Identification of cuproptosis-related subtypes, the establishment of a prognostic model, and exploration of drug candidates in urothelial carcinoma

Abstract Background Urothelial carcinoma (UC) originates from the urinary tract and can arise from the renal pelvis, ureter, urinary bladder, and urethra. This new kind of programmed cell death, called cuproptosis, is linked to tumor advancement and microenvironment homeostasis. Nevertheless, the biological functions of cuproptosis-related genes (CRGs) in UCs are yet unknown. Methods The three Gene Expression Omnibus datasets (GEO) datasets and The Cancer Genome Atlas (TGCA) database were utilized to obtain data from 972 UC patients. R software was utilized to identify different cancer clusters and comprehensively depicted their relationships with clinical pathological characteristics and tumor microenvironment. Results Firstly, 972 UC patients underwent classification to obtain three CRG clusters based on 17 prognostic CRGs, showing significantly different prognoses, clinical features, and immune infiltration statuses (P < 0.001). Specifically, we found that cluster C displayed more like an immune-inflamed phenotype and cluster B was more consistent with the immune-desert pattern. Also, three cuproptosis gene clusters were identified according to differentially expressed genes (DEGs), displaying significantly different prognoses (P < 0.001). Secondly, we developed and validated a cuproptosis risk model and the formula we utilized was: CRG_score = (-0.0887 * HSD17B2) + (0.2014 * KDELR3) + (0.1125* EFEMP1) + (0.1118 * TMEM45A). All samples were randomly divided into the training and validation sets, and the high-risk score group was strongly correlated with poor prognosis in both sets. Besides, the two risk groups displayed significant differences in the tumor microenvironment, mutation landscape, and anti-tumor agent sensitivity. Finally, a nomogram was constructed to predict individual risk, with AUC values of 0.78, 0.79, and 0.81 for 1-, 3-, and 5-year periods, correspondingly. Conclusions This study revealed the underlying biological patterns of cuproptosis-related genes in UC. We developed a CRG risk score that could act as a practical prognostic model to predict the clinical outcome of individual UC patients.