Abstract 4106: High throughput chemical profiling of chemoresistant triple negative breast cancer

Abstract

Abstract Triple-negative breast cancer (TNBC) makes up approximately 10-20% of all breast cancers, is characterized by a lack of targeted therapies, and in approximately 40-50% of cases a significant residual cancer burden remains after neoadjuvant chemotherapy, thus presenting a need to identify actionable targets in both treatment naive and resistant settings. To address the unmet treatment needs of chemo-resistant patients, our multi-institutional collaborative of clinical oncologists and basic researchers are assembling a panel of patient-derived xenograft (PDX) models of TNBC as part of an ongoing IRB-approved clinical trial at MD Anderson Cancer Center (ARTEMIS; NCT02276443). PDX models are generated from serial fine needle aspirates of primary TNBC from patients (1) prior to treatment, and in non-responding tumors (2) after four cycles of Adriamycin plus cyclophosphamide (AC), and (3) after a 3-month course of targeted therapy. We first developed a standardized intake procedure that simultaneously optimized cell seeding densities in 384-well plates, tested multiple cell culturing medias, determined baseline phenotypic characteristics, and established robust assay controls. After optimization, we performed high-throughput drug screens with a panel of 634 mechanistically annotated chemical probes that are FDA approved, late stage investigational or mechanistic probes. A bioluminescent reporter quantifying the amount of ATP (CellTiter Glo, Promega) was used as a surrogate readout of cell viability, which is well suited for the majority of PDX cultures that grew as non-adherent multi-cellular aggregates as well as being integrable into larger datasets run in a similar format. Screening assays are performed with 3-4 technical replicates over a three log concentration range (10.0, 1.0, 0.1 uM). The effectiveness of each drug is summarized as an area under the fitted concentration-response curve, simultaneously capturing alterations in potency and efficacy. In total, we have tested 21 PDX models isolated from 14 unique patients using this strategy. From these data, we show that PDX models generated from individual patients have unique pharmacologic profiles with little correspondence to molecularly defined sub-classifications, which may suggest the need for the development of an actionable sub-classification strategy to better stratify patients. Despite the significant heterogeneity across pharmacologic profiles, we identified a consolidated set of drugs targeting specific epigenetic processes and protein homeostasis that have increased activity in a subset of chemo-resistant samples. We are currently collecting orthogonal data including next-generation sequencing and transcriptomics to further elucidate the mechanism of these targets and identify potential biomarkers. Citation Format: Reid Trenton Powell, Abena Redwood, Lei Guo, Shirong Cai, Yizheng Tu, Yan Jiang, Xuan Liu, Xinhui Zhou, Clifford Stephan, Peter Davies, Jeffrey T. Chang, Stacey Moulder, William F. Symmans, Helen Piwnica-Worms. High throughput chemical profiling of chemoresistant triple negative breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4106.