Leveraging a disulfidptosis-related signature to predict the prognosis and immunotherapy effectiveness of cutaneous melanoma based on machine learning

Disulfidptosis is a recently discovered programmed cell death pathway. However, the exact molecular mechanism of disulfidptosis in cutaneous melanoma remains unclear. In this study, clustering analysis was performed using data from public databases to construct a prognostic model, which was subsequently externally validated. The biological functions of the model genes were then investigated through various experimental techniques, including qRT-PCR, Western blotting, CCK-8 assay, wound healing assay, and Transwell assay. We constructed a signature using cutaneous melanoma (CM) data, which accurately predicts the overall survival (OS) of patients. The predictive value of this signature for prognosis and immune therapy response was validated using multiple external datasets. High-risk CM subgroups may exhibit decreased survival rates, alterations in the tumor microenvironment (TME), and increased tumor mutation burden. We initially verified the expression levels of five optimum disulfidptosis-related genes (ODRGs) in normal tissues and CM. The expression levels of these genes were further confirmed in HaCaT cells and three melanoma cell lines using qPCR and protein blotting analysis. HLA-DQA1 emerged as the gene with the highest regression coefficient in our risk model, highlighting its role in CM. Mechanistically, HLA-DQA1 demonstrated the ability to suppress CM cell growth, proliferation, and migration. In this study, a novel signature related to disulfidptosis was constructed, which accurately predicts the survival rate and treatment sensitivity of CM patients. Additionally, HLA-DQA1 is expected to be a feasible therapeutic target for effective clinical treatment of CM.