Abstract 3142: Patient-derived models of breast cancer for drug screening and precision medicine approaches

Abstract Introduction: The breast tumor microenvironment (TME) consists of a complex and heterogeneous network of cancer cells, cancer-associated fibroblasts (CAFs), immune cells, and an acellular extracellular matrix (ECM). These components critically impact treatment responses, emphasizing the utilization of patient-derived models that closely mimic the TME. Such models provide valuable insights into chemotherapy responses and facilitate the development of personalized treatment strategies. Natural products represent a vast and diverse source for anti-tumor drug discovery due to their structural diversity and wide range of biological activities. This study aims to utilize patient-derived models to screen a library of natural products to identify lead compounds with potential for personalized breast cancer therapy. Methods: A primary screen was performed using a patient-derived co-culture system comprising breast cancer cells and matched CAFs. A library of 61 natural products and 6 standard chemotherapy drugs was evaluated for their impact on cell viability, with IC50 values calculated to assess efficacy. Positive hits were further investigated in a secondary screen using 3D patient-derived organoid (PDO) models derived from four breast cancer subtypes. These PDOs included patient-derived cancer cells, matched CAFs, and autologous tumor-infiltrating lymphocytes (TILs), representing the tumor immune microenvironment. The anti-tumor effects of these compounds on PDO viability, growth, TIL recruitment, TIL activity and induction of apoptosis were assessed. Results: The patient-derived co-culture and PDO models retained characteristics of their respective primary breast cancers and showed varied responses to different chemotherapeutics. Among standard drugs, doxorubicin and camptothecin showed uniform IC50 values across all PDOs, while paclitaxel primarily inhibited proliferation rather than inducing significant cell death. Conversely, 5-fluorouracil and etoposide displayed limited efficacy. In the natural product screen, Emodin (an anthraquinone) and Platycodin D (PD, a triterpenoid steroid) demonstrated potent anti-tumor activity. Both compounds caused a dose-dependent reduction in PDO size and a significant increase in cell death. Apoptosis assays revealed enhanced tumor cell and CAF apoptosis following treatment. Additionally, Emodin and PD increased TIL cytotoxicity, evidenced by elevated Granzyme B expression in CD4+ and CD8+ T cells. Conclusion: This study highlights the utility of patient-derived models in precision medicine and drug screening, providing a valuable platform for identifying effective therapeutic agents for breast cancer. Emodin and PD emerge as promising candidates with broad efficacy across multiple breast cancer subtypes, highlighting their potential for future development in breast cancer therapies. Citation Format: Rui Li,Xiao Hu,Ning Ge,Michael Kerin,Laura Barkley. Patient-derived models of breast cancer for drug screening and precision medicine approaches [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 3142.