Novel prognostic biomarkers of gastric cancer based on gene expression microarray: COL12A1, GSTA3, FGA and FGG

Abstract

Gastric cancer (GC) is the fifth most common malignancy and the third leading cause of cancer‑associated mortality in the world. However, its mechanisms of occurrence and development have not been clearly elucidated. Furthermore, there is no effective tumor marker for GC. Using DNA microarray analysis, the present study revealed genetic alterations, screened out core genes as novel markers and discovered pathways for potential therapeutic targets. Differentially expressed genes (DEGs) between GC and adjacent normal tissues were identified, followed by pathway enrichment analysis of DEGs. Next, the protein‑protein interaction (PPI) network of DEGs was built and visualized. Analyses of modules in the PPI network were then performed to identify the functional core genes. Finally, survival analysis of core genes was conducted. A total of 256 genes were identified as DEGs between the GC samples and normal samples, including 169 downregulated and 87 upregulated genes. Through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis, the present study identified a total of 143 GO terms and 21 pathways. Six clusters of functional modules were identified, and the genes associated with these modules were screened out as the functional core genes. Certain core genes, including collagen type 12 α1 chain (COL12A1), glutathione S‑transferase α3 (GSTA3), fibrinogen α chain (FGA) and fibrinogen γ chain (FGG), were the first reported to be associated with GC. Survival analysis suggested that these four genes, COL12A1 (P=0.002), GSTA3 (P=3.4x10‑6), FGA (P=0.00075) and FGG (P=1.4x10‑5), were significant poor prognostic factors and therefore, potential targets to improve diagnosis, optimize chemotherapy and predict prognostic outcomes.

Figure 1.

Normalized expression value data. The black line in each box represents the median of each set of data, which determined the degree of standardization of data through its distribution. (A) Box plots of expressed value data prior to normalization. (B) Box plots of expressed value data following normalization. The gene expression value is presented on the y-axis and the samples on the x-axis.

Figure 2.

Heatmap of the DEGs. Overview of the associations between DEGs from normal and GC tissues. Each row represents the tissues (10 GC and 10 normal). Red coloration indicated 10 GC tissues. Blue coloration indicated 10 adjacent normal tissues. Each column represents the gene IDs. Log ratio scale bar for the Treeview color change was also presented. The red in the log ratio scale bar indicates highly expressed genes in GC tissues when compared with the normal tissues, while the blue represents the low expression of genes. The number in each block with its color in the log ratio scale bar represents the expression level of each gene in different samples. GC, gastric cancer; DEGs, differentially expressed genes.

Figure 3.

Volcano plot overview of the distribution of DEGs. Red spots indicate upregulated genes, while green spots indicate downregulated genes. Black spots represent genes that were not DEGs. The-log10 (adj. P-value) on the x-axis is the P-value following correction and the log2FC value on the y-axis represents the fold change of genes. DEGs, differentially expressed genes.

Figure 4.

PPI network of DEGs. The network contained 116 nodes and 197 edges. The 39 red nodes were the upregulated genes and the 76 green nodes were the downregulated genes. The one pink node indicates a gene that was automatically generated and was potential associated with DEGs. The edges represented the correlation between two genes. PPI, protein-protein interaction; DEGs, differentially expressed genes.

Figure 5.

Functional modules involved in the PPI network. The red nodes represent the upregulated genes and the green nodes represent the downregulated genes. The edges indicate the correlation between two genes. (A) Module A was comprised of 8 nodes and 28 edges. (B) Module B was comprised of 6 nodes and 9 edges. (C) Module C was comprised of 3 nodes and 3 edges. (D) Module D was comprised of 3 nodes and 3 edges. (E) Module E was comprised of 3 nodes and 3 edges. (F) Module F was comprised of 3 nodes and 3 edges. PPI, protein-protein interaction.

Figure 6.

Kaplan-Meier survival curves of four genes generated from Kaplan-Meier plotter. Curves of (A) COL12A1, (B) GSTA3, (C) FGA and (D) FGG. COL12A1, collagen type 12 α1 chain; GSTA3, glutathione S-transferase α3; FGA, fibrinogen α chain; FGG, fibrinogen γ chain; HR, hazard ratio; CI, confidence interval.