BNIP3+ fibroblasts associated with hypoxia and inflammation predict prognosis and immunotherapy response in pancreatic ductal adenocarcinoma.

Background: Pancreatic ductal adenocarcinoma (PDAC) is one of the most malignant tumors that lacks effective treatment options. Cancer-associated fibroblasts (CAFs), an important component of the tumor microenvironment, associated with tumor progression, prognosis, and treatment response. This work aimed to explore the novel CAFs-associated target to improve treatment strategies in PDAC.

Methods: The PDAC single-cell sequencing data (CRA001160, n = 35) were downloaded and integrated based on GSA databases to classify fibroblasts into fine subtypes. Functional enrichment analysis and coexpression regulatory network analysis were used to identify the functional phenotypes and biological properties of the different fibroblast subtypes. Fibroblast differentiation trajectories were constructed using pseudochronological analysis to identify initial and terminally differentiated subtypes of fibroblasts. The changes in the proportions of different fibroblast subtypes before and after PDAC immunotherapy were compared in responsive and nonresponding patients, and the relationships between fibroblast subtypes and PDAC immunotherapy responsiveness were determined based on GSA and GEO database. Using molecular biology methods to confirm the effects of BNIP3 on hypoxia and inflammation in CAFs. CAFs were co cultured with pancreatic cancer cells to detect their effects on migration and invasion of pancreatic cancer.

Results: Single-cell data analysis divided fibroblasts into six subtypes. The differentiation trajectory suggested that BNIP3+ Fibro subtype exhibited terminal differentiation, and the expression of genes related to hypoxia and the inflammatory response increased gradually with differentiation time. The specific overexpressed genes in the BNIP3+ Fibro subtype were significantly associated with overall and disease progression-free survival in the patients with PDAC. Interestingly, the greater the proportion of the BNIP3+ Fibro subtype was, the worse the response of PDAC patients to immunotherapy, and the CRTL treatment regimen effectively reduced the proportion of the BNIP3+ Fibro subtype. After knocking out BNIP3, the hypoxia markers and inflammatory factors of CAFs were inhibited. Co-culture of CAFs with pancreatic cancer cells can increase the migration and invasion of pancreatic cancer, but this could be reversed by knocking out BNIP3.

Conclusions: This study revealed the BNIP3+ Fibro subtype associated with hypoxia and inflammatory responses, which was closely related to the poor prognosis of patients with PDAC, and identified signature genes that predict the immunotherapy response in PDAC.