Development and validation of a model based on immunogenic cell death related genes to predict the prognosis and immune response to bladder urothelial carcinoma.

BACKGROUND: Immunogenic cell death (ICD) has been categorized as a variant of regulated cell death that is capable of inducing an adaptive immune response. A growing body of evidence has indicated that ICD can modify the tumor immune microenvironment by releasing danger signals or damage-associated molecular patterns (DAMPs), potentially enhancing the efficacy of immunotherapy. Consequently, the identification of biomarkers associated with ICD that can classify patients based on their potential response to ICD immunotherapy would be highly advantageous. Therefore the goal of the study is to better understand and identify what patients with bladder urothelial carcinoma (BLCA) will respond to immunotherapy by analyzing ICD signatures and investigate ICD-related prognostic factors in the context of BLCA. METHODS: The data obtained from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases regarding BLCA and normal samples was categorized based on ICD-related genes (IRGs). Specifically, we conducted an immunohistochemical (IHC) experiment to validate the expression levels of Calreticulin (CALR) in both tumor and adjacent tissues, and evaluated its prognostic significance using the Kaplan-Meier (KM) curve. Subsequently, the samples from TCGA were divided into two subtypes using consensus clustering. To obtain a more comprehensive comprehension of the biological functions, we utilized Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA). The calculation of immune landscape between two subtypes was performed through ESTIMATE and CIBERSORT. Risk models were constructed using Cox and Lasso regression and their prognosis predictive ability was evaluated using nomogram, receiver operating characteristic (ROC), and calibration curves. Finally, Tumor Immune Dysfunction and Exclusion (TIDE) algorithms was utilized to predict the response to immunotherapy. RESULTS: A total of 34 IRGs were identified, with most of them exhibiting upregulation in BLCA samples. The expression of CALR was notably higher in BLCA compared to the adjacent tissue, and this increase was associated with an unfavorable prognosis. The differentially expressed genes (DEGs) associated with ICD were linked to various immune-related pathways. The ICD-high subtypes exhibited an immune-activated tumor microenvironment (TME) compared to the ICD-low subtypes. Utilizing three IRGs including CALR, IFNB1, and IFNG, a risk model was developed to categorize BLCA patients into high- and low-risk groups. The overall survival (OS) was considerably greater in the low-risk group compared to the high-risk group, as evidenced by both the TCGA and GEO cohorts. The risk score was identified as an independent prognostic parameter (all p < 0.001). Our model demonstrated good predictive ability (The area under the ROC curve (AUC), AUC(1)-year= 0.632, AUC(3)-year= 0.637, and AUC(5)-year =0.653). Ultimately, the lower risk score was associated with a more responsive immunotherapy group. CONCLUSION: The potential of the ICD-based risk signature to function as a marker for evaluating the prognosis and immune landscape in BLCA suggests its usefulness in identifying the suitable population for effective immunotherapy against BLCA.