Abstract SY43-03: Screening for triple negative breast cancer vulnerabilities

Abstract

Abstract There is no targeted therapy for triple negative breast cancers (TNBC), which have the worst prognosis of human breast cancers. TNBCs are prone to relapse and metastasize after cytotoxic drug treatment. This group of cancers, defined by low or absent expression of estrogen, progesterone and Her2 receptors, are heterogeneous in genetic, epigenetic and phenotypic features, making it difficult to identify specific drug targets suitable for this group of cancers as a whole. About half or more of TNBCs have an epithelial phenotype (classified based on gene expression profiling as basal-like or basal-A) and a large minority of the remaining tumors are mesenchymal (classified as basal-B). To identify selective genetic dependencies of the basal-like subtype of TNBC that are not required for survival of normal breast epithelial cells, which might be good targets for targeted drug therapy, we first performed a genome-wide siRNA lethality screen that compared two human breast epithelial TNBC cell lines transformed with the same genes in different media - basal-like BPLER and myoepithelial HMLER. BPLER are highly malignant and enriched for tumor-initiating cells, forming tumors in nude mice after injection of just 50 cells, while the more differentiated HMLER cell line only forms tumors when more than 50,000 cells are injected. The screen identified 154 genes on which BPLER, but not HMLER, depended. Expression of the 154 BPLER dependency genes correlated with poor prognosis in breast, but not lung or colon, cancer. Of note, no oncogenic kinases, which are often chosen as cancer drug targets, were identified in the screen. Proteasome genes were overrepresented hits. Basal-like TNBC lines as a group were selectively sensitive to proteasome inhibitor drugs relative to normal epithelial, luminal breast cancer and mesenchymal TNBC lines. Proteasome inhibition reduced growth of established basal-like TNBC tumors in mice and blocked tumor-initiation and metastasis. Tumor-initiating cells (T-ICs) within basal-like cancers did not survive proteasome inhibition. Developing ways to target T-ICs is a priority of cancer drug development. Proteasome addiction in basal-like TNBCs was mediated by NOXA and linked to MCL-1 dependence, which was also a common dependency gene of basal-like TNBCs. Because of the heterogeneity of TNBCs, we next assessed how many of the BPLER dependencies were shared with other breast cancers. A targeted screen was performed comparing the viability of 10 breast cancer cell lines (3 luminal A, 5 basal-A and 2 basal-B TNBCs) and telomerase-immortalized normal breast epithelial cells after knockdown of the 154 BPLER dependency genes. This screen confirmed the selective dependence of basal-A TNBCs on the proteasome and MCL-1. This screen identified 17 additional hits that were shared dependencies in at least 4 of 5 basal-like TNBC cell lines. A novel hit, required for survival of 4 of 5 basal-A TNBC cell lines, but not normal breast epithelial cells, was the spindle checkpoint regulator NDC80. This work was supported by Department of Defense grants to FP and JL and by a grant from the Harvard Stem Cell Institute. The authors have no financial conflicts of interest to disclose. Citation Format: Fabio Petrocca, Gabriel Altschuler, Shen Mynn Tan, Marc L. Mendillo, Haoheng Yan, D. Joseph Jerry, Andrew L. Kung, Winston Hide, Tan A. Ince, Judy Lieberman. Screening for triple negative breast cancer vulnerabilities. [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 SY43-03. doi:10.1158/1538-7445.AM2014-SY43-03