Abstract 49: Genetic, genomic, and pharmacologic characterization of a new cohort of ancestrally diverse patient-derived xenograft models of cancer

Abstract Introduction: Patient derived xenografts (PDX) are important preclinical models of cancer as they faithfully recapitulate many aspects of human tumors. Currently, models derived from patients of non-European genetic ancestry are underrepresented, limiting the potential of precision medicine advancements to equitably improve patient outcomes. The United for Health Equity Living PDX Program (U4HELPP) seeks to address this need by generating new PDX models representative of the diverse population served by the VCU Health System. This study aims to 1) characterize the histologic, genomic, and transcriptomic features of a new cohort of ancestrally diverse PDX models of breast, pancreatic, colon, and lung cancer, and 2) identify drug sensitivities and response markers that will be targeted in ongoing studies utilizing these models. Methods: Tumor fragments were implanted in the bilateral mammary glands or subcutaneous flank of NSG mice, and growth rates were defined. Tissue was fixed and embedded for histological staining to identify the cancer type and subtype. Whole genome and bulk RNA sequencing were performed to assess the mutational profile, genetic ancestry, and global transcriptomic landscape. Single cell RNA sequencing was utilized to capture the transcriptomic heterogeneity within each model. Finally, high throughput drug screening with the NCI NExT Oncology Interrogation Tools Set of 555 agents was performed in vitro to assess the drug sensitivities of each PDX. Results: Whole genome sequencing and bulk RNA sequencing positively identified distinct genetic ancestries within the cohort. Histopathologic analysis revealed a rare metaplastic breast carcinoma, and immunohistochemical staining stratified breast cancer models into estrogen receptor positive and triple negative subtypes. Bulk RNA sequencing was used to subtype pancreatic ductal adenocarcinoma models as basal-like, classical, or intermediate, and single-cell RNA sequencing confirmed the presence of mixed subtype populations within these samples. High throughput drug screening identified sensitivities that correlated with mutations detected through whole genome sequencing; overall, several classes of drugs including proteasome and cyclin-dependent kinase inhibitors were found to be highly cytotoxic in all cancer types tested. Conclusion: This study integrates multiple modalities to characterize the histological features, molecular signatures, and cellular heterogeneity of this new cohort of PDXs, which will provide researchers with extensive data on which to base future studies utilizing these models. Furthermore, it identifies promising therapeutic agents and explores genomic markers of drug sensitivity which could be used to predict patient response to novel treatments. Citation Format: Nina M. Dashti-Gibson, Amy L. Olex, Katarzyna M. Tyc, Carson J. Walker, Oniya M. Smith, Rachel K. Myrick, Bin Hu, Praveen Bhoopathi, Madhavi Puchalapalli, Li Wang, Yue Zhang, Brydon P. Wall, Valentina Robila, Mikhail G. Dozmorov, Jinze Liu, Jennifer E. Koblinski, Jose G. Trevino, J Chuck Harrell. Genetic, genomic, and pharmacologic characterization of a new cohort of ancestrally diverse patient-derived xenograft models of cancer [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 49.