Single-cell transcriptomic analysis reveals key cell types and pathogenic genes associated with metastasis in pancreatic adenocarcinoma

Pancreatic ductal adenocarcinoma (PAAD) metastasis is driven by complex tumor-microenvironment interactions. Here, we integrated single-cell and bulk transcriptomic analyses of 104,855 cells from 10 patients to delineate the cellular and molecular landscape of primary versus metastatic PAAD. We identified metastasis-associated epithelial (LMO7⁺, TOP2A⁺, PIGR⁺), fibroblast (IGKC⁺, RGS5⁺), and M2-like macrophage (APOE⁺, CD14⁺, FOLR2⁺, SPP1⁺) subpopulations, validated via bulk deconvolution. Functional analyses revealed upregulated Wnt signaling, epithelial-mesenchymal transition, and angiogenesis in metastatic epithelial and fibroblast compartments. Intercellular communication analysis highlighted SPP1-mediated macrophage-epithelial/fibroblast crosstalk involving key receptor-ligand pairs, contributing to immune suppression and metastatic niche formation. Integrating gene expression and cell proportions, we developed a prognostic model with high predictive accuracy (C-index > 0.85), stratifying patients into risk groups with distinct immune landscapes. Furthermore, PTK6 was identified as a driver of PAAD proliferation, migration, and invasion. Collectively, our study elucidates TME-driven mechanisms of PAAD metastasis, identifies prognostic and therapeutic targets, and provides a framework for precision intervention.