Front Oncol. 2026 ;16
1754912
Background: Amino acid metabolism is integral to tumor proliferation, redox control, and immune regulation. Yet, studies in bladder cancer have largely centered on single amino acids, leaving the broader metabolic gene network insufficiently characterized.
Methods: Transcriptomic and clinical data from TCGA-BLCA, GSE13507, and GSE32894 were integrated with 32 MSigDB amino acid metabolism gene sets. Differential analysis, enrichment profiling, and consensus clustering defined metabolic subtypes. WGCNA and survival filtering identified candidates for a prognostic model, which was optimized using the MIME platform. Immune features and drug sensitivities were evaluated through multiple deconvolutions and pharmacogenomic resources. Single-cell data (GSE222315) were used to trace the cellular origin of model genes. PSPH expression and function were validated in tissues and bladder cancer cell lines.
Results: A total of 144 dysregulated amino acid metabolism-related genes were identified and used to define two distinct metabolic subtypes. One subtype was marked by coordinated upregulation of glutamine, branched-chain amino acid, tryptophan, and serine metabolic programs, accompanied by higher grade and stage, significantly worse survival, and dense but functionally impaired immune infiltration. From 24 candidate genes, a 16-gene metabolic signature was constructed and consistently validated across TCGA, GSE13507, and GSE32894, showing strong and stable prognostic performance superior to several published models. High-risk group displayed activation of cell-cycle, DNA-replication, mTORC1, and inflammatory-stress pathways, together with predicted sensitivity to PI3K/mTOR inhibitors, DNA-damaging agents, and selected epigenetic or cytoskeletal drugs. In the IMvigor210 cohort, the high-risk group showed a greater likelihood of responding to PD-1/PD-L1 blockade. Single-cell profiling localized signature expression predominantly to malignant epithelial cells. PSPH, a core model gene, was overexpressed in tumor tissues and cell lines, and functional assays demonstrated its role in promoting proliferation, migration, invasion, and survival of bladder cancer cells.
Conclusions: This study highlights the central role of amino acid metabolic networks in shaping bladder cancer heterogeneity and provides a metabolically grounded framework for risk stratification and therapeutic development.
Keywords: PSPH; amino acid metabolism; bladder cancer; machine learning; molecular subtype; prognostic signature