Individualized treatment and key prognostic biomarkers based on folate metabolism in patients with pancreatic cancer.

[BACKGROUND] Folate metabolism plays a pivotal role in tumor proliferation. However, the relationship between folate metabolism-related genes (FMGs) and the tumor immune microenvironment (TIME) in pancreatic cancer (PC) remains unclear. This study aimed to identify key FMGs and investigate whether FMGs are related to TIME in PC.

[METHODS] Transcriptomic data from 732 PC patients were obtained from public databases, and 78 FMGs were gathered from the Molecular Signature Database (MSigDB). Pan-cancer analysis of genetic alterations in FMGs was performed. Patients with PC were stratified into two subtypes using non-negative matrix factorization (NMF). Clinical data and pathological sections from 150 PC patients were retrospectively collected as a validation cohort. The levels of dihydrofolate reductase (DHFR), FAP and α-SMA in cancer-associated fibroblasts (CAFs), CD8 tumor-infiltrating lymphocytes (TILs), Foxp3 TILs, CD206 tumor-associated macrophages (TAMs) were analyzed using immunohistochemistry.

[RESULTS] Copy number variation (CNV), single nucleotide variation (SNV), methylation, and risk levels of FMG in pan-cancer were confirmed. Compared to Cluster 1, Cluster 2 demonstrated significantly poorer overall survival (OS) (P<0.05), increased sensitivity to chemotherapy drugs, lower immune cell counts, and more immunosuppressive cells in TIME. DHFR was identified as the folate metabolism-driving gene in PC. DHFR was an independent predictor of poor prognosis (P=0.001). DHFR expression was strongly associated with CD206 TAMs (P<0.001) and Foxp3 T cells (P<0.05).

[CONCLUSIONS] FMG expression heterogeneity significantly impacts PC prognosis and TIME. DHFR, a central folate metabolism enzyme, demonstrates critical associations with TIME modulation and clinical outcomes in PC.