Abstract PR011: Digenic CRISPR screening identifies novel KRAS/FGFR combination therapies for pancreatic cancer

Abstract

Adaptive signaling remains a major barrier to durable responses with targeted therapy in pancreatic ductal adenocarcinoma (PDAC), where oncogenic KRAS drives the vast majority of tumors. To identify clinically actionable resistance nodes, we developed RAS+, a digenic CRISPR knockout library that systematically interrogates pairwise genetic interactions among 144 genes spanning RAS/MAPK, PI3K, YAP/TAZ, receptor tyrosine kinase, adaptor, and cell-cycle signaling networks. The library evaluates 10,296 gene pairs and was screened across 12 cancer cell lines, enabling context-specific mapping of digenic dependencies. Across genetic backgrounds, RAS+ screening reveals marked lineage and oncogene specificity. KRAS-mutant models are dominated by KRAS-containing dependencies, a pattern distinct from those of BRAF- and PIK3CA-mutant models, which show more heterogeneous interaction maps. In KRAS-mutant PDAC, combined loss of KRAS and FGFR pathway components emerges as a uniquely selective vulnerability. Dual knockout of KRAS and FRS2 (a pan-FGFR signaling adaptor) is the most differentially effective gene pair in KRAS-mutant PDAC cell lines, with notably greater selectivity and synergy versus other clinically evaluated combinations, such as KRAS/EGFR or KRAS/SHP2. Pharmacologic studies demonstrated that combined KRAS and FGFR inhibition is highly synergistic in KRAS-mutant PDAC. We combined the mutant-selective KRAS G12D inhibitor zoldonrasib with the pan-FGFR inhibitor futibatinib, demonstrating substantial synergy and enhanced cell killing in KRAS G12D PDAC models. In xenograft studies, zoldonrasib/futibatinib was well-tolerated and induced deeper tumor regressions and delayed tumor regrowth compared with zoldonrasib monotherapy. Mechanistically, prolonged KRAS inhibition induced adaptive FGFR pathway activation, with associated reactivation of MAPK and PI3K signaling. PDAC models with acquired resistance to zoldonrasib showed upregulation of FGFR paralogs, and combined KRAS/FGFR inhibition overcame resistance. Cancer-associated fibroblasts (CAFs) may be a significant source of FGF ligands in the pancreatic tumor microenvironment (TME), and CAF conditioned media activated FGFR signaling and promoted KRAS inhibitor resistance, nominating the pancreatic TME as a paracrine source of adaptive resistance. These findings establish RAS-pathway focused combinatorial CRISPR screening as a robust approach for identifying clinically actionable combination vulnerabilities and nominate FGFR signaling as a lineage-specific mediator of KRAS inhibitor resistance in PDAC. Combined KRAS/FGFR inhibition represents a rational therapeutic strategy to deepen responses and delay resistance in KRAS-mutant pancreatic cancer.