DN is the main cause of renal failure worldwide, it is quite clear that glomerulopathy and tubulopathy played a central role in the progression of DN. Recently, diabetic nephropathy has been pay more attention to acquire earlier diagnosis and additional targeted invention . Increasing evidence indicated that autophagy, inflammation, oxidative stress and apoptosis, mitochondrial dysfunction , and abnormal activation of the renin angiotensin system  were closely involved in diabetic nephropathy. Although extensive efforts have been made, the underlying mechanisms of diabetic nephropathy remain elusive. The development of high throughput microarray technology and bioinformatic methods allows us to identify important genes associated with diabetic nephropathy, thus gaining a deeper understanding of its pathogenesis, and to seek molecular diagnostic markers and new therapeutic approaches through relevant important genes.
In this study, a total of 7564 genes were included after data consolidation and normalization. Limma package of R could provide valuable information on large-scale genes with a relatively normal fold change value. We identified 91 DEGs between renal tissues of DN patients and normal controls based on five microarray expression data sets. Among them, 42 differentially expressed genes were up-regulated and 49 were down-regulated. Protein-protein interraction coexpression network showed closely connected genes among DEGs. GO annotation result of DEGs demonstrated that extracellular matrix structural constituent, glycosaminoglycan binding, heparin binding, serine-type peptidase activity, sulfur compound binding, SMAD binding and serine hydrolase activity were chiefly enriched. The GO enrichment annotation results are 23 of the 42 up-regulated genes with differentially expressed genes, such as the extracellular matrix structural constituent, the extracellular matrix structural constituent conference ring tensile strength, etc. No GO enrichment annotation results were found among 49 down-regulated genes with differentially expressed genes. Most of the enrichment results were about molecular function. As an important lesion of DN, tubulointerstitial fibrosis is featured by accelerated deposition of extracellular matrix . Many of extracellular matrix structural constituents have been implicated in the pathogenesis of renal disease  such as COL4A3 , decorin, biglycan and lumican.Enrichment analysis of DEGs by KEGG showed that many of these genes were mapped to protein digestion and absorption, Rap1 signaling pathway, ECM-receptor interaction, platelet activation, PI3K-Akt signaling pathway, complement and coagulation cascade pathway, etc. In 42 up-regulated differentially expressed genes in KEGG enrichment has 5 results. Such as, Protein digestion and absorption, pertussis, Complement and coagulation cascades, ECM -receptor interaction and viral protein interaction with cytokine and cytokine receptor. The pathogenesis of diabetes is due to the abnormal insulin action, and then the long-term blood glucose abnormality leads to the occurrence of glucose kidney. Abnormal insulin is a typical protein digestion and absorption disorder, indicating that protein digestion and absorption must be involved in the process of DN. However, no studies have shown that collagen genes such as COL6A3/COL1A2/COL15A1 are related to protein digestion and absorption. There is also considerable evidence that activated complement systems and coagulant events may contribute to the development and progression of DN, such as complement C3, C1, and C7 [24–26].The expression of C7 gene was significantly up-regulated in the renal tissue of patients with DN ; renal immunohistochemistry detection of C1q and C3c deposition was associated with severe renal damage in DN. Clusterin (CLU), an extracellular chaperone, is a ubiquitously expressed protein that can be identified in various body fluids and tissues. Expression of CLU can lead to various processes including suppression of complement system . Clusterin is involved in apoptosis, inflammation, lipid transportation, cell-to-cell interactions and aging . Additionally, it plays a role in diabetic conditions. Several previous reports showed that clusterin attenuates angiotensin-induced fibrosis and unilateral urethral obstruction (UUO) induced renal fibrosis [29, 30]. However, it has not been shown that CLU gene is related to the complement system related pathways in DN, thus providing us with research thought. The ECM-receptor interaction is a micro-environmental pathway that maintains cell and tissue structure and function. Usually, the ECM consists of a complex mixture of structural and functional macromolecules, which mainly include collagen, fibronectin (FN) and laminin. KEGG enrichment analysis was not found among the 49 down-regulated genes with differential expression.
After combining the two analysis methods, a total of 14 DEGs were distinguished as hub genes, including COL1A2, COL3A1, COL6A3, COL15A1, LUM, CLU, EGF, VEGFA, ALB, SST, FOS, FOSB, EGR1 and EGR2. The upregulation of key genes, including collagen type I α2 chain (COL1A2) and collagen type III α1 chain (COL3A1) are closely related to the severity of renal fibrosis in DN[32, 33]. In other diseases, over-expression of miR-29 has been found to inhibit COL1A1/COL3A1 mRNA and Collagen I/III protein, but not in kidney diseases [34, 35]. The down–regulation of key genes such as Epidermal Growth Factor (EGF) is an active polypeptide used to promote the repair and regeneration of the damaged epidermis . Soluble EGF binds to EGFR which leads to inner kinase activation and phosphorylation of tyrosine residues[37, 38], and then excessive activation of EGFR start multiple intracellular signaling pathways, including extracellular signal conditioning TGF-β1-smad signaling pathways, phosphoinositide 3 kinase (PI3K)/Akt pathway and extracellular regulated protein kinases(ERK) pathway[38, 39], so as to regulate kidney diseases such as obstructive nephropathy, diabetic nephropathy and hypertensive nephropathy.Vascular endothelial growth factor A (VEGFA), an important endogenous angiogenic factor ,which is synthesized and secreted by tubular cells. Also, previous studies identified that vascular endothelial growth factor (VEGF) A is thought to be associated with glomerular endothelial cell dysfunction and tubular cell damage in diabetic nephropathy[41, 42]. Here we found that mRNA expression of VEGFA and EGF was positively correlated with GFR, negatively correlated with SCR, and negatively correlated with VEGFA and proteinuria. This suggests that VEGFA and EGF may be a potential treatment for diabetic nephropathy. Moreover, VEGFA has targeted effects with Mir-29, Mir-199a, Mir-377, Mir-186, Mir-1, Mir-206, Mir-383 and Mir-134. Albumin (ALB) encodes serum albumin, which regulates the plasma colloid osmotic pressure. It is clear that albumin as a basic indicator of renal disease is not specific in clinic. Somatostatin(SST), a natural cyclic peptide inhibitor, which plays the same beneficial role as angiotensin converting enzyme inhibitor(ACEI) in most parameters of diabetic nephropathy. FOS and FOSB are members of the FOS family genes .Under the condition of high glucose, FOS, proto-oncogenes, significantly up-regulated and Jun protein form a dimer with activator protein-1(AP-1) which serve an important role in DN. However, the specific exact mechanism is yet unclear .Over-expression of microrNA-29b can inhibit the activation of hepatic stellate cells by inhibiting the expression of C-FOS mRNA in liver fibrosis . Therefore, whether mir29 can alleviate renal fibrosis by inhibiting FOS expression in diabetic nephropathy? Early growth response proteins (EGRs), as a transcriptional regulatory family, are involved in the process of cell growth, differentiation, apoptosis, and even carcinogenesis. EGR1 and EGR2 are members of the EGRs family genes . It has been proved for the first time that EGR1 is a transcriptional activator of oxidative stress in diabetic nephropathy . Meanwhile, early growth reactive protein-1 (EGR1) plays a key role in DKD by promoting mesangial cell proliferation and the production of extracellular matrix (ECM) .
So far, there is no report on the association between such hub genes and DN as EGR2, FOSB, COL15A1, COL6A3 and LUM. EGR2, which has the most extensive role in the body, also reduces the expression of fibroblast genes in human fibroblast cell lines .Among them, FOSB is an ap-1 transcription factor. In small intestinal neuroendocrine tumors, down-regulated expression of Mir-1 upregulates the expression of mir-1 gene targeted oncogene FOSB, leading to disease progression . At present, there have been no relevant reports about Mir-206 and FOSB. Studies have shown that COL6A3 and COL15A1 may be involved in tubulointerstial injury, but whether they cause glomerulopathy is unknown . And how to participate in the mechanism of tubule injury is not clear. In our study, COL6A3 and COL15A1 were reversely correlated with GFR. In addition, both COL6A3 and COL15A1 were in targeted correlation with the MIR29 family. Therefore, it may be possible to inhibit the expression of COL6A3 and COL15A1 by interfering with the content of MIR29 in vivo, so as to delay the progression of DN and play a role of targeted therapy. Lumican(LUM) is a small leucine-rich proteoglycan (SLRPs) that, like decorin, binds TGF-β in the extracellular matrix to modulate renal fibrosis [53, 54]. In our work, the mRNA expression of LUM in DN patients was negatively correlated with GFR and positively correlated with SCR. Of course, there are many shortcomings of this study. First, although 5 original microarray data sets of diabetic nephropathy were integrated and the total number of samples was sufficient, diabetic nephropathy did not account for the majority of the samples. Second, the lack of detailed demographic information makes it difficult to obtain a more convincing correlation between the selected genes and the severity of diabetic nephropathy using the same samples. Third, due to the lack of online data, the information of selected genes and related clinical indicators is insufficient, resulting in the unclear specific role of some genes. Therefore, further clinical and basic studies are needed in the future to verify our results and clarify the biological role of these genes and related microRNA in diabetic nephropathy.
In summary, the purpose of this study was to identify key molecules associated with diabetic nephropathy and to seek targeted therapies. Through transcriptome data analysis, 91 DEGs, 14 hub genes and 11 micrornas were screened out, among which MIR29 can target COL6A3 and COL15A1 to regulate their expression, thereby delaying the progress of DN. This may provide new ideas for the treatment of DN in the future.