Abstract 3424: Consensus molecular subtype of triple negative breast cancer to implicate in chemotherapy response

Abstract

Objective: The heterogeneity of triple negative breast cancer (TNBC) confers the difficulties in chemotherapy and induces poor outcome. Subtype classification of TNBC using gene expression profile could achieve to identify molecular markers to suggest therapeutic guidance. Methods: In this study, we collected gene expression profiles of 957 TNBC patients from GEO and integrated into one meta-data using meta analysis. We identified subtypes using nonnegative matrix factorization (NMF) and explored for comprehensive characteristics of consensus molecular subtype (CMS) by investigating key pathway activity, tumor microenvironment, stemness and so on. In addition, we computed the drug response score (DRscore) from MSigDB chemical perturbation signature gene set and examined drug associations with key pathways of CMS. Drug candidates were validated from independent two data sets; our patients’ expression profile (n = 38) and biobank TNBC organoids (n=64). Results: We classified four different TNBC subtypes displaying gene expression patterns including mesenchymal-like (MSL), luminal-AR (LAR), immunomodulatory (IM), and stem-like (SL). MSL were activated pathways with epithelial-to-mesenchymal (EMT) and TGF-beta signaling whereas SL was up-regulation of cell cycle and WNT signaling pathway. The LAR subtype was activated of androgen and estrogen receptor pathway. Although metastatic TNBC generally shared equal activity of key pathways with primary, coagulation, toll-like receptors, TNF, and Jak-STAT signaling pathways were dysregulated in metastasis. Especially, SPP1 gene expression to induce metastasis was associated with poor prognosis. DRscores were discriminative in CMS of meta-data and biobank expression profile. We found subtype-specific 18 drugs as therapeutic candidates and these drugs were also correlated with key pathways. In a case of cisplatin, DRscores appear resistant in MSL while response in SL. Our patients’ expression profiles were shown to consistent result as well as biobank. Conclusion: These finding might facilitate understanding heterogeneity of TNBC. Our novel approach to explore drug response suggested functional intervention landscape of drug response without in-vivo experiment. Taken together, we propose biomarkers for diagnosis and also suggest therapeutic strategy depending on CMS. Citation Format: Jihyun Kim, Doyeong Yu, Jiyoon Noh, Wooyeong Jang, Hanna Yang, Youngmee Kwon, Keun Seok Lee, Sung Hoon Sim, Sun-Young Kong, In Hae Park, Charny Park. Consensus molecular subtype of triple negative breast cancer to implicate in chemotherapy response [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 3424.