Identification of ferroptosis-related prognostic models and FDFT1 as a potential ferroptosis driver in colorectal cancer

Abstract

Abstract Background : Prediction of colorectal cancer (CRC) prognosis is challenging. Ferroptosis constitutes a newly reported kind of cell death, and its association with CRC prognosis remains unexplored. Herein, we aimed to develop ferroptosis-related gene (FRG) signatures to predict overall survival (OS) along with disease-free survival (DFS) in individuals with CRC. Methods : The clinical data and mRNA expression were extracted from the TCGA web data resource. The Lasso algorithm was utilized to construct the OS and DFS prediction signatures. Independent data from GSE38832 were used for verification. Results : Our findings revealed there was a discrepancy in the expression of 85% of FRGs between CRC and healthy tissues. Among them, 11 prognostic genes were identified using UniCox analysis. Predicted risk scores from the two models stratified patients into low- as well as high-risk groups and were demonstrated as independent prognostic factors using MultiCox analysis. The efficacy of the models was verified using ROC curve analysis. Functional enrichment analysis indicated that cancer-linked pathways were abundant in the high-risk group, and that immune status differed between the two risk groups. The CMap web data resource helped in identifying a total of sixteen potential drugs. In addition, FDFT1 was proved to play an anti-tumor role in CRC and may promote ferroptosis by regulating the expression of ISCU. Conclusions : Our FRG-based prognostic models are reliable predictive tools for CRC patients, suggesting that FRGs may be potential targets for CRC therapy.