bioRxiv2025
Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest malignancies with a 5-year survival rate of less than 10%. Chemotherapy is the current standard-of-care (SOC) for advanced PDAC; however, treatment resistance driven by a complex and immunosuppressive tumor microenvironment (TME) limits its effectiveness. To address mechanisms of resistance, we studied cabozantinib (cabo), a multi-kinase inhibitor approved for several solid tumors. Cabo has shown promise in preclinical PDAC models and clinical trials, particularly when combined with immunotherapy, but its mechanisms of action within the human PDAC TME and its potential to overcome therapy resistance remain unclear. Using CosMx Spatial Molecular Imaging (CosMx SMI) and Orion Multiplex Immunofluorescence (MxIF), we analyzed PDAC tissues collected after first line systemic chemotherapy and the Whipple surgical procedure, as well as from liver or lung metastatic sites from patients with PDAC. To functionally model the TME, we established matched patient-derived organoid (PDO) co-cultures harboring cancer associated fibroblasts (CAFs), and autologous immune cells (IMM) (PDO/CAF/IMM). These models were used to evaluate the effects of cabo and pembrolizumab, an anti-PD1 immune checkpoint inhibitor, with benchmarking of findings to the patients TME. Spatial analysis of post-chemotherapy and metastatic PDAC tissues revealed heterogeneous cellular neighborhoods within the TME, including enrichment of Schwann cells, CAFs, T regulatory cells, and cancer stem cells (CSCs). Distinct niches were observed in metastatic liver tissues, characterized by mesenchymal stem cells, fibroblasts, including myCAFs and iCAFs, and CSCs expressing CD44 and TROP2. In PDO/CAF/IMM co-cultures, treatment with cabo in combination with pembrolizumab enhanced cancer cell death by depleting myeloid-derived suppressor cells (MDSCs) and promoted cytotoxic T lymphocyte proliferation. Across both patient tissues and treated co-cultures, a persistent SOC-resistant cancer stem cell population emerged that expressed CD44 variant 9 (CD44v9). Taken together, these integrative spatial and organoid-based studies demonstrate that cabo can remodel the PDAC TME and potentiate PD-1 immunotherapy in preclinical models, while resistance associates with a CD44v9+ CSC population, revealing a potential therapeutic target.