高通量离体TCR-T细胞筛选鉴定出KRAS突变型PDAC中聚集自噬驱动的治疗耐药克隆

Background: Pancreatic ductal adenocarcinoma (PDAC) is characterized by frequent KRAS‑G12 oncogenic mutations, low immune cell infiltration, and early metastatic spread, leading to poor outcomes. Selective pressure from systemic therapies promotes the emergence of resistant clones and relapse. Patient-derived models that capture tumor heterogeneity are needed to optimize T cell receptor-engineered T cell (TCR‑T) strategies and anticipate mechanisms of resistance. Methods: We developed a high-throughput ex vivo screening platform to test cell-based immunotherapies in patient-derived PDAC spheroids. KRAS‑G12V-specific CD8⁺ TCR‑T cells were co-cultured with KRAS-mutant 3D tumor spheroids under graded levels of cytotoxic pressure and monitored using automated image cytometry. Microcavity plates enabled parallel tracking of thousands of spheroids across two sequential TCR‑T challenges. Surviving cells were profiled by single-cell RNA sequencing (scRNA‑seq), followed by 2D regrowth assays and orthotopic transplantation into NSG mice to assess tumor-initiating capacity. Results: High-throughput resistance screening of over 33,000 microcavities revealed that 6.6% of 3D-PDAC spheroids remained viable after repeated TCR-T cell challenge, enriching for rare resistant subclones. After the first challenge, most cells exhibited an antiviral-like inflammatory response, characterized by increased levels of MX1, MX2, RIG-I, and interferon-stimulated gene expression. scRNA-seq after re-challenge revealed proliferating cells with upregulation of MHC II molecules and TUBA1B-associated aggrephagy pathways, implicating aggrephagy in immune escape. Resistant cells retained tumor-initiating potential and generated distant metastases in orthotopic NSG models ( n = 10). In vitro , paclitaxel, a microtubule disassembly inhibitor, reduced colony formation by TCR‑T-escaped clones, supporting a rationale for combination therapy. Conclusions: High-throughput ex vivo cytotoxicity screening of patient-derived PDAC models reveals aggrephagy-enriched, therapy-resistant clones and nominates rational combination partners such as paclitaxel. This platform may help anticipate patient-specific resistance to KRAS‑targeted TCR‑T therapy and guide the design of preventative combination strategies.