Integrative genomics analysis of hub genes and their relationship with prognosis and signaling pathways in esophageal squamous cell carcinoma

Abstract

The main purpose of the present study was to recognize the integrative genomics analysis of hub genes and their relationship with prognosis and signaling pathways in esophageal squamous cell carcinoma (ESCC). The mRNA gene expression profile data of GSE38129 were downloaded from the Gene Expression Omnibus database, which included 30 ESCC and 30 normal tissue samples. The differentially expressed genes (DEGs) between ESCC and normal samples were identified using the GEO2R tool. Gene Ontology and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to identify the functions and related pathways of the genes. The protein‑protein interaction (PPI) network of these DEGs was constructed with the Search Tool for the Retrieval of Interacting Genes and visualized with a molecular complex detection plug‑in via Cytoscape. The top five important modules were selected from the PPI network. A total of 928 DEGs, including ephrin‑A1 (EFNA1), collagen type IV α1 (COL4A1), C‑X‑C chemokine receptor 2 (CXCR2), adrenoreceptor β2 (ADRB2), P2RY14, BUB1B, cyclin A2 (CCNA2), checkpoint kinase 1 (CHEK1), TTK, pituitary tumor transforming gene 1 (PTTG1) and COL5A1, including 498 upregulated genes, were mainly enriched in the 'cell cycle', 'DNA replication' and 'mitotic nuclear division', whereas 430 downregulated genes were enriched in 'oxidation‑reduction process', 'xenobiotic metabolic process' and 'cell‑cell adhesion'. The KEGG analysis revealed that 'ECM‑receptor interaction', 'cell cycle' and 'p53 signaling pathway' were the most relevant pathways. According to the degree of connectivity and adjusted P‑value, eight core genes were selected, among which those with the highest correlation were CHEK1, BUB1B, PTTG1, COL4A1 and CXCR2. Gene Expression Profiling Interactive Analysis in The Cancer Genome Atlas database for overall survival (OS) was applied among these genes and revealed that EFNA1 and COL4A1 were significantly associated with a short OS in 182 patients. Immunohistochemical results revealed that the expression of PTTG1 in esophageal carcinoma tissues was higher than that in normal tissues. Therefore, these genes may serve as crucial predictors for the prognosis of ESCC.

Figure 1.

Heatmap of DEGs in ESCC and normal esophageal samples. The heatmap shows the DEGs between the ESCC and normal samples. When |logFC| was set to >2.5 and the P-value was adjusted to <0.01, 52 differential genes, including 30 upregulated and 22 downregulated genes, were identified. The clustering of the samples is shown above the dendrogram and the level of gene expression signal us indicated by the color (red and green for high and low expression levels, respectively). ESCC, esophageal squamous cell carcinoma; DEGs, differentially expressed genes.

Figure 2.

Volcano plot of all genes. A volcano plot of the differentially expressed genes between esophageal squamous cell carcinoma and normal samples is shown. The red dots on the left indicate upregulated genes, the red dots on the left and right and black dots indicate upregulated, downregulated and unregulated genes, respectively. FC, fold-change.

Figure 3.

GO term enrichment analysis. (A) Top five clusters of DEGs in BPs (red). (B) Top five clusters of DEGs in CCs (blue). (C) Top five clusters of DEGs in MFs (orange). These groups are ranked and presented as bar plots according to their Fisher's exact P-value. DEGs, differentially expressed genes; GO, Gene Ontology; BP, biological process; CC, cellular component; MF, molecular function.

Figure 4.

Top five modules from the protein-protein interaction network. (A) Module 1, (B) enriched pathways of module 1; (C) module 2, (D) enriched pathways of module 2; (E) module 3, (F) enriched pathways of module 3; (G) module 4, (H) enriched pathways of module 4; (I) module 5, (J) enriched pathways of module 5. Green and red nodes represent downregulated and upregulated genes, respectively.

Figure 5.

Box plots of ESCC and normal tissue expression levels in The Cancer Genome Atlas database. Box plots indicate the expression of different genes in tumor and normal tissues. The expression levels of (A) PTTG1 and (B) EFNA1 in ESCC tissues were higher than those in normal tissues, although there was no statistical significance. (C) COL4A1, (D) ACOX2, (E) CXCR2, (F) ADRB2, (G) CHEK1 and (H) BUB1B exhibited statistically significant differences. *P<0.05 between ESCC and normal tissues. ESCC tissues are shown in red, normal tissues are shown in black. Y-axis: |log2FC| cut-off. X-axis: num (T)=182, num (N)=286. ESCC, esophageal squamous cell carcinoma; FC, fold-change; T, tumor; N, normal tissues.

Figure 6.

Prognostic value of eight hub genes in GSE38129. Relationships between gene expression and OS in 182 patient prognostic values in the GSE38129 of (A) PTTG1, (B) EFNA1, (C) COL4A1, (D) ACOX2, (E) CXCR2, (F) ADRB2, (G) CHEK1 and (H) BUB1B were obtained from Gene Expression Profiling Interactive Analysis. The results showed that the patients with high expression levels of EFNA1 (P=0.0026) and COL4A1 (P=0.037) had a poorer prognosis than those with low expression levels of EFNA1 and COL4A1. *P<0.05. OS, overall survival; HR, hazard ratio; TPM, number of transcripts per million reads.

Figure 7.

Protein expression of PTTG1 in ESCC and normal tissues. (A) Expression of PTTG1 in ESCC, cytoplasmic staining (magnification, ×100, ×200, ×400). (B) Expression of PTTG1 in normal tissues, cytoplasmic staining (magnification, ×100, ×200, ×400). (C) Expression of PTTG1 in ESCC tissues (n=36) was significantly higher than that in normal tissues (n=35), and the difference was statistically significant (P=0.002). (D) PTTG1 IHC staining score in ESCC and normal tissues using independent-samples t-test analysis; each bar represents the mean ± SD. ***P<0.001. ESCC, esophageal squamous cell carcinoma; IHC, immunohistochemistry.