A novel model of anoikis-related gene for predicting prognosis and assessing immune status with hepatocellular carcinoma

Abstract

Abstract Background Hepatocellular carcinoma (HCC) is one of the highly malignant and aggressive gastrointestinal tumors. Anoikis is a specific form of cell death that is closely related to malignant aggressive behavior of tumors. The role and significance of anoikis-related genes (ANRGs) in HCC deserve to be explored. Methods Here, transcriptome profiling and relevant clinical data needed for analysis were collected from public databases. Prognostic model of ANRGs was constructed by using Lasso regression algorithm. Then, patients were given a reasonable risk grouping, and survival analysis was conducted to compare the different survival rates in each risk group. Receiver operating characteristic (ROC) curve was employed to examine the predictive accuracy of the prognostic model. The single sample gene set enrichment (ssGSEA) was carried out to investigate important disease characteristics of each risk group, such as immune status profile and tumor microenvironment differences. The gene set enrichment analysis (GSEA) method was also implemented to complete functional and pathway enrichment analysis. In addition, drug sensitivity analysis and exploration of single cell data for HCC were completed with the aid of online analytical databases. Results We successfully created a prognostic model containing 14 ANRGs, namely: ANXA5, BSG, SKP2, BAK1, PHLDA2, CDKN3, SFN, EZH2, HMGA1, PBK, NRAS, SLC2A1, MAD2L1 and CASP2, and observed a lower overall survival in high-risk group. The ROC curve confirmed good performance of this new model in predicting prognosis. The ssGSEA revealed significant differences in tumor immune microenvironment between different risk groups, with higher activity about cancer related pathways in high-risk group. The expression level of immune checkpoint and m6A genes also differed between risk subgroups. These prognostic genes were also be related to chemotherapy susceptibility. Conclusion The novel prognostic model identified with ANRGs can be applied to prediction prognostic and assessment immune status profile, tumor microenvironment differences and chemosensitivity in HCC. Rational use of the prognostic new model may provide an important reference for individualized treatment of HCC.