Background: Cancer-associated fibroblasts (CAFs) are a vital constituent of the tumor microenvironment (TME) and have several activities, but the effect of CAF heterogeneity on the molecular features and clinical outcomes of pancreatic ductal adenocarcinoma (PDAC) remains unknown.Methods: An algorithm "scFrac" based on single-cell sequencing data from the Gene Expression Omnibus was introduced to emulate the enrichment of CAF subtypes in a TCGA-PDAC cohort and their prognostic influence, and confirmed by an external validation group (66 patients with PDAC) with multiplex immunohistochemistry staining. A comprehensive analysis including metabolic profile and transcription factor regulon activity was carried out among CAF subtypes.Results: Three distinct CAF populations were confirmed: myofibroblast (myCAF), inflammatory CAF (iCAF), and antigen-presenting CAF (apCAF). These subtypes expressed distinct metabolic profiles and transcriptional regulon activity. KEGG pathway annotation demonstrated that complement and coagulation cascades, as well as cytokine-cytokine receptor interaction were dominant in iCAFs, and pathways related to focal adhesion, and ECM-receptor interaction showed dominance in myCAFs, while antigen processing and presentation were the top enriched pathways in apCAFs. iCAFs trended to glycolysis with CREB3L1, EGR2 and SOX4 activation, whereas myCAFs depend on the tricarboxylic acid cycle and its derivatives with NRF2, CEBPD and YBX1 activation. iCAF is a protective factor associated with an inflammatory phenotype, but myCAF is an important factor in the poor prognosis of PDAC. Conclusions:We identified distinct molecular characteristics of 3 CAF subtypes in PDAC and plotted their metabolism profile. We introduced a novel algorism, scFrac, for exploring how CAF subgroups dysregulate cancer biology, and also shed a new therapeutic light on targeting the CAF subtype in TME.

No abstract