Analysis of Ferroptosis-Mediated Modification Patterns and Tumor Immune Microenvironment Characterization in Uveal Melanoma

Uveal melanoma (UVM) is an intraocular malignancy in adults in which approximately 50% of patients develop metastatic disease and have a poor prognosis. The need for immunotherapies has rapidly emerged, and recent research has yielded impressive results. Emerging evidence has implicated ferroptosis as a novel type of cell death that may mediate tumor-infiltrating immune cells to influence anticancer immunity. In this study, we first selected 11 ferroptosis regulators in UVM samples from the training set (TCGA and GSE84976 databases) by Cox analysis. We then divided these molecules into modules A and B based on the STRING database and used consensus clustering analysis to classify genes in both modules. According to the Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA), the results revealed that the clusters in module A were remarkably related to immune-related pathways. Next, we applied the ESTIMATE and CIBERSORT algorithms and found that these ferroptosis-related patterns may affect a proportion of TME infiltrating cells, thereby mediating the tumor immune environment. Additionally, to further develop the prognostic signatures based on the immune landscape, we established a six-gene-regulator prognostic model in the training set and successfully verified it in the validation set (GSE44295 and GSE27831). Subsequently, we identified the key molecules, including ABCC1, CHAC1, and GSS, which were associated with poor overall survival, progression-free survival, disease-specific survival, and progression-free interval. We constructed a competing endogenous RNA network to further elucidate the mechanisms, which consisted of 29 lncRNAs, 12 miRNAs, and 25 ferroptosis-related mRNAs. Our findings indicate that the ferroptosis-related genes may be suitable potential biomarkers to provide novel insights into UVM prognosis and decipher the underlying mechanisms in tumor microenvironment characterization.