Bioinformatics profiling identifies seven immune-related risk signatures for hepatocellular carcinoma

Abstract

Background Density of tumor infiltrating lymphocytes (TIL) and expressions of certain immune-related genes have prognostic and predictive values in hepatocellular carcinoma (HCC); however, factors determining the immunophenotype of HCC patients are still unclear. In the current study, the transcript sequencing data of liver cancer were systematically analyzed to determine an immune gene marker for the prediction of clinical outcome of HCC. Methods RNASeq data and clinical follow-up information were downloaded from The Cancer Genome Atlas (TCGA), and the samples were assigned into high-stage and low-stage groups. Immune pathway-related genes were screened from the Molecular Signatures Database v4.0 (MsigDB) database. LASSO regression analysis was performed to identify robust immune-related biomarkers in predicting HCC clinical outcomes. Moreover, an immune gene-related prognostic model was established and validated by test sets and Gene Expression Omnibus (GEO) external validation sets. Results We obtained 319 immune genes from MsigDB, and the genes have different expression profiles in high-stage and low-stage of HCC. Univariate survival analysis found that 17 genes had a significant effect on HCC prognosis, among them, 13 (76.5%) genes were prognostically protective factors. Further lasso regression analysis identified seven potential prognostic markers (IL27, CD1D, NCOA6, CTSE, FCGRT, CFHR1, and APOA2) of robustness, most of which are related to tumor development. Cox regression analysis was further performed to establish a seven immune gene signature, which could stratify the risk of samples in training set, test set and external verification set (p < 0.01), and the AUC in both training set and test set was greater than 0.85, which also greater compared with previous studies. Conclusion This study constructed a 7-immunogenic marker as novel prognostic markers for predicting survival of HCC patients.

Keywords

• Bioinformatics
• Immunity
• Prognostic markers
• TCGA