Abstract PO2-16-01: Molecular correlates of drug response to guide therapy in triple-negative breast cancer

Abstract TNBC accounts for 10-20% of total breast cancer cases in the US and is more aggressive, higher grade, and has a poorer prognosis than other forms of breast cancer. TNBC is more likely to metastasize within 3-5 years than other forms of breast cancer and has a median survival of 17.6-21.3 months after metastasis. Advances in immunotherapy are promising, but recent trial results of immunotherapy-chemo combination have resulted in a 3-year overall survival rate under 36% in a PD-L1 based selected cohort. Thus, there is an unmet need for innovations that lead to new treatment options and improve outcomes for patients with TNBC in a personalized manner beyond only one or a few markers. In our recently published study (PMID: 31655920), we tested the dose response for a panel of 78 investigational drugs and plotted this readout against molecular characterization to identify statistically significant DNA, RNA, and protein predictors of drug efficacy in TNBC. We have expanded on this published study by identifying correlates from 3D cell culture. According to the literature, 3D culture systems are more representative of in vivo activity. We have increased our sample number (now 27 cell lines and 8 ex vivo PDXs) and expanded our drug library to 387 compounds, selected for their diverse mechanisms and promise in other cancers. Following screening, we prioritized 35 compounds based on their elevated activity and variance of response. From these 35 prioritized compounds, we identified molecular correlates of drug response for 27 compounds (77% of prioritized compounds yielding molecular correlates compared to 50% in our previously published study). Focusing on a known active drug in TNBC that is approved for patients with metastatic TNBC, we identified multi-omic (protein expression/activity and RNA expression) molecular correlates of response to a chemotherapeutic- SN-38, the payload of Sacituzumab and the active metabolite of irinotecan. We utilized a sub-panel of RNA correlates to characterize treatment-naive PDXs and predict response to SN-38. We then tested SN-38 response ex vivo for 5 PDXs to show that we could accurately identify sensitive/resistant models to SN-38 therapy. From significant protein correlates, we identified that the autophagy and AKT pathways were elevated in models with poor response to SN-38. Co-targeting the autophagy or AKT pathway with SN-38 was an effective strategy to identify rational, synergistic therapeutic combinations at therapeutically relevant levels (effective dose (ED) 50 and ED75). We evaluated irinotecan in vivo using a cell line with sensitivity in our 3D assay, demonstrating strong activity in vivo to validate that our screening assay accurately predicts in vivo response. We have further applied this methodology to targeted inhibitors with potential activity in TNBC. We have identified an ATR inhibitor, berzosertib, that targets the DNA damage response (DDR) pathway and correlates with protein expression and activity of Aurora kinases (AKs). Furthermore, combining the pan-AK inhibitor danusertib with berzosertib results in synergistic killing of TNBC cells. As a future direction, we are evaluating this combination in vivo utilizing an ethnically diverse patient-derived xenograft library grown in mice to evaluate this novel combination in a mock clinical trial and to identify molecular correlates of response and resistance that could be used to guide future human trials. Citation Format: Nathan Merrill, Nathalie Vandecan, Athena Apfel, Habib Serhan, Peter Ulintz, Liwei Bao, Xu Cheng, Aki Morikawa, Sofia Merajver, Matthew Soellner. Molecular correlates of drug response to guide therapy in triple-negative breast cancer [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-16-01.