Construction and Comprehensive Analysis of a Stratification System Based on AGTRAP in Patients with Hepatocellular Carcinoma.

Abstract

Background With the development of sequencing technology, several signatures have been reported for the prediction of prognosis in patients with hepatocellular carcinoma (HCC). However, the above signatures are characterized by cumbersome application. Therefore, the study is aimed at screening out a robust stratification system based on only one gene to guide treatment. Methods Firstly, we used the limma package for performing differential expression analysis on 374 HCC samples, followed by Cox regression analysis on overall survival (OS) and disease-free interval (PFI). Subsequently, hub prognostic genes were found at the intersection of the above three groups. In addition, the topological degree inside the PPI network was used to screen for a unique hub gene. The rms package was used to construct two visual stratification systems for OS and PFI, and Kaplan-Meier analysis was utilized to investigate survival differences in clinical subgroups. The ssGSEA algorithm was then used to reveal the relationship between the hub gene and immune cells, immunological function, and checkpoints. In addition, we also used function annotation to explore into putative biological functions. Finally, for preliminary validation, the hub gene was knocked down in the HCC cell line. Results We discovered 6 prognostic genes (SKA1, CDC20, AGTRAP, BIRC5, NEIL3, and CDC25C) for constructing a PPI network after investigating survival and differential expression genes. According to the topological degree, AGTRAP was chosen as the basis for the stratification system, and it was revealed to be a risk factor with an independent prognostic value in Kaplan-Meier analysis and Cox regression analysis (P < 0.05). In addition, we constructed two visualized nomograms based on AGTRAP. The novel stratification system had a robust predictive value for PFI and OS in ROC analysis and calibration curve (P < 0.05). Meanwhile, AGTRAP upregulation was associated with T staging, N staging, M staging, pathological stage, grade, and vascular invasion (P < 0.05). Notably, AGTRAP was overexpressed in tumor tissues in all pancancers with paired samples (P < 0.05). Furthermore, AGTRAP was associated with immune response and may change immune microenvironment in HCC (P < 0.05). Next, gene enrichment analysis suggested that AGTRAP may be involved in the biological process, such as cotranslational protein targeting to the membrane. Finally, we identified the oncogenic effect of AGTRAP by qRT-PCR, colony formation, western blot, and CCK-8 assay (P < 0.05). Conclusion We provided robust evidences that a stratification system based on AGTRAP can guide survival prediction for HCC patients.