Abstract
OBJECTIVE: To explore biomarkers of diabetic nephropathy (DN) and predict upstream miRNAs. METHODS: The data sets GSE142025 and GSE96804 were obtained from Gene Expression Omnibus database. Subsequently, common differentially expressed genes (DEGs) of renal tissue in DN and control group were identified and protein-protein interaction network (PPI) was constructed. Hub genes were screened from in DEGs and made an investigation on functional enrichment and pathway research. Finally, the target gene was selected for further study. The receiver operating characteristic (ROC) curve was used to evaluate the diagnostic efficiency of target gene and predicted its upstream miRNAs. RESULTS: 130 common DEGs were obtained through analysis, and 10 Hub genes were further identified. The function of Hub genes was mainly related to extracellular matrix (ECM), collagen fibrous tissue, transforming growth factor (TGF) -β, advanced glycosylation end product (AGE) -receptor (RAGE) and so on. Research showed that the expression level of Hub genes in DN group was significantly higher than that in control group. (all P<0.05). The target gene matrix metalloproteinase 2 (MMP2) was selected for further study, and it was found to be related to the fibrosis process and the genes regulating fibrosis. Meanwhile, ROC curve analysis showed that MMP2 had a good predictive value for DN. miRNA prediction suggested that miR-106b-5p and miR-93-5p could regulate the expression of MMP2. CONCLUSION: MMP2 can be used as a biomarker for DN to participate in the pathogenesis of fibrosis, and miR-106b-5p and miR-93-5p may regulate the expression of MMP2 as upstream signals.