Ferroptosis-Related lncRNAs Signature is a Prognostic Factor in Hepatocellular Carcinoma: A Study Based on TCGA Data
Abstract Background Hepatocellular carcinoma (HCC) is the most common type of liver cancer and is the fourth leading cause of cancer-related death worldwide. Ferroptosis is a form of iron-dependent programmed cell death, and is characterized by intracellular accumulation of reactive oxygen species (ROS). Long non-coding RNAs (lncRNAs), as valuable prognostic factors for HCC patients, play a vital role in regulating ferroptosis. Methods RNA-sequencing datasets and ferroptosis-related genes were retrieved from The Cancer Genome Atlas (TCGA) database and the Molecular Signature Database. we performed Pearson correlation analysis between the lncRNAs and ferroptosis-related genes, and subsequently used regression analysis (univariate Cox analysis, multivariate Cox regression analysis, and Lasso regression analysis) to screen the ferroptosis-related lncRNAs with prognostic value in HCC, the prognostic ferroptosis-related lncRNAs signature (FRLS) was finally constructed. In addition, we reevaluated the model in terms of survival, clinical characteristics, and immune microenvironment. Results Univariate Cox regression analysis revealed 34 differently expressed ferroptosis-related lncRNAs related to the prognosis of HCC. Among them, 12 ferroptosis-related lncRNAs (LUCAT1, LINC01224, THUMPD3-AS1, AC116025.2, LINC00942, SNHG10, AC131009.1, POLH-AS1, MKLN1-AS, LINC01138, LNCSRLR, AL031985.3) were regarded as independent prognosis predictors of HCC, and were incorporated into the construction of the prognostic FRLS. Patients were divided into two groups based on the prognostic FRLS. Kaplan–Meier survival plot showed that patients in the high-risk groups exhibited shorter overall survival (OS) than those in low-risk groups (P < 0.001). Compared with clinical data, the area under curve (AUC) values of the risk factors, decision curve analysis (DCA), the AUC values of different years and multivariate Cox regression suggested that the signature had better predictive power. Gene set enrichment analysis (GSEA) revealed the potential pathways of 12 ferroptosis-related lncRNAs, including sphingolipid-metabolism, mTOR signaling pathway, notch signaling pathway, homologous recombination, endocytosis, cell cycle, etc. Immune microenvironment including tumor-infiltrating immune cells, immune-related functions, checkpoint-related genes and N6-methyladenosine (m6A)-related mRNA were also significantly different between the two risk groups. Conclusions This study constructed 12 FRLS for HCC patients to predict survival, which may provide promising targets for the therapy of HCC.