Characterization of an Autophagy-Immune Related Genes Score Signature and Prognostic Model and its Correlation with Immune Response for Bladder Cancer

Abstract

The study aimed to identify an autophagy-related molecular subtype and characterize a novel defined autophagy-immune related genes score (AI-score) signature and prognosis model in bladder cancer (BLCA) patients using public databases. Methods: The transcriptome cohorts downloaded from TCGA and GEO database were carried out with genomic analysis and unsupervised methods to obtain autophagy-related molecular subtypes. The single-sample gene-set enrichment analysis (ssGSEA) was utilized to perform immune subtype clustering. We defined a novel autophagy subtype and evaluated the role in TME cell infiltration. Then, the principal-component analysis (PCA) was applied to construct an AI-score signature. Subsequently, two immunotherapeutic cohorts were used to evaluate the predictive value in immunotherapeutic benefits and immune response. Finally, univariate, Lasso and multivariate Cox regression algorithm were used to construct and evaluate an autophagy-immune-related genes prognosis model. Also, qRT-PCR and IHC was applied to validate the expression of the 6 genes in the model. Results: Three distinct autophagy clusters and immune-related clusters were identified, and a novel autophagy-related molecular subtypes were defined. Furthermore, the roles in TME cell infiltration and clinical traits for the autophagy subtypes were characterized. Meanwhile, we constructed an AI-score signature and demonstrated it could predict genetic mutation, clinicopathological traits, prognosis, and TME stromal activity. We found that it could accurately predict the clinicopathological characteristics and immune response of individual BLCA patients and provide guidance for selecting immunotherapy. Ultimately, we constructed and verified an autophagy-immune-related prognostic model of BLCA patients and established a prognostic nomogram with a good prediction accuracy. Conclusion:We constructed AI-score signatures and prognosis risk model to characterize their role in clinical features and TME immune cell infiltration. It revealed that the AI-score signature and prognosis model could be a valid predictive tool, which could accurately predict the prognosis of BLCA patients and contribute to choosing effective personalized immunotherapy strategies.

Construction of the model for predicting prognosis by key genes regulating EGFR-TKI resistance

“…2) The expression data profiles of these five datasets were combined, and then we used "ComBat" function from the "sva" package to remove the batch effect (Leek et al, 2012). The 'sva' package and "ComBat" function were used in multiple studies to eliminate the batch effects (Li et al, 2020;Tang et al, 2020;J. Yu et al, 2022).…”