Integrated analysis of metabolic gene features and the immune microenvironment: identification of DPYD-mediated prognostic model and therapeutic targets in pancreatic cancer

Pancreatic ductal adenocarcinoma (PAAD) is a malignant digestive system tumor characterized by extremely poor prognosis and significant molecular heterogeneity. Metabolic reprogramming, a key hallmark of cancer, is considered a vital mechanism underlying immune evasion and malignant progression in pancreatic cancer. RNA-sequencing data and clinical information of PAAD patients were retrieved from TCGA and GTEx datasets. Metabolism-related genes were curated from previous literature (n = 712). Differential expression analysis between TCGA-PAAD tumors and GTEx normal pancreas identified 681 metabolism-associated differentially expressed genes (MAGDEGs), from which 19 prognosis-related metabolism genes (MAGCOXs) were further screened using univariate Cox regression. Consensus clustering determined metabolic subtypes. A prognostic model was constructed using StepCox forward + GBM algorithm. Kaplan–Meier survival curves, log-rank test, and ROC curves validated predictive performance. Differences in immune infiltration, TMB, IPS, TIDE, and anticancer drug sensitivity related to risk scores were analyzed. Single-cell clustering and pseudotime analyses further elucidated the significance of metabolism-related genes. DPYD expression in PAAD tissues and cell lines was assessed via qRT-PCR, WB, and IHC. Functional assays and proteomic analysis were used to explore the role of DPYD in PAAD. The model reliably distinguished high- and low-risk patient groups, demonstrating robust predictive stability and generalizability across independent GEO validation cohorts. The high-risk group showed enhanced immune evasion characteristics and distinct drug sensitivity profiles. DPYD overexpression correlated with poor prognosis; DPYD knockdown inhibited colony formation and migration in BxPC-3 cells, whereas its overexpression promoted these capacities. Proteomics analysis revealed DPYD deficiency disrupted key metabolic pathways including fatty acid metabolism, TCA cycle, and amino acid metabolism. Enrichment analysis indicated DPYD involvement in pancreatic cancer and related diseases. We identified metabolism-associated DEGs and further screened 19 prognosis-related metabolism genes, based on which a metabolism-associated prognostic model with excellent prognostic value for PAAD patients was constructed. The high-risk group displayed stronger immune evasion characteristics and distinct drug sensitivity compared to the low-risk group. External validation confirmed the model’s predictive stability and generalizability. DPYD was predominantly enriched in macrophages and T-cell subsets, and its high expression positively correlated with poor prognosis. Elevated DPYD expression was validated in PAAD tissues and cell lines, and functional assays demonstrated that DPYD knockdown inhibited tumor cell clonogenicity, migration, and disrupted key metabolic pathways.