Construction and validation of a prognostic model of metabolism-related genes driven by somatic mutation in bladder cancer

Abstract

Abstract Background Bladder cancer (BLCA) is a highly malignant disease in the urinary system. Somatic mutation is a key feature in cancer occurrence, development, and treatment. Moreover, altered metabolism contributes to patient prognosis. However, the role of metabolism-related genes (MRGs) driven by somatic mutations in BLCA remains unclear. Methods The data were gathered from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. The differentially expressed metabolism-related genes (DEMRGs) between normal and BLCA patients were first identified and the functions were discussed. Then The risk model was construct by the DEMRGs with mutation frequency. The accuracy of the risk model was verified by risk curves, Kaplan-Meier (K-M) curves, and Receiver Operating Characteristic (ROC) curves. Subsequently, the correlation of risk score and clinical traits was also researched. Gene Set Enrichment Analysis (GSEA), immune checkpoints, immune microenvironment, and chemotherapeutic drug sensitivity were performed in high- and low-risk groups. And the scRNA-seq revealed that the expression pattern of prognostic biomarkers and cellular heterogeneity. Eventually, the mRNA expression levels of biomarkers were validated by quantitative real-time PCR (qRT-PCR). Results A total of 201 DEMRGs were retrieved, and the DEMRGs were significantly enriched in alcohol metabolic process, cellular modified amino acid metabolic process, and purine metabolism. Then the 24 DEMRGs of the mutation frequency greater than 3% were further analyzed, and a risk model was constructed by 5 biomarkers (FASN, ABCC4, ATP2B4, ATP8B2, and MTHFD1L). Moreover, the AUCs were all greater than 0.6, indicating the risk model had good efficacy. Meanwhile, the riskScore, T-pathologic, age, and N-pathologic were regarded as independent prognostic indicators. The DEMRGs were enriched in OXIDATIVE_PHOSPHORYLATION. Three immune checkpoints, four types of immune cells, and 146 drugs were substantially different in the two risk groups. And the scRNA-seq further disclosed relationships between genes regulation and tracked the development trajectories of distinct cell lineages. Finally, qRT-PCR results showed the expression levels of FASN and MTHFD1L were significantly higher in carcinoma tissue. Conclusion In brief, this study constructed a novel biomarkers, which could improve the prediction of independent prognosis indicators and guide individualized treatment of BLCA patients.