Identification of DEGs
By analyzing the GEO2R online tool in the GEO database, we obtained 756 DEGs from the GSE25070 dataset, including 254 upregulated genes and 502 downregulated genes (Table 1), and visualized the results of the DEGs as a volcano plot (Fig. 1a). Heatmaps of DEGs was plotted using Heml 1.0.3.7 software (Fig. 1b).
Functional Enrichment Analysis Of Degs
We used the DAVID database to obtain GO functional and KEGG pathway enrichment analyses of DEGs. The GO analysis revealed that 254 upregulated genes were primarily involved in inflammatory response, extracellular matrix organization, and neutrophil chemotaxis biological processes (Fig. 2a, Table 2), whereas 502 downregulated genes were primarily involved in cadmium ion response, retinol metabolic process, and cellular zinc ion homeostasis (Fig. 2b, Table 2). For cellular fractions, upregulated genes were mainly enriched in the extracellular space, extracellular region, and extracellular matrix (Fig. 2a, Table 2), while downregulated genes were enriched in the extracellular exosome, apical plasma membrane, and extracellular space (Fig. 2b, Table 2). In terms of molecular function, the results showed that upregulated genes were mainly enriched in extracellular matrix structural constituent, chemokine activity, and extracellular matrix structural constituent conferring tensile strength (Fig. 2a, Table 2), while the downregulated genes were mainly enriched in structural constituent of muscle, steroid dehydrogenase activity, and retinol dehydrogenase activity (Fig. 2b, Table 2).
KEGG pathway enrichment analysis showed that upregulated genes were mainly involved in Rheumatoid arthritis, IL-17 signaling pathway, and Cytokine-cytokine receptor interaction (Fig. 2c, Table 3), while downregulated genes were mainly involved in Metabolic pathways, Drug metabolism - cytochrome P450, and Bile secretion (Fig. 2d, Table 3).
Analysis Of Ppi Networks And Hub Genes
Using the STRING database, we obtained PPI networks for 254 upregulated genes and 502 downregulated genes. The PPI network for the upregulated gene contains 203 nodes and 1258 edges, and the PPI network for the downregulated gene includes 400 nodes and 1146 edges. We visualized the results using Cytoscape 3.9.0 software (Fig. 3a-b). We also screened 5 hub genes each for upregulated and downregulated genes using the cytoHubba plugin (Fig. 3c-d), including VEGFA, IL1B, MMP9, CXCL8, CCND1, MAPK3, ADH1A, SLC26A3, ADH1C, and UGT1A8. Specific information about the 10 hub genes is shown in Table 4.
We further performed GO and KEGG enrichment analysis on 10 hub genes using WebGestalt online tool. GO enrichment analysis showed that the 10 hub genes were mainly enriched in cytokine-mediated signaling pathway, leukocyte migration, myeloid leukocyte migration, and positive regulation of cell migration for BP (Fig. 4a, Table 5); and secretory granule, secretory vesicle, and pseudopodium for CC (Fig. 4b, Table 5); and alcohol dehydrogenase activity, zinc-dependent, alcohol dehydrogenase (NAD) activity, retinol dehydrogenase activity, and molecular function regulator for MF (Fig. 4c, Table 5). KEGG pathway analysis revealed 10 hub genes mainly involved in bladder cancer, AGE-RAGE signaling pathway in diabetic complications, IL-17 signaling pathway, and human cytomegalovirus infection (Fig. 4d, Table 6).
Expression And Survival Analysis Of Hub Genes
We performed an analysis using the GEPIA 2 database to further compare the expression of the 10 hub genes between CRC patients and normal patients. The findings revealed that 5 upregulated genes (VEGFA, IL1B, MMP9, CXCL8, and CCND1) were significantly more expressed in colorectal cancer than in normal controls, while 5 downregulated genes (MAPK3, ADH1A, SLC26A3, ADH1C, and UGT1A8) were the inverse (Fig. 5a).
Next we performed survival analysis for the 5 upregulated genes, including overall survival and disease-free survival (Fig. 5b-c). The results showed that IL1B and CXCL8 genes were significantly associated with OS in colorectal cancer patients (p < 0.05), and Patients with high expression of IL1B and CXCL8 had a greater survival advantage (Fig. 5b), so we tentatively speculate that IL1B and CXCL8 are potential biomarkers of CRC.