CXCL12 derived from cancer-associated fibroblasts mediates dysfunctional intratumoral adaptive immunity in diabetic pancreatic adenocarcinoma

Diabetes mellitus (DM) affects the occurrence and progression of pancreatic adenocarcinoma (PAAD), but the DM-related immune characteristic in PAAD remains unclear. This study aimed to elucidate the immune landscape and underlying mechanism in diabetic PAAD. RNA-sequencing data from primary sorted cancer-associated fibroblasts (CAFs) and TCGA database were used to screen out differentially expressed genes (DEGs), and enrichment analyses were performed to identify key pathways. xCell, Timer, Estimate, Quantiseq, and MCPcounter algorithms were applied to assess intratumoral immune infiltration, which was meanwhile verified by multicolor immunofluorescence of clinical samples. In vitro experiments, including proliferation assay and immune phenotype identification, were carried out to confirm the impact of the key gene on immune reprogramming. An animal model of mixed injection of pancreatic stellate cells (PSCs) and pancreatic cancer cells was constructed to validate in vivo effects of involved pathway in diabetic pancreatic cancer. Screened DEGs were enriched in several immunomodulatory pathways. The immunity in diabetic pancreatic cancer was shown to be seemingly flourishing but functionally defective, characterized by CD22+ B cells and TIGIT+ CD8+ T cells. Correlation analyses indicated that CXCL12 was significantly associated with the abundance and immune phenotype transformation of B and CD8+ T cells, with consistent results across multiple algorithms. More CXCL12High CAFs were detected in diabetic PAAD patients and mice. Meanwhile, the treatment of CXCL12High PSCs promoted the immune-suppressive phenotype of B and CD8+ T cells in both in vitro and in vivo experiments, and the blockade of CXCL12–CXCR4 axis by plerixafor showed therapeutic effects in subcutaneous tumor models. Our results delineated a distinct immune-suppressive landscape mediated by CAF-derived CXCL12 in diabetic PAAD individuals, which may help guide treatment decisions and predict immunotherapy response.