Disulfidptosis-related genes and long non-coding RNAs define a prognostic model for liver cancer.

BACKGROUND

Liver cancer is a prevalent malignancy with high mortality. Owing to the liver's capacity to conceal tumors until advanced stages, it is often challenging to detect them early. This study investigated a newly discovered form of cell death, disulfidptosis, and its prognostic value for liver cancer through analysis of disulfidptosis-related genes and long non-coding RNAs (lncRNAs).

METHODS

Data from 406 liver cancer cases and 58 normal cases obtained from The Cancer Genome Atlas (TCGA) database were analyzed with R and Perl software for differential expression, survival analysis, and Cox regression analysis. Candidate genes were further validated using the Human Protein Atlas (HPA) and Clinical Proteomic Tumor Analysis Consortium (CPTAC) databases. The real-time quantitative polymerase chain reaction (RT-qPCR) was used to verify the messenger RNA (mRNA) expression of candidate genes in liver cancer cell lines. Co-expression analysis was conducted to identify disulfidptosis-related lncRNAs, and a prognostic risk model was constructed via least absolute shrinkage and selection operator (LASSO)-Cox regression. The predictive performance of the model was evaluated by internal cohort validation, time-dependent receiver operating characteristic (ROC) curves, concordance index (C-index), and a prognostic nomogram.

RESULTS

Initially, 113 disulfidptosis-related genes were screened, among which 45 showed significant results. After Cox regression analysis, 26 genes were retained, and 739 lncRNAs co-expressed with these genes were detected. Six disulfidptosis-related genes (AAAS, IPO7, NCKAP1, PCBP2, RNH1, and SMPD4) were identified as independent prognostic factors. Protein-level expression of these genes was validated using the HPA and CPTAC datasets. RT-qPCR results demonstrated that NCKAP1, PCBP2, RNH1 and SMPD4 were highly expressed in liver cancer cell lines (MHCC97-L, MHC97-H, and HCC-LM3) in contrast to normal hepatocytes (THLE2). A lncRNA-based risk-score prognostic model was further constructed. Three pivotal lncRNAs were screened to build the prognostic signature. This signature effectively stratified patients into high- and low-risk groups with distinct survival outcomes and was confirmed to be an independent prognostic indicator.

CONCLUSIONS

This study suggested that these six genes could potentially serve as valuable prognostic indicators for liver cancer, and the developed risk assessment framework using correlated lncRNAs might assist in the clinical evaluation and personalized treatment of liver cancer.