A novel CD8+ T cell-related gene signature for predicting the prognosis and immunotherapy efficacy in bladder cancer.

ABSTRACT

OBJECTIVE: To identify CD8+ T cell-related molecular clusters and establish a novel gene signature for predicting the prognosis and efficacy of immunotherapy in bladder cancer (BCa). METHODS: Transcriptome and clinical data of BCa samples were obtained from the Cancer Genome Atlas (TCGA) and GEO databases. The CD8+ T cell-related genes were screened through the CIBERSORT algorithm and correlation analysis. Consensus clustering analysis was utilized to identified CD8+ T cell-related molecular clusters. A novel CD8+ T cell-related prognostic model was developed using univariate Cox regression analysis and Lasso regression analysis. Internal and external validations were performed and the validity of the model was validated in a real-world cohort. Finally, preliminary experimental verifications were carried out to verify the biological functions of SH2D2A in bladder cancer. RESULTS: A total of 52 CD8+ T cell-related prognostic genes were screened and two molecular clusters with notably diverse immune cell infiltration, prognosis and clinical features were developed. Then, a novel CD8+ T cell-related prognostic model was constructed. The patients with high-risk scores exhibited a significantly worse overall survival in training, test, whole TCGA and validating cohort. The AUC was 0.766, 0.725, 0.739 and 0.658 in the four cohorts sequentially. Subgroup analysis suggested that the novel prognostic model has a robust clinical application for selecting high-risk patients. Finally, we confirmed that patients in the low-risk group might benefit more from immunotherapy or chemotherapy, and validated the prognostic model in a real-world immunotherapy cohort. Preliminary experiment showed that SH2D2A was capable of attenuating proliferation, migration and invasion of BCa cells. CONCLUSIONS: CD8+ T cell-related molecular clusters were successfully identified. Besides, a novel CD8+ T cell-related prognostic model with an excellent predictive performance in predicting survival rates and immunotherapy efficacy of BCa was developed.