Identification of CFHR4 as a Potential Prognosis Biomarker Associated With lmmune Infiltrates in Hepatocellular Carcinoma

Abstract

Background: Complement factor H-related 4 (CFHR4) is a protein-coding gene that plays an essential role in multiple diseases. However, the prognostic value of CFHR4 in hepatocellular carcinoma (HCC) is unknown.

Methods: Using multiple databases, we investigated CFHR4 expression levels in HCC and multiple cancers. The relationship between CFHR4 expression levels and clinicopathological variables was further analyzed. Various potential biological functions and regulatory pathways of CFHR4 in HCC were identified by performing a Gene Ontology (GO) analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis and Gene Set Enrichment Analysis (GSEA). Single-sample gene set enrichment analysis (ssGSEA) was performed to confirm the correlation between CFHR4 expression and immune cell infiltration. The correlations between CFHR4 expression levels in HCC and N6-methyladenosine (m6A) modifications and the competing endogenous RNA (ceRNA) regulatory networks were confirmed in TCGA cohort.

Results: CFHR4 expression levels were significantly decreased in HCC tissues. Low CFHR4 expression in HCC tissues was significantly correlated with the patients' sex, race, age, TNM stage, pathological stage, tumor status, residual tumor, histologic grade and alpha fetal protein (AFP) level. GO and KEGG analyses revealed that differentially expressed genes related to CFHR4 may be involved in the synaptic membrane, transmembrane transporter complex, gated channel activity, chemical carcinogenesis, retinol metabolism, calcium signaling pathway, PPAR signaling pathway, insulin and gastric acid secretion. GSEA revealed that the FCGR-activated reaction, PLK1 pathway, ATR pathway, MCM pathway, cascade reactions of PI3K and FGFR1, reactant-mediated MAPK activation and FOXM1 pathway were significantly enriched in HCC with low CFHR4 expression. Moreover, CFHR4 expression was inversely correlated the levels of infiltrating Th2 cells, NK CD56bright cells and Tfh cells. In contrast, we observed positive correlations with the levels of infiltrating DCs, neutrophils, Th17 cells and mast cells. CFHR4 expression showed a strong correlation with various immunomarker groups in HCC. In addition, high CFHR4 expression significantly prolonged the overall survival (OS), disease-specific survival (DSS) and progression-free interval (PFI). We observed a substantial correlation between the expression of CFHR4 and multiple N6-methyladenosine genes in HCC and constructed potential CFHR4-related ceRNA regulatory networks.

Conclusions: CFHR4 might be a potential therapeutic target for improving the HCC prognosis and is closely related to immune cell infiltration.

Keywords: CFHR4; biomarker; hepatocellular carcinoma; immune Infiltrate; prognosis.

Figure 1

Differences in the expression of CFHR4 and CFHR4-associated DEGs. (A) CFHR4 expression levels in different cancer tissues compared to normal tissues (TCGA). (B–D) CFHR4 expression in HCC samples. (E) CFHR4 expression was detected in WRL68, BEL7402, SK-Hep1, HCCLM3, HepG2, and Huh7 cell lines using Western blotting. (F) CFHR4 protein expression in 30 paired adjacent noncancerous tissues and HCC tissues. (G) CFHR4 expression was detected in WRL68, BEL7402, SK-Hep1, HCCLM3, HepG2, and Huh7 cell lines using PCR. (H) CFHR4 mRNA expression in 30 paired adjacent noncancerous tissues and HCC tissues. (I) ROC curves were created to investigate the value of CFHR4 in identifying HCC tissues. (J, K) Volcano plots of the DEGs and heatmap showing the top 10 DEGs. *p < 0.05, **p < 0.01, ***p < 0.001, NS, no significance.

Figure 2

Functional enrichment analyses of CFHR4-related genes in HCC. (A, B) The enriched terms in GO categories in HCC. (C, D) KEGG pathway analysis based on CFHR4-associated DEGs. (E–J) GSEA enrichment plots, including FCGR, activated reaction, PLK1 pathway, reactant FCERI-mediated MAPK activation, ATR pathway, MCM pathway, cascade reactions of PI3K and FGFR1, reactant-mediated MAPK activation and FOXM1 pathway.

Figure 3

PPI network enrichment analysis. (A) The PPI network was built based on PPI pairs identified by the STRING dataset. (B–F) Hub gene clusters were selected from the PPI network (criteria of total scores ≥ 5,000). (G) Top 7 hub genes in the PPI network.

Figure 4

Integrative analysis of CFHR4 expression in the infiltrating immune microenvironment. (A) The forest plot depicts the relationship between the level of CFHR4 expression and the relative abundances of 24 immune cells. (B–H) Scatter plots showing the differentiation of Th17 cells, Th2 cells, DCs, NK CD56bright cells, neutrophils, TFH cells and mast cells infiltration levels between high and low groups of CFHR4 expression. (I) The SCNA showed that CFHR4 expression correlated with the level of immune cell infiltration. (J) Scatter plots showing the correlations between 24 immune cells and CFHR4 expression levels. *p < 0.05, **p < 0.01, ***p < 0.001, NS, no significance.

Figure 5

Correlation of CFHR4 expression with clinicopathological characteristics. (A) Sex. (B) Age. (C) Race. (D) Histologic grade. (E) T stage. (F) N stage. (G) M stage. (H) AFP level. (I) Pathological stage. (J) Tumor status. (K) Residual tumor. (L) Vascular invasion. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 6

The prognostic value of CFHR4 in HCC. (A–C) Survival curves showing a comparison of OS, DSS and PFI between patients with HCC presenting high and low CFHR4 expression. (D–F) OS, DSS and PFI survival curves for Asian patients with HCC presenting high and low CFHR4 expression. (G–I) OS, DSS and PFI survival curves for patients with HCC aged ≤60 years presenting with high and low CFHR4 expression. (J–L) Univariate survival analysis of OS, PFI, and DSS in patients from different subgroups stratified according to TNM stage, pathological grade, tumor status, and CFHR4 expression levels. (M) For patients with HCC, a nomogram was constructed to estimate the probability of 1-, 3-, and 5-year OS. (N) Nomogram calibration plots for determining the probability of OS at 1, 3, and 5 years.

Figure 7

Analysis of the association between the CFHR4 expression level and the expression of m6A-related genes in HCC. (A) Correlation of CFHR4 expression levels with m6A gene expression in HCC. (B–P) Scatter plot showing the relationship between CFHR4 and m6A genes. (Q) Correlation of m6A genes in the CFHR4 high and low expression groups of HCC tumor samples. **p < 0.01, ***p < 0.001, NS, no significance.

Figure 8

Prediction of the ceRNA network in HCC. (A) Venn diagram showing the results for CFHR4 targets predicted using the TargetScan, DIANA-microT and RNAinter databases. (B–E) Scatter plots were generated to show miRNAs-mRNAs with significant correlations. TargetScan prediction of the potential binding sites in CFHR4 for the target miRNAs. (F–H) The lncRNAs that bind to target miRNAs were predicted using the miRNet and starBase online databases and displayed in a Venn diagram, including hsa−miR−146−5p, hsa−miR−361−5p and hsa−miR−580−3p. (I) Sankey diagram showing the CFHR4-related ceRNA regulatory network.