Screening of DEGs involved in diabetic tubulointerstial injury. To identify DEGs related to diabetic tubulointerstial injury, the mRNA expression microarray (GSE30122) was downloaded from GEO. After normalization of the raw microarray data (Figure 1a), 166 DEGs associated with diabetic tubulointerstial lesions were identified using limma package as shown in the volcano plot (Figure 1b). Among them, 159 genes were upregulated and 7 genes were downregulated.
Identification of transcription factor genes and regulatory network construction of top 10 transcription factors. To determine transcription factor genes related to diabetic tubulointerstial injury, UCSC_TFBS online tool on DAVID was employed to identify transcription factor genes that regulate DEGs. As shown in Figure 2a, a total of 38 transcription factor genes were indicated to be involved in diabetic tubulointerstial injury. Top 10 transcription factors and their target DEGs were applied to create the regulatory network via Cytoscape software. The regulatory network consisted of 500 interactions between 10 transcription factors and 116 DEGs (Figure 2b).
GO enrichment analysis of targeted DEGs. To investigate biological roles of DEGs moduled by 38 transcription factors, GO enrichment analysis was conducted via DAVID. Myelination (P=5.43E-05), glomerulus development (P=5.64E-05), lung development (P=6.01E-05), extracellular matrix organization (P=6.74E-05) and heart development (P=2.41E-04) were the top5 significant enrichment of biological process (Figure 3a). Extracellular space (P=3.04E-08), extracellular exosome (P=6.00E-06), cell surface (P=1.71E-05), slit diaphragm (P=3.34E-05) and proteinaceous extracellular matrix (P=1.07E-04) were the top5 significant enrichment of cell component (Figure 3a). Glycosaminoglycan binding (P=5.39E-04), heparin binding (P=6.42E-04), extracellular matrix binding (P=0.0016), phospholipase inhibitor activity (P=0.0043) and tropomyosin binding (P=0.0069) were the top5 significant enrichment of molecular function (Figure 3a).
KEGG pathway analysis of targeted DEGs. To explore signaling pathways of DEGs moduled by identified transcription factors, KEGG pathway analysis was performed via DAVID. Figure 3b showed that these DEGs were primarily enriched in complement and coagulation cascades (P=5.60E-04), tight junction (P=0.0105) and cell adhesion molecules (CAMs) (P=0.0129).
The mRNA expression pattern of transcription factor genes in diabetic renal tubulointerstium. To find out the mRNA expression pattern of selected transcription factor genes, relevant analysis was performed by Nephroseq v5 online platform. The results demonstrated that the mRNA expression of CDC5, CEBPA, FAC1, HFH1, IRF1, NFE2 and TGIF1 increased in renal tubulointerstium of DN patients compared with normal controls while that of CEBPB and FOXO4 decreased in renal tubulointerstium of DN patients compared with normal controls (Figure 4).
Association between mRNA expression of transcription factor genes in renal tubulointerstium and clinical features of DN. To explore clinical significance of identified transcription factors in DN, correlation analysis between transcription factor genes and clinical features of DN was conducted by Nephroseq v5 online tool. Firstly, the results showed that mRNA expression of AP1, BACH1, CDC5, FAC1, FOXJ2, IRF1, POU3F2, SOX5, SOX9 and TGIF1 in renal tubulointerstium reversely correlated with GFR in DN patients (Figure 5), suggesting that those transcription factor genes may contribute to the progression of DN. Meanwhile, the mRNA expression of FOXO1 and FOXO4 in renal tubulointerstium positively correlated with GFR in DN patients (Figure 5), indicating that the two transcription factor genes may play a renoprotective role in DN. Secondly, the mRNA expression of AP1, BACH2, FOXD1, FOXJ2 and IRF1 in renal tubulointerstium positively correlated with SCR in DN patients (Figure 6), suggesting that those transcription factor genes may promote the progression of DN. Besides, the mRNA expression of FOXO4, RSRFC4 and S8 in renal tubulointerstium negatively correlated with SCR in DN patients (Figure 6), indicating that the three transcription factor genes may have renoprotective roles in DN. Thirdly, the mRNA expression of CDC5 in renal tubulointerstium negatively correlated with proteinuria in DN patients (Figure 7a). Besides, the mRNA expression of CDC5 and FOXO4 in renal tubulointerstium negatively correlated with weight of DN patients (Figures 7b-c). Moreover, the mRNA expression of HFH1 in renal tubulointerstium positively correlated with body mass index in DN patients (Figure 7d).