The present study is designed to determine potential target genes involved in hepatocellular carcinoma (HCC) and provide possible underlying mechanisms of action. Several studies (GSE112790, GSE87630, and GSE56140) from the GEO database looking at molecular characteristics in HCC were screened and analyzed by GEO2R, which led to the identification of a total of 93 differentially expressed genes (DEGs). From the protein–protein interaction (PPI) network, we selected 13 key genes with high degree of variability in expression in HCC. Expression of three key genes (NQO1, CYP2C9, and C6) presented with poor overall survival (OS) in HCC patients by UALCAN. C6, which is a complement component, was found by ONCOMINE and TIMER to have low expression in many solid cancers including HCC. Besides, Kaplan-Meier plotter and UALCAN database analysis to access diseases prognosis suggested that low expression of C6 is significantly related to worse OS in LIHC patients, especially in advanced HCC patients. Finally, the TIMER analysis suggested that the C6 expression showed significant negative correlation with infiltrating levels of six immune cells. The somatic copy number alterations (SCNAs) of C6 were associated with CD4+ T cell infiltration in HCC. Taken together, these results together identified C6 as a potential key gene in the diagnosis and prognosis of HCC.
Liver cancer has been ranked the fourth leading cause of tumor-related mortality worldwide, and about 841,000 new cases and 782,000 deaths worldwide are reported each year [
Recently, several HCC clinical samples have been analyzed with high-throughput sequencing technologies and extensive bioinformatics. The rapid advancement of genome sequencing technology and microarray has facilitated our understanding of cancer genomics, allowing us to explore more effective therapeutic targets for HCC. As a result, precision medicine has been extensively explored in HCC and other tumors [
We chose three HCC gene chip datasets from Gene Expression Omnibus (GEO). The differentially expressed genes (DEGs) comparing nontumor and tumor tissue were assessed by GEO2R. We then utilized DAVID database for gene annotation and enrichment pathway analysis of DEGs. From the DEGs’ PPI network, we identified key genes in common between the datasets. Overall survival (OS) associated with expression of the identified genes was assessed by UALCAN database. We found three key genes whose expression levels were related to poor survival in HCC patients. Literature retrieval results showed that expression of all genes identified, including NQO1, CYP2C9, and C6, except for the sixth complement component (C6), had been associated with poor HCC prognosis. In this study, C6 expression and its association with prognosis of various cancer including HCC was comprehensively investigated with the ONCOMINE, TIMER, Kaplan-Meier plotter, and UALCAN database. To further investigate the biological function of C6, we correlated its expression with tumor-infiltrating immune cells in HCC microenvironments by TIMER. These findings illustrated a vital role of C6 in HCC and provided a potential relationship between C6 expression and tumor-immune interactions.
We chose three gene expression profiles GSE112790 [
DEGs between the HCC groups and the nontumor groups in the three datasets were determined by the online tool GEO2R, using following cutoff criteria: adjusted
Gene ontology analysis (GO), which provides information on the molecular functions (MF), biological processes (BP), and cellular components (CC), was utilized for annotation of DEGs. Besides, KEGG (Kyoto Encyclopedia of Genes and Genomes), which is an online bioinformatics resource, was used to perform the pathway analysis of DEGs. DEGs annotated with both GO and KEGG were analyzed and summarized with DAVID (the Database for Annotation, Visualization and Integrated Discovery).
To identify PPI (the protein–protein interaction) among the DEGs and construct the DEGs’ interactome, we used STRING (search tool or the retrieval of interacting Genes). The cutoff values were set as confidence score above 0.4. Furthermore, the Cytoscape software (version 3.6.0) was used that automatically integrates and presents the results from STRING. The CytoHubba app in Cytoscape may be helpful for analyzing DEGs.
UALCAN is an online database based on the transcriptome results from RNA-seq and clinical data in TCGA. It can, therefore, provide information on the relative expression of genes in tumor tissues compared with normal tissues. Furthermore, the database also provides information on the influence of gene expression level on patient survival with multiple clinicopathologic profiles [
For further analysis, we chose one hub gene, namely C6, which had rarely been reported before in the context of HCC. First, the expression level of C6 across multiple cancers was obtained from the two independent databases in ONCOMINE which contained 715 cancer-related microarray datasets and in TIMER which based on TCGA database [
The online tool TIMER that includes 10,897 samples among diverse cancers in the TCGA database provides a comprehensive information on the relationship between gene expression and six different types of tumor-infiltrating immune cells including B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells [
Four HCC cell lines were generously donated by the Institute for Viral Hepatitis, Chongqing Medical University. HCC cell lines were cultured in DMEM (Hyclone) medium supplemented with 1% penicillin and streptomycin and 10% fetal bovine serum in a 5% CO2 atmosphere at 37°C.
Total RNA from HCC cells was extracted with Trizol (Life technology, USA) following the manufacturer’s protocol. RT-PCR Kit (TaKaRa, China) was used to amplify 2
All statistical analyses were performed with the SPSS 17.0 software, and continuous variables were described by
Figure
(a) The overall workflow of our work. (b) The data of Venn diagram about DEGs is extracted from the three GSE datasets (GSE112790, GSE87630, GSE56140). A total of 93 overlapping DEGs in the three datasets were obtained, as the following criteria: the absolute value of log
Characteristics of the three datasets.
Dataset ID | Country | Number of samples | Platform |
---|---|---|---|
GSE112790 | Japan | 183T 15N | GPL570 |
GSE87630 | Korea | 64T 30N | GPL6947 |
GSE56140 | USA | 35T 34N | GPL18461 |
GSE: Gene Expression Omnibus Series; GPL: Gene Expression Omnibus Platform; T: tumor samples; N: normal samples.
To further evaluate the biological function of DEGs, GO and KEGG enrichment analyses were performed using the DAVID database. With respect to KEGG pathway analysis, the most enriched pathways were chemical carcinogenesis and retinol metabolism pathways (Figure
Based on the information provided by STRING, which is a public database, we created the PPI network of DEGs by Cytoscape. Top 13 key genes with degree connectivity above 10 were selected by CytoHubba plug-in. The 13 genes were TAT, F9, MBL2, SPP2, FETUB, NQO1, C8A, HGFAC, KLKB1, ALDH8A1, CYP2E1, C6, and CYP2C9 (Figure
We then investigated the association between the expression of the 13 key genes, and HCC prognosis from the available database UALCAN was investigated. Based on UALCAN, the mRNA expression of C6 (
The relative expression of the NQO1 (a), CYP2C9 (b), and C6 (c) in HCC by UALCAN database. The relationship between the three key gene ((d) NQO1; (e) CYP2C9; (f) C6) expression and overall survival in HCC by UALCAN database. Color images are acquired according to the online UALCAN database.
To explore the difference of the C6 expression between the tumor and normal tissue, we analyzed C6 mRNA expression levels across various malignancies by ONCOMINE. Interestingly, we found that the C6 expression was significantly lower in most solid cancers except in leukemia, based on 1 dataset (Figure
The expression of C6 across different cancer types was explored by the ONCOMINE (a) and TIMER (b) (
Based on the results in both ONCOMINE and TIMER databases, we found low C6 expression in BRCA (breast invasive carcinoma), HNSC (head and neck cancer), COAD (colon adenocarcinoma), LUAD (lung adenocarcinoma), LUSC (lung squamous cell carcinoma), and LIHC (liver hepatocellular carcinoma). We then used the UALCAN Database to investigate whether the C6 expression in these selected cancer types was correlated with diseases prognosis. The results demonstrated that the low C6 mRNA expression was correlated with relatively low OS in LUAD (
The correlation of the C6 mRNA expression and overall survival in other four cancer types with UALCAN database.
Cancer types | C6 mRNA expression | Overall survival | |
---|---|---|---|
Expression level | |||
BRCA | Downregulation | 0.15 | |
COAD | Downregulation | 1.55 | 0.9 |
HNSC | Downregulation | 0.77 | |
LUSC | Downregulation | 0.21 |
BRCA: breast invasive carcinoma; HNSC: head and neck cancer; COAD: colon adenocarcinoma; LUAD: lung adenocarcinoma.
To explore the potential mechanisms of the C6 expression level in HCC in more depth, we correlated the C6 expression with both disease prognosis and clinical features in Kaplan-Meier plotter databases. Low C6 expression was related to significantly poor OS in male patients (
Overall survival curves in LIHC with association between the C6 expression and different clinicopathological feature by Kaplan-Meier plotter databases. The high and low C6 expression with different gender in male (a) (
Cancer cells within the tumor microenvironment (TME) and neighboring tumor-associated noncancerous cells play important role in tumor biology. TIMER was used to investigate the potential associations of the C6 expression with tumor homogeneity and infiltrating immune cells. The results suggested that the C6 expression in HCC was not associated with tumor homogeneity (
Correlation between the C6 mRNA expression and the abundance of immune infiltration level in LIHC (a). Each dot indicates a sample from TCGA dataset. Association of C6 SCNAs with immune infiltration in LIHC (b). The infiltration abundance in every SCNA category was compared to the diploid/normal.
The process of hepatocarcinogenesis involves progressive accumulation of genetic alterations. The advent of high-throughput technologies has revolutionized cancer genomic research by focusing on genetic alterations and providing an effective approach to identify new biomarkers involved in HCC development and progression. Despite these advances, the molecular dysregulations leading to HCC remain unknown. In this study, we identified a total of 93 DEGs using GEO2R from the three GEO databases (GSE112790, GSE87630, and GSE56140). The GO pathways enriched with the DEGs included oxidoreductase activity, acting on paired donors, with incorporation or reduction of molecular oxygen, organelle membrane, and epoxygenase P450 pathway. Interestingly, these pathways were found to be mainly involved in chemical carcinogenesis and retinol metabolism. Among these DEGs, we chose 13 key genes. Using UALCAN, the expressions of these three key genes (NQO1, CYP2C9 and C6) were found to be associated with poor OS in HCC patients.
We further elucidated the progression of the three key genes in HCC. One of the three genes, NQO1 (NAD(P)H quinone dehydrogenase 1), is an antioxidant enzyme, which is known to be upregulated in HCC patients. A recent study has shown that elevated NQO1 could activate both the PI3K/Akt and MAPK/ERK pathways and promote metabolic adaptation in HCC [
The third gene, C6, is a member of the complement system and is involved in innate immunity. Complete primary structure and function of C6 was elucidated in 1989. C6 is located on chromosome 5p13 and synthesized mainly in the liver in the form of glycoproteins [
Based on this analysis, the C6 expression between normal tissues and cancer was found to be similar in diversified solid cancers. According to the results of ONCOMINE database, the C6 expression presented a significant decline in many solid cancers compared with normal tissues, with only one study showing an opposite result: the high C6 expression found in leukemia patients. Compared with normal adjacent tissues, the TIMER based on TCGA further indicated that the C6 expression was lower in 14 cancer types including HCC but higher only in KIRC. Moreover, the prognostic impact of the low C6 expression in LUAD and LIHC by UALCAN was consistent, demonstrating a reduced C6 expression associated with shorter overall survival (OS) in HCC and LUAD. Nevertheless, analysis of data from Kaplan-Meier plotter database suggested that low level of the C6 expression was strongly related to worse prognosis only in LIHC. In addition, the low C6 expression was found to be related with worse OS of HCC patients in stages 3 and 4, OS of male patients, and those in Asian. It is interesting to find the gender distinction between the C6 expression and OS in HCC patients. Previous findings have postulated that sex-based distinction in the immune system could be associated with the natural course of chronic inflammatory including cancer. Cancer mortality rates were higher in male than in female from the vast majority of cancers [
A growing body of research has shown that the tumor-infiltrating immune cells influence the clinical outcomes and efficacy of chemotherapy and immunotherapy in HCC [
In summary, the decreased C6 expression in HCC is associated with worse prognosis and elevated immune cell infiltration. This finding suggests that C6 is likely to participate in immune cell infiltration and may be a potential biomarker for HCC diagnosis and prognosis.
The open available data sources or tools are as following: (1)
The authors declare that there is no conflict of interests regarding the publication of this paper.
DM did the RT-PCR, written the first draft of the manuscript, and did literature review. FQ and BSL were responsible for the data collection and data analysis. QZ designed the study and edited the draft and intellectual inputs.
The authors would like to thank the institute for viral hepatitis, Chongqing Medical University, for providing the four HCC cell lines. This study was supported by Medical Science Research Project of Chongqing Health and Family Planning Commission, China (NO. 2016MSXM080).
Supplementary Table 1: compared with normal tissues, a total of 93 DEGs were obtained from three profile datasets, including 10 upregulated and 83 downregulated genes in HCC tissue.