PP2 Ameliorates Renal Fibrosis by Regulating the NF-κB/COX-2 and PPARγ/UCP2 Pathway in Diabetic Mice

Renal fibrosis is characterized by glomerulosclerosis and tubulointerstitial fibrosis in diabetic nephropathy (DN). We aimed to evaluate the effects of PP2 on renal fibrosis of DN. GSE33744 and GSE86300 were downloaded from the GEO database. Firstly, 839 DEGs were identified between nondiabetic and diabetic mice renal glomerular samples. COX-2 was selected to assess the effects of PP2 on renal glomerulosclerosis. In db/db mice, PP2 decreased the expression of COX-2, phosphorylated p65, and fibrotic proteins, accompanied with attenuated renal glomerulosclerosis. In cultured glomerular mesangial cells, high glucose- (HG-) induced p65 phosphorylation and COX-2 expression were attenuated by PP2 or NF-κB inhibitor PDTC. PP2, PDTC, or COX-2 inhibitor NS-398 ameliorated abnormal proliferation and expression of fibrotic proteins induced by HG. Secondly, 238 DEGs were identified between nondiabetic and diabetic mice renal cortex samples. UCP2 was selected to assess the effects of PP2 on renal tubulointerstitial fibrosis. In db/db mice, PP2 decreased the expression of PPARγ and UCP2, accompanied with attenuated renal tubulointerstitial fibrosis and EMT. In cultured proximal tubular cells, HG-induced PPARγ and UCP2 expression was inhibited by PP2 or PPARγ antagonist GW9662. PP2, GW9662, or UCP2 shRNA ameliorated HG-induced EMT. These results indicated that PP2 ameliorated renal fibrosis in diabetic mice.


Introduction
Diabetic nephropathy (DN), a major microvascular complication of diabetes mellitus, is the leading cause of end-stage renal disease. The pathogenesis and development of DN are extremely complicated. Clinically, it is initially characterized by an increase in urinary albumin excretion, followed by a gradual decline in glomerular filtration rate, eventually progressing to end-stage renal disease [1]. Morphologically, it is featured by thickening glomerular basement membrane and renal tubular basement membrane, widening mesangial matrix, glomerular sclerosis, podocyte loss, tubular atrophy and increased apoptosis, renal inflammatory infiltration, and renal interstitial fibrosis. In the meantime, periosteal capillaries become sparse, and the wall of efferent arteriole and afferent arteriole is hyaline degeneration, especially the efferent arteriole [2]. DN affects approximately one third of patients with either type I or type II diabetes, contributing to increased mortality and economic burden [3]. Therefore, it is extremely important to understand the molecular mechanisms of DN and identify novel therapeutic targets.
Renal fibrosis is the final common pathway for a variety of chronic kidney diseases, including DN. Chronic progressive renal fibrosis leads to loss of renal function and, eventually, end-stage renal disease. Generally, renal fibrosis is a complex process and refers to glomerulosclerosis and tubulointerstitial fibrosis, especially the increase of matrix protein synthesis and the inhibition of matrix degradation contributing to abnormal accumulation of extracellular matrix (ECM, e.g., collagen fibers and fibronectin) [4]. Many cells in the kidney have been reported to participate in this pathogenesis, especially renal mesangial cells and tubular epithelial cells. On the one hand, abnormal and excessive growth of mesangial cells plays a prominent role in the pathophysiologic processes of glomerulosclerosis in DN [5]. Under the diabetic conditions, mesangial cells are activated, leading to cell proliferation, glomerular hypertrophy, and excess renal ECM deposition in the glomerular mesangial region. Moreover, mesangial cells also secrete various inflammatory cytokines, adhesion molecules, chemokines, and enzymes, all of which contribute significantly to the process of renal glomerulosclerosis and reduced kidney function eventually [6]. On the other hand, tubular injury is another crucial determinant of progressive renal injury in DN, while epithelial-to-mesenchymal transition (EMT) of tubular cells contributes to tubulointerstitial fibrosis and glomerular tubular dissociation [7]. EMT, characterized by loss of epithelial phenotype and gain of profibrotic features that are characteristic of mesenchymal cells, enables cell cycle arrest of renal tubular epithelial cells, elevated migration capacity, resistance to apoptosis, the ability to produce inflammatory cytokines, and even direct production of ECM [8]. Nevertheless, the exact role of mesangial cells and tubular epithelial cells on the renal fibrosis and pathogenesis of DN has not been completely understood yet.
PP2 [4-amino-5-(4-chlorophenyl)-(t-butyl)pyrazolo [3,4-d]pyrimidine] is a selective Src family kinase inhibitor. c-Src is a member of the Src tyrosine kinase family that is involved in many cellular events such as mitosis, cell growth, and tumorigenesis [9,10]. In response to a variety of external stimuli, c-Src can be activated (phosphorylated at Tyrosine 416) and subsequently mediates intracellular signal transduction by phosphorylating tyrosine residues of numerous cellular cytosolic, nuclear, and membrane proteins [2]. We and other investigators have verified the critical role of c-Src in pathological conditions including DN. In glomerular mesangial cells, astaxanthin treatment or overexpression of Cx43 attenuated the high glucose-(HG-) induced interaction between Nrf2 and c-Src, thereby enhancing the Nrf2 accumulation in the nuclei [11,12]. In podocytes, c-Src mediated the translocation of diacylglycerol kinase α and contributed to the amelioration of DN upon stimulation by epigallocatechin gallate [13]. In addition, pretreatment of the differentiated podocytes with the c-Src kinase inhibitor, PP2, not only inhibited phosphorylation of c-Src but also restricted phosphorylation of EGFR and the activation of TGFβ-Smad2/3 signaling pathway in response to HG [14]. In tubular epithelial cells, we have previously demonstrated that PP2 reduces c-Src activity and therefore inhibits cell apoptosis, which at least in part, is due to suppressed p38 MAPK activation [15]. However, the exact role of c-Src on renal fibrosis is yet to be elucidated.
In the present study, we aimed to investigate the role of c-Src on proliferation and fibrosis of mesangial cells as well as EMT of tubular epithelial cells in DN and identify its novel downstream genes. Therefore, we performed two separate studies to address these items. In the first study, by performing bioinformatics analysis based on Gene Expression Omnibus (GEO) data, we identified the differentially expressed genes (DEGs) between nondiabetic and diabetic renal glomerular samples. We identified COX-2 as the common hub gene of DM. Then, we performed a serial experiment to investigate the correlation between c-Src activation and COX-2 expression and evaluated the effects of c-Src/COX-2 pathway on proliferation and fibrosis of mesangial cells under in vivo and in vitro diabetic conditions. In the second study, by performing bioinformatic analysis based on GEO data, we identified the DEGs between nondiabetic and diabetic mice renal kidney cortex samples. We identified UCP2 as the common hub gene of DN. Then, we performed a serial experiment to investigate the correlation between c-Src activation and UCP2 expression and evaluated the effects of c-Src/UCP2 pathway on EMT of tubular epithelial cells under in vitro and in vivo diabetic conditions. The present study may provide an effective therapeutic target for DN.

Bioinformatic Analysis.
Microarray gene expression profiles GSE33744 [16] and GSE86300 [17] were used to perform bioinformatic analysis, respectively. Probe IDs of two datasets were matched to gene symbol on the platform. Probe IDs matching to none or multiple genes symbol were deleted. If multiple IDs match a single gene, the highest expression value of these probes would be set for this gene. The limma package of R software program (version 4.0.3, https://www.r-project.org/) was used to normalize and screen for DEGs in each dataset, respectively. Adjusted p value <0.05 and |log FC | ≥1 were selected as the cut-off criteria of DEGs. The upregulated and downregulated DEGs were saved for subsequent analysis.
In order to explore the biological function and pathway of the above DEGs, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were performed with the Database for Annotation, Visualization, and Integrated Discovery (DAVID v6.8, https://david.ncifcrf.gov/). The results of enrichment analyses were downloaded for subsequent analysis. p value <0.05 was considered statistically significant.
The STRING database (https://www.string-db.org/) was used to construct protein-protein interaction (PPI) network for the identified DEGs with the restriction of a minimum required interaction score of 0.400. The Cytoscape software (http://www.cytoscape.org/) was used to visualize the PPI network. The CytoHubba plugin in Cytoscape software was employed to identify the top 50 hub genes among DEGs. The Molecular Complex Detection (MCODE) plugin in Cytoscape software was employed to select the significant modules.

Animals and Groups.
The animal experiment was performed as previously described [15]. Briefly, male diabetic db/db mice and nondiabetic db/m mice were purchased from the Model Animal Research Center of Nanjing University. According to the guidelines of National Institute of Health, all experimental animals were housed in the specific pathogen-free conditions with free access to food and water. At eight weeks of age, the animals were randomly divided into three groups: control group (db/m mice, n = 10), diabetes group (db/db mice, n = 10), and PP2 group (db/db mice treated with PP2, n = 10). The PP2 group received 2 mg/kg PP2 dissolved in DMSO every other day by i.p. injection. Mice in the control group and diabetes group received the same amount of normal saline. At the age of 16 weeks, the animals that were sacrificed by exsanguination and the kidneys were harvested for further analysis. Human renal tubule epithelial (HK-2) cells were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in DMEM medium supplemented with 10% fetal bovine serum, 1% streptomycinpenicillin mixture in a 95% air and 5% CO 2 atmosphere at 37°C. The culture medium was replaced every two days. After reaching 70-80% confluence, the cells were washed once with serum-free DMEM medium, and then growtharrested in serum-free DMEM medium for 24 h to synchronize the cell growth. The media was then changed to fresh serum-free media containing NG (5.6 mmol/l) or HG (30 mmol/l) for an additional 48 h. PP2 (10 μM) and GW9662 (20 μM) were added to the culture medium 2 h prior to HG. Stable transfections of HK-2 cells with pIRES2-ZsGreen1-UCP2 plasmid or PLV-UCP2-shRNA plasmid were performed with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions.

Protein Extraction and Western
Blotting. Whole protein extraction was performed according to the method used in the previous research [15]. In detail, after the treatment, renal cortex tissues and cells were washed twice with ice cold phosphate-buffered saline (PBS), lysed in protein lysis buffer with protease and phosphatase inhibitors for 30 min on ice. The cell lysates were centrifuged at 12,000 rpm for 20 min at 4°C. Protein concentration was determined with the BCA protein assay kit with BSA as the standard. Protein samples (40 mg/lane) were separated with SDS-PAGE and transferred to PVDF membranes which were then blocked with 5% (v/v) nonfat milk at 37°C for 2 h. The membranes were subsequently incubated overnight at 4°C with primary antibodies against c-Src, phospho-Src (Tyr416), p65, phospho-p65 (Ser536), COX-2, PPARγ, UCP2, fibronectin, collagen IV, E-cadherin, α-SMA, DRP1, FIS1, and β-actin, respectively. The next morning, the membranes were rinsed with TBST and incubated with secondary antibody for 1 h at room temperature. Imaging was performed with the Odyssey Fc System (LICOR, USA). The intensity of the bands was performed with NIH ImageJ software.
2.6. Histology and Immunohistochemistry. Kidneys were fixed in 4% paraformaldehyde overnight and embedded in paraffin. Paraffin sections were cut at 2 mm and processed for periodic acid Schiff (PAS) and Masson's trichrome staining according to the standard protocol. For immunohistochemical analysis, paraffin sections were cut at 4 mm and deparaffinized with xylene and rehydrated in graded ethanol. Internal peroxidase was inactivated with 3% hydrogen peroxide in 100% methanol for 30 min. Antigen retrieval was performed with microwave oven in 10 mM citrate buffer for 15 min. 10% normal goat serum in PBS was then added to the sections for 30 min at room temperature to block nonspecific antibody binding. The sections were then incubated with primary antibodies for fibronectin, collagen IV, E-cadherin, α-SMA, and COX-2 (diluted 1 : 300) overnight at 4°C. After rinsed in PBS, the sections were then incubated with biotinylated secondary antibody and horseradish peroxidase-conjugated streptavidin. Labeling was visualized 3 Oxidative Medicine and Cellular Longevity with diaminobenzidine to produce a brown color, and sections were counterstained with hematoxylin. For negative controls, the primary antibody was omitted. The area percent of staining was calculated using ImageJ software.

Transmission Electron Microscopy.
The ultramicrostructure of glomerulus was observed with electron microscopy. Kidney tissue was immerged in 2.5% (v/v) glutaraldehyde and 1% (v/v) osmium tetroxide. Then, the kidney tissue was dehydrated in graded alcohol solutions and embedded in Epon 812 resin. Ultrathin sections (70-80 nm) were cut and double-stained with uranyl acetate and lead citrate. Sections in ultrastructure were observed and photographed using Hitachi H7500 transmission electron microscope (Hitachi, Tokyo, Japan).
2.8. Cell Proliferation Assay. Cell proliferation of MMCs was estimated with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium (MTT) assay according to the manufacturer's instructions (Beyotime Biotechnology, Shanghai, China). In brief, MMCs (2000 cells/well) were inoculated into 96-well plates. After adherence, the medium was replaced with serum-free medium for 24 h. The media was then changed to fresh serum-free media containing NG (5.6 mmol/l) or HG (30 mmol/l) for an additional 24 h. PP2 (10 μM), PDTC (10 μM), or NS-398 (20 μM) was added to the culture medium 2 h prior to HG. Next, 10 μl MTT solution (5 mg/ml) was added to each well and incubated for 4 h at 37°C. Then, 100 μl formazan solution was added to each well and incubated for 4 h at 37°C to fully dissolve the crystals. Finally, the absorbance was measured at 570 nm using BioTek's Microplate spectrophotometer (Gene Company Limited, USA). All conditions were carried out in triplicate (n = 6), and the experiments were replicated three times.  2.11. Statistical Analysis. Data are expressed as means ± SD. Statistical analysis was performed by one-way ANOVA. Statistical significance was defined as p < 0:05. Based on the criteria of adjusted p value <0.05 and |log FC | ≥1, a total of 839 DEGs were identified from GSE33744. Among them, 579 were upregulated, and 260 were downregulated in diabetic mice renal glomerular. The DEGs are visualized by the cluster heat map (Figure 1(a)) and the volcano plot ( Figure 1 Table 1. In the upregulated DEGs, the most enriched GO terms of biological processes, cellular components, and molecular functions were associated with signal transduction, membrane, and hydrolase activity, respectively. In the downregulated DEGs, the most enriched GO terms of biological processes, cellular components, and molecular functions were associated with oxidation-reduction process, membrane, and metal ion binding, respectively. KEGG pathway enrichment analysis demonstrated that the DEGs were enriched in 23 pathways (Figure 1(g)). The detailed information on the KEGG enrichment analysis of DEGs are listed in Table 2.

Identification of COX-2 as a DEG between
PPI network of DEGs was constructed with Cytoscape software, and the results are shown in Figure 1(h). Furthermore, 23 functional modules were identified from the PPI network with MCODE plugin of Cytoscape, and the 6 most important modules are shown in Figure 1(i). Afterward, the top 50 hub genes were identified from the PPI network with the CytoHubba plugin of Cytoscape based on their degree value, and the results are shown in Table 3. The top 2 hub genes were epidermal growth factor receptor (EGFR) and COX-2 (also known as prostaglandin-endoperoxide synthase 2, PTGS2). The role of the c-Src/EGFR signaling pathway in DN has been previously reported [18]. COX-2 is an inducible enzyme responsible for prostaglandin synthesis and primarily associated with inflammation. There is also increasing evidence that COX-2 is upregulated in pathological conditions [19]. The GO pathway enrichment showed that COX-2 was related to positive regulation of apoptotic process, inflammatory response, positive regulation of cell proliferation, response to lipopolysaccharide, oxidationreduction process, response to oxidative stress, angiogenesis, negative regulation of cell proliferation, regulation of blood pressure, cellular response to mechanical stimulus, memory, regulation of cell proliferation, maintenance of blood-brain 4 Oxidative Medicine and Cellular Longevity      Oxidative Medicine and Cellular Longevity 7 Oxidative Medicine and Cellular Longevity sensory perception of pain, positive regulation of smooth muscle cell proliferation, cyclooxygenase pathway, positive regulation of fever generation, response to lithium ion, negative regulation of calcium ion transport, prostaglandin metabolic process, membrane, endoplasmic reticulum, intracellular membrane-bounded organelle, neuron projection, organelle membrane, metal ion binding, oxidoreductase activity, heme binding, and prostaglandinendoperoxide synthase activity. Therefore, COX-2 pathway was selected for further investigation.

Inhibition of c-Src Ameliorates ECM Accumulation and
Fibrosis via NF-κB/COX-2 Pathway in Renal Glomerulus of Diabetic Mice. Our previous study has shown that PP2 treatment can attenuate diabetes induced c-Src activation and therefore ameliorates renal tubular epithelial cells apoptosis in db/db mice [15]. To understand whether c-Src inhibition has a protective effect in renal fibrosis, we first explored the role of PP2 treatment on the histopathological changes of kidney. As shown in Figure 2(a), HE staining showed that compared with the nondiabetic mice, glomerular hypertrophy, partial glomerulosclerosis, mesangial cells hyperplasia and hypertrophy, and inflammatory cell infiltration were significantly increased in diabetic mice. These pathological changes were significantly attenuated by PP2 treatment. Meanwhile, the results of PAS staining showed that there was prominent mesangial expansion in diabetic mice compared with the nondiabetic mice, which was ameliorated with PP2 treatment. Masson's trichrome staining and quantitative assessment confirmed that PP2 treatment resulted in a significant reduction in collagen deposition of renal glomerular when compared with the diabetic mice. Likewise, immunohistochemistry staining showed that ECM components, including fibronectin and collagen IV, were markedly increased in db/db mice than that in the db/m mice, whereas PP2 treatment significantly reversed these changes (Figure 2(b)). Western blotting results were consistent with the immunofluorescence observation (Figure 2(c)). These results indicate that c-Src plays an important role in the renal fibrosis and injury of diabetic mice.
As mentioned above, the results of the bioinformatic analysis indicated that COX-2 showed significantly different expression between nondiabetic and diabetic mice renal glomerular,and was implicated in the biological processes of positive regulation of NF-κB import into nucleus. Then, we evaluated the relationship between c-Src inhibition and NF-κB/COX-2 pathway in db/db mice. Western blotting showed that the phosphorylation levels of NF-κB p65 (activation) and expression of COX-2 were markedly increased in db/db mice than that in the db/m mice, whereas PP2 treatment significantly reversed these changes (Figure 2(d)). Likewise, immunohistochemistry staining showed that the COX-2 expression was statistically increased in renal glomeruli of db/db mice than in the db/m mice, whereas PP2 treatment statistically reversed these changes (Figure 2(e)). Collectively, these data indicate that c-Src inhibition can attenuate renal ECM accumulation and fibrosis through regulation of NF-κB activation and COX-2 expression in diabetic mice.

Inhibition of c-Src Ameliorates HG-Induced ECM Accumulation via NF-κB/COX-2 Pathway in MMCs.
To determine whether the activation of c-Src is indispensable for HG-induced NF-κB activity and COX-2 expression, we treated MMCs with the c-Src inhibitor PP2 or NF-κB inhibitor PDTC. As shown in Figure 3(a), compared with NG groups, exposure of MMCs to HG for 48 h significantly increased phosphorylation of c-Src at Tyr416 and NF-κB p65 at Ser536, which were significantly inhibited by PP2 or PDTC, respectively. PP2 and PDTC also markedly reduced the HG-induced COX-2 expression. We further assessed whether the c-Src/NF-κB/COX-2 pathway was involved in HG-induced ECM accumulation in MMCs. As shown in Figure 3(b), compared with NG groups, exposure of MMCs to HG for 48 h significantly increased the expression of fibronectin and collagen IV, which could be significantly    ITGB1, CDKN1A, LAMA1, LPAR1, ADCY1, LPAR4, RASGRP2, PTGS2,  EGFR, GLI3, ADCY5, FGF6, GNG2, BID, RUNX1T1, FZD3, FZD2,  PTCH1, MMP2, PTCH2, VEGFB, BMP2, TRAF4, CCNE1,   NS-398 treatment, respectively. Collectively, inhibition of c-Src/NF-κB/COX-2 pathway significantly inhibited HGinduced ECM accumulation, inflammation, and proliferation in mesangial cells. In addition, as a solvent control for PP2, DMSO had no effect on activation of c-Src (Figure 3(d)). Based on the criteria of adjusted p value <0.05 and |log FC | ≥1, a total of 238 DEGs were identified from GSE33744. Among them, 112 were upregulated, and 126 were downregulated in diabetic mice renal cortex. The DEGs are visualized by the cluster heat map (Figure 4(a)) and the volcano plot (Figure 4(b)). GO functional enrichment analysis of DEGs was categorized based on biological processes, cellular components, and molecular functions. The top 10 enriched GO terms of biological processes, cellular components, and molecular functions for the upregulated and downregulated DEGs were displayed in Figures 4(c)-4(f). The detailed information on the GO analysis of the upregulated and downregulated DEGs are list in Table 4. In the upregulated DEGs, the most enriched GO terms of biological processes, cellular components, and molecular functions were associated with signal transduction, extracellular exosome, and oxidoreductase activity, respectively. In the downregulated DEGs, the most enriched GO terms of biological processes, cellular components, and molecular functions were associated with lipid metabolic process, membrane, and ATP binding, respectively. KEGG pathway enrichment analysis demonstrated that the DEGs were enriched in 12 pathways (Figure 4(g)). The detailed information on the KEGG enrichment analysis of DEGs are listed in Table 5.

Identification of UCP2 as a DEG between
PPI network of DEGs was constructed using Cytoscape software, and the results are shown in Figure 4(h). Furthermore, 7 functional modules were identified from the PPI network using MCODE plugin of Cytoscape, and the 6 most important modules are shown in Figure 4(i). Afterward, the top 50 hub genes were identified from the PPI network with the CytoHubba plugin of Cytoscape based on their degree value, and the results are shown in Table 6. In these hub genes, UCP2 was annotated as an inner mitochondrial membrane protein that belongs to the uncoupling protein family and has recently been demonstrated in the control of mitochondria dysfunction [20]. Considering the key role of mitochondrial dysfunction in DN, we focused our attention to UCP2. The GO pathway enrichment showed that UCP2 was related to the biological processes of negative regulation of apoptotic process, liver regeneration. Therefore, the UCP2 pathway was selected for further investigation.

Inhibition of c-Src Ameliorates EMT and
Tubulointerstitial Fibrosis via PPARγ/UCP2 Pathway in Renal Tubule of Diabetic Mice. Our previous studies have indicated that EMT in renal tubular epithelial cells, characterized by loss of epithelial phenotype and gain of profibrotic features that are characteristic of mesenchymal cells, is directly involved in the progression of tubulointerstitial fibrosis that eventually contributed to renal fibrosis in DN [7,21]. To understand whether c-Src inhibition has a protective effect in renal EMT and tubulointerstitial fibrosis, we first explore the role of PP2 treatment on the histopathological changes of kidney. As shown in Figure 5(a), HE staining showed that compared with the nondiabetic mice, interstitial fibrous tissue proliferation and inflammatory cell infiltration, tubular ectasia, and cell swelling were significantly increased in diabetic mice. These pathological changes were significantly attenuated by PP2 treatment. Masson's trichrome staining and quantitative assessment confirmed that PP2 treatment resulted in a significant reduction in interstitial fibrosis when compared with the diabetic mice. Additionally, compared with the nondiabetic mice, transmission electron microscopy showed nuclear volume reduction (pyknosis), chromatin condensation, and nuclear membrane shrinkage in renal tubular epithelial cell of db/db mice. PP2 treatment could partially reverse these morphological changes of apoptosis in renal tubular epithelial cells. Likewise, immunohistochemistry staining showed that the expression of epithelial marker Ecadherin was markedly decreased, and mesenchymal marker α-SMA was markedly increased in db/db mice than that in the db/m mice, whereas PP2 treatment significantly reversed these changes ( Figure 5(b)). Western blotting results were consistent with the immunofluorescence observation ( Figure 5(c)). These results indicate that c-Src plays an important role in the renal EMT and tubulointerstitial fibrosis of diabetic mice.
As mentioned above, the results of the bioinformatic analysis indicated that UCP2 showed significantly different expression between nondiabetic and diabetic mice renal cortex. Interestingly, PPARγ was previously highlighted as the major regulator of UCP2 under physiological conditions [22,23]. Therefore, we evaluated the relationship between c-Src inhibition and PPARγ/UCP2 pathway in db/db mice. Western blotting showed that the expression of PPARγ and UCP2 was markedly increased in db/db mice than that in the db/m mice, whereas PP2 treatment significantly reversed these changes ( Figure 5(d)). Moreover, western blotting also showed that the expression of DRP1 and FIS1, two mitochondrial fission-related proteins, was markedly increased in db/db mice than that in the db/m mice, whereas PP2 treatment significantly reversed these changes ( Figure 5(e)). Collectively, these data indicate that c-Src inhibition can attenuate renal EMT and tubulointerstitial fibrosis through regulation of PPARγ and UCP2 expression in diabetic mice.

Inhibition of c-Src Ameliorates HG-Induced EMT via
PPARγ/UCP2 Pathway in HK-2 Cells. To determine whether the activation of c-Src is indispensable for HG-induced PPARγ and UCP2 expression, we treated HK-2 cells with the c-Src inhibitor PP2 or PPARγ antagonist GW9662. Compared with NG groups, exposure of HK-2 cells to HG for 48 h significantly increased PPARγ and UCP2 expression, which were significantly inhibited by PP2 or GW9662, respectively (Figure 6(a)). We further assessed whether the c-Src/PPARγ/UCP2 pathway was involved in HG-induced EMT in HK-2 cells. As shown in Figure 6(b), compared with NG groups, exposure of HK-2 cells to HG or UCP2 overexpression plasmid for 48 h significantly decreased the expression of E-cadherin and increased the expression of α-SMA, respectively. Furthermore, HGinduced EMT could be significantly attenuated by transfection with UCP2 shRNA plasmid. Next, to elucidate the mechanisms by which blockage of UCP2 ameliorated HGinduced EMT, we examined the role of UCP2 on the MMP in HK-2 cells by the confocal microscope. As shown in   14 Oxidative Medicine and Cellular Longevity Figure 6(c), HG or UCP2 overexpression plasmid significantly induced decreased MMP. HG-induced MMP loss was rescued by cotreatment with UCP2 shRNA plasmid. We further explored the role of UCP2 on the mitochondrial morphology using Mitotracker. As shown in Figure 6(d), HG or UCP2 overexpression plasmid induced mitochondrial fragmentation, indicating the increased mitochondrial fission and dysfunction. In contrast, UCP2 shRNA plasmid treatment attenuated this trend induced by HG. Collectively, inhibition of c-Src/PPARγ/UCP2 pathway significantly inhibited HG-induced EMT and mitochondrial dysfunction in HK-2 cells.

Discussion
DN is the leading cause of end-stage renal disease and contributes to increased mortality and economic burden worldwide. Multiple factors, such as metabolic abnormalities, inflammation, immunoregulation, and genetic predisposition, are involved in the process of DN, but the exact mechanism is not quite clarified [24]. It is now widely accepted that renal fibrosis is the most vital process for a variety of chronic kidney diseases that eventually turn into end-stage renal disease, including DN. Renal fibrosis, characterized by the abnormal accumulation of ECM, can manifest as glomerulosclerosis in the glomerulus and tubulointerstitial fibrosis mainly in the tubules [25,26]. c-Src, a member of the Src kinase family, is activated by a number of stresses and implicated in the regulation of cellular processes in multiple tissues. c-Src can be activated in the kidneys of diabetic animals, glomerular mesangial cells, podocytes, and tubular epithelial cells exposed to HG [14,15,27]. Indeed, accumulating evidence suggested a strong correlation between c-Src and DN pathogenesis. Our previous study has demonstrated that c-Src activity was increased in kidneys of db/db mice and involved in renal tubular epithelial cells apoptosis in response to hyperglycemia in DN [15]. However, the role and molecular mechanisms of c-Src on renal fibrosis are yet to be elucidated. In this study, we explored the effect of c-Src inhibition on glomerulosclerosis and tubulointerstitial fibrosis in vivo and in vitro, respectively. In vivo, c-Src inhibition improved diabetes-induced glomerular hypertrophy, partial glomerulosclerosis, mesangial cells hyperplasia and hypertrophy, inflammatory cell infiltration, mesangial expansion, collagen deposition, tubular ectasia and cell swelling, and interstitial fibrous tissue proliferation in the  These findings suggest that inhibition of c-Src can ameliorate diabetes-induced renal fibrosis. Bioinformatic analysis is an efficient and reliable tool to identify potential therapeutic targets and diagnostic biomarkers for precision medicine. With the widespread appli-cation of high-throughput sequencing technology and public databases, bioinformatic analysis has been gradually used to study the molecular mechanisms of kidney diseases, such as membrane nephropathy, lupus, IgA nephropathy, and DN [28]. In this study, we performed bioinformatic analysis to screen the key genes and pathways related to DN and elucidated their molecular mechanism based on two GEO datasets. The GSE33744 dataset was used to identify the DEGs between nondiabetic and diabetic mice renal glomerular. A total of 839 DEGs were identified from GSE33744, including 579 upregulated genes and 260 downregulated genes. GO functional enrichment analysis showed that the upregulated   18 Oxidative Medicine and Cellular Longevity genes were mainly enriched in signal transduction, membrane, and hydrolase activity, and the downregulated genes were mainly enriched in oxidation-reduction process, membrane, and metal ion binding. KEGG pathway enrichment analysis showed that the DEGs were mainly enriched in pathways in cancer, PI3K-Akt signaling pathway, neuroactive ligand-receptor interaction, MAPK signaling pathway, and Rap1 signaling pathway. Following PPI network analysis identified 23 functional modules and top 50 hub genes. In these 50 hub genes, the COX-2 pathway was selected for further investigation based on GO and KEGG enrichment analyses. Furthermore, the GSE86300 dataset was used to identify the DEGs between nondiabetic and diabetic mice renal cortex. A total of 238 DEGs were identified from GSE86300, including 112 upregulated genes and 126 downregulated genes. GO functional enrichment analysis showed that the upregulated genes were mainly enriched in signal transduction, extracellular exosome, and oxidoreductase activity, and the downregulated genes were mainly enriched in lipid metabolic process, membrane, and ATP binding. KEGG pathway enrichment analysis showed that the DEGs were mainly enriched in pathways in metabolic pathways, complement and coagulation cascades, Bile secretion, and PPAR signaling pathway. Following PPI network analysis identified 7 functional modules and top 50 hub genes. In these 50 hub genes, the UCP2 pathway was selected for further investigation based on GO and KEGG pathway enrichment analyses. COX-2 is an inducible enzyme that catalyzes the conversion of arachidonic acid into prostaglandins which in turn cause pain, fever, and inflammation responses. Its expression can be induced by a number of inflammatory cytokines, mitogenic factors, and physical stimuli [29]. Previously, numerous studies revealed that COX-2 was increased in podocyte, glomerular mesangial cells, and renal tubular cells under HG condition, which contributed to the pathogenesis of DN [30][31][32][33]. Interestingly, the promoter region of the COX-2 gene contains several putative binding motifs for the binding of NF-κB [34]. NF-κB is a family of inducible transcription factors that mediates signal-induced expression of numerous genes involved in different biological processes, including immune response, inflammation, cellular transformation, cell proliferation, angiogenesis, invasion, and metastasis [35]. The activated and phosphorylated form of NF-κB subunits (p65 and p50) translocates to the nucleus and thereby induces the expression of target genes related to DN [36]. However, almost nothing is known about the role of NF-κB in modulating the COX-2 expression in DN. In this study, we found that the phosphorylation of NF-κB subunit p65 and its downstream target gene COX-2 was significantly elevated in diabetic kidneys. These changes were reversed by treatment of c-Src inhibitor PP2. Meanwhile, c-Src inhibitor PP2 also inhibited HG-induced the phosphorylation of NF-κB p65 and COX-2 expression in MMCs. In addition, NF-κB inhibitor PDTC not only reduced HGinduced COX-2 expression but also dramatically abrogated HG-induced phosphorylation of c-Src. These results reveal c-Src or NF-κB could act as an upstream regulator of COX-2 and a complex crosstalk between NF-κB and c-Src. We then further examined the role of the c-Src/NF-κB/COX-2 pathway on HG-induced ECM accumulation in MMCs. Our data demonstrated that pharmacological inhibition of c-Src with PP2, NF-κB with PDTC, or COX-2 with NS-398 also resulted in the decreased expression of fibronectin and collagen IV. On the mechanism, prostaglandin E2 (PGE2) was reported to be involved in numerous physiological and pathological processes of the kidney via its four receptors [37]. PGE2 receptor (EP1 or EP2) deficiency markedly suppressed the extracellular matrix accumulation and COX-2 expression in mouse mesangial cells induced by TGF-β1 which is a pivotal mediator of renal fibrosis and glomerulosclerosis in diabetic nephropathy [38,39]. Importantly, inhibition of COX-2 via a specific inhibitor celecoxib markedly blocked PGE2 production [40]. These findings suggested not only the important role of COX-2 in the expressions of extracellular matrix proteins in mesangial cells but also a complex crosstalk among COX-2, PGE2 and PGE2 receptors. Collectively, these results suggest that inhibition of the c-Src pathway ameliorated renal ECM accumulation and fibrosis at least in part through modulating the activity of the NF-κB and expression of COX-2.
UCP2 is a mitochondrial inner membrane protein that uncouples the oxidative phosphorylation from ATP synthesis by dissipating the proton gradient generated across the mitochondrial inner membrane, therefore regulating glucose and lipid metabolism, immune cell activation, and ROS formation by mitochondria [41]. Consistently, previous studies have demonstrated that downregulation of UCP2 exacerbated oxidative stress damage as well as diabetes or HGinduced renal tubular epithelial cells apoptosis [42,43]. However, a conflicting report suggested that UCP2 may act as a double-edged sword in renal tubular epithelial cell injury. Increased mitochondrial UCP-2 expression in renal proximal tubular cells increased O 2 consumption and reduced O2 availability which subsequently contributed to diabetes-induced progressive kidney damage [44]. In this study, we found that the expression of UCP2 and its upstream gene PPARγ was significantly elevated in diabetic kidneys. These changes were reversed by treatment with c-Src inhibitor PP2. Meanwhile, PP2 also inhibited HGinduced UCP2 and PPARγ expression in HK-2 cells. In addition, PPARγ antagonist GW9662 also reduced the HG-induced UCP2 expression. We further examined the role of the c-Src/PPARγ/UCP2 pathway on HG-induced    Figure 7: Schematic illustration of the mechanisms responsible for the PP2 effects on renal fibrosis in diabetic mice. PP2 decreased the expression of COX-2, phosphorylated p65, and fibrotic proteins, accompanied with attenuated renal glomerulosclerosis. PP2 also decreased the expression of PPARγ and UCP2, accompanied with attenuated renal tubulointerstitial fibrosis and EMT. 22 Oxidative Medicine and Cellular Longevity