Pancreatic ductal adenocarcinoma (PDAC) is driven by oncogenic KRAS mutations in more than 90% of patients. Although direct KRAS inhibitors have recently entered clinical development, durable therapeutic responses remain limited. Identifying effective combination strategies remains challenging, partly due to preclinical models that poorly capture the multicellular tumor microenvironment. We aimed to develop a defined, heterotypic 3D PDAC model that allows cell-type–specific tracking and therapeutic profiling.

We established a multicellular PDAC spheroid model composed of pancreatic cancer cells, pancreatic stellate cells, and endothelial cells embedded in a defined extracellular matrix primarily based on collagen I, supplemented with collagen III and IV, fibronectin, and laminin. Model robustness was validated across three pancreatic cancer cell lines. To enable cell-type discrimination, selected populations were fluorescently labeled (GFP-expressing stellate cells and dsRed-labeled cancer cells). Therapeutic responses were evaluated following 48 hour treatment with selective KRAS^G12D inhibitors (zoldonrasib, MRTX1133, daraxonrasib). Based on potency and reproducibility, MRTX1133 was selected for further experiments. To explore the platform’s utility for therapeutic discovery, clinically tested potassium channel modulators were screened in combination with KRAS inhibition. Membrane potential dynamics were assessed in complementary 2D assays using the voltage-sensitive reporter rEstus.

The defined ECM supported stable spheroid formation across multiple PDAC cell lines and enabled spatial discrimination of distinct cellular compartments. Fluorescent tracking revealed dynamic cellular organization, with stellate cells showing early outward migration while a distinct subpopulation remained within the spheroid core, indicating spatial heterogeneity. KRAS inhibition selectively reduced viability in KRAS-mutant spheroids (Panc-1, AsPc-1), whereas KRAS wild-type cells (BxPc-3) were largely unaffected. High-dose MRTX1133 reduced spheroid viability by 63%, while 1 nM had minimal effect. Screening of potassium channel modulators identified compounds that enhanced sensitivity to KRAS inhibition. Notably, combining 1 nM MRTX1133 with the Kv7 inhibitor linopirdine reduced spheroid viability by 41%. Complementary assays showed that dual KRAS and Kv7 inhibition induced membrane hyperpolarisation, reduced PDAC cell migration, and decreased TGF-β secretion by 2–4.8-fold.

This defined heterotypic 3D PDAC spheroid model captures key tumor–stroma interactions and enables functional testing of therapeutic combinations. As a proof of concept, the platform identified potassium channel modulation as a previously underexplored strategy to enhance KRAS inhibitor response.