Deciphering a hydrogen sulfide-related signature to supervise prognosis and therapeutic response in colon adenocarcinoma

Abstract

Abstract Background: Hydrogen sulfide (H2S) is a critical molecule that participates in various molecular, physiological, and pathophysiological processes in biological systems. Emerging evidence has revealed that H2S is implicated in the progression of colon cancer and immune escape. Against this backdrop, the present study aimed to construct a prognostic risk feature for colon adenocarcinoma (COAD) by leveraging H2S-related genes (HSRG). Methods: Transcriptomic data and corresponding clinical-pathological information of colon cancer were obtained from TCGA and GEO databases. Univariate Cox regression analysis was employed to assess the prognostic relevance of HSRG. Consensus clustering was utilized to perform molecular subtyping of COAD, followed by comparison of immune cell infiltration, drug sensitivity, and immune therapy response between subtypes. Differential expression gene and gene set enrichment analyses were conducted between subtypes. Univariate, lasso, and multivariate Cox regression analyses were applied to construct a prognostic model derived from HSRG. A nomogram model for predicting COAD prognosis was constructed and evaluated. Results: In this study, we identified 12 HSRGs that were associated with COAD prognosis. Consensus clustering analysis revealed 3 COAD molecular subtypes that exhibited significant differences in terms of prognosis, tumor immune cell infiltration, drug sensitivity, and immune therapy response. Gene set enrichment analysis demonstrated that immunoregulatory processes were significantly suppressed in the poor-prognosis subtype while Wnt-related pathways and processes were significantly upregulated. Based on the differentially expressed genes between subtypes, we constructed a risk model comprising 11 genes that effectively distinguished high-risk patients from low-risk patients with significant associations with patient survival outcomes, drug treatment, pathological staging, and T staging. The HSRG-derived risk feature was an independent prognostic factor for COAD in drug treatment and pathological staging and could be integrated into a nomogram for prognosis prediction. Calibration curve, receiver operating characteristic curve (ROC), and decision curve analysis demonstrated excellent performance of the nomogram in evaluating COAD prognosis. Conclusion: Our study systematically assessed the prognostic significance of HSRG in COAD, identified HSRG-based molecular subtypes and risk features, and highlighted their potential utility in predicting prognosis and treatment response.