A 3D patternoid model for the reproducible characterization of invasive phenotypes and drug sensitivity in PDAC

Pancreatic ductal adenocarcinoma (PDAC) is a highly invasive and heterogeneous malignancy, posing challenges for reproducible modeling and functional phenotypic analysis. To address these limitations, we developed a standardized 3D patternoid platform using collagen-based microcavity arrays to enhance organoid formation consistency and quantify subtype-specific invasion mechanisms. We utilized murine primary PDAC cells stratified by epithelial-mesenchymal transition (EMT) into three subtypes: epithelial (E-9591), hybrid EMT (Mlow-8028), and mesenchymal (M-16992). The platform's sensitivity was verified by a strong correlation between EMT scores and invasive phenotypes, as well as responses to physiological concentrations of the protease inhibitor batimastat. Key invasion parameters-including invasive area, maximum invasion distance, and branching complexity-were measured under both genomic and drug-induced conditions. The platform demonstrated high inter-organoid reproducibility, with precise control over initial cell numbers ensuring batch-to-batch comparability. Invasion dynamics analysis revealed that epithelial cells (E-9591) primarily relied on spatial constraints within the microcavity to invade. Batimastat drug sensitivity assays further distinguished invasion dependencies of the mesenchymal subtypes, confirming that M-16992 patternoids exhibit a stronger sensitivity towards MMP inhibition compared to Mlow-8028 patternoids. Concurrentlty, both subtypes experienced a shift towards epithelial-like spatial constraint triggered invasion morphology, reflecting the plasticity of PDAC invasiveness. This scalable and adaptable 3D patternoid platform enables high-throughput analysis of invasive behaviors and therapeutic responses, offering significant potential for preclinical cancer research and personalized medicine.