Crucial genes associated with diabetic nephropathy explored by microarray analysis

Diabetic nephropathy (DN) is a complication correlated with both type 1 and type 2 diabetes and is characterized by glomerulosclerosis due to accumulation of extracellular matrix [1]. Despite great attention from both clinicians and basic scientists, the morbidity of end-stage renal disease and glomerulosclerosis in diabetic patients is increasing dramatically [2].

In the past years, many molecules associated with DN have been uncovered. For example, the mammalian target of rapamycin (mTOR) complex 1 activation has a key role in podocyte dysfunction in diabetic mice [3, 4]. Rooney et al. have shown that connective tissue growth factor/CCN family protein 2 (CTGF/CCN2) can activate the Wnt signaling in mesangial cells through low density lipoprotein receptor-related protein 6 (LRP6), which may be implicated in the pathogenesis of DN [5]. In diabetic mice, Glo1 (Glyoxalase 1) overexpression completely inhibits diabetes-induced increases in methylglyoxal modification of glomerular proteins, and promotes the development of diabetic kidney disease (DKD) [6]. Furthermore, there is evidence that miR-192 enhances collagen expression by regulating the E-box repressors Zeb1/2, and locked nucleic acid–anti-miR-192 alleviates proteinuria in the diabetic mice [7]. Based on a gene expression profiling of DN, Hans et al. have shown that some genes in glomeruli from patients with DN are down-regulated, such as bone morphogenetic protein 2, fibroblast growth factor 1, vascular endothelial growth factor, nephrin and insulin-like growth factor binding protein 2, suggesting that progression of DN might be due to diminished tissue repair ability [8]. However, the pathways which the differentially expressed genes (DEGs) participate in and regulators that target these genes remain unknown.

In this study, the microarray dataset GSE1009 deposited by Hans [8] was used to identify DEGs between diabetic glomeruli samples and healthy controls. Gene Ontology (GO) and pathway enrichment analyses were then performed for the up- and down-regulated DEGs. Furthermore, microRNAs (miRNAs) and transcription factors (TFs) regulating DEGs were predicted, and miRNA-DEG-TF regulatory network was constructed. These findings may contribute to a better understanding of the nosogenesis of DN.

In the present study, a set of 444 DEGs in the dataset GSE1009 were identified from diabetic glomeruli samples, including 14 up-regulated ones and 430 down-regulated ones, compared with healthy glomeruli samples. Among them, the expression of 143 DEGs (one upregulated gene and 142 downregulated genes) were validated by another dataset GSE30528. According to the GO functional enrichment analysis, a set of DEGs were related to the function of cytoskeleton organization, such as MTSS1, ACTN4 and CALD1.

MTSS1 encodes metastasis suppressor 1, which is also known as MIM (missing in metastasis). It is an actin and membrane binding protein that was originally identified as a potential tumor metastasis suppressor [16]. MTSS1 can induce actin-rich protrusions at the plasmalemma and promote disintegration of actin stress fibers [17], indicating that it may be crucial in regulating cytoskeletal dynamics. Renal tubular cell and podocyte apoptosis is an inevitable event in the progression of glomerulosclerosis [18, 19], and major modifications of the cytoskeleton are involved in the apoptosis progress, including dynamic membrane blebbing, nuclear disintegration, chromatin condensation and cell fragmentation [20]. Therefore, MTSS1 may play a pivotal role in DN, via participation in the regulation of cytoskeleton organization. ACTN4 encodes an actinin, which participates in cytoskeleton action. Previous studies have been reported that mutations in ACTN4 cause focal segmental glomerulosclerosis [21‐23]. There is evidence that two single-nucleotide polymorphisms in ACTN4 are associated with DN in women [24]. Moreover, the up-regulation of ACTN4 was observed during the proteinuria stage [25]. Thereby, ACTN4 may be also important in the progression of DN. CALD1 encodes a caldesmon that plays a key role in the regulation of smooth muscle and nonmuscle contraction [26]. A previous study has shown that CALD1 is over-expressed in fibroblasts from the diabetic patients with nephropathy, compared with those from the controls [27]. Conway et al. have demonstrated that caldesmon is a candidate susceptibility gene for DN [28]. In this study, CALD1 was found to be regulated by several TFs, such as LEF1. LEF1 (lymphoid enhancer-binding factor 1) is a nuclear transcription factor modulated by Wnt, and it expedites epithelial to mesenchymal transition (EMT) when its activity is activated by β-catenin [29]. Rooney et al. have reported that overexpression of CCN2 during the progression of DN likely results in the activation of Wnt signaling and subsequent initiation of TCF/LEF transcription [5]. Collectively, CALD1 may play an essential role in the development of DN likely through the regulation of LEF1.

Furthermore, a series of downregulated DEGs were discovered to be enriched in the pathways of arrhythmogenic right ventricular cardiomyopathy and hypertrophic cardiomyopathy in our study, such as ITGB5. Diabetic cardiomyopathy is a frequent event in diabetic patients due to hyperglycemia [30]. ITGB5 encodes integrin beta 5 [31], and it has been found to be differentially expressed in DKD glomeruli and tubuli [9]. Furthermore, integrins are the primary receptors for intercellular adhesion molecule 1 (ICAM1), which has been reported to have a close relationship with DN [32]. Additionally, in our study, ITGB5 was also regulated by LEF1. These results suggest that ITGB5 may have a significant function in the development of DN, via participating in cardiomyopathy pathways.

Additionally, a set of genes were markedly enriched in complement and coagulation cascades, such as C1S and C1R. These two genes both encode members of the human complement subcomponent C1, which is involved in immune response [33]. Migration of immune cells into the renal is a feature of early DN, and it is implicated in the development of this complication [34‐37]. Previous studies also have found C1S and C1R to be differentially expressed in DN [38, 39]. In our study, C1S and C1R were regulated by LEF1 and hsa-miR-33a. MiR-33a has been reported to be implicated in diabetes due to its regulation of insulin signaling and fatty acid metabolism [40, 41]. Therefore, C1S and C1R may be crucial in the development of DN.

However, this study has some limitations. For example, these predictions were not validated by experiments. The number of samples used for analysis is small. In further studies, more samples will be analyzed, and the predictions will be validated by experimental data.

In conclusion, a total of 14 up-regulated genes and 430 down-regulated ones were identified. Some DEGs related to cytoskeleton organization (e.g. MTSS1, ACTN4 and CALD1), cardiomyopathy (e.g. ITGB5) and immune response (e.g. C1S and C1R), as well as some regulators (e.g. LEF1 and hsa-miR-33a) might play pivotal roles in the progression of DN. These findings may contribute to our better understanding of DN pathogenesis, and provide a theoretical basis for further experimental studies.