Identifying key genes involved in HBV-related hepatocellular carcinoma: diagnose, prognosis, interaction and immune analysis

Abstract

Background: Hepatitis B virus associated hepatocellular carcinoma (HBV-HCC) have been a serious global health problem. This study aimed to uncover the key genes in HBV-HCC, and clarity their function, interaction, diagnostic and prognostic value, impacts on immune infltration and potential drugs targeting these genes.

Methods: Four gene expression datasets totally containing 117 paired tumor tissues and adjacent control tissues were selected from the GEO database and used to screen the differentially expressed genes (DEGs). Function analysis were performed by using GO and KEGG enrichment. STRING and cytoscape were used to analyze protein-protein interaction (PPI) and screen hub gene. Survival analysis and receiver operator characteristic (ROC) curve were used to explore the prognostic and diagnostic value of key genes. Immune infiltration analysis were performed by CIBERSORT algorithm. Drug-Gene Interaction Database (DGIdb) was used to screen the potential drug that affect hub genes.

Results: Overall, 234 shared DEGs were screened from four GSE datasets, which were mainly enrichment in cell growth regulation, epoxygenase P450 pathway, cellular response to multiple ion, xenobiotic metabolic process and complement activation. Six hub genes (HMMR, NDC80, CDK1, EZH2, ESR1, FOXM1) were screen by PPI analysis. ESR1 was down-regulated and associated with favorable prognosis in HBV-HCC, while HMMR, NDC80, CDK1 and EZH2 were up-regulated and correlation with shorter overall survival. Furthermore, ROC analysis and nomogram demonstrated the high diagnostic performance of NDC80, CDK1 and EZH2. Immune infiltration analysis showed that there were significant difference of several immune cell types between tumor and control tissues, including T cells, monocyte/macrophage and dendritic cells. There were significant correlation between hub genes with immune infiltration. Finally, DGIdb analysis showed there were several approved or new drugs that interaction with HMMR, CDK1, ESR1 and EZH2.

Conclusion: Six hug-genes are closely related to the HBV-HCC development, which involved in multiple biological progress and immune infiltration. Among them, NDC80, CDK1, EZH2 could severed as markers with good diagnostic and prognostic value. Notably, several approved drugs interaction with hub genes might be potential drug used for HBV-HCC therapy.

Keywords: Diagnose; Differentially expressed genes; HBV-HCC; Immune infiltration; Prognosis.

Fig. 1

DEGs screening and function analysis based on GSE datasets. A The DEGs of GSE47197 dataset. B The DEGs of GSE55092 dataset. C The DEGs of GSE84402 dataset. D The DEGs of GSE121248 dataset. E Venn diagram showed the number of shared DEGs among four datasets. F The correlation matrix of 234 shared DEGs among four datasets. G–I GO analysis for 234 shared DEGs. J KEGG analysis for 234 shared DEGs

Fig. 2

PPI network construction and hub genes screening. A–C PPI network construction based on Betweenness, Closeness and Degree algorithm using Cytoscape. D Venn diagram showed the number of shared hub genes. E The expression changes of hub genes in paired tumor and adjacent control tissues based on GSE datasets. F The expression changes of hub genes in TCGA database

Fig. 3

Prognostic value of hub genes in HBV-HCC patients. A–F Overall survival of patients with high or low expression of HMMR, NDC80, CDK1, EZH2, ESR1 and FOXM1. G–L Progression Free Survival of patients with high or low expression of HMMR, NDC80, CDK1, EZH2, ESR1 and FOXM1

Fig. 4

Diagnostic value of hub genes. A–F ROC analysis showed the diagnostic performance of HMMR, NDC80, CDK1, EZH2, ESR1 and FOXM1. G, H Nomogram showed the diagnostic value of multi-gene combined diagnostic model

Fig. 5

The interaction network of hub genes. A Gene-gene network of hub genes was analyzed via GeneMANIA. B GO enrichment of genes interaction with hub genes. C–H Networkanalyst showed the TFs-gene network of HMMR, NDC80, CDK1, EZH2, ESR1 and FOXM1

Fig. 6

Immune infiltration analysis of HBV-HCC. A The difference of 22 types of immune cells infiltration between tumor tissues and adjacent control tissues based on four GSE datasets. B–G The correlation between hub genes expression (HMMR, NDC80, CDK1, EZH2, ESR1 and FOXM1) and immune cell infiltration in tumor tissues. (H) The correlation between the expression of hub genes and immune checkpoint