Bioinformatics Analysis of Hub Genes of Diabetic Foot Ulcer and Their Biofunctions

  Objective  To explore the hub genes associated with the pathogenesis and healing of diabetic foot ulcer (DFU) and their biological functions through bioinformatics analysis of transcriptome sequencing data.

  Methods  The transcriptome sequencing datasets of DFU were selected from Gene Expression Omnibus (GEO) database, and the data were regrouped and normalized for bioinformatics analysis. The skin transcriptome sequencing datasets of DFU patients were compared with those of normal controls and the transcriptome sequencing datasets of skin from ulcerous wound edge of DFU patients were compared with those of non-ulcerous skin of DFU patients so that differentially expressed genes were identified, pathway enrichment and protein-to-protein interaction (PPI) analyses were performed, hub genes were found through nodal analysis, and receiver operating characteristic (ROC) curve was applied to a testing dataset to validate the diagnostic efficiency of the hub genes related to DFU. The intersecting genes from the two sets of analyses were again subjected to pathway enrichment and PPI analyses to screen for hub genes associated with DFU wound healing. What’s more, gene set enrichment analysis (GSEA) was carried out on relevant samples to probe for the possible functions and pathway of non-significant genes in DFU.

  Results  A total of 620 up-regulated differentially expressed genes and 196 down-regulated differentially expressed genes were identified in the training dataset which compared DFU patients with non-diabetic patients. The functions of these genes were enriched in the metabolism of terpenoids and polyketides, signaling molecules and interaction, phospholipase D signaling pathway, propanoate metabolism, PI3K-Akt signaling pathway, Toll-like receptor signaling pathway, pyrimidine metabolism, IL-17 signaling pathway, Rap1 signaling pathway, etc. A total of 10 hub genes were identified with the PPI network. Among them, BGN’s value of the area under the curve of ROC analysis was 0.714 and CCND1’s was 0.712. In the sequencing analysis of ulcerous wound edge of DFU patients and non-ulcerous skin of DFU patients, 4072 up-regulated genes and 911 down-regulated genes were identified, of which, 372 genes were also detected in the differentially expressed genes of DFU. The functions of these differentially expressed genes were enriched in phospholipase D signaling pathway, xenobiotics biodegradation and energy metabolism, glutathione metabolism, pyrimidine metabolism, ErbB signaling pathway, melanin production, etc. A total of 7 hub genes were identified from PPI network. In GSEA analysis, pathways including pentose and glucuronate interconversions and homologous recombination, nicotinate and nicotinamide metabolism, neuroactive ligand receptor interaction, maturity-onset diabetes of the young, butanoate metabolism, lysine degradation, pantothenate and coenzyme A biosynthesis, riboflavin metabolism, steroid hormone biosynthesis, and valine, leucine and isoleucine degradation showed significant expression differences between DFU patients and normal controls.

  Conclusion  Bioinformatics analysis results suggest that BGN and CCND1 are potential biomarkers for predicting DFU; CXCL12, TLR4, JAK2, PPARA, UBC, DCN, KDR, and ARNTL are the hub genes of DFU, while CXCL8, CXCL12, TXN, SLIT3, KRT14, KIT, and NEO1 are the hub genes related to wound healing of DFU.