Abstract DL-1: Leveraging Preclinical Models of Breast Cancer

Abstract Jeffrey M. Rosen1, Kevin Roarty1, Se Jin Kim1, Charles M. Perou2, Sendurai Mani3, Paul Ik Sun Kim4 and Xiang Zhang1,4 1Department of Molecular & Cellular Biology and 1,4Lester and Sue Smith Breast Center, Baylor College of Medicine, Houston, TX, 2University of North Carolina, Chapel Hill, NC and 3MD Anderson Cancer Center, Houston, TX Breast cancer is a disease that displays both inter- and intra-tumoral heterogeneity. More than a decade ago we asked if a subpopulation of breast cancer stem cells might be resistant to chemotherapy and responsible for relapse in breast cancer patients given neoadjuvant therapy. These clinical studies provided evidence for population of chemotherapy-resistant breast cancer-stem cells. The gene expression signature derived from these comparisons overlapped with an identified “claudin-low” molecular subtype characterized by the low to absent expression of luminal differentiation markers, high enrichment for many EMT-associated genes, and immune-response genes. The claudin-low subtype also most closely resembled a subset mammary epithelial stem cells. The “claudin-low” signature was enriched in residual tumors remaining after either endocrine therapy or chemotherapy treatment. Double positive cells expressing mesenchymal and epithelial markers were enriched in resistant tumors. This suggests the presence of an intermediate or “partial EMT”, which has been proposed to be a hybrid E/M phenotype between the epithelial to mesenchymal transition. Recently we have developed a set of EMT/MET sensors to identify cells undergoing this transition, and we have used these to screen an FDA approved compound library. Genetically engineered mouse (GEM) syngeneic p53-null mammary tumor models that closely mimic several of the subtypes in human breast cancer are being used as preclinical models to study the response to both novel targeted therapies and chemotherapy. We have developed, extensively characterized and “credentialed” a bank of these tumors, which represent the different subtypes of human breast cancer including a subset of these tumors with a similar gene expression signature as the human claudin-low tumors. These murine claudin-low tumors showed high expression of EMT inducers, low expression of members of the miR-200 family and resistance to most standard-of-care therapies. Re-expression of miR-200 family members reversed EMT, decreased the CSC population and sensitized cells to chemotherapy. The ZEB1/miR-200 axis has additional, non-cell autonomous roles in cancer pathogenesis, and has recently been shown to affect immune recognition of cancer cells whereby ZEB1 suppression of the miR-200 family leads to upregulation of PD-L1, a direct miR-200 family target. Therefore, the EMT transition also may play a role in immunosuppression exacerbating treatment response. Accordingly, immunoprofiling has identified unique subsets of immunosuppressive neutrophils and macrophages in mice with the different subtypes of p53 null tumors. Hierarchical clustering revealed "extrinsic" subtypes of breast cancers in terms of their innate immune cell profiles. Finally, we have employed specific Wnt and Stat3 pathway reporters to better understand the signaling pathways involved in intratumoral heterogeneity and metastasis. Citation Format: Rosen JM. Leveraging Preclinical Models of Breast Cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr DL-1.