Identification of crucial genes through WGCNA in the progression of gastric cancer

Abstract

Background: To explore the hub gene closely related to the progression of gastric cancer (GC), so as to provide a theoretical basis for revealing the therapeutic mechanism of GC. Methods: The gene expression profile and clinical data of GSE15459 in Gene Expression Omnibus (GEO) database were downloaded. The weighted gene co-expression network analysis (WGCNA) was used to screen the key modules related to GC progression. Survival analysis was used to assess the influence of hub genes on patients' outcomes. CIBERSORT analysis was used to predict the tissue infiltrating immune cells in patients. Immunohistochemical staining was conducted to further verify the expression of hub genes. Results: Through WGCNA, a total of 26 co-expression modules were constructed, in which salmon module and royalblue module had strong correlation with GC progression. The results of enrichment analysis showed that genes in the two modules were mainly involved in toll-like receptor signaling pathway, cholesterol metabolism and neuroactive ligand-receptor interaction. Six hub genes (C1QA, C1QB, C1QC, FCER1G, FPR3 and TYROBP) related to GC progression were screened. Survival analysis showed overall survival in the high expression group was significantly lower than that in the low expression group. CIBERSORT analysis revealed that immune characteristics difference between patients in early stage and advanced stage. Immunohistochemical results confirmed that C1QB, FCER1G, FPR3 and TYROBP were significantly associated with disease progression in GC. Conclusion: Our study identified that C1QB, FCER1G, FPR3 and TYROBP played important roles in the progression of GC, and their specific mechanisms are worth further study.

Keywords: Gastric cancer; Hub gene; Immunohistochemistry; Tumor progression; WGCNA.

Figure 1

Module-trait relationship. Each row indicates a module eigengene and each column indicates a trait.

Figure 2

The functional enrichment analysis of gene in the two clinically important modules. (A) Top 10 GO terms in each category of GO enrichment analysis. (B) Significantly enriched KEGG pathways. FDR: false discovery rate.

Figure 3

Hub gene identification between early stage and advanced stage of gastric cancer. (A)The scatterplot describing the relationship between MM and GS in salmon module. (B) The scatterplot describing the relationship between MM and GS in the royalblue module. (C) Fourteen hub genes expression between early stage and advanced stage of gastric cancer in GSE15459 cohort.

Figure 4

The correlation between C1QA, C1QB, C1QC, FCER1G, FPR3, TYROBP expression and the prognosis of gastric cancer was analyzed using GSE15459 cohort.

Figure 5

GSEA for samples with high hub gene expression and low hub gene expression. (A) The enriched gene sets in HALLMARK collection by samples of high C1QA, C1QB, C1QC, FCER1G, FPR3, TYROBP expression, respectively. (B) Gene set enriched in the interferon gamma response (p.adjust = 4.545e-10, NES =3.33, p-value = 1e-10). (C) Gene set enriched in the oxidative phosphorylation (p.adjust = 4.545e-10, NES =-2.39, p-value = 1e-10). NES: normalized enrichment score. GSEA: Gene Set Enrichment Analysis.

Figure 6

Estimation of tissues infiltrating immune cell types between early stage and advanced stage patients in the GSE15459 cohorts via CIBERSORT. (A) Stacked barplots show the relative composition of 22 immune cell subsets in 191 gastric cancer patients. (B) The boxplots show tissues infiltrating immune cell difference between early stage and advanced stage gastric cancer patients. Data were assessed via the method of wilcox test. * p-value < 0.05, ** p-value < 0.01, ns, no significance. (C) Heatmap of 6 hub genes and tissues infiltrating immune cell. Data were assessed via the method of Spearman analysis. * p-value < 0.05, ** p-value < 0.01.

Figure 7

Experimental verification of six hub genes in different gastric cancer stage tissues. (A) Statistical analysis of immunohistochemistry results in early-stage tissues (n = 9) and advanced stage tissues (n = 9). Data were assessed via the method of Mann-Whitney U test. * p-value < 0.05, ns, no significance. (B) Representative images of immunohistochemical staining for C1QA, C1QB, C1QC, FCER1G, FPR3, and TYROBP between early stage and advanced stage gastric cancer patients. Scale bars = 100 μm.