process. This hypothesis was strengthened by the observation that platelets from P2Y12-deficient mice significantly reduced the invasiveness of LLC cells. Moreover, P2Y12 deficiency reduced the capability of the cancer cells to stimulate the production of active TGFβ1 from platelets. This, in turn, resulted in the prevention of platelet-induced epithelial-mesenchymal transition (EMT) of the tumor cells (Wang et al., 2013), a process known to contribute pathologically to cancer progression and metastasis.
|
study
| 99.9 |
In agreement with these data, Guillem-Llobat et al. (2016) recently reported that the interaction of platelets with HT29 human colon carcinoma cells leads to the induction of EMT in tumor cells associated to enhanced cell mobility. When the cells were co-cultured with platelets in the presence of ticagrelor, both the down-regulation of E-cadherin, an epithelial cell marker, and the enhanced migratory capacity of HT29 cells were inhibited (Guillem-Llobat et al., 2016). Ticagrelor also caused the simultaneous inhibition of TXB2 (the stable hydrolysis product of TXA2) and of prostaglandin (PG)E2 production, suggesting an inhibitory effect on the release of arachidonic acid from platelet membrane phospholipids (Guillem-Llobat et al., 2016).
|
study
| 100.0 |
Ovarian cancer has been shown to be a potential target for P2Y12 inhibitors. Ticagrelor, given by daily gavage, reduced the growth of primary tumors in rodent models of ovarian cancer (Cho et al., 2017). The drug caused a significant decrease in Ki67 immunostaining (a proliferation marker) in tumors resected form ticagrelor treated mice compared to controls. A direct role of P2Y12 in ovarian cancer cells was ruled out since Western blot analysis did not detect the receptor at the protein levels, and knocking down P2Y12 by siRNA or CRISPR-Cas9 techniques did not affect the cancer cell proliferation either in vitro or in vivo (Cho et al., 2017).
|
study
| 99.94 |
P2Y12 seems to mediate also bone loss under pathological conditions including cancer. In a mouse model of bone metastasis, clopidogrel increased bone mineral density and trabecular bone volume compared to controls (Su et al., 2012). Accordingly, in a mouse model of tumor-associated bone loss, the P2Y12 deficiency protected the animals from trabecular bone loss. Moreover, in the P2Y12-/- mice, the number and surface of osteoclasts significantly increased in the tumor-bearing wild-type animals, was similar to the controls (Su et al., 2012).
|
study
| 99.94 |
Combination of antiplatelet agents has been poorly evaluated. In a mouse model of chronic immune-mediated hepatitis B that progresses to hepatocellular carcinoma, the treatment with low-dose aspirin and clopidogrel caused a marked reduction in the development of hepatomas. Also, the overall mass of the hepatomas resulted to be lower in mice treated with the combination therapy. At the time when 75% of vehicle-treated mice were found dead, only 20% of the animals treated with aspirin plus clopidogrel had died. Notably, the combined antiplatelet treatment did not cause significant bleeding side effects in these animals (Sitia et al., 2012).
|
study
| 99.9 |
Although the few published preclinical studies suggest a potential role for P2Y12 antagonists in chemoprevention and/or in potentiating the effect of cytotoxic drugs (Sarangi et al., 2013; Dasari and Tchounwou, 2014; Pandey et al., 2014) there are no results from randomized clinical trials (RCTs) aimed to assess the effects of these drugs on cancer and metastasis.
|
review
| 99.8 |
Concerns have been raised on the possible association between P2Y12 antiplatelet therapy and solid tumor growth or metastatic dissemination, even if the evidence for this harmful association was not sufficient to modify the clinical practice (Serebruany et al., 2015). A recent population-based cohort study has been published comparing the association of cancer risk between treatment with aspirin alone and aspirin in combination with clopidogrel. At least 3 years of follow-up was guaranteed by the study and patients with a diagnosis of cancer within the first year were excluded. Newly diagnosed cases of cancer, with the exception of melanoma skin tumor, were 21.977 out of 184.781 subjects, the primary endpoint being the time until first diagnosis (Leader et al., 2017). Breast, colorectal, prostate, and lung cancer were the most common cancer types. The study showed that there was not a higher risk for cancer in subjects assuming the combined treatment compared to aspirin used alone (HR 0.92, 95% CI: 0.86–0.97) and suggested that clopidogrel could even reduce cancer incidence (Leader et al., 2017). Similarly, an FDA meta-analysis, carried out to assess the effects of clopidogrel on death rates from all causes, showed that the dual antiplatelet therapy with aspirin and clopidogrel, given for 12 months or longer, was safe. Indeed, there was no increased risk for cancer-related deaths compared to aspirin and clopidogrel administered for 6 months or less, or to aspirin alone (FDA, 2015).
|
review
| 99.9 |
A systematic review and meta-analysis were performed with the aim to verify whether thienopyridines increased cancer mortality and cancer events. Nine studies have been analyzed including six RCTs and three retrospective cohort studies for a total number of 282,084 participants. All the studies reported on clopidogrel, whereas only two specifically reported on prasugrel. The exposure to clopidogrel showed no association with increased odds of cancer (OR 0.70, 95% CI: 0.66–0.75, n = 1). Furthermore, no significant difference in cancer event rate was pointed out for prasugrel when compared to clopidogrel (OR 1.10, 95% CI: 0.89–1.37, n = 2). Similar results were obtained on the analysis of cancer mortality. On the whole, these data do not support concerns for a class effect of thienopyridines in increasing the cancer event rate and/or mortality (Kotronias et al., 2017).
|
review
| 99.9 |
Several lines of evidence show that the cross-talk of cancer cells with stromal cells (such as immune cells and fibroblasts) and platelet-derived products induce a novel phenotype, which allows them to invade the healthy tissue around, to enter the bloodstream and to colonize distant tissues. The role of platelets in conditioning tumor microenvironment, by releasing several bioactive mediators and microvesicles, is increasingly recognized. Platelets and cancer cells have been shown to release a significant amount of ATP and ADP (Beigi et al., 1999; Pellegatti et al., 2008; Burnstock and Di Virgilio, 2013).
|
review
| 99.75 |
Given the site-specific nature of cancer, along with the specific mechanisms that different cancer cells can develop to activate platelets and viceversa the final ATP/ADP effect will depend on the P2 receptor subtypes expressed and activated (Burnstock and Di Virgilio, 2013). Thus, an improved knowledge of the signaling induced by the activation of P2Y12R during the interaction of platelets with cancer cells is needed. Interestingly, the PLATelet inhibition and patient Outcomes (PLATO) study showed that ticagrelor exerts an anti-inflammatory effect higher than clopidogrel (Thomas and Storey, 2015). Ticagrelor can inhibit the adenosine transporter ENT1 (Cattaneo et al., 2014). This increases the extracellular concentration of the nucleoside, which is a recognized modulator of inflammatory and immune responses. Thus, clinical research, focused on targeted biomarkers, is needed to clarify the role of P2Y12R-antagonists in these mechanisms, which are relevant in cancer development and metastasis.
|
review
| 99.75 |
Moreover, an in-depth characterization of P2Y12R expression in tumor cells is necessary to enlighten potential novel direct therapeutic activities of P2Y12 antagonists. On the other hand, further findings on P2Y12R activity in non-malignant cells should be useful, since they could help monitoring the possible adverse effects caused by the systemic administration of P2Y12-targeted drugs.
|
other
| 99.5 |
In conclusion, although emerging evidence suggests a rationale for targeting P2Y12R to constrain tumor progression, metastasis development, and pain and to potentiate the responses to conventional therapies, a definitive evidence of the anticancer effect of P2Y12R antagonists is lacking. Further studies of basic and clinical research are urgently needed in this setting to put a final word on the role of the purinergic signaling in cancer development and metastasis.
|
review
| 99.9 |
Conception of work: PB and PP; design of work: PB and PP; drafting of manuscript: PB, MD, ST, and AB; critical revision of manuscript: PB and PP; final approval of work: PB, MD, ST, AB, and PP; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: PB, MD, ST, AB, and PP.
|
other
| 99.94 |
The episodic memory deficits experienced by older adults are characterized not only by increased forgetting, but also greater susceptibility to false memories of events that did not occur (see Devitt & Schacter, 2016 for review). One particularly robust example of false memory among older adults is the false recognition of novel objects that are perceptually similar to studied objects in a recognition memory test. Despite this increased tendency to incorrectly identify similar foils as having been studied previously, the ability to correctly identify previously studied targets as old, and identify novel and perceptually distinct foils as new, is typically unaffected (Koutstaal, 2003; Toner, Pirogovsky, Kirwan, & Gilbert, 2009; Yassa et al., 2011; Holden, Toner, Pirogovsky, Kirwan, & Gilbert, 2013). This deficit has proven to be resistant to a number of task manipulations, including those that aim to enhance attention to perceptual detail during encoding and those that encourage more strict responding during retrieval (Koutstaal, Schacter, Galluccio, & Stofer, 1999; Stark, Stevenson, Wu, Rutledge, & Stark, 2015). Despite the frequent emergence and stability of this pattern, our current understanding of the basis for elevated rates of false recognition of lures among older adults remains limited.
|
review
| 99.9 |
One reason that a Yes/No recognition test with similar lures reveals greater age-related effects than do typical tests of item recognition may be related to the increased demands it places on recollection-based retrieval strategies (Migo, Montaldi, Norman, Quamme, & Mayes, 2009; Norman, 2010). In particular, because targets and foils are perceptually similar and therefore both highly familiar, it is difficult to reliably distinguish targets from lures using a strength-based criterion alone. Instead, participants must use a recollection-based retrieval strategy (i.e., recall-to-reject) to support performance (Migo et al., 2009). This strategy describes the process of disqualifying an exemplar as having been studied by first recalling details of the studied target, and then detecting a mismatch between the target and the lure (Brainerd, Reyna, Wright, & Mojardin, 2003; Gallo, 2004). Critically, the ability to use this strategy successfully requires both that sufficient information about the target is available to disqualify the lure as having been studied, as well as the ability to selectively retrieve and evaluate stored details about the target. Thus, age-related increases in false recognition could plausibly arise due to declines in either of these factors, or a combination of the two.
|
study
| 99.94 |
The relative contribution of each of these factors to elevated rates of lure false recognition among older adults remains unclear. On the one hand, existing evidence suggests that aging negatively affects the ability the implement controlled and strategic retrieval processes, relative to more spontaneous and automatic processes (Jennings & Jacoby, 1993; Yonelinas, 2002). For example, older adults typically exhibit disproportionate deficits in memory performance under conditions that place similar demands on recollection-based retrieval strategies, such as rejecting recombined pairs in an associative recognition test (Castel & Craik, 2003; Cohn, Emrich, & Moscovitch, 2008), or rejecting studied items from a nontarget source in an exclusion paradigm (Gallo, Bell, Beier, & Schacter, 2006; Jennings & Jacoby, 1993). Conversely, age differences are typically absent when performance can be supported based on the presence or absence of item familiarity, such as endorsing studied pairs and rejecting experimentally novel foils (Cohn et al., 2008; Gallo et al., 2006; Yonelinas, 2002). As the ability to strategically retrieve and evaluate stored details appears to be impaired with age, this factor alone could account for elevated rates of false recognition among older adults.
|
study
| 99.94 |
However, it could also be the case that aging is associated with declines in the availability of object details that are necessary to disqualify lures as having been studied. Evidence for this possibility has been mixed. Previous work has found that despite impaired explicit memory for object details, implicit memory remained intact with age (Koutstaal, 2003), suggesting that object details may be available even when they are not retrieved successfully. In contrast, other work has identified age-related impairments in perceptual discrimination tasks that require participants to distinguish between stimuli sharing overlapping features, even though such tasks minimize mnemonic demands (Lee, Smith, Grady, Hoang, & Moscovitch, 2014; Ryan et al., 2012; Yeung, Ryan, Cowell, & Barense, 2013). Notably, if the availability of object details is reduced with age, this could also be the sole factor driving age-related increases in false recognition. That is, if object details that can disambiguate targets and foils are not available, a mismatch between targets and foils cannot be successfully detected, rendering a recall-to-reject strategy unsuccessful.
|
study
| 99.7 |
Critically, in a typical Yes/No object recognition test with similar foils, the availability of object details and the ability to retrieve and evaluate these details are confounded, making it difficult to tease apart the relative contributions of these two factors to false recognition. One way of overcoming this limitation is to hold demands on the availability of object details constant, while varying demands on strategic retrieval processes. Existing empirical and modeling evidence indicates that this can be done by comparing performance in a typical Yes/No recognition memory format to performance in a Forced Choice test format, wherein targets and corresponding foils are presented simultaneously (Guerin, Robbins, Gilmore, & Schacter, 2012; Holdstock et al., 2002; Migo et al., 2009, 2014; Norman, 2010). In particular, whereas the item-wise presentation of targets and foils in the Yes/No test places considerable demands on recollection-based processes (i.e., recall-to-reject), the simultaneous presentation of targets and their corresponding foils in a Forced Choice test can be supported by judgments of relative familiarity differences between the two choices (Norman, 2010).
|
study
| 99.94 |
In support of this proposal, behavioral work in younger adults has found that high rates of lure false recognition in a Yes/No test can be reduced substantially by presenting targets and corresponding foils simultaneously at test (Guerin et al., 2012). This pattern suggests that the presentation of corresponding foils increases the accessibility of stored details, likely by reducing demands on strategic retrieval processes. Consistent with this possibility, previous work in younger adults using a modified Remember/Know procedure has found that successful performance in the Yes/No test format is supported predominantly by the use of a ‘recall-to-reject’ strategy, whereas accurate performance in the Forced Choice format can be supported by relying on familiarity alone (Migo et al., 2009). Similarly, previous work in older adults has identified distinct relationships between independent standardized measures of recall and recognition memory, and performance in the Yes/No and Forced Choice tests, respectively (Migo et al., 2014).
|
study
| 99.7 |
Collectively, these observations indicate that the simultaneous presentation of targets and foils in the Forced Choice test considerably reduces demands on strategic retrieval, thereby enabling a more direct assessment of the availability of object details and the contribution of this factor to increases in false recognition. In the present study, we directly compare older and younger adults’ performance across test formats to characterize the relative contribution of these two factors to elevated rates of lure false recognition with age. If older adults exhibit selective deficits in Yes/No performance, coupled with intact Forced Choice performance, then this would suggest that age-related increases in false recognition are driven primarily by impairments in strategic retrieval processes. However, if older adults also exhibit deficits in Forced Choice performance, this would suggest that a decline in the availability of object details that are necessary to disambiguate targets and foils contributes to increases in lure false recognition with age, and may be the primary factor driving impaired performance.
|
study
| 100.0 |
The primary aim of this experiment was to investigate the degree to which age differences in Yes/No recognition memory with similar foils are ameliorated when demands on recall-to-reject are reduced through the simultaneous presentation of targets and corresponding foils at test. To this end, we compared older and younger adults’ recognition memory performance across Yes/No and Forced Choice test formats. If age-related increases in false recognition in the Yes/No test arise solely as a result of impairments in the strategic retrieval and evaluation of item information, we should observe deficits in Yes/No performance coupled with intact Forced Choice performance. In contrast, age differences in Forced Choice performance would suggest that a reduction in the availability of disambiguating object details contributes to false memory errors among older adults.
|
study
| 100.0 |
A second aim of this experiment was to gain further support for the proposal that successful performance in the Yes/No test and Forced Choice test differentially relies on the use of recollection and familiarity. To this end, we included the modified Remember-Know (RK) judgments that were used in previous work (Migo et al., 2009) to provide an indication of the strategy participants use to make their decisions. Specifically, participants were asked to provide a ‘remember’ response if their decision was based on retrieval of specific target details, indicating the use of a recall-to-reject or recall-to-accept strategy, and a ‘familiar’ response if their decision was based on the relative familiarity of the presented exemplar, in the absence of retrieval of specific stimulus details. We predicted that performance in the Yes/No test would rely primarily on the successful use of a recall-to-reject strategy, whereas performance in the Forced Choice test could be more successfully supported by familiarity-based judgments, as described in previous work (Migo et al., 2009). Moreover, we predicted that age differences in the ability to successfully execute a recollection-based retrieval strategy would be greater than age differences in accurate familiarity-based responding (Koen & Yonelinas, 2016).
|
study
| 100.0 |
Thirty-two younger adults aged 18 to 28 years (M = 22.66, SD = 3.04) and 32 older adults aged 60–80 years (M = 70.47, SD = 4.59) participated. All participants were native English speakers. The younger adults were students from the University of Cambridge and the older adults were healthy, community-dwelling volunteers. The groups were matched with respect to years of formal education (t < 1) and the older adults outperformed the younger adults on the Vocabulary subtest of the Shipley Institute of Living Scale (Shipley, 1986; t(62) = 3.47, p < .005, d = 0.881). Older adults were additionally screened for cognitive impairment using the Montreal Cognitive Assessment (MoCA; Nasreddine et al., 2005) and all participants performed within the normal range (M = 28.03, SD = 1.03). A summary of demographic information can be found in Table 1. Participants in all experiments provided written informed consent prior to beginning the experiment using methods approved by the Cambridge Psychology Research Ethics Committee, and received monetary compensation at a rate of £7.50 per hour for participation.
|
study
| 100.0 |
A total of 800 color images of everyday objects were used as stimuli. This set consisted of 400 unique pairs of everyday object exemplars, collected from a combination of online sources, including Google Image Search (Mountain View, CA) and the stimulus sets available from the Konkle Lab (http://konklab.fas.harvard.edu). Each exemplar pair shared the same basic-level name (e.g., umbrella) and possessed a high degree of feature overlap (e.g., shape, color, pattern) such that targets and foils could not be discriminated without a detailed representation of each object (see Figure 1). To minimize any effects of pairwise variability in target-foil similarity on performance, an independent sample of participants rated exemplar pairs on perceptual similarity. We then created stimulus lists with equivalent levels of target-foil similarity on average, and counterbalanced the assignment of stimulus lists to test format and study block across participants.
|
study
| 100.0 |
Each session began with a practice block in which participants completed an abbreviated version of the task that provided feedback on performance accuracy. This ensured that all participants understood the nature of the memory test, including the importance of memory for specific perceptual details of each stimulus for successful test performance. Each participant then completed two study-test blocks, with the procedure identical for each block of the experiment. A 5-min break divided the first and second block during which participants performed the Vocabulary subtest of the Shipley Institute of Living Scale. Older adults additionally completed the MoCA at the end of the testing session.
|
study
| 100.0 |
During each study phase, participants were presented with 200 pictures of everyday objects for 3000 ms each and asked to judge the pleasantness of each object. Participants were instructed to make this judgment based on the physical attributes of the stimulus (e.g., color, shape, pattern, and texture) to direct attention toward the perceptual features of each object, and to equate as much as possible the way in which stimuli were processed during the study phase across participants. After a 60-second filled interval during which participants counted backward by sevens from a random 3-digit number, participants completed a recognition memory test.
|
study
| 100.0 |
One half of the test comprised a Forced Choice format, wherein a target and its corresponding foil were presented simultaneously, one on the left of the screen and one on the right. The other half of the test was in a Yes/No format, wherein a single exemplar was presented in the center of the screen, which could be either a target or a foil. To equate the length of the Forced Choice and Yes/No tests, in the Yes/No test, either the target or its corresponding foil was presented. Accordingly, half of the stimuli were tested with the studied item as the test cue, and the other half tested with the corresponding foil as the cue, with this assignment counterbalanced across participants. Prior to the beginning of each test format, participants were reminded of the instructions and response options for that test. The order of the test formats was consistent across both blocks and counterbalanced across participants.
|
study
| 99.94 |
During each test phase, participants followed a modified Remember/Know procedure (Migo et al., 2009), indicating their recognition decision and the nature of their memory for the object by selecting from four response options. In the Forced Choice test, participants were instructed to select a ‘remember left’ or ‘remember right’ response if they could recall specific details of the exemplar they judged to have been previously studied, and a ‘familiar left’ or ‘familiar right’ response if their decision was based on greater familiarity of one exemplar over the other. In the Yes/No test, participants were instructed to select a ‘remember’ response if they recalled specific details of a studied exemplar and used these details to either accept a target (‘remember old’) or to reject a foil (‘remember new’). Participants were instructed to select a ‘familiar’ response if they were unable to recall specific details of a studied exemplar, and instead based their decision on the presence (‘familiar old’) or absence (‘unfamiliar new’) of familiarity for the presented object.
|
study
| 100.0 |
We first compared recognition memory performance across age groups in the Forced Choice and Yes/No test formats, collapsing across remember and familiar responses (see Figure 2; raw proportions are shown in Table 2). We did this by computing d′ scores from the proportion of correct responses in the Forced Choice test and the proportion of hits and false alarms in the Yes/No test (Macmillan & Creelman, 1991), and submitting these scores to a 2 × 2 mixed ANOVA with Test Format (FC, YN) as a within-subjects factor and Age (young, old) as a between-subjects factor. The ANOVA revealed that both older and younger adults performed significantly better on the Forced Choice test relative to the Yes/No test, F(1, 62) = 45.61, p < .001, ηp2 = 0.424, and that older adults performed significantly worse than younger adults across both test formats, F(1, 62) = 11.84, p < .005, ηp2 = 0.160, with the size of this deficit equivalent across test formats (F < 1).
|
study
| 100.0 |
We next sought to test the prediction that successful performance in the Yes/No test is driven primarily by the use of a recall-to-reject strategy, and that this strategy is impaired with age. To this end, we computed a ‘recall-to-reject’ measure calculated as the proportion of ‘remember new’ responses minus the proportion of ‘remember’ misses, and a ‘recall-to-accept’ measure calculated as the proportion of ‘remember old’ responses minus the proportion of ‘remember’ false alarms (Migo et al., 2009). A Response Type (reject, accept) × Age ANOVA revealed a main effect of Response Type, F(1, 62) = 6.51, p < .05, ηp2 = 0.095, indicating a greater proportion of accurate responses were associated with the use of the ‘recall-to-reject’ strategy relative to a ‘recall-to-accept’ strategy. This did not interact with Age (F < 1), indicating that older adults’ performance also benefited from the use of a recall-to-reject strategy. However, a main effect of Age, F(1, 62) = 15.32, p < .001, ηp2 = 0.198, indicated that older adults used both strategies less successfully than younger adults (reject: t(62) = 3.40, p < .001, d = 0.86; accept: t(62) = 3.35, p < .001, d = 0.85). Age-related deficits in the ability to use these strategies were driven by both an increase in ‘remember’ false alarms, t(62) = 3.21, p < .005, d = 0.82, and a reduction in the proportion of ‘remember’ correct rejections, t(62) = 3.23, p < .005, d = 0.82.
|
study
| 100.0 |
Next, we tested the prediction that the Forced Choice test format reduces demands on recollection-based retrieval strategies relative to the Yes/No test, enabling performance to be more successfully supported by familiarity-based judgments. To this end, we first conducted a 2 × 2 × 2 mixed ANOVA with Test Format (YN, FC) and Response Type (Remember, Familiar) as within-subjects factors and Age as a between-subjects factor on the proportion of correct responses to old items (i.e., hits in the Yes/No test, correct responses in the Forced Choice test). This revealed a main effect of Response Type, F(1, 62) = 295.58, p < .001, ηp2 = 0.827, with more responses associated with ‘remember’ than ‘familiar’ responses across groups, which was qualified by a Test × Response Type interaction, F(1, 62) = 6.89, p < .05, ηp2 = 0.100, that did not vary with age (F < 1). This reflected a greater contribution of recollection to correct ‘old’ responses in the Yes/No test than the Forced Choice test, t(63) = 2.38, p < .05, d = 0.57, but a greater contribution of familiarity to correct responses in the Forced Choice test than the Yes/No test, t(63) = 2.71, p < .01, d = 0.68. The main effect of Age and the Age × Response Type interaction were not significant (Fs < 1), indicating older adults made similar proportion of correct remember and familiar responses to old items as did younger adults.
|
study
| 100.0 |
Next we conducted the same ANOVA on the proportion of incorrect old responses to similar foils (i.e., false alarms in the Yes/No test, incorrect responses in the Forced Choice test). This revealed a main effect of Test Format, such that more false alarms were made in the YN test than the FC test, F(1, 62) = 305.18, p < .001, ηp2 = 0.831, and Response Type, such that false recognition was more often associated with ‘remember’ than ‘familiar’ responses, F(1, 62) = 4.93, p < .05, ηp2 = 0.074, overall. This was qualified by a Test × Response Type interaction, F(1, 62) = 25.42, p < .001, ηp2 = 0.291, reflecting a larger increase in the tendency to make ‘remember’ false alarms, t(63) = 14.49, p < .001, d = 3.65, than familiarity-based false alarms, t(63) = 9.05, p < .001, d = 2.28, in the Yes/No test relative to the Forced Choice test. There was also a main effect of Age, F(1, 62) = 13.78, p < .001, ηp2 = 0.182, indicating older adults made more false alarms overall, although the interaction with Response Type did not reach significance (p = .11). Nonetheless, independent t tests indicated that older adults made more remember-based false alarms across test formats (YN: t(62) = 3.21, p < .005, d = 0.82; FC: t(62) = 2.59, p < .05, d = 0.66) but did not differ with respect to the number of familiarity based false alarms in either test format (YN: t(62) = 1.60, p = .11, d = 0.41; FC: t < 1).
|
study
| 100.0 |
The results of Experiment 1 revealed high rates of lure false recognition in the Yes/No test across both groups, which were reduced substantially by the simultaneous presentation of targets and foils in the Forced Choice test format. This observation is consistent with previous findings (Guerin et al., 2012; Migo et al., 2009), and with proposals that false recognition in the Yes/No test is often driven by a failure to retrieve stored details. This idea was further supported by the modified RK judgments, which indicated that accurate responding in the Yes/No test relies primarily on the successful application of a recall-to-reject strategy, whereas familiarity can more successfully support Forced Choice decisions, again consistent with previous work (Migo et al., 2009).
|
study
| 100.0 |
Interestingly, we found that incorrect responses were more often associated with ‘remember’ judgments across age groups. This pattern suggests that memory errors were typically associated with illusory recollection, wherein the subjective experience of remembering accompanies inaccurate responses. This tendency is not uncommon in recognition memory tests with highly similar foils, and has been observed previously among healthy younger adults (Kim & Yassa, 2013). This may arise from an increased likelihood for participants to think a recollected detail is diagnostic of the target, when in fact is it shared by targets and foils (Migo et al., 2009). Alternatively, participants may have tended to misattribute a spontaneously activated prototypical feature of an item as having been studied, or erroneously recombined studied features from one object and another (Doss, Bluestone, & Gallo, 2016). Notably, the incidence of illusory recollection was considerably greater in the Yes/No test than the Forced Choice test, consistent with increased demands on postretrieval monitoring and evaluation of retrieved details in this test format.
|
study
| 100.0 |
As predicted, older adults were significantly impaired in the ability to execute a recall-to-reject strategy, as well as a recall-to-accept strategy. This was driven by an increase in lure false alarms, coupled with an intact hit rate to studied items, replicating previous work (Koutstaal, 2003; Stark et al., 2015; Toner et al., 2009; Yassa et al., 2011). These findings are consistent with existing proposals that aging is associated with declines in the ability to use a recall-to-reject strategy in other domains, such as source exclusion tasks (Gallo et al., 2006) and associative recognition tests (Cohn et al., 2008). Interestingly, older adults’ false alarms were primarily associated with incorrect ‘remember’ responses, consistent with previous observations that increased false recognition with age is more often accompanied by high confidence, illusory recollection than increased reliance on familiarity (Dodson, Bawa, & Slotnick, 2007; see McCabe, Roediger, McDaniel, & Balota, 2009 for review). This tendency is thought to reflect impairments in postretrieval monitoring and evaluation processes with age (Dodson, Bawa, & Krueger, 2007; Gallo et al., 2006; Wong, Cramer, & Gallo, 2012), and may arise, at least in part, as a result of reductions in the availability of object information that can be used to detect a mismatch between targets and lures.
|
study
| 93.1 |
Consistent with this possibility, although older adults displayed similar benefits to performance from the reinstatement of target details in the Forced Choice test as younger adults, age differences in this test format were still observed. This deficit indicates that impairments in strategic retrieval processes alone cannot account for older adults’ performance, and suggests an age-related decline in availability of stimulus details that can disambiguate targets and foils, as suggested by previous work (Burke et al., 2010, 2011; Ryan et al., 2012; Yeung et al., 2013). These results raise an important possibility. If object details that can successful disqualify lures are less available to older adults, this factor alone could be driving age-related impairments in the use of a recall-to-reject strategy in the Yes/No test by reducing older adults’ ability to detect a mismatch between targets and foils. Alternatively, it may be the case that even if the availability of object details were equated across groups, older adults would continue to exhibit deficits in Yes/No performance due to impairments in the ability to strategically retrieve and evaluate these details. We aimed to tease apart these possibilities in Experiment 2.
|
study
| 100.0 |
Existing empirical and modeling work suggests that the ability to discriminate between perceptually similar object exemplars can be impacted by the presence of interference from objects sharing common lower-level features (Cowell et al., 2006). For example, studies in healthy younger adults have found that Forced Choice object recognition is reduced after viewing visual interference containing objects, but not analogous interference comprised of scenes (Watson & Lee, 2013; O’Neil, Watson, Dhillon, Lobaugh, & Lee, 2015). In contrast, existing work in healthy older and younger adults has found that varying the number of objects between study and test in a Yes/No recognition memory test did not affect performance in either age group (Stark et al., 2015). These findings suggest that Forced Choice performance may be more directly impacted by the presence of interference from objects sharing lower level features than Yes/No performance. Furthermore, older adults may be more susceptible to feature-level interference than younger adults (Burke et al., 2012; Newsome et al., 2012; Ryan et al., 2012; Yeung et al., 2013), raising the possibility that this factor exacerbated age differences in performance in Experiment 1.
|
study
| 99.94 |
To our knowledge, no work to date has compared the effects of interference across test formats, or how this might impact the presence of age differences in each case. In Experiment 2, we explored this question by assessing older and younger adults’ performance in each test format while varying the number of objects in the study list across study-test cycles. We varied the length of each study list differently according to age group, such that the longer study list for older adults was the same length as the shorter list in younger adults. This enabled us to compare performance across groups when older adults faced an equivalent amount of interference relative to younger adults, as well as when they faced reduced interference relative to younger adults.
|
study
| 100.0 |
If Forced Choice performance is disproportionately affected by interference from viewed objects, and older adults are more vulnerable to interference than younger adults, reducing the number of studied objects may ameliorate age differences in this test format. If so, this will enable us to assess whether age differences in Yes/No performance continue. Persistent deficits in Yes/No performance would suggest that age differences cannot be explained solely by reductions in the availability of object details, implicating additional contributions of impaired strategic retrieval processes. In contrast, if performance improves similarly across test formats, this would suggest a single factor, namely reductions in the availability of object details, drives increased false recognition across test formats.
|
study
| 100.0 |
A new group of 34 younger adults aged 18 – 28 years (M = 21.74, SD = 2.39) and 48 older adults aged 60–80 years (M = 70.29, SD = 5.86) participated in this experiment. All participants were native English speakers. The younger adults were students from the University of Cambridge and the older adults were healthy, community-dwelling volunteers. Older and younger adults did not differ with respect to years of formal education (t < 1) and older adults scored significantly higher on the Vocabulary subtest of the Shipley Institute of Living Scale, t(77) = 5.77, p < .001, d = 1.315. Three older adults were excluded from the analyses because they performed below the normal range (<26) on the MoCA, leaving 45 healthy older adults who performed well within the normal range (M = 28.11, SD = 1.19). A summary of demographic information can be found in Table 1.
|
study
| 100.0 |
A total of 960 color images of everyday objects were used as stimuli. This set consisted of the 800 images used in Experiment 1, plus an additional 160 images obtained from similar sources to produce 480 unique pairs of object exemplars. As in Experiment 1, each exemplar in a pair served equally often as the studied target and unstudied foil. These object pairs were divided into eight 60-item lists, with the allocation of each list to the short and long study lists and to the Forced Choice and Yes/No test format counterbalanced across participants. The length of the short and long lists differed for each age group, such that younger adults studied 180 items in their short block and 300 items in their long block, whereas older adults studied 60 items in their short block and 180 items in their long block. This resulted in a reference block of the same length completed by both groups, coupled with a block that was shorter or longer than the reference block for older and younger adults, respectively. The length of each list was selected so as to create two lists that were maximally different in length, where the difference in length was equivalent across age groups (in this case the lists differed by 120 items), with the reference block length as close as possible to the list length used in Experiment 1.
|
study
| 100.0 |
The procedure for this experiment was identical to that of Experiment 1, with the following exceptions. Participants completed two study-test blocks of unequal length, with length scaled for each age group, as described above. The order of the short and long blocks was counterbalanced across participants. A 10-min break was provided between blocks one and two during which participants were asked to rest quietly, to minimize the possibility of carry-over effects of interference from block one to block two. The test phase was again divided into a Forced Choice Test and a Yes/No Test, with test order consistent across blocks and counterbalanced across participants. Unlike Experiment 1, participants made simple Left/Right and Yes/No decisions in the Forced Choice and Yes/No formats, respectively. The modified Remember/Know judgments were removed from this test to simplify the response options.
|
study
| 100.0 |
As in Experiment 1, we computed d′ scores to compare older and younger adults’ recognition memory performance in the Forced Choice and Yes/No Tests at each block length (Macmillan & Creelman, 1991). Before exploring the effects of interference on performance, we first assessed performance when our two groups faced an equivalent amount of interference, namely for the reference block containing 180 stimuli, which is depicted in Figure 3 (left; raw scores in Table 3). To this end, we conducted a 2 × 2 mixed ANOVA with Test Format (FC, YN) as a within-subject factor and Age (young, old) as a between-subjects factor. The results replicated those observed in Experiment 1, with (a) both older and younger adults performing significantly better in the Forced Choice test relative to the Yes/No test, F(1, 77) = 15.15, p < .001, ηp2 = 0.164, (b) older adults performing significantly worse than younger adults across both test formats, F(1, 77) = 10.13, p < .005, ηp2 = 0.116, and (c) the size of this age effect being equivalent across test formats (F < 1).
|
study
| 100.0 |
We next examined the effects of increasing interference, that is, increasing block length from 60 items to 180 items in older adults and from 180 to 300 items in younger adults (see Figure 3, right; raw scores in Table 3). To do so, we submitted participants’ d′ scores to a 2 × 2 × 2 mixed ANOVA with Block Length (short, long) and Test Format as within-subject factors and Age as a between-subjects factor. The ANOVA revealed significant main effects of Block Length, F(1, 77) = 18.80, p < .001, ηp2 = 0.1960, Test Format, F(1, 77) = 42.77, p < .001, ηp2 = 0.357, and a marginal effect of Age, F(1, 77) = 3.48, p = .066, ηp2 = 0.043. These main effects were qualified by a significant Block Length × Test Format interaction, F(1, 77) = 8.49, p < .005, ηp2 = 0.099, and Test Format × Age interaction, F(1, 77) = 5.38, p < .05, ηp2 = 0.065. To investigate how the effects of Block Length and Age varied across Test Format, we conducted follow-up Block Length × Age ANOVAs separately for Forced Choice and Yes/No Test performance.
|
study
| 100.0 |
In the Forced Choice Test, we observed a significant main effect of Block Length, F(1, 77) = 26.57, p < .001, ηp2 = 0.257, that did not differ with age, F(1, 77) = 1.59, p = .211, ηp2 = 0.020. Critically, the effect of age on recognition memory performance was not significant (F < 1), and this was true for both short (t < 1) and long, t(77) = 1.37, p = .17, block lengths. In contrast, Yes/No performance did not decline significantly as block length increased (F < 1), and this was true across both age groups, with no evidence of an interaction, F(1, 77) = 1.67, p = .20, ηp2 = 0.021. However, significant age differences in Yes/No performance persisted, F(1, 77) = 7.63, p < .01, ηp2 = 0.090.
|
study
| 100.0 |
The results of Experiment 2 revealed dissociable effects of the list length manipulation across test formats. In both groups, performance in the Forced Choice test was modulated by the number of studied items, whereas performance in the Yes/No test remained stable across list lengths. This observation is consistent with previous work identifying effects of object interference on Forced Choice performance (O’Neil et al., 2015; Watson & Lee, 2013), but not on Yes/No performance (Stark et al., 2015), and lends support to the proposal that distinct mechanisms support memory performance across test formats. In particular, this observation suggests that Forced Choice performance is more directly related to the availability of object information, and therefore affected by the presence of interference from increased exposure to objects sharing common features. In contrast, Yes/No performance may be constrained by one’s ability to successfully execute recollection-based retrieval processes, which may place an upper boundary on performance. Similarly, previous work has suggested that familiarity-based memory performance may be more sensitive to the effects of interference than recollection-based memory (Sadeh, Ozubko, Winocur, & Moscovitch, 2016), consistent with the proposal that familiarity and recollection differentially contribute to performance in the Forced Choice and Yes/No test formats, respectively (Migo et al., 2009, 2014; Norman, 2010).
|
study
| 99.94 |
Notably, we found that reducing list length eliminated age differences in Forced Choice performance, whereas age difference in Yes/No performance persisted. The observation that reducing exposure to objects sharing overlapping features enhanced older adults’ Forced Choice performance may reflect an increased vulnerability to interference with age, as suggested by previous work (Burke et al., 2012; Newsome et al., 2012; Ryan et al., 2012; Yeung et al., 2013), which is consistent with an age-related decline in the availability of object information. In support of this possibility, as the number of studied objects increased, older adults exhibited a similar decline in performance as younger adults, but in the face of a considerably smaller amount of object interference. Accordingly, when we examined conditions that were analogous to those of Experiment 1, in which list length was matched across groups, older adults exhibited deficits across both test formats, replicating our prior results. Importantly, the observation of persistent deficits in Yes/No performance suggests that this single factor is unlikely to fully account for age differences in the Yes/No test. Instead, the current results point to an additional contribution of impaired strategic retrieval processes to age differences in Yes/No performance. This observation is consistent with existing evidence for disproportionate age differences in memory performance when demands on these strategies (i.e., recall-to-reject) are high (Cohn et al., 2008; Gallo et al., 2006; Luo & Craik, 2009).
|
study
| 100.0 |
An important caveat to these interpretations is that we did not examine the effects of different types of interference on memory performance across test formats. Accordingly, we cannot be certain that the effects of increasing list length on Forced Choice performance is related to an increase in interference from objects sharing lower-level features. Moreover, the list length manipulation not only altered the amount of object exposure, but also affected the memory load and duration of the study phase. Thus, we cannot rule out the possibility that other factors, such as increased attentional demands or fatigue associated with studying more items, contributed to this observation. However, the selective effect of the list length manipulation on performance in the Forced Choice test argues against these explanations, as such effects would be expected to impact both test formats in a similar fashion. Irrespective of the specific mechanism that led to the pattern of results observed here, the selective effect of the list length manipulation on Forced Choice performance lends support to the possibility that partially distinct factors determine performance across test formats. Future work should assess whether the current pattern also extends to different forms of interfering information, or is specific to objects that share common features.
|
study
| 100.0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.