Search

Background: Aromatase inhibitors (AIs) are highly effective at reducing recurrences and mortality in postmenopausal patients with estrogen receptor positive breast cancer (ER+ BC). Poor anti-proliferative (Ki67) response or ER+ BCs to AIs after 2 weeks is associated with worse long-term outcomes. Factors that relate to the degree of the response may identify markers and/or mechansims of resistance. Methods: The PeriOperative Endocrine Therapy for Individualizing Care (POETIC) trial randomized 4,480 with ER+ BC to 2 weeks’ AI before surgery or no presurgical treatment. All patients within the bottom 15% of Ki67 responders to AI (poor responders [PRs]; n=177 with RNA extracted) were selected from and matched to good responders (GRs) within the 50% showing the best response (n=190). Matching was based on baseline Ki67 levels as measured by immunohistochemistry (IHC). Response to AI was measured by the percentage change in Ki67 after 2 weeks’ treatment. PRs were further divided into groups expressing high ESR1 (PRs ESR1HIGH; n=119) and low ESR1 (PRs ESR1LOW; n=58) levels since there were very few GRs with low ESR1. RNAseq, targeted exome DNA sequencing of 87 BC/resistance related genes and measurement of plasma estradiol levels by mass spectrometry were performed to understand mechanisms of de novo resistance. Intrinsic subtypes were estimated from RNAseq data. Results: More than 90% of PRs ESR1LOW were non-luminal subtypes with low expression of estrogen-responsive genes. In contrast, 11% of PRs ESR1HIGH were non-luminal compared to 4% of GRs but only HER2-enriched subtypes were significantly higher in PR ESR1HIGH (p=0.05, Fisher exact). While AI treatment had limited impact on Ki67 IHC values in PRs ESR1HIGH, PGR expression was more than 2-fold lower after 2 weeks of AI. Gene-set enrichment analysis showed significantly lower expression of estrogen-response genes in PRs ESR1HIGH compared to GRs (FDR< 10-9) at baseline despite similar percentage of Luminal subtypes in PRs ESR1HIGH and GRs. Plasma estradiol levels were correlated with expression of estrogen-response genes (FDR=0.01) and levels were significantly lower in PRs ESR1HIGH compared to GRs (p=0.003, Mann Whitney). PRs ESR1HIGH had significantly more mutations in RB1, TP53, ARID1B and DNAH11 genes (p< 0.05, Fisher exact). TP53 mutations were significantly enriched in Luminal-A PRs ESR1HIGH compared to GRs (22% and 3% respectively; p=0.003, Fisher exact), but not in Luminal-B tumors (23% and 15% mutated respectively). Discussion and conclusions: In approximately 33% of PRs, de novo AI resistance was associated with and most likely due to low expression of ER/ESR1 and estrogen-responsive genes in non-luminal tumors. In the remaining tumors, AI treatment still impacted some estrogen responsive genes but had limited downstream impact on suppressing proliferation. This might be due to mutations including in TP53 that limit suppression of proliferation downstream of estrogen signaling. The proportion of Luminal tumors in GRs and PRs ESR1HIGH was similar, suggesting better outcome of Luminal-A tumors on AI is likely due to their better intrinsic prognosis rather than better response to endocrine therapy. Citation Format: Eugene F. Schuster, Elena López-Knowles, Anastasia Alataki, Lila Zabaglo, Elizabeth Folkerd, David Evans, Kally Sidhu, Holly Tovey, Perry Maxwell, Nicholas Turner, Stephen Johnston, Manuel Salto-Tellez, Maggie Chon U Cheang, John Robertson, Ian Smith, Judith Bliss, Mitch Dowsett. PD10-07 Low plasma estradiol, low expression of estrogen responsive genes and TP53 mutations are associated with poor anti-proliferative response to aromatase inhibitors [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD10-07.

Background: Aromatase inhibitors (AIs) are one of the main treatment strategies for the clinical management of estrogen receptor-positive (ER+) breast cancer (BC). Despite prolonged time to recurrence and initial clinical responses, >20% of patients eventually relapse, and previous studies have shown an association of poor anti-proliferative response to AIs and worse outcome. High immune activity in ER+ tumors may be associated with worse outcome, in contrast to ER-negative BC where immune infiltration is a feature associated with better outcome. Our work focused on understanding the correlations between immune cell infiltration and response to AI. Methods: All patients with ER+ HER2- tumors within the bottom 15% of Ki67 anti-proliferative responders to AIs (poor responders [PRs]; n=177) were selected from the PeriOperative Endocrine Therapy for Individualizing Care (POETIC) trial and matched on baseline Ki67 levels to good responders (GRs) within the 50% showing the best response (n=190). Response to AI was measured by the Ki67 percentage change after 2 weeks of treatment. PRs were further divided into groups expressing high ESR1 (PRs ESR1HIGH; n=119) and low ESR1 (PRs ESR1LOW; n=58) levels to represent PR subgroups that showed partial or no response to AIs. The percentage of stromal tumor-infiltrating lymphocytes (TILs) was assessed. Multiple immunofluorescence was performed for ER, CD3, CD20, CD68, FOXP3, and CD3/FOXP3 in 15 baseline samples from each of the GR, PR ESR1HIGH, and PR ESR1LOW populations and immune cell density in stromal or tumor compartments was estimated. Spearman correlations of TILs with Consensus tumor microenvironment (TME) deconvolution and Molecular Signatures Database hallmark gene sets were conducted. The relationship between the immune markers’ density and genes, hallmark gene sets and Consensus TME was assessed. Results: The percentage of TILs was significantly higher in the PR ESR1HIGH and PR ESR1LOW compared to the GRs (adjusted p< 0.05). As expected, TILs were highly correlated with T cells (particularly T-regulatory cells) and immune hallmark gene sets. There was a tendency for higher density of each of the immune markers in PRs compared to GRs, with significant differences being observed in stromal B-cell marker CD20 density (p< 0.05). Analysis showed a significant correlation between TILs and stromal FOXP3 marker density (FDR< 0.05), and stromal biomarker density was highly correlated to the gene expression of the encoding genes of the same tumors (CD3/CD3D, FOXP3/FOXP3, and CD20/MS4A1) (FDRs< 0.05). There was also a strong and significant correlation between the stromal expression of CD20, CD3, FOXP3, and CD3/FOXP3 with the immune hallmark gene sets (FDRs< 0.05). Finally, the immune phenotyping showed the expected correlations with TME deconvolution, with particularly strong correlations of CD20 and CD3 with B- and T-cell gene signatures, respectively (FDRs< 0.05). Conclusions: Different immune features indicated a broad involvement of several immune cell types in PRs to AIs, suggesting that the immune system might be associated with resistance of ER+ breast tumors to AI treatment. Spatial gene expression profiling is ongoing to characterize these tumors further and investigate potential mechanisms of AI resistance. Citation Format: Anastasia Alataki, Gene Schuster, Lila Zabaglo, Perry Maxwell, Elena López-Knowles, Elizabeth Folkerd, David Evans, Kally Sidhu, Holly Tovey, Nicholas Turner, Stephen Johnston, Maggie Chon U Cheang, John Robertson, Manuel Salto-Tellez, Ian Smith, Judith Bliss, Mitch Dowsett. PD10-08 Immune cell infiltration associated with poor anti-proliferative response to aromatase inhibitors in postmenopausal women with primary ER-positive HER2-negative breast cancer [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD10-08.

Background: POETIC was a phase III clinical trial, with patients randomised 2:1 to 2-week perioperative aromatase inhibitor (POAI) vs control for postmenopausal women with oestrogen receptor positive (ER+) early breast cancer (BC) (Smith et al., Lancet Oncology 2020). Our previous study on POETIC trial patients with ER+ human epidermal growth factor receptor 2 positive (HER2+) BC suggested both HER2 enriched subtype (HER2-E) and immune enrichment pre-POAI (baseline, B) are main drivers of poor early response to POAI (Bergamino et al., 2022). However, some patients with HER2-E or immune enriched BC at B still showed good response to POAI. In this study, we aim to further investigate a sub-cohort of ER+ HER2+ BC from the POETIC trial, including a subset of aforementioned HER2-E tumours, to further explore the multi-modal molecular characteristics of the tumours resistant to POAI. Methods: Proliferation rate was assessed as percentage of cancer cells stained by Ki67. Patient POAI response was determined by Ki67 reduction at 2 weeks of treatment. A sub-cohort of 37 patients were selected based on response and classified as poor responders (PR, reduction < 30%, n=18), good responders (GR, reduction > 90%, n=11) and good responders with HER2-E BC at B (GR, reduction > 65%, n=8). Paired B and post-POAI (surgery, S) samples were taken from each patient of the sub-cohort. Multiplex immunofluorescence (mIF) was performed on these samples, measuring the immune cell densities in stroma and tumour compartments using five biomarkers: CD3 (all T cells), CD20 (B cells), CD68 (Macrophages), FOXP3 (regulatory T cells), and CD3 FOXP3 co-expression. The samples were also profiled using Breast Cancer 360TM (NanoString, BC360), covering the expressions of 758 genes and 46 biological signatures. Wilcoxon test, hierarchical clustering and spearman correlation test were performed to compare the tumour characteristics of GR and PR. Results: In this study, two B and four S samples were not achievable for mIF experiments due to low tumour content. At B (n = 35), among the five mIF biomarker measurements in stroma and tumour, only the stromal CD3 density was significantly different between GR (median = 0.0013) and PR (median = 0.0003, p = 0.041). In GR, HER2-E BC at B were separated into immune-high and immune-low groups with mIF biomarkers at B; the immune-high group was more likely to change into luminal subtypes post-POAI, while the immune-low group remained HER2-E. After POAI, the density changes in five mIF biomarkers in stroma and CD68 in tumour were all significantly higher in PR than GR (Table 1, n of paired samples = 62). The BC360 signatures of BC p53 (p < 0.001), BC proliferation (p < 0.001), LumB (p < 0.001) and HER2-E correlation coefficients (p < 0.001) were significantly downregulated in GR after POAI, while LumA correlation coefficients (p < 0.001) were notably increased. Conclusions: Our results suggest that for this sub-cohort, increased stromal immune response is associated with poor response to 2-week POAI in ER+ HER2+ early BC. HER2-E GR display visible immune heterogeneity at B. Lower-risk BC characteristics were exhibited in GR after the 2-week treatment. Further integrating mIF imaging data and additional digital spatial profiling are ongoing to reveal additional characteristics of ER+ HER2+ BC and tumour microenvironment predicting POAI resistance. Table 1: List of medians of log2 fold changes in mIF biomarker densities between GR and PR among the 62 paired samples, and Wilcoxon test p-values. Citation Format: Xixuan Zhu, Hui Xiao, Elena López-Knowles, Milana A. Bergamino Sirvén, Anastasia Alataki, Perry Maxwell, Holly Tovey, Lucy Kilburn, Chris Holcombe, Anthony Skene, Ian Smith, John Robertson, Katherine A Hoadley, Roberto Salgado, Judith Bliss, Nicholas Turner, Manuel Salto-Tellez, Gene Schuster, Mitch Dowsett, Maggie Chon U Cheang. Deconstructing the molecular characteristics of ER+ HER2+ early breast cancer in the POETIC trial using multiplex immunofluorescence and gene expression profiles [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P2-03-08.

Introduction: Immune checkpoint blockade therapy is a new paradigm in cancer treatment with durable tumor regression and prolonged stabilization of disease in patients with advanced cancers, including non-small cell lung cancer (NSCLC). However, clinical assessment using immunohistochemistry (IHC) remains complex and ambiguous. Methods: PD-L1 IHC staining was performed as a routine clinical reflex test on 802 NSCLC cases over an 18 month period using an automated staining system with PD-L1 clone SP263. PD-L1 expression was assessed using image analysis (IA) by 2 individuals and concordance was evaluated between the recorded manual clinical score and the digitally generated score. Additionally, a subset of 330 samples, representing cases with either varying PD-L1 expression or difficulty in assessment, were stained by multiplex immunofluorescence. Sensitivity and specificity was compared between manual assessments of both chromogenic and multiplex IHC. Results: PD-L1 expression by IA showed significant concordance (R²=0.8248) to manual assessment. Sensitivity and specificity was 86.8% and 91.4%, respectively. Discordant cases were mainly due to spurious or poor staining quality, difficult classification by IA and overabundant macrophages and a low IA detection threshold. Evaluation of IA scores 10-49% and >70% revealed 96.8% concordance to manual assessment. Multiplex assessment enabled PD-L1+/CD68+ macrophages to be readily identified within PD-L1+/CK+ or PD-L1-/CK+ tumor nests. Manual assessment of multiplex vs. chromogenic IHC was highly concordant, with a sensitivity and specificity respectively of 97.8% and 91.8%. Conclusions: Deployment of IA in routine PD-L1 IHC assessment of NSCLC represents a beneficial time saving process and could aid case triage saving >50% of pathologist review time. However cases close to diagnostic thresholds still require a more comprehensive pathological evaluation. Multiplex IHC provides a higher level of confidence in PD-L1 assessment and could be offered for challenging cases by centers with access to appropriate expertise and specialist equipment. Citation Format: Matthew P. Humphries, Victoria Bingham, Fatima Abdullahi Sidi, Stephanie G. Craig, Stephen McQuaid, Jacquline James, Manuel Salto-Tellez. The impact of image analysis and multiplex hybridization on PD-L1 diagnostic case triage [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5307.

The surface marker CD44 has been identified as one of several markers associated with cancer stem cells (CSC) in solid tumors, but its ubiquitous expression in many cell types, including hematopoietic cells, has hindered its use in targeting CSCs. In this study, 28 paired primary tumor and adjacent nontumor gastric tissue samples were analyzed for cell surface protein expression. Cells that expressed pan-CD44 were found to occur at significantly higher frequency in gastric tumor tissues. We identified CD44v8-10 as the predominant CD44 variant expressed in gastric cancer cells and verified its role as a gastric CSC marker by limiting dilution and serial transplantation assays. Parallel experiments using CD133 failed to enrich for gastric CSCs. Analyses of another 26 primary samples showed significant CD44v8-10 upregulation in gastric tumor sites. Exogenous expression of CD44v8-10 but not CD44 standard (CD44s) increased the frequency of tumor initiation in immunocompromised mice. Reciprocal silencing of total CD44 resulted in reduced tumor-initiating potential of gastric cancer cells that could be rescued by CD44v8-10 but not CD44s expression. Our findings provide important functional evidence that CD44v8-10 marks human gastric CSCs and contributes to tumor initiation, possibly through enhancing oxidative stress defense. In addition, we showed that CD44v8-10 expression is low in normal tissues. Because CD44 also marks CSCs of numerous human cancers, many of which may also overexpress CD44v8-10, CD44v8-10 may provide an avenue to target CSCs in other human cancers. Cancer Res; 74(9); 2630–41. ©2014 AACR.

Apoptosis is essential for chemotherapy responses. We previously developed the mathematical models APOPTO-CELL and DR_MOMP to estimate the magnitude of tumor cells resilience to undergo and execute apoptosis in response to chemotherapy, and demonstrated that these models are prognostic markers of clinical responses in stage 3 colorectal cancer patients [Salvucci M et al., Clin Cancer Res 2017; Lindner AU et al., Gut 2018]. However, the heterogeneity of colorectal cancer tissue represents a hurdle that we have yet to overcome on our way to entirely understand differences in response and clinical outcome in patients. Quantification of protein levels as input for systems studies routinely reflects a local cell population average rather than protein levels in individual cells. In this study we aim to gain a better understanding of apoptosis resilience in tumors with respect to tumor cell heterogeneity and cell composition and local aspects of the tumor microenvironment. Two to three core slices of formalin-fixed and paraffin-embedded primary colorectal cancer tissue were collected from 165 stage II-IV patients from a previously used cohort [Salvucci M et al. 2017; Lindner AU et al. 2018]. The levels of 25 protein markers for cell type and immune status, apoptosis, metabolism and cell proliferation, as well as cell segmentation markers, were measured on TMA slides using a multiplexed tissue immunofluorescence imaging platform (Cell DIVE [Gerdes et al., Proc Natl Acad Sci USA 2013]). Batch correction between slides was performed using linear regression (using markers for regulatory T cell in patient tissues as housekeeping cells. Formalin-fixed gene deficient cells were used in parallel for antibody validation. The layout allowed us to create protein profiles for on average 13,079 (± 2,282 SD) cells per core per patient, and to categorize these as epithelial-like cells, regulatory T cells, T helper cells, cytotoxic T cells, other leukocytes and other cells. The data will be used to identify cells contributing most to the signal measured with traditional methods (such as reverse protein phase arrays) that were used to predict patients clinical response to therapy in our previous studies. Further, we will assess spatial differences in distribution and frequency of different cells and their contribution to clinical outcome in colorectal cancer patients. Specifically, we are investigating metrics to characterize the spatial features and layout found in each core in terms of textures, degree of lacunarity, neighboring cells composition and patterns of protein expression (diffuse vs. patchy) and co-expression among cell types. Further, we will characterize the degree of intra- and inter-patient spatial heterogeneity and will examine the association between the identified spatial motifs with macroscopical characteristics of the tumor and outcome. Citation Format: Andreas Urlich Lindner, Manuela Salvucci, Xanthi Stachtea, Steven Carberry, Philip D. Dunne, Anup Sood, Elizabeth McDonough, Sanghee Cho, Pierre Laurent-Puig, Sandra Van Schaeybroeck, Manuel Salto-Tellez, John F. Graf, Markus Rehm, Mark Lawler, Daniel B. Longley, Fiona Ginty Fiona, Jochen H. Prehn. Exploratory multiplex tissue image analysis of the impact of heterogeneity in the microenvironment of primary colorectal cancer on apoptosis markers in patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr LB-088.

Identifying robust predictive biomarkers to enable stratification of colorectal cancer (CRC) patients, based on their response to immune checkpoint therapy, is an area of unmet clinical need. Genetic algorithms represent an exciting branch of artificial intelligence which can be used to extract meaningful associations from ‘big data’ now emerging more frequently in oncological research. We have employed Atlas Correlation Explorer (ACE), a user-friendly workbench that utilises a genetic algorithm to mine data deposited in The Cancer Genome Atlas (TCGA). Our aim was to establish common intersections between gene expression analyses in ACE using nine well established immune checkpoint markers (CD274, PDCD1, CTLA4, LAG3, TIM3, TIGIT, ICOS, IDO1 and BTLA). We observed IL2RB to be the common gene associated with immune checkpoints in both microarray and RNA sequencing data from the TCGA (7/9 gene lists). Assessment of IL2RB indicates that it is highly expressed on CD56+ natural killer cells and is associated with an increased infiltration of cytotoxic lymphocytes and a decreased infiltration of fibroblasts. It is also significantly enriched in the immune consensus molecular subtype group CMS1. We next demonstrated that patients with high IL2RB gene expression have better relapse free survival in the TCGA CRC cohort (n = 322, log-rank p = 0.011) and an all stage CRC validation cohort GSE39582 (n = 519, log-rank p = 0.006). It is also an independent prognostic factor by multivariate analysis (p = 0.01). We next observed strong correlations between IL2RB gene expression in CRC and previously published predictive gene signatures for anti-PD1 therapies in other solid tumours (Pearson correlation, R = 0.88). Finally, we optimised assessment of IL2RB immunohistochemistry in a large CRC cohort (n=661) using a digital pathology approach with the open-source QuPath software. To conclude, we have validated IL2RB as prognostic biomarker and have provided evidence to demonstrate that IL2RB expression could be used for CRC patient stratification in future immunotherapy based clinical trials. Citation Format: Matthew Alderdice, Stephanie Craig, Matt Humphries, Alan Gilmore, Victoria Bingham, Nicole Johnston, Stephen McQuaid, Manuel Salto-Tellez, Mark Lawler, Darragh G. McArt. Artificial intelligence approach identifies IL2RB as a common prognostic and potential predictive biomarker associated with immune checkpoints in colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2787.

Copy number alterations of the tumor-suppressor gene PTEN occur frequently in primary prostate cancer and are prognostic for relapse post-radiotherapy (RT). Preclinical modelling supports elevated CXC-chemokine signaling as a critical mediator of PTEN-depleted disease progression, and therapeutic resistance. Our objective was to establish the correlation of PTEN-deficiency with CXC-chemokine signaling and its association with clinical outcome following RT. Techniques employed included gene cluster and survival analysis of prostate cancer datasets; in vitro biologic assays extending to RT-PCR, immunoblotting, clonogenic survival assays and flow cytometry; and the use of two in vivo PTEN-deficient xenograft models. Analysis of gene expression data from the MSKCC cohort characterized a PTENLOW/CXCR1HIGH/CXCR2HIGH cluster of tumors that associates with earlier time to biochemical recurrence (HR 5.87, p90 months). Kaplan-Meier analysis confirmed that PTENLOW/CXCR1HIGH/CXCR2HIGH tumors were associated with a significantly reduced time to biochemical recurrence (HR 2.65, p Citation Format: Chris W.D. Armstrong, Jonathan A. Coulter, Chee Wee Ong, Pamela J. Maxwell, Karl T. Butterworth, Oksana Lyubomska, Melissa J. LaBonte, Silvia Berlingeri, Rebecca Gallagher, Steven M. Walker, Joe M. O’Sullivan, Suneil Jain, Ian G. Mills, Manuel Salto-Tellez, Timothy Illidge, Richard D. Kennedy, Kevin M. Prise, David J.J. Waugh. Radio-resistance of PTEN-deficient prostate tumors is enhanced by treatment-induced chemokine signaling and is associated with biochemical recurrence and development of metastasis [abstract]. In: Proceedings of the AACR Special Conference: Prostate Cancer: Advances in Basic, Translational, and Clinical Research; 2017 Dec 2-5; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(16 Suppl):Abstract nr B035.

Background: Accurate identification of Gleason score (GS) 3+4 prostate cancer patients with low risk of disease progression is an unmet challenge in treatment decision making. We investigated the potential of using integrative genomic profiling to improve stratification of GS 3+4 prostate cancer patients to aid decision making. Methods: High-throughput gene expression, methylation and mutational analysis were performed on a subset of 28 GS 3+4 cases selected from a cohort (N = 69) of intermediate risk prostate cancer cases. Preselection was based on immunohistochemistry (IHC) expression of 6 major drivers (PTEN, MYC, RB1, TP53, AURKA and AR) of aggressive disease in prostate cancer. Fuzzy clustering and unsupervised hierarchical clustering were utilized to determine molecular subgroups. Genes with high expression differences between subgroups were identified through group difference statistics. A prognostic index based on Cox proportional hazard model was generated for individual and combinations of differentially expressed genes and was tested in GS 3+4 cases (total N = 215, median follow-up = 65.8 months) extracted from 3 prostate cancer datasets available at Gene Expression Omnibus (GSE21032, GSE16560 and GSE40272). Hazard risk (HR) and 95% confidence interval (CI) of association with outcomes (recurrence and overall survival) were analyzed using Cox modelling and log-rank analysis. Results: We found that GS 3+4 prostate cancer cases could be classified as "complex" (differential expression of more than one driver) or "simple" (differential expression of only one) based on IHC analysis. Focusing on the "simple" cases, gene expression and methylation PTEN expression were tightly correlated and not observed in other drivers (R = 0.41, P = 0.03). Group difference statistics showed 35 genes were highly expressed in PTEN-high subgroup and were evaluated individually and in combination for prognostic value in 3 independent cohorts. The prognostic potential of the combinatorial signature was observed in all 3 cohorts: GSE21032 (HR: 6.95, 95% CI: 2.73-17.54, P < 0.001), GSE40272 (HR: 6.40, 95% CI: 2.28-17.96, P < 0.001) and GSE16560 (HR: 1.77, 95% CI, 1.29-2.41; P = 0.003) datasets. Individual gene-by-gene analysis showed that 4 genes (ACTA2, ACTG2, MYH11 and TPM2) were prognostic (all P < 0.05) in all 3 cohorts. Compared to the combination of 35 genes (AUC = 0.728), the prognostic potential of combining these 4 genes was still of significant value (AUC = 0.689). Conclusion: This study shows that by using IHC as an upfront stratifier of intermediate risk prostate cancers, it is possible to identify, through subsequent differential gene expression profiling, a geneset with prognostic power across multiple cohorts. This strategy has not been used previously and the signature, subject to further validation, has the potential to impact on treatment decisions in GS 3+4 prostate cancer patients. Citation Format: Chee Wee Ong, Pamela Maxwell, Muhammad A. Alvi, Stephen McQuaid, Ian G. Mills, Manuel Salto-Tellez, David J. Waugh. A gene signature associated with PTEN activation defines good outcomes in intermediate-risk prostate cancer cases [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1555.

Background:Tumour infiltrating lymphocytes (TILs) have been associated with improved prognosis in cancer, including oropharyngeal squamous cell carcinoma (OPSCC); however, the survival benefit appears modest and favours HPV positive OPSCC. The current literature presents a mixed view on the best method to score TILs and if it would be beneficial to score all cancers of the head and neck. Use of immunotherapies to stimulate immune cell activation is increasing in the clinic and as such a robust method of assessment is needed if we want to apply this routinely to OPSCC. This study investigates the prognostic value of assessing TILs by H&E in HNSCC using digital pathology solutions against human eye visual assessment and by comparing TIL estimates in H&Es against markers of adaptive immunity (CD3, CD4,CD8 and FOXP3) using digital image analysis. Methods:This study utilised 190 HNSCC arising from the oropharynx. TMAs constructed, in triplicate, from these samples were stained with either H&E or IHC for CD3, CD4, CD8 and FOXP3. Visual estimates of TIL counts by human eye from H&E stained slides were assessed on the whole core at x10 magnification based on published thresholds by two observers to form a consensus score. Image analysis was performed on whole cores using image analysis software QuPath on H&E and IHC stained slides. HPV status was determined by p16 IHC and high risk HPV RNAScope. Statistical analyses were performed using R. Results:Scoring of TILs in H&Es by visual estimates was prone to misestimation and reduced survival estimates when using predetermined thresholds compared to machine scoring by image analysis. Comparison of digital pathology solutions assessing the TIL population in H&E and IHC stained slides found the percentage of TILs present to be significantly higher in the HPV positive's compared to the HPV negative's in the H&E, CD3 and CD8 immunostained slides but not CD4 or FOXP3. When dichotomised into high and low TILs only H&E based TIL counts and CD4/FOXP3 ratios were found to be independently predictive of survival when adjusted for age and smoking history. Conclusion:Image analysis creates continuous variables from which patient TIL counts can be reliably assessed. The adaptive immune landscape of HPV negative OPSCC is significantly different to HPV positive OPSCC of which only total TIL counts (H&E) and CD4/FOXP3 ratios appears to be predictive of five-year overall survival. Citation Format: Stephanie G. Craig, Pete Bankhead, Victoria Bingham, Stephen McQuaid, Maria Tumelty, Simon S. McDade, Manuel Salto-Tellez, Jacqueline A. James. Tumor-infiltrating lymphocytes and CD4/FOXP3 ratios reliably predict survival using digital image analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3142.

Introduction: Colorectal cancer (CRC) is a heterogeneous malignancy, treatment for which has principally been cytotoxic chemotherapy. Herein we describe the assessment of immune-biomarkers in a large patient cohort, demonstrating the therapeutic potential of immunotherapy in CRC. Methods: By immunohistochemistry, employing REMARK criteria, using well-validated antibodies, 4 biomarkers (CD3, CD4, CD8, FOXP3) in over 1000 cases, represented across 33 TMAs, were assessed for their protein expression. Digital pathological assessment of immune biomarker densities were quantified via the application of QuPath, an open source digital pathology platform, developed at Queens University, Belfast (https://qupath.github.io.). Biomarker densities were dichotomised by R.O.C analysis. Kaplan-Meier curves (Log-rank p values) were used to assess the impact of immune cell densities on overall survival (OS) and progression-free survival (PFS). Hazard ratios were assessed using the Cox proportional hazards regression model for univariate analysis. Spearman's rank was used to assess correlations of the biomarkers. Results: Expression of two biomarkers were found to vary by stage and three significantly affected OS. Across the patient cohort CD3, CD4 and CD8 were individually shown to be positively prognostic for OS alone (log-rank p = 0.0007, p = 8.65e-06, p = 0.006, respectively). CD4 demonstrated a more significant impact on OS in later stage disease (Log-rank stage II p = 0.0017, stage III p = 0.0002), while conversely CD8 only showed significance in earlier stage disease (Log-rank stage II p = 0.009, stage III p = 0.546). Although significantly associated with OS and within all stages, CD3 showed no differential expression in a particular stage. Correlative analysis demonstrated a statistically significant association between biomarkers. Upon univariate analysis, all biomarkers remained significantly prognostic for OS. All analyses were performed using R. Conclusions: These data demonstrate the prognostic significance of adaptive immune markers in a large CRC cohort on OS. Moreover, stage specific expression maybe clinically relevant and assessment within stages II and III may provide important stratifications for specific immunotherapy treatment regimes. We additionally add weight to the beneficial utility of digital pathology and its potential for integration into routine tissue biomarker assessment. Additional analysis of immune checkpoint and adaptive immune biomarkers in CRC is warranted. Citation Format: Matthew P. Humphries, Stephanie Craig, Victoria Bingham, Maurice Loughrey, Helen Coleman, Liam Murray, Maughan Tim, Stephen McQuaid, Jacqueline James, Manuel Salto-Tellez. Assessment of immune biomarkers by digital pathological analysis across a large colorectal cancer patient cohort predicts patient outcome and may provide a clinically relevant therapeutic index for immunotherapeutic treatment stratification [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4049.

Ligand-induced activation of peroxisome proliferator-activated receptor γ (PPARγ) inhibits proliferation in cancer cells in vitro and in vivo; however, the downstream targets remain undefined. We report the identification of a peroxisome proliferator response element in the promoter region of the Na+/H+ transporter gene NHE1, the overexpression of which has been associated with carcinogenesis. Exposure of breast cancer cells expressing high levels of PPARγ to its natural and synthetic agonists resulted in downregulation of NHE1 transcription as well as protein expression. Furthermore, the inhibitory effect of activated PPARγ on tumor colony-forming ability was abrogated on overexpression of NHE1, whereas small interfering RNA–mediated gene silencing of NHE1 significantly increased the sensitivity of cancer cells to growth-inhibitory stimuli. Finally, histopathologic analysis of breast cancer biopsies obtained from patients with type II diabetes treated with the synthetic agonist rosiglitazone showed significant repression of NHE1 in the tumor tissue. These data provide evidence for tumor-selective downregulation of NHE1 by activated PPARγ in vitro and in pathologic specimens from breast cancer patients and could have potential implications for the judicious use of low doses of PPARγ ligands in combination chemotherapy regimens for an effective therapeutic response. [Cancer Res 2009;69(22):8636–44]

A tumor suppressor function has been attributed to RUNX3, a member of the RUNX family of transcription factors. Here, we examined alterations in the expression of three members, RUNX1, RUNX2, and RUNX3, and their interacting partner, CBF-β, in breast cancer. Among them, RUNX3 was consistently underexpressed in breast cancer cell lines and primary tumors. Fifty percent of the breast cancer cell lines (n = 19) showed hypermethylation at the promoter region and displayed significantly lower levels of RUNX3 mRNA expression (P < 0.0001) and protein (P < 0.001). In primary Singaporean breast cancers, 9 of 44 specimens showed undetectable levels of RUNX3 by immunohistochemistry. In 35 of 44 tumors, however, low levels of RUNX3 protein were present. Remarkably, in each case, protein was mislocalized to the cytoplasm. In primary tumors, hypermethylation of RUNX3 was observed in 23 of 44 cases (52%) and was undetectable in matched adjacent normal breast epithelium. Mislocalization of the protein, with or without methylation, seems to account for RUNX3 inactivation in the vast majority of the tumors. In in vitro and in vivo assays, RUNX3 behaved as a growth suppressor in breast cancer cells. Stable expression of RUNX3 in MDA-MB-231 breast cancer cells led to a more cuboidal phenotype, significantly reduced invasiveness in Matrigel invasion assays, and suppressed tumor formation in immunodeficient mice. This study provides biological and mechanistic insights into RUNX3 as the key member of the family that plays a role in breast cancer. Frequent protein mislocalization and methylation could render RUNX3 a valuable marker for early detection and risk assessment. (Cancer Res 2006; 66(13): 6512-20)

We have previously identified a DNA damage response deficient (DDRD) subgroup in breast cancer, associated with loss of the Fanconi Anemia/BRCA DNA repair pathway. The 44-gene DDRD signature developed to prospectively identify this subgroup has been validated as predictive of response to DNA damaging chemotherapy in the treatment of breast cancer. This subgroup is defined by immune signalling, with immune genes, such as the chemokines CXCL10 and CCL5, and immune checkpoints IDO1 and PD-L1 (CD274) overexpressed within the signature and subgroup. Here we report innate immune pathway activation in this subgroup, and association with PD-L1 expression. Methods: We used isogenic cell line models to identify and model pathways constitutively active in DDRD cancer cells. CD4+, CD8+ T-cell infiltration and PD-L1 expression was identified by immunohistochemistry (IHC) on a previously described breast TMA of 191 breast tumor samples, of which 65 were in the DDRD subgroup. DNA microarray data from FFPE samples was used to identify other immune checkpoints associated with this subgroup. Results: CD4+ and CD8+ T-cell infiltration, and PD-L1 expression were significantly associated with DDRD breast tumors on IHC analysis. In addition, expression of immune checkpoints such as IDO1, TIM-3 and LAG-3 were associated with the DDRD subgroup. Increased migration of peripheral blood mononuclear cells was identified into media conditioned by DDRD cells, this was dependent on chemokines secreted by DDRD cells. Endogenous activation of the innate immune pathway STING/TBK1/IRF3 was identified in DDRD cancer cells, and required for expression of chemokines CXCL10 and CCL5. This chemokine expression was associated with cytosolic DNA detected by cGAS and activation of the innate immune pathway STING/TBK1/IRF3. Importantly, this pathway was cell cycle related with upregulation of chemokine expression in S-phase of the cell cycle. Chemokine expression could be induced by S-phase DNA damaging chemotherapy but not by taxanes. Similarly, PD-L1 was induced by treatment with DNA damaging agents in cancer cells dependent on STING. Conclusions: We have identified constitutive activation of the innate immune STING pathway in a DNA damage response deficient subgroup of breast cancer, which is associated with CD4+, CD8+ infiltration and PD-L1 expression. We propose activation of the STING pathway as an important signal for lymphocytic infiltration, independent of tumor-associated neoantigens. The DDRD signature could identify patients with innate immune activation in response to endogenous DNA damage, which may allow stratification for immune checkpoint targeted treatments. Moreover, S-phase DNA damaging agents activate PD-L1 expression, and a combination approach of immune checkpoint inhibitors and S-phase specific chemotherapy may be synergistic in the clinic. Citation Format: Eileen E. Parkes, Steven M. Walker, Nuala McCabe, Laura E. Taggart, Laura Hill, Niamh E. Buckley, Kienan I. Savage, Manuel Salto-Tellez, Stephen McQuaid, Mary T. Harte, Paul B. Mullan, D Paul Harkin, Richard D. Kennedy. A DNA damage response deficiency (DDRD) group in breast cancer is associated with activation of the STING innate immune pathway and PD-L1 expression. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4000.

With 1.4 million new cases every year, colorectal cancer (CRC) is the fourth most common cancer worldwide [Globocan 2012, WHO]. Despite therapeutic advances and improvements in overall care, TNM staging remains the best prognostic indicator for CRC patients’ clinical outcomes and is pivotal for deciding on use of adjuvant chemotherapy after resection of the tumour. Adjuvant chemotherapy is not recommended for many stage II patients and mostly high-risk patients receive chemotherapy. However, there is a lack of robust biomarkers for identifying patient response to chemotherapy, recurrence and mortality risk. We developed a system model of the BCL2 family of proteins (DR_MOMP) to assess the sensitivity of cells to genotoxic stress and to induce apoptosis triggered by chemotherapy. It calculates the stress dose required to induce mitochondrial outer membrane permeabilization (MOMP) based on absolute protein levels and the interaction of pro- and anti-apoptotic BCL2 family proteins. Cells predicted to require a high stress dose showed decreased cell death rates after being exposed to 5FU and Oxaliplatin. Profiles of BAK, BAX, BCL2, BCL(X)L and MCL1 were determined by Reverse Phase Protein Array (RPPA) technology in FFPE primary tumours collected from two distinct cohorts: stage III CRC patients who underwent adjuvant 5FU-based chemotherapy (n = 128), and stage II CRC patients from a completed clinical trial with patients randomised to adjuvant 5FU-based chemotherapy or observation only (n = 138). Protein profiles were inputted into DR_MOMP to determine chemotherapy sensitivity and to classify patients into high- or low risk categories. Findings were validated on the TCGA COAD cohort using both protein (RPPA) and mRNA (SeqV2 RSEM) expression data. Stage II patients classified as high-risk by DR_MOMP and randomised to observation only had approximately 2-fold increased risk of death from CRC compared to those classified as low-risk or received chemotherapy (HR 2.4; 95% CI 1.2-4.8; p-value = 0.0199). Among stage III patients treated with FOLFOX, those classified as high- versus low-risk had a more than 10-fold increased risk of death from CRC (HR 10.6; 95% CI 2.4-46.3; p-value < 0.0001). We validated this finding in 261 stage II-IV patients of the TCGA COAD cohort (HR 10.6; 95% CI 1.2-12.5; p-value = 0.0125). DR_MOMP predicted mortality risk independent of TNM staging and KRAS mutation status. Our system delivers a novel predictive and prognostic biomarker that could be combined with TNM staging when assessing initial risk and subsequent clinical management of CRC patients. Citation Format: Andreas U. Lindner, Manuela Salvucci, Mattia Cremona, Naser Monsefi, Sarah Curry, Clare Morgan, Alexa Resler, Robert O’Byrne, Orna Bacon, Michael Stuehler, Lorna Flanagan, Richard Wilson, Patrick G. Johnston, Manuel Salto-Tellez, Sophie Camilleri-Broët, Deborah A. McNamara, Bryan T. Hennessy, Elaine W. Kay, Pierre Laurent-Puig, Sandra Van Schaeybroeck, Jochen H.M. Prehn. Retrospective evaluation of a system model of the BCL2 family of proteins as a predictive and prognostic biomarker for the clinical outcome of stage II-IV colorectal cancer patients. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4924.

Introduction - MErCuRIC is a Phase Ib/II clinical trial study using a combined MEK1/2 - cMET inhibition in RAS MT and RAS WT (with aberrant c-MET) colorectal cancer patients. As part of the discovery efforts on the cases enrolled in Phase I, we aimed to analyze the mutation status of 10 genes that could potentially be associated to the doublet MEK/MET inhibition. This study compared the MEK/MET-specific panel with the Ion Torrent 50 gene panel, aiming to compare: long-established, commercially available panels against this newly developed panel; the Ion Torrent PGM2 platform against Illumina MiSeq v.3 600 bp chemistry; hot-spot-based against full exomes-designed; and to compare the use of different bioinformatics reporter systems. The overlapping genes between the panels were: EGFR (n = 3); BRAF (n = 4); KRAS (n = 8); NRAS (n = 1); MET (n = 8); PIK3CA (n = 6); and ERBB2 (n = 5). Summary of Method NGS Design - The Multiplicom - MErCuRIC specific MASTR assay includes PTEN, MAP2K1 (MEK), EGFR, KRAS, NRAS, BRAF, PIK3CA, ERBB2, MET and PIK3R1, involving 257 amplicons in 4 plexes covering all coding exons of the 10 genes. Of the 257, 21 are control amplicons to allow for gene deletions/amplifications. The average length of the amplicons is 198 bp ranging from 124 bp to 255 bp. Validation - From a pool of 120 routine tumour samples characterised with a 50 gene mutation panel (Ion Torrent PGM2) and confirmed by Sanger sequencing and/or COBAS (Roche) QPCR, 24 FFPE cases were selected representing colorectal cancer and 4 other solid tumour types; all included a variety of DNA quality, and DNA concentration was standardized prior to library preparation. Pre-analytical handling was in accordance with established protocols in a laboratory, clinically-accredited in the UK for tissue-based, anatomical pathology testing. Results The evaluation of this MEK/MET-specific panel (Illumina MiSeq platform) resulted in 100% coverage of all targets, a higher than 97% on target reads and higher than 99% of all amplicons within 20% of mean coverage. The design minimized the areas of low coverage. A small part of exon 9 of ERBB2 was covered suboptimally: the low covered region is 30 bp in size located at the 5’ end of exon 9. It is unlikely that this low coverage led to false negative results since no mutations are reported in the COSMIC database for this DNA fragment. The results were concordant in relation to mutations involved in the genes stated above. Importantly, the percentages of allele frequency between both methods were similar, with variations ranging from 0.2% to 11.5% with an average variation of 5.2%. Insertion/deletion (Indel) detection however, required alternative bioinformatics pathways. Conclusion After combining well-established quality metrics to cover pre-analytical aspects with suitable technologies such as the MiSeq platform (Illumina) and appropriate bioinformatics, we recognize that this MEK/MET-specific NGS panel is fit for purpose. Citation Format: Perry Maxwell, Jurgen Del Favero, Marc-Aurel Fuchs, Josep Tabernero, Tim Maughan, Mark Middleton, Richard Addams, Christian Rolfo, Brian Hennessey, Pierre Laurent-Puig, Alberto Bardelli, Thierry Andre, Vlad Popovici, Patrick Johnstone, Richard Wilson, Mark Lawler, Sandra Van Schaeybroeck, Manuel Salto Tellez. Validation of a MEK/MET-specific NGS panel for early phase trial interrogation. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1396.

Background: The c-MET proto-oncogene is frequently overexpressed (50-60%), amplified (1-3%), and mutated (1-3%) in colorectal cancer (CRC). Hepatocyte growth factor (HGF)-dependent and independent activation of c-MET has been associated with increased survival and resistance to targeted therapies. This study aimed to investigate the role of the HGF/c-MET axis in regulating migration/invasion in CRC, using pre-clinical models and clinical samples. Methods: In order to model CRC tumour cell invasion, we have generated invasive CRC subpopulations using Boyden Invasion chambers. To model the CRC microenvironment, we have used a range of co-culture techniques with CRC cells and colon fibroblasts. Migration/invasion was determined using xCELLigence System (Roche). c-MET expression in parental and invasive cell lines was measured using Western blotting and qRT-PCR. c-MET expression in CRC FFPE tissues was measured using IHC and RNAScope®. Results: We identified marked upregulated expression of c-MET at both the protein and transcript levels in our invasive CRC cell line models. Importantly, both parental and invasive subpopulations were found to be inherently dependent on c-MET for migration, as RNAi against c-MET abrogated migration/invasion in both parental and invasive models. We also demonstrated that stimulation of CRC cells with rh-HGF resulted in increased CRC cell migration/invasion. In addition, co-culture of CRC cells with colonic myofibroblasts, resulted in marked increases in migratory and invasive capacity, and this was dependent on HGF/c-MET signaling. Interestingly, stimulation with myofibroblast conditioned medium or HGF promotes rapid degradation of c-MET at the protein level, followed by recycling, while MET transcript remains unaltered, illustrating a dynamic expression of c-MET protein in response to activation. We further showed that MET is transcriptionally upregulated in tumour budding foci at the invasive front of a cohort of stage III CRC tumors. Intriguingly, c-MET protein levels do not correlate with the transcript, most likely due to a similar protein degradation process observed in our aforementioned in vitro models. Conclusions: We show for the first time a key role for transcriptional upregulation of MET as a molecular driver of tumour invasion, both in vitro and in stage III CRC tumours. Citation Format: Conor Bradley, Philip Dunne, Stephen McQuaid, Victoria Bingham, Mark Lawler, Manuel Salto-Tellez, Patrick Johnston, Sandra Van Schaeybroeck. Upregulation of the MET transcript is consistently associated with invasion and tumor budding in colorectal cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 924.

Small bowel accounts for only 0.5% of cancer cases in the US; a third of which are adenocarcinomas. But incidence rates have been rising at a rate of 2.4% per year over the last decade. Because of the rarity of this cancer, little is known about its molecular pathology and there are no molecular markers for diagnosis, predicting prognosis or therapeutic intervention. The aim of this study was therefore to look into this disease at a molecular level to better understand its biology and identify biomarkers and potential points of therapeutic intervention. Using a retrospective 28 patient matched normal-tumor cohort next generation sequencing (NGS) was performed using a 50 gene cancer hotspot panel, gene expression arrays were used to profile ∼29,000 RNA transcripts and 450k CpG methylation arrays were used to carry out DNA methylation analysis. We also looked at microsatellite instability (MSI), HER2 and p53 expression. NGS identified novel mutations in IDH1, CDH1, KIT, NRAS, FGFR2, FLT3, NPM1, PTEN, MET, AKT1, RET, NOTCH1 and ERBB4. Previously known mutations such as high-frequency KRAS and TP53 and low-frequency HER2 were also confirmed in our cohort. The average patient had 2.6 mutations with eight patients having only a single mutation to one having seven. Array data revealed 17% of CpGs and 5% of RNA transcripts assayed to be differentially methylated and expressed respectively (p This study has for the first time highlighted the extent of molecular changes taking place in SBA. The clinical potential of TP53 mutations and Kazald1 hypomethylation as prognostic biomarkers and CHN2 as a diagnostic biomarker are focus areas for further research by our group. Citation Format: Muhammad A. Alvi, Darragh G. McArt, Paul Kelly, Marc-Aurel Fuchs, Matthew Alderdice, Clare M. McCabe, Victoria Bingham, Claire McGready, Stephen McQuaid, Perry Maxwell, Peter Hamilton, Jacqueline A. James, Richard Wilson, Manuel Salto-Tellez. Comprehensive molecular pathology analysis of small bowel adenocarcinoma reveals novel targets with clinical utility. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4792. doi:10.1158/1538-7445.AM2015-4792

The c-MET proto-oncogene is frequently overexpressed (50-60%), amplified (5-10%), and mutated (5%) in colorectal cancer (CRC). Hepatocyte growth factor (HGF)-induced c-MET activation has been linked with migration, survival, invasion, and resistance to targeted therapy, and has been suggested as a possible prognostic biomarker for CRC. We are currently investigating the role of c-MET and microenvironment derived HGF in the acquisition of an invasive and migratory phenotype, and resistance to targeted therapy in CRC. In order to model CRC tumour cell invasion and metastasis, we have generated invasive subpopulations from CRC cells using Boyden Invasion chambers. Invasive cell lines were characterised for protein expression/activity by Western blotting, and analysed for migratory and invasive potential using the xCELLigence System (Roche). To model the CRC microenvironment, we have utilised a range of co-culture techniques with CRC cell lines and colon fibroblasts. c-MET expression in FFPE tissues was measured using IHC in a tissue microarray (TMA) derived from early stage CRC patients. HCT116 and LoVo invasive subpopulations showed an EMT-like, mesenchymal, migratory/invasive phenotype. In addition, increased expression and activation levels of c-MET were found in these sublines, which was determined to be ligand independent. Inhibition of c-MET using RNAi abrogated both basal and HGF-induced migration and invasion in CRC cell lines. Co-culture of CRC cells with HGF-expressing colon-derived fibroblasts leads to activation of c-MET, inducing CRC migration, invasion, and resistance to MEK inhibition, and this phenotype could be diminished using a HGF neutralising antibody. Significant increased expression levels of c-MET were also found in a CRC TMA compared to matched normal tissues, and also at the invasive edge of some CRC tumours, where it may be driving invasive biology. The identification of key pathways driving metastasis has a huge potential to change treatment strategies in CRC. We intend to further investigate the role of c-MET in regulation of migration, invasion, and resistance to therapy in CRC. We hypothesise that patients with high tumoural c-MET at the invasive front may benefit in particular from anti- c-MET therapies. Citation Format: Conor A. Bradley, Philip Dunne, Darragh McArt, Ken Arthur, Stephen McQuaid, Manuel Salto-Tellez, Patrick Johnston, Sandra Van Schaeybroeck. The role of c-MET/HGF signaling as a critical mediator of an invasive and resistant phenotype in colorectal cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4018. doi:10.1158/1538-7445.AM2015-4018

Loss of RUNX3 expression is suggested to be causally related to gastric cancer as 45% to 60% of gastric cancers do not express RUNX3 mainly due to hypermethylation of the RUNX3 promoter. Here, we examined for other defects in the properties of RUNX3 in gastric cancers that express RUNX3. Ninety-seven gastric cancer tumor specimens and 21 gastric cancer cell lines were examined by immunohistochemistry using novel anti-RUNX3 monoclonal antibodies. In normal gastric mucosa, RUNX3 was expressed most strongly in the nuclei of chief cells as well as in surface epithelial cells. In chief cells, a significant portion of the protein was also found in the cytoplasm. RUNX3 was not detectable in 43 of 97 (44%) cases of gastric cancers tested and a further 38% showed exclusive cytoplasmic localization, whereas only 18% showed nuclear localization. Evidence is presented suggesting that transforming growth factor-β is an inducer of nuclear translocation of RUNX3, and RUNX3 in the cytoplasm of cancer cells is inactive as a tumor suppressor. RUNX3 was found to be inactive in 82% of gastric cancers through either gene silencing or protein mislocalization to the cytoplasm. In addition to the deregulation of mechanisms controlling gene expression, there would also seem to be at least one other mechanism controlling nuclear translocation of RUNX3 that is impaired frequently in gastric cancer.

Background Colorectal cancer (CRC) is the third most common cancer worldwide. Among patients with early stage CRC, there are subsets of patients who will not benefit from adjuvant 5-FU or 5-FU/oxaliplatin treatment, highlighting the need for improved therapy options for early stage disease. In addition mutations in the KRAS proto-oncogene are found in 40%-45% of CRC patients and this mutation predicts resistance to EGFR mAb therapies. More recently it has been shown that treatments, such as bevacizumab or cetuximab, which prolong survival in the advanced disease setting, cannot be translated into successful adjuvant treatment strategies. The aim of this study was to identify therapy options which will be effective in the adjuvant setting and also whose knock-down is synthetically lethal with KRAS mutation in CRC models. Method A novel systems biology approach combined profiling of primary KRASMT/WT CRC tissues and isogenic paired KRASMT/WT CRC xenografts models. Metacore pathway analysis identified individual genes from novel KRAS-dependent pathways for incorporation into a primary RNAi screen. This screen was crossed with receptor tyrosine kinase analysis of our previously published pre-clinical adjuvant invasive cell line models. Our findings were validated by in vitro screening and expression analysis in CRC FFPE sections was measured using IHC. Results Our comprehensive approach identified 30 signalling pathways and 160 genes associated with KRASMT CRC cancer. Our siRNA screen confirmed a number of novel KRAS synthetic lethal target genes with a critical role in EMT and cell adhesion/migration. Down-regulation of members of the Eph/Ephrin family, in particular EpHA2, was found to be lethal in KRASMT but not KRASWT cells. This lethality was further synergistically enhanced in the presence of chemotherapy treatment. Overexpression of EphA2 was also identified in each of our invasive isogenic cell line models, regardless of genetic background, and EpHA2 silencing inhibited migration and invasion in a panel of CRC cells. EpHA2 expression has previously been reported as an essential component in cancer cells's ability to form tumors in vivo and in-line with this our EpHA2 overexpressing invasive subpopulations displayed an increased ability to form colonies. Importantly, high EpHA2 mRNA and protein expression were found to be associated with poor overall survival in both early stage CRC tissues and in advanced disease. Conclusion Using our novel comprehensive systems biology approach we have identified novel KRAS synthetic lethal targets and pathways. Combining these findings with targets which are effective against our pre-clinical adjuvant models we have identified EpHA2 as a key driver of invasion and migration and a synthetically lethal target in KRASMT CRC. In addition, we show that EpHA2 is a poor prognostic marker and an important novel target for KRASMT CRC tumors in both the adjuvant and advanced disease setting. Citation Format: Philip D. Dunne, Darragh G. McArt, Jaine K. Blayney, Sonali Dasgupta, Manuel Salto-Tellez, Patrick G. Johnston, Sandra Van Schaeybroeck. EpHA2 is an essential driver of invasion and a novel target in KRAS mutant colorectal cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2079. doi:10.1158/1538-7445.AM2014-2079

Breast cancer remains a frequent cause of female cancer death despite the great strides in elucidation of biological subtypes and their reported clinical and prognostic significance. We have defined a general cohort of breast cancers in terms of putative actionable targets, involving growth and proliferative factors, the cell cycle, and apoptotic pathways, both as single biomarkers across a general cohort and within intrinsic molecular subtypes. We identified 293 patients treated with adjuvant chemotherapy. Additional hormonal therapy and trastuzumab was administered depending on hormonal and HER2 status respectively. We performed immunohistochemistry for ER, PR, HER2, MM1, CK5/6, p53, TOP2A, EGFR, IGF1R, PTEN, p-mTOR and e-cadherin. The cohort was classified into luminal (62%) and non-luminal (38%) tumors as well as luminal A (27%), luminal B HER2 negative (22%) and positive (12%), HER2 enriched (14%) and triple negative (25%). Patients with luminal tumors and co-overexpression of TOP2A or IGF1R loss displayed worse overall survival (p=0.0251 and p=0.0008 respectively). Non-luminal tumors had much greater heterogeneous expression profiles with no individual markers of prognostic significance. Non-luminal tumors were characterised by EGFR and TOP2A overexpression, IGF1R, PTEN and p-mTOR negativity and extreme p53 expression. Our results indicate that only a minority of intrinsic subtype tumors purely express single novel actionable targets. This lack of pure biomarker expression is particular prevalent in the triple negative subgroup and may allude to the mechanism of targeted therapy inaction and myriad disappointing trial results. Utilising a combinatorial biomarker approach may enhance studies of targeted therapies providing additional information during design and patient selection while also helping decipher negative trial results. Citation Format: Manuel Salto-Tellez, David P. Boyle, Darragh McArt, Gareth Irwin, Charlotte Charlotte Wilhelm-Benartzi, Tong G. LIoe, Martha Minter, Stephen McQuaid, Paul Mullan, Richard D. Kennedy, Peter Hamilton, D P. Harkin. Defining a therapeutic classification of breast cancer by actionable targets. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1905. doi:10.1158/1538-7445.AM2014-1905

BACKGROUND: Next Generation Sequencing (NGS) has the potential of becoming an important tool in clinical diagnosis and therapeutic decision-making in oncology owing to its enhanced sensitivity in DNA mutation detection, fast-turnaround of samples in comparison to current gold standard methods and the potential to sequence a large number of cancer-driving genes at the one time. We aim to test the diagnostic accuracy of current NGS technology in the analysis of mutations that represent current standard-of-care, and its reliability to generate concomitant information on other key genes in human oncogenesis. METHODS: Thirteen FFPE clinical samples (8 lung adenocarcinomas, 3 colon carcinomas and 2 malignant melanomas) already genotyped for EGFR, KRAS and BRAF mutations by current standard-of-care methods (Sanger Sequencing and q-PCR), were analyzed for detection of mutations in the same three genes using two NGS platforms and an additional 43 genes with one of these platforms. The results were analyzed using closed platform-specific proprietary bioinformatics software as well as open third party applications (including, among others, the ‘Open’ CLC Genomics Workbench). RESULTS: Our results indicate that the existing format of the NGS technology performed well in detecting the clinically relevant mutations stated above, with apparent improved sensitivity, but may not be reliable for a broader unsupervised analysis of the wider genome in its current design. They also illustrate the relevance of different bioinformatics curation approaches, allowing different degrees of stringency, for the avoidance of potential false positives in the broader choice of genes selected, when compared with the expected mutation frequencies from sources such as the COSMIC database, for the tumor types analyzed. The systems are also able to detect other mutations outside the expected ones representing the current main "standard-of-care," the significance of which are uncertain at this point. CONCLUSIONS: Our study represents a diagnostically lead validation of the major strengths and weaknesses of this technology in its current state of development before consideration for diagnostic use. Citation Format: Clare McCourt, Perry Maxwell, Mark Catherwood, Manuel Salto-Tellez. Validation of Next Generation Sequencing technologies in comparison to current diagnostic gold standards for BRAF, EGFR and KRAS mutational analysis. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 57. doi:10.1158/1538-7445.AM2013-57

PTEN is a haplo-insufficient tumor suppressor and the most frequently mutated gene in prostate cancer (CaP). Moreover, loss of PTEN expression arising from allele deletion or promoter methylation is prevalent early disease. The objective of our study was to characterize how the genetic instability and/or stress-induced signaling arising from PTEN loss modulated the expression of clinically-relevant pro-inflammatory chemokines. Transient (siRNA) or prolonged (shRNA) repression of PTEN resulted in increased transcription-mediated increases in expression of CXCL8 and its receptors, CXCR1 and CXCR1 in DU145 and 22Rv1 cells. Stress-induced expression of CXCL8, CXCR1 and CXCR2 was greater in magnitude and duration in PTEN-null PC3 and LNCaP cells, in contrast to the dampened response observed in PTEN-expressing cells. PTEN had no effect on basal or stress-induced expression of CCL2, CXCL12 or their receptors in prostate cancer cells. Moroever, the efficacy of autocrine CXCL8 signaling to induce HIF-1 and NF-κB transcription was also greater in magnitude and duration in PTEN-null prostate cancer cells or following siRNA-mediated PTEN knockdown in DU145 and 22Rv1 cells. Consistent with in vitro data, we observed increased expression of the orthologous murine chemokine KC in atypical cytological features of PTEN+/− prostate tissue relative to normal epithelium in PTENWT glands. Finally, siRNA or shRNA-mediated attenuation of CXCL8 signaling selectively decreased the viability of CaP cells harboring partial or complete PTEN-loss. Furthermore, the knockdown of CXCL8 signaling potentiated the sensitivity of PTEN-depleted prostate cancer cells to clinicallyrelevant doses of ionozing radiation. We conclude that there is a selective upregulation of CXCL8 signaling in PTEN-deficient prostate epithelium, which contributes to the growth, survival and reduced therapeutic sensitivity of these cells. Citation Format: Jonathan Coulter, Pamela Maxwell, Melanie McKechnie, Jessica Neisen, Richard Kennedy, Manuel Salto-Tellez, Chris Albanese, David Waugh. PTEN loss induces selective upregulation of CXCL8 signaling to modulate survival and resistance to radiation [abstract]. In: Proceedings of the AACR Special Conference on Advances in Prostate Cancer Research; 2012 Feb 6-9; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2012;72(4 Suppl):Abstract nr A70.

Cellular Fas-associated death domain (FADD)-like interleukin 1β-converting enzyme (FLICE) inhibitory protein (c-FLIP) is a key inhibitor of caspase 8 (FLICE)-promoted apoptosis. We have previously shown c-FLIP to be a transcriptional target of the androgen receptor (AR) and NF-κB. Accordingly, we proposed that cells with elevated expression of c-FLIP may remain viable and withstand androgen-deprivation. Expression of c-FLIP was assessed in prostatectomy tissue while in vitro experiments were performed to determine its importance in regulating response to the androgen receptor (AR)-targeted therapeutic bicalutamide. c-FLIP expression was increased in pre-malignant prostatic intra-epithelial neoplasia (PIN) (P Citation Format: Clare McCourt, Pamela Maxwell, Roberta Mazzucchelli, Manuel Salto-Tellez, Joe O'Sullivan, Rodolfo Montironi, Marina Scarpelli, Daniel Longley, David Waugh. Elevated expression of c-FLIP in castrate-resistant prostate cancer antagonizes response to androgen receptor-targeted therapy [abstract]. In: Proceedings of the AACR Special Conference on Advances in Prostate Cancer Research; 2012 Feb 6-9; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2012;72(4 Suppl):Abstract nr B4.

Purpose: High-throughput DNA methylation analysis of the entire non-neoplastic colonic mucosa (NNCM) at different distances from tumor was performed to explore the extent and distribution (“topography”) of methylation-mediated field defects in colorectal cancer (CRC). Methylation in NNCM of CRC was compared with that of colonic mucosa from non-CRC patients. Materials and Methods: Seventy-one samples (NNCM and tumor) were obtained from rectosigmoid specimens of 3 CRCs (case 1, 2 stage IV; case 3, stage II). Entire NNCM from resection specimens was completely stripped from underlying layers, divided into equally-sized quadrants and numbered with first and last being proximal and distal to tumor respectively. Distances from tumor center to center of each NNCM quadrant were recorded. Eight unequivocally normal colonic mucosal samples from diverticulosis (n=3) and solitary rectal ulcer (n=1) were also analyzed. NNCM and tumor DNA were subjected to Illumina's Golden Gate DNA methylation analysis for 1506 CpG sites. Probes were defined as methylated, if beta-value > 0.297 and unmethylated, if ≤ 0.297, based on methylated (M) and unmethylated (HSD) controls at 5% false discovery rate. Unsupervised hierarchical cluster analysis was performed with methylation as both continuous and binarized variable. Average levels and frequency of methylation for all probes was used for correlating methylation with distance of NNCM from tumor at 5% significance level. Results: Tumor of case-1 was significantly methylated followed by that of case-2 and case-3 (average β-values: 0.415, 0.181 and 0.144). In case-1, NNCM samples approximately 4-5cm from tumor were highly methylated and clustered with tumor samples. Beyond 4-5 cms, levels (correlation coefficient: −0.418 and p=0.017) and frequency of methylation (correlation coefficient:-0.493 and p=0.004) in NNCM significantly decreased with increasing distance away from tumor. In case-2, the NNCM samples within 4-5 cms from tumor were less methylated and clustered with the tumor, which was equally methylated and although not significant, levels and frequency of methylation tend to increase with increasing distance away from tumor. In contrast, in case-3, neither the tumor clustered with the immediately adjacent NNCM nor was any correlation detected between methylation and distance from the tumor. Overall, NNCM of CRCs was significantly methylated than that of non-CRC patients. Significant methylation of HDAC9, ACTG2 and HRASLS (all p< 0.001) were detected in NNCM of CRCs than non-CRC patients. Conclusions: Methylation-mediated field cancerization was detected in NNCM of advanced CRCs usually within 4-5cms radius of the tumor. NNCM of CRCs was significantly methylated compared to colonic mucosa of non-CRC patients. Our data supports the notion that methylation-mediated field defects may be due to presence of a frankly invasive malignant tumor rather than a pre-existing event. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 1477.

Cancer stem cells (CSCs) have been discovered in cancers of many tissues but they have not been explored for gastric cancer. A candidate approach was taken to isolate putative gastric cancer stem cells from primary gastric cancer specimens because surface markers have been used to describe CSCs from different cancers. Cells dissociated from biopsy samples extracted from primary tumour and non-tumour sites of the same patient were analyzed by flow cytometry. Cell surface markers surveyed include CD45 for hematopoietic cells, CD31 for endothelial cells, CD44, CD133, CD34, CD117, CD166, CD90 and EpCAM as potential cancer stem cell markers. EpCAM levels were found to be higher in cells isolated from the tumor than non-tumor sites. Using median fluorescence intensity, it was found that cells at the tumor site express 1.2-1.7 fold higher level of EpCAM. This was observed in histopathologically distinct types of gastric tumors, namely intestinal subtypes that are well, moderately and poorly differentiated. Data from diffused type gastric tumors were not as consistent mainly because of sampling difficulties. 100,000 EpCAM-positive cells form xenograft tumors in NOD/SCID mice whereas 500,000 EpCAM-negative cells from a fraction that was not hematopoietic (CD45-negative), endothelial (CD31-negative) nor fibroblastic (CD140b-negative) were unable to do so. To further enrich for gastric cancer stem cells, the EpCAM-positive cells were sorted based on EpCAM+CD44+ or EpCAM+CD133+ cells. To expand the tumor material, we generated xenograft tumors directly from fresh tumor pieces. Histological analyses verified that the xenograft tumors recapitulate those in the patient samples. Cells dissociated from the xenograft tumor were then subjected to CSC identification assays. To date, three lines of xenograft tumors were generated from tumor tissues of well, moderately and poorly differentiated histological subtype of gastric cancers. As few as 1000 EpCAM+CD44+ cells were able to initiate tumors in NOD/SCID mice whereas 20,000 EpCAM+CD44- cells were required to do so. We are currently doing limiting dilution assays to estimate the number of cancer initiating cells in each fraction we isolate from the xenografts. We were unable to see differential tumorigenic potentials in EpCAM+CD133+ and EpCAM+CD133- fractions. In conclusion, we have identified a subpopulation of gastric cancer cells that potentially mark gastric cancer stem cells as characterized by their cancer-initiating potential. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 3371.