Mediator kinase inhibition reverses castration resistance of advanced prostate cancer

Mediator kinases CDK19 and CDK8, pleiotropic regulators of transcriptional reprogramming, are differentially regulated by androgen signaling, but both kinases are upregulated in castration-resistant prostate cancer (CRPC). Genetic or pharmacological inhibition of CDK8 and CDK19 reverses the castration-resistant phenotype and restores the sensitivity of CRPC xenografts to androgen deprivation in vivo. Prolonged CDK8/19 inhibitor treatment combined with castration not only suppressed the growth of CRPC xenografts but also induced tumor regression and cures. Transcriptomic analysis revealed that Mediator kinase inhibition amplified and modulated the effects of castration on gene expression, disrupting CRPC adaptation to androgen deprivation. Mediator kinase inactivation in tumor cells also affected stromal gene expression, indicating that Mediator kinase activity in CRPC molded the tumor microenvironment. The combination of castration and Mediator kinase inhibition downregulated the MYC pathway, and Mediator kinase inhibition suppressed a MYC-driven CRPC tumor model even without castration. CDK8/19 inhibitors showed efficacy in patient-derived xenograft models of CRPC, and a gene signature of Mediator kinase activity correlated with tumor progression and overall survival in clinical samples of metastatic CRPC. These results indicate that Mediator kinases mediated androgen-independent in vivo growth of CRPC, supporting the development of CDK8/19 inhibitors for the treatment of this presently incurable disease.

Genomic Mapping of Epidermal Growth Factor Receptor and Mesenchymal-Epithelial Transition-Up-Regulated Tumors Identifies Novel Therapeutic Opportunities

BACKGROUND

The identification of proteins in the cellular membrane of the tumoral cell is a key to the design of therapeutic agents. Recently, the bi-specific antibody amivantamab, targeting the oncogenic membrane proteins EGFR and MET, received regulatory approval for the treatment of adult patients with locally advanced or metastatic NSCLC.

METHODS

The authors interrogated several publicly available genomic datasets to evaluate the expression of both receptors and PD-L1 in most of the solid and hematologic malignancies and focused on prostate adenocarcinoma (PRAD) and pancreatic adenocarcinoma (PAAD).

RESULTS

In PAAD, EGFR highly correlated with PD-L1 and MET, and MET showed a moderate correlation with PD-L1, while in PRAD, EGFR, MET and PD-L1 showed a strong correlation. In addition, in tumors treated with immune checkpoint inhibitors, including anti-PD(L)1 and anti-CTLA4, a high expression of EGFR and MET predicted detrimental survival. When exploring the relationship of immune populations with these receptors, the authors observed that in PAAD and PRAD, EGFR moderately correlated with CD8+ T cells. Furthermore, EGFR and MET correlated with neutrophils in PRAD.

CONCLUSIONS

The authors identified tumor types where EGFR and MET were highly expressed and correlated with a high expression of PD-L1, opening the door for the future combination of bi-specific EGFR/MET antibodies with anti-PD(L)1 inhibitors.

Analyses of Genes Critical to Tumor Survival Reveal Potential 'Supertargets': Focus on Transcription

The identification of mechanisms that underlie the biology of individual tumors is aimed at the development of personalized treatment strategies. Herein, we performed a comprehensive search of genes (termed Supertargets) vital for tumors of particular tissue origin. In so doing, we used the DepMap database portal that encompasses a broad panel of cell lines with individual genes knocked out by CRISPR/Cas9 technology. For each of the 27 tumor types, we revealed the top five genes whose deletion was lethal in the particular case, indicating both known and unknown Supertargets. Most importantly, the majority of Supertargets (41%) were represented by DNA-binding transcription factors. RNAseq data analysis demonstrated that a subset of Supertargets was deregulated in clinical tumor samples but not in the respective non-malignant tissues. These results point to transcriptional mechanisms as key regulators of cell survival in specific tumors. Targeted inactivation of these factors emerges as a straightforward approach to optimize therapeutic regimens.

Analysis of circulating extracellular vesicle derived microRNAs in breast cancer patients with obesity: a potential role for Let-7a

BACKGROUND

The incidence of obesity, a known risk factor for several metabolic and chronic diseases, including numerous malignancies, has risen sharply in the world. Various clinical studies demonstrate that excessive Body Mass Index (BMI) may worsen the incidence, prognosis, and mortality rates of breast cancer. Thus, understanding the link tying up obesity and breast cancer onset and progression is critically important, as it can impact patients' survival and quality of life. Recently, circulating extracellular vesicle (EV) derived miRNAs have attracted much attention for their diagnostic, prognostic and therapeutic potential in oncology research. Although the potential role of EV-derived miRNAs in the early detection of breast cancer has been repeatedly mentioned, screening of miRNAs packaged within serum EVs has not yet been reported in patients with obesity.

METHODS

Circulating EVs were isolated from normal weight (NW), and overweight/obese (OW/Ob) breast cancer patients and characterized by Transmission Electron Microscopy (TEM), Nanoparticle Tracking Analysis (NTA), and protein marker expression. Evaluation of EV-associated miRNAs was conducted in a screening (RNA-seq) and a validation (qRT-PCR) cohort. Bioinformatic analysis was performed to uncover significantly enriched biological processes, molecular functions and pathways. ROC and Kaplain-Meier survival analyses were used for clinical significance.

RESULTS

Comparison of serum EV-derived miRNAs from NW and OW/Ob patients detected seven differentially expressed miRNAs (let-7a-5p, miR-122-5p, miR-30d-5p, miR-126-3p, miR-27b-3p, miR-4772-3p, and miR-10a-5p) in the screening cohort. GO analysis revealed the enrichment of protein phosphorylation, intracellular signal transduction, signal transduction, and vesicle-mediated transport among the top biological processes. In addition, the target genes were significantly enriched in pathways related to PI3K/Akt, growth hormones, and insulin signalings, which are all involved in obesity-related diseases and/or breast cancer progression. In the validation cohort, qRT-PCR confirmed a significant down-regulation of EV-derived let-7a in the serum of OW/Ob breast cancer patients compared to NW patients. Let-7a levels also exhibited a negative correlation with BMI values. Importantly, decreased let-7a miRNA expression was associated with higher tumor grade and poor survival in patients with breast cancer.

CONCLUSION

These results suggest that serum-EV derived miRNAs may reflect a differential profile in relation to a patient's BMI, which, once validated in larger cohorts of patients, could provide insights into novel specific biomarkers and innovative targets to prevent the progression of obesity-mediated breast cancer.

Abstract P5-02-04: Immune Signatures Display Subtype-Specific Activation in Breast Cancer

Background. The development of new anti-HER2 therapies for the treatment of HER2-positive breast cancer (BC) is changing the concept of HER2 dichotomization to select and treat this patient population. In addition, different immunotherapies tested in HER2-low BC are gaining continuous interest. However, a comprehensive characterization of HER2 BC subgroups and patients is required to identify the best treatment approach. Materials and Methods. Using 123 BC samples with gene expression and IHC/FISH determined HER2 status we determined cutoff values to identify HER2 positive, HER2 low, and HER2 ultralow cohorts. With the inclusion of hormone receptor (HR) status six clinically relevant cohorts were defined (HR+/HER2+, HR+/HER2-low, HR+/HER2-ultralow, and HR-/HER2+, HR-/HER2-low, HR-/HER2-low). An ingegrated database of 7,624 BC cases were assigned to the six subtypes. Prognosis determination was based on relapse-free survival (RFS), distant-metastasis-free survival (DMFS), and overall survival (OS). Clinical parameters evaluated include MKI67 expression, lymph node status and grade. All together 17 immune signatures resembling immune genes and related activated pathways were tested against the six molecular BC subgroups. Results. We defined a robust cutoff for HER2 expression levels to define six distinct HER2 BC molecular subgroups (>3034 for HER2 positivity and < 1780 for HER2 ultralow). Regardless the HR positivity, the overall distribution of HER2-low, and HER2-ultralow was 23% and 52%, respectively. In the HR+ subgroups the HER2-low showed a better prognosis as compared to the HER2-ultralow and HER2+ (RFS and DMFS P = 0.0048 and 0.0015, respectively) while there was no prognostic effect of HER2 expression in the HR- subgroups. Not surprisingly, an association with higher grade was demonstrated in all HR- as compared to the HR+ subgroups regardless of HER2 status. Overall, all HR- subgroups showed a higher involvement of immune genes as compared to the three HR+ subgroups. Of interest, HER2-low (HR+ and HR-) and HR-/HER2+ showed a significant overlap expression of immune signatures (71%). While the HR+/HER2-ultralow and HR+/HER2+ displayed minimal activation of immune pathways, the HR-/HER2-ultralow was the group most significantly associated with the activation of immune signaling including IFN signaling (67% percent of genes in the panel with altered expression), T cell active cytokines (34% of genes hit), and cytotoxic effector molecules (48% of genes hit). Conclusions. Our study supports a further molecular stratification of breast cancer based on HER2 status. The different tumor-immune background of these BC subgroups highlights that selected patient cohorts may derive benefit from targeted immunotherapy.

Citation Format: Gyongyi Munkacsy, Libero Santarpia, Balazs Gyorffy. Immune Signatures Display Subtype-Specific Activation in 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 P5-02-04.

Clinical sequencing identifies potential actionable alterations in a high rate of urachal and primary bladder adenocarcinomas

OBJECTIVE

Administration of targeted therapies provides a promising treatment strategy for urachal adenocarcinoma (UrC) or primary bladder adenocarcinoma (PBAC); however, the selection of appropriate drugs remains difficult. Here, we aimed to establish a routine compatible methodological pipeline for the identification of the most important therapeutic targets and potentially effective drugs for UrC and PBAC.

METHODS

Next-generation sequencing, using a 161 cancer driver gene panel, was performed on 41 UrC and 13 PBAC samples. Clinically relevant alterations were filtered, and therapeutic interpretation was performed by in silico evaluation of drug-gene interactions.

RESULTS

After data processing, 45/54 samples passed the quality control. Sequencing analysis revealed 191 pathogenic mutations in 68 genes. The most frequent gain-of-function mutations in UrC were found in KRAS (33%), and MYC (15%), while in PBAC KRAS (25%), MYC (25%), FLT3 (17%) and TERT (17%) were recurrently affected. The most frequently affected pathways were the cell cycle regulation, and the DNA damage control pathway. Actionable mutations with at least one available approved drug were identified in 31/33 (94%) UrC and 8/12 (67%) PBAC patients.

CONCLUSIONS

In this study, we developed a data-processing pipeline for the detection and therapeutic interpretation of genetic alterations in two rare cancers. Our analyses revealed actionable mutations in a high rate of cases, suggesting that this approach is a potentially feasible strategy for both UrC and PBAC treatments.

In Vitro Insertional Mutagenesis Screen Identifies Novel Genes Driving Breast Cancer Metastasis

Metastasis, a complex, multistep process, is responsible for the overwhelming majority of cancer-related deaths. Despite its devastating consequences, it is not possible to effectively treat cancer that has spread to vital organs, the mechanisms leading to metastasis are still poorly understood and the catalogue of metastasis promoting genes is still incomprehensive. To identify new driver genes of metastasis development, we performed an in vitro Sleeping Beauty transposon-based forward genetic screen in non-metastatic SKBR3 human breast cancer cells. Boyden chamber- based matrix invasion assays were used to harvest cells that acquired a de novo invasive phenotype. Using targeted RNA-sequencing data from eighteen pools of invasive cells, we carried out a gene-centric candidate gene prediction and identified established and novel metastasis driver genes. Analysis of these genes revealed their association with metastasis related processes and we further established their clinical relevance in metastatic breast cancer. Two novel candidate genes, GIT2 and MUSK, were functionally validated as metastasis driver genes in a series of in vitro and in vivo experimental metastasis models. We propose that our robust and scalable approach will be a useful addition to the toolkit of methodological resources used to identify genes driving cancer metastasis. Implications: Novel metastasis drivers were identified in a human breast cancer cell line by performing an in vitro, Sleeping Beauty transposon-based forward genetic screen and an RNA fusion-based candidate gene prediction.

Abstract P4-07-20: Survival analysis using the entire transcriptome to pinpoint biomarkers with the highest prognostic power

Introduction. Extensive research is directed to uncover new biomarkers capable to stratify breast cancer patients into clinically relevant cohorts. However, the overall performance ranking of such marker candidates compared to other genes is virtually absent. Here, we present the ranking of all survival related genes in chemotherapy treated basal and estrogen positive/HER2 negative breast cancer. Methods. We searched the GEO repository to uncover transcriptomic datasets with available follow-up and clinical data. After quality control and normalization, samples entered an integrated database. Molecular subtypes were designated using gene expression data. Survival analysis was performed using Cox proportional hazards regression. False discovery rate was computed to combat multiple hypothesis testing. Kaplan-Meier plots were drawn to visualize the best performing genes. Results. The entire database includes 7,830 unique samples from 55 independent datasets. Of those with available relapse-free survival time, 3,382 samples were estrogen receptor-positive and 696 were basal. In chemotherapy treated ER positive/ERBB2 negative patients the significant prognostic biomarker genes achieved hazard rates between 1.76 and 3.33 with a p value below 5.8E-04. The significant prognostic genes in adjuvant chemotherapy treated basal breast cancer samples reached hazard rates between 1.88 and 3.61 with a p value below 7.2E-04. Finally, an online accessible platform utilizing the entire database was set up enabling the validation of future biomarker candidates. Conclusions. A reference ranking for all genes in two chemotherapy treated breast cancer cohorts is presented. The results help to neglect those with unlikely clinical significance and to focus future research on the most promising candidates.

Citation Format: Balazs Gyorffy. Survival analysis using the entire transcriptome to pinpoint biomarkers with the highest prognostic power [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P4-07-20.

Concomitant activation of GLI1 and Notch1 contributes to racial disparity of human triple negative breast cancer progression

Mortality from triple negative breast cancer (TNBC) is significantly higher in African American (AA) women compared to White American (WA) women emphasizing ethnicity as a major risk factor; however, the molecular determinants that drive aggressive progression of AA-TNBC remain elusive. Here, we demonstrate for the first time that AA-TNBC cells are inherently aggressive, exhibiting elevated growth, migration, and cancer stem-like phenotype compared to WA-TNBC cells. Meta-analysis of RNA-sequencing data of multiple AA- and WA-TNBC cell lines shows enrichment of GLI1 and Notch1 pathways in AA-TNBC cells. Enrichment of GLI1 and Notch1 pathway genes was observed in AA-TNBC. In line with this observation, analysis of TCGA dataset reveals a positive correlation between GLI1 and Notch1 in AA-TNBC and a negative correlation in WA-TNBC. Increased nuclear localization and interaction between GLI1 and Notch1 is observed in AA-TNBC cells. Of importance, inhibition of GLI1 and Notch1 synergistically improves the efficacy of chemotherapy in AA-TNBC cells. Combined treatment of AA-TNBC-derived tumors with GANT61, DAPT, and doxorubicin/carboplatin results in significant tumor regression, and tumor-dissociated cells show mitigated migration, invasion, mammosphere formation, and CD44⁺/CD24^- population. Indeed, secondary tumors derived from triple-therapy-treated AA-TNBC tumors show diminished stem-like phenotype. Finally, we show that TNBC tumors from AA women express significantly higher level of GLI1 and Notch1 expression in comparison to TNBC tumors from WA women. This work sheds light on the racial disparity in TNBC, implicates the GLI1 and Notch1 axis as its functional mediators, and proposes a triple-combination therapy that can prove beneficial for AA-TNBC.

A quantitative polymerase chain reaction based method for molecular subtype classification of urinary bladder cancer-Stromal gene expressions show higher prognostic values than intrinsic tumor genes

Transcriptome-based molecular subtypes of muscle-invasive bladder cancer (MIBC) have been shown to be both prognostic and predictive, but are not used in routine clinical practice. We aimed to develop a feasible, reverse transcription quantitative polymerase chain reaction (RT-qPCR)-based method for molecular subtyping. First, we defined a 68-gene set covering tumor intrinsic (luminal, basal, squamous, neuronal, epithelial-to-mesenchymal, in situ carcinoma) and stromal (immune, extracellular matrix, p53-like) signatures. Then, classifier methods with this 68-gene panel were developed in silico and validated on public data sets with available subtype class information (MD Anderson [MDA], The Cancer Genome Atlas [TCGA], Lund, Consensus). Finally, expression of the selected 68 genes was determined in 104 frozen tissue samples of our MIBC cohort by RT-qPCR using the TaqMan Array Card platform and samples were classified by our newly developed classifiers. The prognostic value of each subtype classification system and molecular signature scores were assessed. We found that the reduced marker set combined with the developed classifiers were able to reproduce the TCGA II, MDA, Lund and Consensus subtype classification systems with an overlap of 79%, 76%, 69% and 64%, respectively. Importantly, we could successfully classify 96% (100/104) of our MIBC samples by using RT-qPCR. Neuronal and luminal subtypes and low stromal gene expressions were associated with poor survival. In conclusion, we developed a robust and feasible method for the molecular subtyping according to the TCGA II, MDA, Lund and Consensus classifications. Our results suggest that stromal signatures have a superior prognostic value compared to tumor intrinsic signatures and therefore underline the importance of tumor-stroma interaction during the progression of MIBC.

ER and immune-related signatures define benefit to palbociclib, trastuzumab, pertuzumab +/- fulvestrant in ER+/HER2+ breast cancer patients in the NA-PHER2 trial

555

Background: In the NA-PHER2 study we assessed the association between biological pathways with pathological complete response (pCR) and Ki67 down-regulation Methods: Patients with centrally confirmed ER+ ( > 10%) HER2+ breast cancer (BC) were treated in two independent, non-randomized cohorts with neoadjuvant trastuzumab, pertuzumab, palbocilib with (Fulv, n = 30) or without (NoFulv, n = 28) fulvestrant (+/-LHRH analogues). We assessed RNA-seq on core-biopsies obtained pre-treatment [n = 53/58 (91.4%)], at day 14 [n = 49/58 (84.5%)], and on residual disease at surgery [n = 42/45 (93.3%)]. We investigated biomarker dynamics and association with pCR or Ki67 down-regulation (centrally evaluated) at day 14 and at surgery. In the overall population and in each cohort, we primarily assessed three pre-defined biomarkers (ER-metagene [from OncotypeDX], a CD8-metagene and ERBB2 expression), and secondarily we explored a pre-defined list of genesets. Continuous and categorical (median cut-point) variables were evaluated. Results: In the biomarker population, pCR rate was 22.5% (28.6% and 16.0% in Fulv and NoFulv cohorts). At baseline, continuous CD8-metagene (OR 1.85 [1.12-3.06], p = 0.016) and ER-metagene (OR 0.56 [0.34-0.90], p = 0.016) associated with higher and lower pCR rate, respectively. High ERBB2 (above median) was marginally associated with pCR (OR 3.83 [0.90-16.3], p = 0.068). Only ER- and CD8-metagenes retained significance in multivariate analysis and were similarly predictive in both cohorts. Combining categorical variables, the groups with high-CD8/low-ER and low-CD8/high-ER had 61.5% and 0% pCR rate respectively, whereas low-CD8/low-ER and high-CD8/high-ER had similar 15% pCR (p = 0.001). The association was significant in both cohorts (p = 0.019 Fulv; p = 0.028 NoFulv). Dynamic assessment of the same biomarkers at day 14 did not improve prediction. Higher ER-metagene at baseline, but not CD8 and ERBB2, was associated with robust down-regulation of Ki67 at day 14 (Ki67 < 2.7%, complete cell cycle arrest) only in Fulv cohort (p = 0.016). ER-metagene also associated with retained Ki67 down-regulation (Ki67 < 10%) at surgery (p = 0.002). Alternative ER- and immune-related signatures provided very similar results. The comprehensive landscape of complex molecular dynamics and exploratory association with outcome will be presented. Conclusions: In ER+/HER2+ BC, low expression of ER-related and high expression of immune-related genes identified patients with very high likelihood of achieving pCR with a chemo-free regimen. In the fulvestrant cohort, the group with high ER-metagene, despite a lower pCR rate, had higher Ki67 down-regulation at day 14, which has been associated with long-term benefit in luminal tumors. These findings provide a potential tool for tailored de-escalation strategies.

Genomic mapping to identify mutations in RYR2 and AHNAK in basal-like breast tumors expressing PD-L1

1027

Background: Basal-like breast cancer is a specific subtype of breast tumors with limited therapeutic options. Although treatment with the anti-PD-L1 antibody atezolizumab has recently shown clinical activity in this setting, not all patients do respond even expressing high levels of PD-L1. In the present article we explored the presence of mutations in breast cancer tumors with high expression of PD1 and PD-L1 with the aim to identify molecular correlates associated with outcome. Methods: We used RNA-seq and mutational data from 971 breast cancer patients using the TCGA dataset, to identify mutations in patients with high levels of PD1 and PD-L1. Data analysis was performed using DESeq R and MAFTools Bioconductor packages. Transcriptomic signatures from the identified mutations were associated with outcome using the Kapplan-Meyer Plotter tool. We correlated the identified transcripts with immune populations using TIMER online tool and correlation between genes with Cancertool online platform. Results: We identified co-occurrent mutations in RYR2 and AHNAK in 8% and 5% of basal like tumors, respectively, in patients with high levels of PD1 and PD-1. The transcriptomic signature of these mutations conferred good prognosis for relapse free survival (RFS) and overall survival (OS). CXCL9 for RYR2 and GBP5, C1QA, IL2RG, CSF2RB and IDO1 for AHNAK were the most relevant genes identified in these signatures. Expression of CXCL9, GBP5, IL2RG and IDO1 correlated with the presence of immune cell populations mainly dendritic cells. This signature, including CXCL9, GBP5, C1QA, IL2RG, CSF2RB and IDO1 classified patients with favorable RFS (HR 0.27 CI 0.2-0.30; p = 1.1e-16) and OS (HR 0.18 CI 0.09-0.34; p = 6.8e-9). This signature showed a stronger prediction capacity compared with already described immunologic signatures. Finally we identify that LAG3 was the only gene commonly present in both signatures and correlated positively with the expression of PD1 and PD-L1. Conclusions: We describe two novel mutations which transcriptomic signatures associated with favorable outcome in basal-like tumors that express elevated levels of PD1 and PD-L1. Future studies should be performed to confirm the role of these mutations and signatures in relation with clinical activity of PD1/PD-L1 inhibitors.

Identification of a stemness-related gene panel associated with BET inhibition in triple negative breast cancer

e13098

Background: Triple negative breast cancers (TNBCs) are enriched in cells bearing stem-like features, i.e., cancer stem cells (CSCs), which underlie cancer progression. Thus, targeting stemness may be an interesting treatment approach. The epigenetic machinery is crucial for maintaining the stemness phenotype. Bromodomain and extra-terminal domain (BET) epigenetic reader family members are emerging as novel targets for cancer therapy, and have already shown preclinical effects in breast cancer. Here, we aimed to evaluate the effect of the BET inhibitor JQ1 on stemness in TNBC. Methods: Transcriptomic, functional annotation and qRT-PCR studies were performed on JQ1-exposed TNBC cells in culture. The results obtained were confirmed in spheroids and spheroid-derived tumours. In addition, limiting dilution, secondary and tertiary tumour sphere formation, matrigel invasion, immunofluorescence and flow cytometry assays were performed to evaluate the effect of JQ1 on CSC features. For clinical outcome analyses, the online tool Kaplan-Meier Plotter and an integrated response database were used. Results: We found that JQ1 modified the expression of stemness-related genes in two TNBC-derived cell lines, MDA-MB-231 and BT549. Among these changes, the CD44/CD24 ratio and ALDH1A1 expression level, i.e., both classical stemness markers, were found to be decreased by JQ1. Using a validated spheroid model to mimic the intrinsic characteristics of CSCs, we found that JQ1 decreased surface CD44 expression, inhibited self-renewal and invasion, and induced cell cycle arrest in G0/G1, thereby altering the stemness phenotype. We also found associations between four of the identified stemness genes, GJA1 , CD24, EPCAM and SOX9, and a worse TNBC patient outcome. The expression of another two of the stemness-related genes was found to be decreased by JQ1, ABCG2 and RUNX2, and predicted a low response to chemotherapy in TNBC patients, which supports a role for RUNX2 as a potential predictive marker for chemotherapy response in TNBC. Conclusions: We identified a stemness-related gene panel associated with JQ1 and describe how this inhibitor modifies the stemness landscape in TNBC. Therefore, we propose a novel role for JQ1 as a stemness-targeting drug. Loss of the stem cell phenotype via JQ1 treatment could lead to less aggressive and more chemo-sensitive tumours, reflecting a better patient prognosis. Thus, the identified gene panel may be of interest for the clinical management of patients with aggressive TNBC.

Abstract P4-05-23: Validation of proteome based molecular stratification of breast cancer using 1,229 patient samples from four independent datasets

Introduction. In routine diagnostics, four genes are used to determine the molecular subtype and the optimal systemic therapy for breast cancer patients, ESR1, PGR, HER2, and in some cases MKI67. While these can be measured by different molecular genetic methods, immunohistochemistry is still the most widely used technique. The aim of present study was to determine how effective recent proteomic technologies are in measuring the expression of these genes.

Methods. A literature search was performed to identify available proteome-level datasets. Differential expression was compared using Mann-Whitney test. Receiver operating characteristics (ROC) was computed to validate discriminative power and to determine the optimal cutoff values. Each dataset was processed separately.

Results. All together four datasets with a total of 1,229 specimens with proteome level-data were identified. One dataset was generated using RPPA (reverse phase protein lysate microarray, n=874) and three using LC-MS (liquid chromatography-mass spectrometry, n=355). In all four datasets combined, a total of 7,419 proteins were measured. It was not possible to analyze all combinations because of limited data availability - HER2 data was available in the RPPA study only, and ESR1/PGR in the LC-MS studies only. The determination of HER2 status by IHC and HER2 protein expression by RPPA show very high correlation (p=2.9E-18). ESR1 protein expression determined by LC-MS show a significant correlation to IHC results (p=0.04). The ROC analysis for HER2 delivered an area under the ROC curve (AUC) of 0.74 (p=1.9e-20). For ESR1, the AUC was 0.61 (p=0.03). ESR1 and PGR show a moderate correlation in LC-MS data (correlation coefficient=0.17, p=0.04). Finally, when comparing MKI-67 expression between normal (n=53) and tumor samples (n=65) within one dataset, tumor samples had significantly higher expression (fold change=2.22, p=3.9E-05).

Conclusions. The RPPA method achieved high correlation with IHC results. Our results suggest that proteomic technologies will be capable to deliver molecular stratification enabling the discrimination of breast cancer subtypes in the future.

Citation Format: Balazs Gyorffy. Validation of proteome based molecular stratification of breast cancer using 1,229 patient samples from four independent datasets [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-05-23.

Low level of exosomal long non-coding RNA HOTTIP is a prognostic biomarker in colorectal cancer

Molecular risk stratification of colorectal cancer can improve patient outcome. A panel of lncRNAs (H19, HOTTIP, HULC and MALAT1) derived from serum exosomes of patients with non-metastatic CRC and healthy donors was analyzed. Exosomes from healthy donors carried significantly more H19, HULC and HOTTIP transcripts in comparison to CRC patients. Correlation analysis between lncRNAs and clinical data revealed a statistical significance between low levels of exosomal HOTTIP and poor overall survival. This was confirmed by multivariate analysis that HOTTIP is an independent prognostic marker for overall survival (HR: 4.5, CI: 1.69-11.98, p = 0.0027). Here, HOTTIP poses to be a valid biomarker for patients with a CRC to predict post-surgical survival time.

Abstract A006: Skp2 associates with Slug and androgen receptor in aggressive prostate cancer

Introduction: The Skp2 F-box protein is the substrate recruiting component of the SCF (Skp1-Cullin 1-F-box) type of E3 ubiquitin-ligase complexes. Skp2 plays an important role in prostate tumorigenesis (e.g., via recently reported stabilization of EZH2 or Twist1), which needs further elucidation.

Patients and Methods: Prostate cancer patient cohort (N=101) was analyzed by immunohistochemistry for the following proteins: Skp2, AR, Ki-67, Slug, E-cadherin, MdmX, Mdm2, and p53. The indicated proteins were also analyzed in seminal vesicles and lymph node metastases (SV=35 and LN=18 patients, respectively), as well as in benign prostatic hyperplasia (BPH=34). Colocalization analysis was performed using Perkin Elmer Opal Multiplex kit, Vectra 3.0 imaging system, and confocal microscope Carl Zeiss LSM 780. Independent prostate cancer cohort (N=30) was evaluated for TMPRSS2-ERG fusion by a triple-color deletion assay and assessed for Skp2, ERG, and NQO1 protein expression. Cox regression analysis and Kaplan-Meier curves were used to assess importance of Skp2, Slug, EZH2, and Twist1 for biochemical relapse in the TCGA prostate cancer dataset (N=488, RNA seq).

Results: Expression of E-cadherin was lower in primary prostate cancer and seminal vesicles compared to BPH and LN, which may be linked to epithelial-mesenchymal transition. Opposite expression profile was observed for Skp2, MdmX, and Slug. High Gleason score (≥8, N=30) was significantly associated with higher Skp2 and lower E-cadherin expression (p&lt;0.001 and 0.011, respectively). Skp2 slightly correlated with AR and Slug in the whole patient cohort (Rs 0.37 and 0.322, respectively). These correlations were stronger in patients with high Gleason score (Rs 0.53 and 0.56, respectively). In patients with metastasis into lymph nodes, Skp2 similarly correlated with Slug and AR (Rs 0.56 and 0.37, respectively) while androgen receptor further correlated with Ki-67 (Rs 0.50). Confocal microscopy revealed colocalization of Skp2 and Slug in prostate cancer cells. In an independent cohort, Skp2 correlated with expression of ERG, NQO1, and percentage of cancer cells positive for TMPRSS2-ERG (Rs 0.503, 0.387 and 0.615, respectively). In the TCGA dataset, high mRNA levels of Skp2 and EZH2 (but not Slug or Twist1) associated with shorter time to biochemical relapse, compared to low mRNA levels [HR 3.62 (2.1 – 6.23), p&lt;0.001].

Conclusion: Immunohistochemistry and colocalization studies revealed association between Skp2 and Slug in aggressive prostate cancer. Other factors such as TMPRSS2-ERG fusion or oxidative stress may contribute to prostate cancer progression.

Citation Format: Jan Bouchal, Gvantsa Kharaishvili, Alena Mickova, Mariam Gachechiladze, Alena Burdova, Martin Mistrik, Milan Kral, Balazs Gyorffy, Zdenek Kolar. Skp2 associates with Slug and androgen receptor in aggressive prostate cancer [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 A006.

Abstract 201: Multiregion joint somatic genotyping in ovarian cancer proves optimal tumor sampling for clinical sequencing

Introduction. Genetic heterogeneity of ovarian tumors is caused by the accumulation of novel mutations through time. A result of this process is a spatially heterogeneous tumor with a fractal-like architecture, comprising of a mixture of spatially separated clones, sub-clones, and single cells. Spatial heterogeneity renders clinical sequencing a challenging task because clinically relevant mutations can go undetected. Here, we compare the genetic composition of single tumors with multiple DNA samples extracted from different regions of the tumor in diverse combinations to examine the effects of spatial separation on tumor heterogeneity.

Methods. We derived tumor samples from five patients with ovarian cancer. Three tumor tissue samples were collected from each patient, in each of these, DNA was extracted from a total of eight segments of different sizes (5x3x8 total segments). Exome sequencing was performed and we examined DNA mutations identified from a biopsy sample, from three merged tissue segments with immediate vicinity to the biopsy, and from segments from spatially distant tissue samples merged into one. We compared identified mutations and copy-number variations called from the exome-seq data.

Results. Three patients had heterozygous germline alterations affecting the BRCA pathway, paired with somatic TP53 mutations, one patient had multiple somatic mutations affecting the PI3K pathway, the last patient had the most common activating mutations in the PIK3CA and KRAS genes. When comparing the different sequencing runs, increasing the size of the tumor sample did not affect the overall quantity of mutations identified. Clonal mutations were identifiable in all samples, while sub-clonal mutations shifted. In case of a hypermutating phenotype, increasing the sequenced sample size strongly decreased the number of identified somatic mutations, most probably due to dilution of sub-clones. These observations are confirmed by an in silico model, in which we shuffled regions from publicly available multi-region sequencing data into known compositions.

Conclusion. In hypermutating tumors the number of detected mutations increases as the size of the sequenced sample decreases. Generally, increasing the tumor sample size did not affect the number of identifiable mutations. Our findings suggest that clinical sequencing using small biopsy samples can generate adequate mutation calling results.

Citation Format: Lorinc Pongor, Gyongyi Munkacsy, Ildiko Vereczkey, Imre Pete, Balazs Gyorffy. Multiregion joint somatic genotyping in ovarian cancer proves optimal tumor sampling for clinical sequencing [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 201.

Immunohistochemical Ki67 after short-term hormone therapy identifies low-risk breast cancers as reliably as genomic markers

Background

The purpose of this study was to test whether immunohistochemical (IHC) Ki67 levels after short-term preoperative hormone therapy (post-Ki67) predict similar numbers of patients with favorable prognoses as genomic markers.

Results

Thirty paired cases (60 samples) were enrolled in this study. Post-Ki67 levels were significantly lower than pre-treatment Ki67 levels (P < 0.001). Post-Ki67 predicted more low-risk cases (83.3%, 25/30) than pre-genomic surrogate signature(GSS) (66.7%: 20/30), but the difference in predictive power was not significant (P = 0.233). Proliferation (MKI67, STK15, Survivin, CCNB1, and MYBL2) and estrogen (ER, PGR, BCL2, and SCUBE2) related signatures were significantly downregulated after therapy (P < 0.001 and 0.041, respectively).

Materials and Methods

Core needle biopsy specimens of primary breast cancer were collected at Okayama University Hospital from hormone receptor-positive and human epidermal growth factor 2-negative patients that subsequently received two weeks of neoadjuvant hormone therapy. Paired post-treatment specimens from surgical samples were also collected. IHC Ki67 levels and GSS were compared between pre- and post-hormone treatment samples. Changes of gene expression pattern in short-term hormone therapy were also assessed.

Conclusions

IHC based post-Ki67 levels may have distinct predictive power compared with the naïve IHC Ki67. Future studies with larger cohorts and longer follow-up periods may be needed to validate our results.

Immune gene signatures in triple-negative breast cancers characterized by varying levels of chromosomal instability

1096

Background: Triple-negative breast cancer (TNBC) is generally characterized by high levels of chromosomal instability (CIN) and an intense immune infiltration. However, the link between these two hallmarks and its implications for clinical practice has not been fully elucidated. Methods: We generated eight immune metagenes representing various immune components: natural killer [NK], dendritic cells [DC], T-cells [TC], B-cells [BC], cytotoxic T-cells [CT], interferon [IFN], nuclear factor-κB [NF-kB], and macrophages [M]. Publicly available gene expression data from forty-two data sets, including 862 TNBC samples, were collected. TNBC tumors were clustered in three main subgroups (Basal-Like [BL1/2], Immunomodulatory [IM], and Mesenchymal/Mesenchymal Stem-Like [MS]) using transcriptomic profiling. The CIN70 signature was used to stratify TNBC patients according to the levels of CIN. Statistical analyses were performed using Mann–Whitney U test and Kaplan-Meier analyses. Results: The majority of TNBC samples showed a high level of CIN (83%), and several immune modules were differentially expressed between CIN-high and CIN-low tumors. Specifically, CT, NK, DC, M, and NF-kB signatures were overexpressed in CIN-low TNBC ( p < 1.0E-04). We then evaluated the distribution of genomic instability among TNBC molecular subgroups. Noteworthy, the CIN-high group was composed by a comparable proportion of BL1/2 (39%) and MS (35%) tumors, while CIN-low TNBCs were consistently enriched for MS cancers (61%). Higher expression of the NK, M, and IFN metagenes lead to better survival in CIN-high tumors ( p = 1.9E-02, p = 2.1E-04, and p = 1.1E-03, respectively). Only IFN had the same correlation to survival in CIN-low ( p = 1.4E-02). Conclusions: TNBCs with low levels of CIN may principally enclose M/MSL tumors, which are characterized by an intense immune infiltration and overall good prognosis. Conversely, the TNBC CIN-high group is more heterogeneous in terms of both biological features and levels of immune infiltrates. Therapeutic strategies to promote and boost immune response in the genomically unstable TNBC subgroup warrant further investigation.

Propagation on Molecular Interaction Networks: Prediction of Effective Drug Combinations and Biomarkers in Cancer Treatment

BACKGROUND

Biomedical sciences use a variety of data sources on drug molecules, genes, proteins, diseases and scientific publications etc. This system can be best pictured as a giant data-network linked together by physical, functional, logical and similarity relationships. A new hypothesis or discovery can be considered as a new link that can be deduced from the existing connections. For instance, interactions of two pharmacons - if not already known - represent a testable novel hypothesis. Such implicit effects are especially important in complex diseases such as cancer.

METHODS

The method we applied was to test whether novel drug combinations or novel biomarkers can be predicted from a network of existing oncological databases. We start from the hypothesis that novel, implicit links can be discovered between the network neighborhoods of data items.

RESULTS

We showed that the overlap of network neighborhoods is strongly correlated with the pairwise interaction strength of two pharmacons used in cancer therapy, and it is also well correlated with clinical data. In a second case study we employed this strategy to the discovery of novel biomarkers based on text analysis. In 2012 we prioritized 10 potential biomarkers for ovarian cancers, 2 of which were in fact described as such in the subsequent years.

CONCLUSION

The strategy seems to hold promises for prioritizing new drug combinations or new biomarkers for experimental testing. Its use is naturally limited by the sparsity and the quality of experimental data, however both of these aspects are expected to improve given the development of current databases.

HOXC10 Expression Supports the Development of Chemotherapy Resistance by Fine Tuning DNA Repair in Breast Cancer Cells

Development of drug resistance is a major factor limiting the continued success of cancer chemotherapy. To overcome drug resistance, understanding the underlying mechanism(s) is essential. We found that HOXC10 is overexpressed in primary carcinomas of the breast, and even more significantly in distant metastasis arising after failed chemotherapy. High HOXC10 expression correlates with shorter recurrence-free and overall survival in patients with estrogen receptor-negative breast cancer undergoing chemotherapy. We found that HOXC10 promotes survival in cells treated with doxorubicin, paclitaxel, or carboplatin by suppressing apoptosis and upregulating NF-κB Overexpressed HOXC10 increases S-phase-specific DNA damage repair by homologous recombination (HR) and checkpoint recovery in cells at three important phases. For double-strand break repair, HOXC10 recruits HR proteins at sites of DNA damage. It enhances resection and lastly, it resolves stalled replication forks, leading to initiation of DNA replication following DNA damage. We show that HOXC10 facilitates, but is not directly involved in DNA damage repair mediated by HR. HOXC10 achieves integration of these functions by binding to, and activating cyclin-dependent kinase, CDK7, which regulates transcription by phosphorylating the carboxy-terminal domain of RNA polymerase II. Consistent with these findings, inhibitors of CDK7 reverse HOXC10-mediated drug resistance in cultured cells. Blocking HOXC10 function, therefore, presents a promising new strategy to overcome chemotherapy resistance in breast cancer. Cancer Res; 76(15); 4443-56. ©2016 AACR.

Abstract PD6-06: Somatic mutation patterns differentially affect survival in breast cancer molecular subtypes

Background: The prognostic effects of somatic gene mutations and correlated gene expression in breast cancer is argument of debate. In this study we analyzed the impact of specific mutations on gene expression and their relevance in the prognosis of breast cancer subtypes.

Materials and methods: Exome sequencing and RNA-seq data obtained from TCGA were analyzed. Data was processed using MuTect, MapSplice and RSEM. All together data from 757 patients (ER-/HER2- [n=143], HER2+ including ER positive and negative patients, [n=136], and ER+/HER2- [n=478]) were included. Univariate Receiver Operating Characteristic (ROC) analysis was performed for the top mutated genes (mutated in at least 5% of patients) using the ROC Bioconductor library in R to identify genes whose expression was significantly associated with a mutation. Then, the mean expression of the significant genes was designated as a metagene for each genotype. We assessed the correlation with survival for each metagene by Cox proportional hazards regression and by plotting Kaplan-Meier survival plots. A significance threshold of p&lt;1E-04 was set for each gene to be considered in the survival analysis, and only the top 100 genes were used when there were more than 100 genes significant.

Results: In the overall population only few mutated genes including TP53 (HR=1.66), CDH1 (HR=0.61), AKT1 (HR=0.54), ATM (HR=1.76), NF1 (HR=0.58), KMT2D (HR=2.32), and UBR5 (HR=1.94) were significantly associated with survival. In ER-/HER2- mutant samples the PIK3CA (HR=2.79) and MAP3K1 (HR=2.98), and in HER2+ mutant samples the ARID1A (HR=0.26) and PIK3CA (HR=0.27) metagenes were associated with survival, respectively. Overall, using the combined metagene the majority of the significant mutated genes retained their prognostic power. Mutations of specific genes impacted their own expression and prognosis. The expression of TP53 (AUC=0.609, p=2.60E-06), and MAP3K1 (AUC=0.617, p=6.07E-03) was higher in samples with a mutation while the expression of CDH1 (AUC=0.684, p=2.72E-07), PTEN (AUC=0.687, p=1.47E-04), and BRCA1 (AUC=0.608, p=2.24E-02) was lower.

Conclusions: Our finding support that specific mutated genes may differentially impact prognosis in breast cancer subtypes. Further efforts are required to understand the biological and prognostic role of specific activating and inactivating mutations across molecular breast cancer subtypes.

Citation Format: Gyorffy B, Pongor L, Szabo A, Bottai G, Pusztai L, Santarpia L. Somatic mutation patterns differentially affect survival in breast cancer molecular subtypes. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr PD6-06.

Abstract P5-04-10: Phosphodiesterase type 5 promotes the invasive potential of breast cancer cells through Rho GTPase activation

The impairment of cyclic guanosine monophosphate (cGMP) signaling by overexpression of PDE5 isoform has been recently described in multiple human carcinomas. In addition, accumulating evidences indicate that PDE5 inhibitors could have direct anti-cancer activities as well as they may enhance the sensitivity of certain types of cancer to standard chemotherapeutic drugs. However, despite these studies, neither the expression of PDE5 in breast cancer subtypes nor the underlying regulatory molecular mechanisms by which PDE5 expression may contribute to breast cancer progression have been deeply studied.

We demonstrated that PDE5 was expressed in different subtypes of breast cancer cell lines at higher levels than in non tumorogenic human epithelial breast cell lines. Increased levels were detected in more aggressive endocrine non responsive basal-like breast cancer cells. Interestingly, PDE5 was expressed at very low levels in luminal A-type breast cancer cell lines, which display low ki67 expression, weak invasive behavior and endocrine responsiveness (MCF-7 and T47D cells) compared to luminal B-like cells (such as ZR-75 cells). These results well correlated with data obtained in immunohistochemistry analyses of human breast cancer tissues, showing PDE5 expression in 30 of 35 tumor entities analyzed, with the highest intensity staining in high-grade tumors. Concomitantly, no cytoplasmic PDE5 staining was observed in non neoplastic tissues examined (n=5). In addition, retrospective analyses (n=1959, median follow-up time: 25 years) showed that high PDE5 expression in breast cancer patients was correlated with a statistically significant poorer survival compared to low PDE5-expressing patients. A more relevant discrimination is achieved in lymphnode-negative patients, suggesting a role of PDE5 for identifying early patients at high risk of rapid progression.

In order to better ascertain the role of PDE5 in breast tumorogenesis, we selected a breast tumor cell line that express low levels of this enzyme, MCF-7 and engineered stable clones for overexpression studies. Both vector- and PDE5-stable MCF-7 clones demonstrated comparable proliferation rates; whereas, cell motility and invasion were dramatically increased in PDE5-overexpressing cells. RNA sequencing to compare the transcriptomes of vector- and PDE5-overexpressing MCF-7 cells identified differential expression of genes involved in cell migration and invasion. Particularly, based on pathway analysis we found marked changes in the expression of Rho GTPase family members, proteins involved in cell cytoskeleton organization, migration, and metastasis dissemination (Rho A, cdc42 and Rac signaling, activation score= 1.9, 1.342, and 0.302, respectively). Indeed, Rho and cdc42 pull-down assays revealed increased Rho GTPase activity in cells overexpressing PDE5. Moreover, the selective ROCK inhibitor Y-27632 as well as the PDE5 inhibitor sildenafil were able to significantly reduce both migration and invasion of PDE5 clones.

Our data reveal that PDE5 expression enhances motility and invasiveness of breast cancer cells through the activation of the Rho family of GTPases, and highlight, for the first time, a novel role for PDE5 as a marker of poor outcome in breast cancer patients.

Citation Format: Barone I, Campana A, Giordano C, Tarallo R, Rinaldi A, Bruno G, Gyorffy B, Lanzino M, Bonofiglio D, Catalano S, Ando' S. Phosphodiesterase type 5 promotes the invasive potential of breast cancer cells through Rho GTPase activation. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-04-10.

Abstract P6-08-31: Linking genotype to clinical outcome in breast cancer by combining NGS and gene chip data

Introduction: Next generation sequencing (NGS) provides the possibility to measure mutational status for any part of any gene. However, because of scarce data available to date, linking these mutations to relevant clinical outcome in a large number of patients is not possible.

Aim: Our goal was to combine available genotype data generated by using NGS with gene expression data generated by gene chips to establish a framework to assess the effect of genotype on clinical outcome.

Methods: NGS data generated by the TCGA consortia and publicly available gene chip data obtained from the GEO and EGA repositories were utilized. NGS data was processed using MuTect, SNPeff, GRCh37 and R. RNA-seq data was normalized using DEseq. Gene chip data was MAS5 normalized. Generation of the transcriptomic fingerprint for mutation status was computed by ROC utilizing the RNA-seq data. In the gene chip data, the average expression of significant genes identified was designated as a metagene for the given genotype. Correlation to survival for this metagene was assessed by computing Cox regression and plotting Kaplan-Meier survival plots. Finally, we have set up an online interface to enable running the analysis for any selected gene.

Results: The database contains 332 NGS samples containing mutational status for 22,938 genes and RNA-seq data for 10,987 genes. The gene chip database contains 5,934 patients with 10,987 genes plus detailed clinical characteristics and survival data. We evaluated correlation to outcome for previously identified genes harboring the ten most common somatic mutations in breast cancer. Of these, TP53 (n of mutations out of 332=93, hazard rate=0.51, p&lt;1E-16), AKT1 (n=18, HR=1.6, p=1.6E-15), PIK3CA (n=119, HR=1.5, p=8.5E-12), MAP3K1 (n=20, HR=1.4, p=1.3E-08), CDH1 (n=34, HR=1.3, p=4.4E-07), and RB1 (n=21, HR=1.3, p=7E-06) reached statistical significance while PI3K, PTEN, CDKN1B and GATA3 were not significant or had insufficient number of mutated samples.

Discussion: By connecting genotype to gene expression signature and employing this signature for survival analysis we have set up a pipeline enabling the functional validation of a discovered mutation for any gene in a large breast cancer cohort.

Citation Format: Balazs Gyorffy, Lorinc Pongor, Mate Kormos. Linking genotype to clinical outcome in breast cancer by combining NGS and gene chip data [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-08-31.

Abstract P6-01-22: PDE5 as a novel biomarker and a potential therapeutic target for breast cancer

Background: Phosphodiesterases are enzymes responsible for regulating second messenger signaling by hydrolyzing 3’-5’ cyclic guanosine monophosphate (cGMP), that activates specific pathways resulting in protein phosphorylation, ion fluxes, or cyclic nucleotide hydrolysis to affect gene expression or other aspects of cellular activity. Previous studies have reported increased PDE5 expression in multiple human carcinomas, including bladder, colon, lung and breast cancers, suggesting a role for PDE5 in tumorigenesis. In addition, several in vitro observations have shown antiproliferative and proapoptotic effects of sildenafil and other PDE5 inhibitors in cancer cell lines. However, very little is known about PDE5 expression in human breast tumours and its potential role in breast cancer initiation and progression. We therefore propose to determine whether PDE5 expression may be predictor of outcome in breast cancer patients, and examine PDE5 impact on breast cancer phenotype in vitro.

Methods: We employed MCF-10A normal breast epithelial cells, estrogen receptor (ER) α-positive (MCF-7/ZR-75/T-47D) and ERα-negative (BT-20/MDA-MB-468/SKBR-3/MDA-MB-435) breast cancer cells. We used RT-PCR, immunoblotting and immunofluorescence analyses for evaluating PDE5 expression. To examine PDE5 impact on breast cancer phenotype, MCF-7 cells were engineered to stably express PDE5 and four clones were selected. Cell proliferation was assessed by MTT and anchorage-independent assays, motility and invasion by wound-healing, transmigration and matrigel-based invasion assays. Retrospective analysis using 1959 breast cancer patients of the Metabric Project was performed to evaluate relationship between PDE5 expression and overall survival by Cox proportional hazard regression.

Results: PDE5 mRNA and protein were constitutively expressed at high levels in all the examined tumor cell lines compared to normal breast cells, except for the less motile and non-invasive MCF-7 cells. Interestingly, higher PDE5 expression was found in more aggressive ER-negative cells. Stable overexpression of PDE5 did not affect proliferation of MCF-7 cells, while it significantly increased motility and invasion of all the stable PDE5-transfected clones tested. Patients having high PDE5 expression had a statistically significant poorer survival compared to patients with low PDE5 expression (p=0.014, HR= 1.2). A more relevant discrimination was achieved in lymph node-negative patients (p=0.0015, HR= 1.6), suggesting that assessing PDE5 levels may be helpful to identify a subgroup of early-stage breast cancer patients who are most likely at the highest risk of progression.

Conclusions: PDE5 expression may enhance cancer cell invasive potential, thereby representing prospectively a potential molecular candidate as prognostic marker and target for breast cancer therapy.

Citation Format: Ines Barone, Antonella Campana, Cinzia Giordano, Marilena Lanzino, Daniela Bonofiglio, Balazs Gyorffy, Stefania Catalano, Sebastiano Andò. PDE5 as a novel biomarker and a potential therapeutic target for breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-01-22.

Abstract P4-15-13: CDK8 inhibition potentiates anti-ER and anti-HER2 therapies in breast cancer

CDK8, along with its paralog CDK19, is a cyclin dependent kinase which, in contrast to other members of the CDK family does not regulate cell cycle progression. CDK8 acts as a pleiotropic transcription regulator potentiating the induction of transcription by several transcription factors. Immunohistochemical staining of breast tissue arrays and bioinformatics analysis of gene expression microarray data of breast cancer patients revealed that CDK8 is overexpressed in breast cancer and that higher CDK8 expression correlates with the failure of systemic therapy. Small-molecule selective inhibitors of CDK8 and CDK19 (Senexin A and Senexin B) inhibited the mitogenic effects of estrogen and estrogen-dependent transcription in estrogen receptor (ER)+ breast cancer cell lines. CDK8/19 inhibitors had a cytostatic effect on different ER+ cell lines, and this growth inhibition was synergistic with the effect of the anti-estrogen fulvestrant, particularly in ER+ cell lines resistant to estrogen deprivation. Some of the ER+ cell lines sensitive to CDK8/19 inhibition also express HER2, and therefore we tested CDK8/19 inhibitors in combination with the HER2 and EGFR tyrosine kinase inhibitor lapatinib and an anti-HER2 monoclonal antibody, a biosimilar of trastuzumab. CDK8/19 inhibition produced a synergistic decrease in cell growth with both HER2 inhibitors; this effect was especially pronounced with a trastuzumab biosimilar. Surprisingly, the synergistic effect with HER2 inhibitors was observed in both ER+ HER2+ and ER-HER2+ cell lines, suggesting an effect on a HER2-complementing molecular target other than ER. Interestingly, CDK8/19 inhibition also synergized with trastuzumab biosimilar in a breast cancer cell line that exhibits innate resistance to trastuzumab, suggesting that CDK8/19 inhibition can overcome trastuzumab resistance in breast cancer. These results suggest that combining anti-estrogen and anti-CDK8 therapy may be more effective than conventional hormone therapy for ER positive breast cancer and that combining anti-HER2 and anti-CDK8 therapy is a rational potential treatment for HER2+ breast cancer, regardless of ER status or sensitivity to trastuzumab.

Citation Format: Martina S McDermott, Chang-uk Lim, Mengqian Chen, Alexander Chumanevich, James F Catroppo, Balazs Gyorffy, David Oliver, Igor B Roninson, Eugenia V Broude. CDK8 inhibition potentiates anti-ER and anti-HER2 therapies in breast cancer [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-15-13.

Abstract P4-07-02: A large integrated-gene profiling analysis identifies prognostic microRNAs and correlated DNA repair genes in estrogen receptor positive and negative breast cancers

Background: MicroRNAs (miRNAs) are small non-coding RNAs involved in the pathogenesis of breast cancer (BC). Estrogen receptor (ER) is one of the most important factors influencing BC outcome. In this study we aim to assess the prognostic value in a cross-project analysis for all known miRNAs in ER-positive (ER+) and ER-negative (ER-) BC separately.

Materials and Methods: We assessed the prognostic value of miRNA expression in three independent BC datasets: TCGA (n=782), Metabric [M] (n=1,293) and GSE40267 (n=181) including 472, 853, and 475 miRNAs, respectively. Overall survival (OS) data was available for all patients. Statistical analysis was performed with R software. Survival analysis was performed using Cox proportional hazards regression in each dataset separately. To define an mRNA expression fingerprint for each miRNA, correlation between miRNA and mRNA expressions was computed using Spearman rank correlation in the Metabric dataset. Subsequently, specific metagenes were established using significant correlated mRNAs ("Metabric mRNA", M-mRNA) for each miRNA. Metagenes were also used in the Cox regression analysis. The entire analysis was performed for each miRNA in ER-positive and ER-negative BCs separately.

Results: In the ER+ BC, 283 miRNAs reached statistical significance in at least one dataset. The best performing and significant miRNAs in all datasets were miR-195 (M: HR=1.73, p=3.5E-06; TCGA: HR=2.17, p=3.5E-06, M-mRNA: HR=0.54, p=2.0E-10), miR-199b (M: HR=1.74, p=3.8E-06; TCGA: HR=0.46, p=0.028, M-mRNA: HR=0.56, p=2.8E-09), and miR-210 (M: HR=0.7, p=0.003; TCGA: HR=0.45, p=0.014, Metabric mRNA: HR=1.34, p=0.002). These 3 miRNAs retained their prognostic significance also in a multivariate analysis paired with HER2 status, lymph node status, size, grade and MKI67 expression. In the ER- group 73 significant miRNAs were identified. The best performing miRNAs in at least two datasets were miR-155 (GSE40267: HR=2.21, p=0.034, M: HR=1.67, p=0.008) and miR-381 (GSE40267: HR=0.41, p=0.026; M: HR=0.67, p=0.038). However, the miRNAs in ER- BC in a multivariate analysis were not independent from tumor size. The metagenes identified were significantly enriched in DNA repair genes (e.g. AURKB, TEK) and exhibited an inverse correlation to survival as compared to miRNA expression.

Conclusions: This is the first large-scale miRNA expression analysis demonstrating different prognostic miRNAs in ER+ and ER- BCs. Correlated miRNA-metagenes are involved in the regulation of DNA repair and genomic stability.

Citation Format: Balazs Gyorffy, Mate Kormos, Giulia Bottai, Kristine Kleivi, George A Calin, Anne-Lise Borresen-Dale, Libero Santarpia. A large integrated-gene profiling analysis identifies prognostic microRNAs and correlated DNA repair genes in estrogen receptor positive and negative breast cancers [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-07-02.

Abstract PR06: Impact of RAL signal transduction on genetic program and growth control in KRAS- and BRAF-mutated colorectal cells and prognostic potential of pathway-responsive genes in cancer patients

The EGFR/RAS signaling system plays an imperative role in tumor pathogenesis and therapy response. Downstream effects of the EGFR/RAS signaling axis are mediated by three major pathway branches: the RAF/MAPK, PI3K and RALA pathways. While the RAF/MAPK and PI3K pathways have been extensively characterized in mechanistic, preclinical and clinical studies, the RAL pathway is less well understood.

Here we compared RALGTPase activities in three colorectal cancer cell lines by active GTPase pull-down assay and analyzed the effects of transient silencing of RALA expression by siRNA. RALA activity was highest in SW480 [KRAS codon 12 mutation] and HCT116 cells [KRAS codon 13 mutation] and also detectable in HT29 colorectal cancer cells [KRAS wild-type, BRAF V600E mutation]. Silencing of RALA expression strongly diminished the active GTP-bound form of the protein in all cell lines. The proliferation of the two KRAS mutated cell lines was significantly reduced. BRAF mutated cells showed a slight increase in cell death only. Next we interrogated microarrays using RNA prepared from the cell lines treated with siRNA and controls. We identified 550 common genes the expression of which was up-regulated or down-regulated after RALA knock-down (“RALA pathway-responsive genes”). None of them were affected when the RAF/MAPK or PI3K pathways were blocked.

To investigate the potential clinical relevance of transcriptional targets regulated by the RALA pathway, we performed a meta-analysis utilizing expression profiles of 1,424 colorectal cancers documented in 8 independent publicly available data sets comprising information on patient survival. Of 19 RALA pathway-responsive genes correlated with progression-free survival, IQGAP1 (IQ-motif containing GTPase-activating protein 1), TOP1 (topoisomerase 1), LGALS1 (lectin galactoside-binding soluble 1), FILIP1L (filamin A-interacting protein 1-like) and TCF4 (transcription factor 4) were identified as the most important RALA pathway targets capable of predicting survival.

In conclusion, the RALA pathway impinges on the transcription of a distinct subset of target genes in colorectal cancer cells independent of the KRAS and BRAF mutational status. RALA pathway-responsive genes were unaffected by RAF/MAPK and PI3K signaling. These findings support the concept of a pathway-sensitive modular organization of the transcriptome. In view of the correlation of RAL pathway-responsive genes and patient survival, further investigations of the diagnostic impact in prospective trials and exploitation of therapeutic approaches are warranted.

This abstract is also presented as Poster B32.

Citation Format: Balazs Gyorffy, Iwona Stelniec-Klotz, Christian Sigler, Attila Szijarto, Yu Qian, Reinhold Schäfer. Impact of RAL signal transduction on genetic program and growth control in KRAS- and BRAF-mutated colorectal cells and prognostic potential of pathway-responsive genes in cancer patients. [abstract]. In: Proceedings of the AACR Special Conference on RAS Oncogenes: From Biology to Therapy; Feb 24-27, 2014; Lake Buena Vista, FL. Philadelphia (PA): AACR; Mol Cancer Res 2014;12(12 Suppl):Abstract nr PR06. doi: 10.1158/1557-3125.RASONC14-PR06

Biomedical hypothesis generation by text mining and gene prioritization

Text mining methods can facilitate the generation of biomedical hypotheses by suggesting novel associations between diseases and genes. Previously, we developed a rare-term model called RaJoLink (Petric et al, J. Biomed. Inform. 42(2): 219-227, 2009) in which hypotheses are formulated on the basis of terms rarely associated with a target domain. Since many current medical hypotheses are formulated in terms of molecular entities and molecular mechanisms, here we extend the methodology to proteins and genes, using a standardized vocabulary as well as a gene/protein network model. The proposed enhanced RaJoLink rare-term model combines text mining and gene prioritization approaches. Its utility is illustrated by finding known as well as potential gene-disease associations in ovarian cancer using MEDLINE abstracts and the STRING database.

Correlations of differentially expressed gap junction connexins Cx26, Cx30, Cx32, Cx43 and Cx46 with breast cancer progression and prognosis

BACKGROUND AND AIMS

Connexins and their cell membrane channels contribute to the control of cell proliferation and compartmental functions in breast glands and their deregulation is linked to breast carcinogenesis. Our aim was to correlate connexin expression with tumor progression and prognosis in primary breast cancers.

MATERIALS AND METHODS

Meta-analysis of connexin isotype expression data of 1809 and 1899 breast cancers from the Affymetrix and Illumina array platforms, respectively, was performed. Expressed connexins were also monitored at the protein level in tissue microarrays of 127 patients equally representing all tumor grades, using immunofluorescence and multilayer, multichannel digital microscopy. Prognostic correlations were plotted in Kaplan-Meier curves and tested using the log-rank test and cox-regression analysis in univariate and multivariate models.

RESULTS

The expression of GJA1/Cx43, GJA3/Cx46 and GJB2/Cx26 and, for the first time, GJA6/Cx30 and GJB1/Cx32 was revealed both in normal human mammary glands and breast carcinomas. Within their subfamilies these connexins can form homo- and heterocellular epithelial channels. In cancer, the array datasets cross-validated each other's prognostic results. In line with the significant correlations found at mRNA level, elevated Cx43 protein levels were linked with significantly improved breast cancer outcome, offering Cx43 protein detection as an independent prognostic marker stronger than vascular invasion or necrosis. As a contrary, elevated Cx30 mRNA and protein levels were associated with a reduced disease outcome offering Cx30 protein detection as an independent prognostic marker outperforming mitotic index and necrosis. Elevated versus low Cx43 protein levels allowed the stratification of grade 2 tumors into good and poor relapse free survival subgroups, respectively. Also, elevated versus low Cx30 levels stratified grade 3 patients into poor and good overall survival subgroups, respectively.

CONCLUSION

Differential expression of Cx43 and Cx30 may serve as potential positive and negative prognostic markers, respectively, for a clinically relevant stratification of breast cancers.

Abstract 2101: Role of CDK8 in estrogen receptor signaling in breast cancers

The majority of breast cancers (BrCa) overexpress estrogen receptor (ER)α, which regulates transcription and drives estrogen-stimulated proliferation of ER+ tumor cells. ER+ patients usually receive adjuvant anti-estrogen therapy based on ER modification, downregulation, or estrogen depletion. Tumors frequently develop resistance to anti-estrogen therapy through various mechanisms, which may involve stimulation of ER itself or of downstream mediators of ER-driven transcription, as well as activation of alternative proliferation pathways, in particular those driven by HER2/EGFR. The treatment efficacy of hormone-resistant BrCa could be greatly augmented by targeting new druggable mediators of ER activity. We have now identified a transcription-regulating oncogenic kinase CDK8 as a potentiator of ER signaling and ER-driven BrCa cell proliferation. Immunohistochemical staining of breast tissue arrays and bioinformatics analysis of gene expression microarray data of breast cancer patients revealed that CDK8 is overexpressed in BrCa and that higher CDK8 expression correlates with the failure of systemic therapy. Among systemically untreated patients, higher CDK8 expression was correlated with shorter relapse-free survival in a subset of ER+ tumors that expressed the lowest levels of ERα. This correlation suggested that CDK8 could play a role in the progression of ER+ breast cancers with reduced levels of ER, possibly by stimulating the mitogenic effects of ER-mediated transcription. Indeed, a selective small-molecule CDK8 inhibitor (Senexin A) decreased estrogen-induced ER-dependent transcription and inhibited estrogen-stimulated proliferation of ER+ BrCa cell lines. Microarray analysis revealed that CDK8 inhibition diminished the induction of genes that show rapid and sustained activation by estrogen in ER+ cells. CDK8 inhibition had an additive effect in combination with anti-estrogens in ER+ BrCa and a synergistic effect with fulvestrant in BrCa cells resistant to estrogen deprivation. Some of these cell lines also express HER2/EGFR, and CDK8 inhibition in combination with a HER2/EGFR inhibitor lapatinib synergistically inhibited the growth of these cells. These results suggest that combining anti-estrogen and anti-CDK8 therapy may be more effective than conventional hormone therapy for ER+ BrCa.

Citation Format: Martina McDermott, Balazs Gyorffy, Chang-uk Lim, Alexander Chumanevich, Zhengguan Yang, Mengqian Chen, James F. Catroppo, Igor Roninson, Eugenia V. Broude. Role of CDK8 in estrogen receptor signaling in breast cancers. [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 2101. doi:10.1158/1538-7445.AM2014-2101