Abstract B058: A Genetically Unbiased In Vivo Platform to Define Therapy-Driven Determinants of Triple-Negative Breast Cancer Brain Metastasis

Abstract Brain metastases from triple-negative breast cancer (TNBC) remain a major cause of mortality and frequently emerge following systemic therapy. Accumulating clinical and preclinical evidence indicates that chemotherapy can increase circulating tumor cells and remodel distant niches, potentially reshaping metastatic trajectories; however, the mechanisms governing therapy-associated relapse and region-specific colonization within the central nervous system (CNS) remain poorly defined. Progress has been limited by a lack of immunocompetent models that capture spontaneous dissemination, clonal evolution, and treatment-imposed selective pressures. We hypothesize that chemotherapy selects distinct tumor subclones that undergo CNS region-specific adaptation to drive metastatic outgrowth. To address this challenge, we established a genetically unbiased in vivo discovery platform based on Sleeping Beauty (SB) transposon mutagenesis in TNBC mammary tumors. SB mutagenesis was combined with a Pik3caH1047R-driven background to generate primary TNBC tumors in immunocompetent mice. Hybrid-capture sequencing of more than 100 SB-accelerated tumors identified recurrent insertion sites in known and novel cancer-associated genes, demonstrating robust clonal selection and subtype-associated transcriptional programs. From these tumors, we generated SB-accelerated primary tumor grafts that are orthotopically transplantable, surgically resectable, and capable of spontaneous metastatic dissemination to distant organs, establishing feasibility for longitudinal modeling of therapeutic pressure. Building on this validated foundation, the platform is designed to model clinically relevant neoadjuvant chemotherapy in TNBC (carboplatin and paclitaxel) and to compare metastatic outcomes arising under treated versus untreated conditions. Disseminated lesions arising within the CNS will be evaluated using integrated SB insertion profiling, single-cell and spatial transcriptomics, and immunophenotypic analyses to define therapy-associated clonal selection and region-specific adaptation programs while preserving tumor-immune interactions. If dissemination extends beyond parenchymal brain lesions, including potential leptomeningeal involvement, these lesions will be characterized using the same analytic framework. Cross-species integration with human TNBC brain metastasis datasets using a machine learning-based multi-omics approach will enable prioritization of conserved drivers and spatially informed biomarkers. In summary, we present a validated, immune-competent in vivo framework that enables interrogation of how chemotherapeutic pressure shapes the genetic and spatial determinants of TNBC brain metastasis. By coupling established feasibility with prospective CNS-focused discovery, this platform provides a foundation for identifying actionable vulnerabilities to intercept metastatic relapse in high-risk TNBC patients. Generative artificial intelligence was used to assist with editing and clarity of this abstract. Citation Format: Zach Seeman, Tao Cheng, Mohammed Qaraad, Christopher Stehn, David Guinovart, Nuri Alpay. Temiz, David Largaespada, Eric Rahrmann. A Genetically Unbiased In Vivo Platform to Define Therapy-Driven Determinants of Triple-Negative Breast Cancer Brain Metastasis abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Brain Cancer; 2026 Mar 23-25; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86 (6Suppl): Abstract nr B058.