Identifying C1QB, ITGAM, and ITGB2 as potential diagnostic candidate genes for diabetic nephropathy using bioinformatics analysis

Background: Diabetic nephropathy (DN), the most intractable complication in diabetes patients, can lead to proteinuria and progressive reduction of glomerular filtration rate (GFR), which seriously affects the quality of life of patients and is associated with high mortality. However, the lack of accurate key candidate genes makes diagnosis of DN very difficult. This study aimed to identify new potential candidate genes for DN using bioinformatics, and elucidated the mechanism of DN at the cellular transcriptional level.

Methods: The microarray dataset GSE30529 was downloaded from the Gene Expression Omnibus Database (GEO), and the differentially expressed genes (DEGs) were screened by R software. We used Gene Ontology (GO), gene set enrichment analysis (GSEA), and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis to identify the signal pathways and genes. Protein-protein interaction (PPI) networks were constructed using the STRING database. The GSE30122 dataset was selected as the validation set. Receiver operating characteristic (ROC) curves were applied to evaluate the predictive value of genes. An area under curve (AUC) greater than 0.85 was considered to be of high diagnostic value. Several online databases were used to predict miRNAs and transcription factors (TFs) capable of binding hub genes. Cytoscape was used for constructing a miRNA-mRNA-TF network. The online database 'nephroseq' predicted the correlation between genes and kidney function. The serum level of creatinine, BUN, and albumin, and the urinary protein/creatinine ratio of the DN rat model were detected. The expression of hub genes was further verified through qPCR. Data were analyzed statistically using Student's t-test by the 'ggpubr' package.

Results: A total of 463 DEGs were identified from GSE30529. According to enrichment analysis, DEGs were mainly enriched in the immune response, coagulation cascades, and cytokine signaling pathways. Twenty hub genes with the highest connectivity and several gene cluster modules were ensured using Cytoscape. Five high diagnostic hub genes were selected and verified by GSE30122. The MiRNA-mRNA-TF network suggested a potential RNA regulatory relationship. Hub gene expression was positively correlated with kidney injury. The level of serum creatinine and BUN in the DN group was higher than in the control group (unpaired t test, t = 3.391, df = 4, p = 0.0275, r = 0.861). Meanwhile, the DN group had a higher urinary protein/creatinine ratio (unpaired t test, t = 17.23, df = 16, p < 0.001, r = 0.974). QPCR results showed that the potential candidate genes for DN diagnosis included C1QB, ITGAM, and ITGB2.

Conclusions: We identified C1QB, ITGAM and ITGB2 as potential candidate genes for DN diagnosis and therapy and provided insight into the mechanisms of DN development at transcriptome level. We further completed the construction of miRNA-mRNA-TF network to propose potential RNA regulatory pathways adjusting disease progression in DN.