Construction and validation of a folate metabolism-related gene signature for prognosis prediction, immune landscape characterization, and molecular subtyping in glioma.

backgroundFolate metabolism plays essential roles in nucleotide synthesis, redox homeostasis, and cellular proliferation, yet its contribution to glioma progression and the tumor immune microenvironment remains incompletely understood.

methodsWe collected transcriptome, genomic, and clinical information of glioma patients from both The Cancer Genome Atlas and the Chinese Glioma Genome Atlas public databases. Key prognostic genes were identified by integrating differential gene expression profiles, weighted gene co-expression network analysis results, and genes linked to folate metabolism (FMRGs). The prognostic model was established through least absolute shrinkage and selection operator regression followed by multivariate Cox regression to select robust gene markers, and validated in two external CGGA cohorts. Subsequent analyses included pathway enrichment, immune landscape assessment (via single-sample gene set enrichment analysis, ESTIMATE, and CIBERSORT), mutation characteristics, drug response prediction, and Tumor Immune Dysfunction and Exclusion-based immunotherapy responsiveness evaluation. Consensus clustering was also employed to explore how the gene signature relates to tumor biology and patient prognosis.

resultsA total of 18 FMRGs showing significant expression variation were identified, of which eight formed a robust prognostic model with excellent predictive performance. High-risk patients exhibited markedly worse survival, enhanced apoptotic and p53-mediated stress pathways, extensive immune activation, elevated stromal and immune scores, and higher expression of immune checkpoints. Low-risk patients demonstrated lower TIDE scores and, in an exploratory external anti-PD-L1-treated cohort, showed higher response rates, suggesting potential immunotherapy-related relevance that requires glioma-specific validation. Using consensus clustering, we stratified the samples into two molecular subgroups characterized by unique transcriptomic landscapes, immune features, and prognostic behaviors. ONCLUSION: This study establishes a novel folate metabolism-related signature that reliably predicts prognosis, delineates immune landscape heterogeneity, and identifies potential therapeutic susceptibilities in glioma. The findings provide important insights into folate-associated tumor biology and offer a framework for risk stratification and personalized therapy.