Abstract 5084: Functional genomic classification of breast cancer using pooled lentivirus shRNA screens

Abstract

Abstract Targeted therapies for most breast cancers are lacking because our knowledge of essential genes driving tumor proliferation is still primitive for each subtype. Although ongoing intensive efforts to genetically characterize large numbers of breast tumors are providing a plethora of new data on genomic abnormalities (e.g., amplifications, deletions, somatic mutations), it can be difficult to access which of these actively drive pathogenesis. Even if an oncogene or tumor suppressor is identified, these often are not amenable to targeted therapy (eg. KRAS, c-MYC, p53, etc). However, unanticipated gene/pathway dependencies can arise as a consequence of these genetic abnormalities in cancer cells (“synthetic lethality”). The recent development of lentiviral-based shRNA libraries enables genome-wide screening of cultured cancer cells in a pooled format, facilitating the identification of genes necessary for cancer cell proliferation and survival in cell culture as well as potential synthetic lethal interactions. The overall objective of this project was to identify subtype-specific targets for human breast cancer using a genome-wide shRNA screen, as well as to compare “functional genomic” and genomic classification schemes. We screened a panel of 53 breast cancer lines using an 80,000 lentiviral shRNA library targeting 16,000 genes in a pooled format. We identified several classes of gene “dropouts,” including general essential genes, which are required for survival or growth in more than 70% of all cell lines, irrespective of subtype. Most of these were enriched in housekeeping functions, although several genes involved in cellular signal transduction also were identified, including mTORC1, RAPTOR, TGFBR2, DDR1 and DDR2. In addition, we observed several “subtype-specific” genes, whose essentiality is restricted to a defined subtype. Furthermore, unsupervised and supervised clustering of our functional screening results identified potential “drivers” unique to each subtype. These include amplified/overexpressed genes such as ERBB2, ERBB3, FOXA1, SPDEF, TFAP2C, and CCDN1 known to be specific to the luminal and HER2 subtypes, respectively. Finally, integration of gene expression, copy number variation, and functional screening results identified potential synthetic lethal interactions with common genetic changes. Our study represents an extensive functional genetic survey of three major breast cancer subtypes, reveals complexities between genomic and functional genomic results, and uncovers several unexpected gene dependencies and potential novel therapeutic target for each subtype. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 5084. doi:1538-7445.AM2012-5084