NaoXinTong Capsules inhibit the development of diabetic nephropathy in db/db mice

NaoXinTong Capsule (NXT), a Chinese medicine, is currently used to treat patients with cardiovascular and cerebrovascular diseases. Clinical observations indicate its anti-diabetic functions with unclear mechanisms. Herein, we report the effect of NXT on diabetic nephropathy (DN). Type 2 diabetic db/db mice were treated with NXT for 14 weeks. In the course of treatment, NXT reduced diabetes-increased glucose levels and improved renal functions. At the end of treatment, we found that NXT ameliorated serum lipid profiles and other biochemical parameters. In the kidney, NXT inhibited mesangial matrix expansion, expression of vascular endothelial growth factor A, fibronectin, advanced glycation end product and its receptor. Meanwhile, it reduced the diabetes-induced podocyte injury by increasing WT1 and nephrin expression. In addition, NXT inhibited accumulation of extracellular matrix proteins by increasing MMP2/9 expression through inactivation of TGFβ/Smad pathway and CTGF expression. Mechanically, NXT activated insulin signaling pathway by increasing expression of INSR, IRS and FGF21, phosphorylation of Akt and AMPKα in the liver, INSR phosphorylation in the kidney, and FGF21 and GLUT4 expression in adipose tissue and skeletal muscle. Taken together, our study demonstrates that NXT inhibits DN by ameliorating glucose/lipid metabolism, maintaining tissue structure integrity, and correcting diabetes-induced renal dysfunctions.


NXT ameliorates serum biochemical parameters in diabetic mice.
To determine if NXT can inhibit DN, we treated db/db mice (~6-week old) with NXT orally for 14 weeks. During the treatment, we routinely determined the bodyweight gain in each group. As shown in Fig. 1a, the average bodyweight of wild type mice was increased from ~20 to ~29 g. In db/db control mice, it was increased from ~33 to ~47 g. However, the NXT treatment reduced the rate of bodyweight gain after 3 weeks treatment indicating a moderate inhibition of obesity in db/db mice by NXT. Meanwhile, we determined the fasting blood glucose levels once in each duration of 16 days. Compared with the stable blood glucose levels in wild type mice, Fig. 1b shows that the blood glucose levels in db/db control mice kept increasing from 13.1 ± 2.0 mM at the beginning of experiment (mice were at ~6-week old) to 30.5 ± 2.5 mM (~2.3-fold) at the end of 14 weeks experiment. In contrast, NXT substantially slowed the increase of glucose levels (22.1 ± 1.7 mM, ~1.6-fold) suggesting that NXT administration improves glucose metabolism. At the end of treatment, we analyzed the fasting blood insulin levels. Similarly, compared with wild type mice, the serum insulin levels were increased ~7.5-fold in db/db control mice. However, the elevation was substantially reduced by NXT treatment (~4.0-fold, Fig. 1c).
To determine the effect of NXT on hyperlipidemia in db/db mice, serum lipid profiles were analyzed at the end of the study (upper panel, Table 1). Compared with wild type mice, much higher total cholesterol (T-CHO) levels (~1.93-fold) were observed in db/db control mice. The increased T-CHO levels were due to increased triglyceride (TG)-rich lipoproteins levels, low-density lipoprotein cholesterol (LDL-CHO, ~3.12-fold) and very low-density lipoprotein cholesterol (VLDL-CHO, ~9.7-fold), particularly the VLDL-CHO levels (Table 1). However, NXT decreased T-CHO, LDL-CHO and VLDL-CHO levels. For instance, both LDL-CHO and VLDL-CHO levels in NXT-treated db/db mice were reduced to that of wild type mice (upper panel, Table 1). Taken together, the results in Fig. 1 and Table 1 suggest that NXT ameliorates serum glucose and lipid levels in db/db mice.

NXT inhibits DN by improving glomerular functions in db/db mouse kidneys.
At the end of treatment, we observed abnormal kidneys with un-matched size and severe lipid accumulation (upper left panel, Fig. 2a) in some db/db control mice (3 of total 10 mice). The renal lipid accumulation was further confirmed by Oil Red O staining kidney cross sections (lower left panel, Fig. 2a). Compared with wild type mice, diabetes also reduced the ratio of kidney weight to bodyweight (right panel, Fig. 2a) indicating the induction of kidney atrophy. However, all the db/db mice receiving NXT treatment had normal kidneys without lipid accumulation (left panel, Fig. 2a), and the kidney atrophy was partially corrected (right panel, Fig. 2a).
The renal dysfunction is the consequence of kidney structural abnormalities associated with DN progress, such as glomerular basement membrane (GBM) thickening, mesangial expansion with ECM accumulation, podocyte injury, glomerulosclerosis and tubulointerstitial fibrosis 28 . To determine the effect of NXT on kidney structure, particularly in glomeruli, kidney cross sections were subjected to HE staining followed by quantification of glomerular area. Compared with wild type mice, the glomerular area was clearly increased in db/db control mice. However, NXT inhibited the glomerular hypertrophy (Fig. 2b,c).
Antibodies against VEGF can improve hyperfiltration and albuminuria in the diabetic animal models suggesting that VEGF can be a potential therapeutic target for DN 29 . In this study, the results of immunohistochemical staining and Western blot demonstrate that NXT decreased VEGFA expression (Fig. 2d,e).
To determine if NXT treatment can ameliorate renal function related parameters, we initially analyzed urea nitrogen and creatinine levels in mouse serum. As shown in Table 1 (lower panel), compared with wild type mice, both serum urea nitrogen and creatinine levels were increased ~50% in db/db control mice but the increases were reduced by NXT treatment. Next, we analyzed levels of nitrogen, creatinine, and microalbumin in urine samples. As shown in Table 2 decreased excreted microalbumin in urine by ~73 and ~90% at day 45 and 70 or 96 of treatment, respectively. Consequently, the UAlb/UCr levels were substantially reduced by NXT treatment ( Table 2). The results of renal function related parameter in both serum and urine indicate that NXT improves glomerular filtration functions.
High glucose induces fibronectin assembly which can make contribution to collagen IV accumulation, and facilitates uncontrolled ECM accumulation during the DN development 30 . Compared with wild type mice, periodic acid-Schiff (PAS) staining shows the accumulation of carbohydrate macromolecules and elevation of the glomerulosclerosis scores in db/db control mice. However, both carbohydrate macromolecule accumulation (upper panel, Fig. 3a) and glomerulosclerosis scores (left panel, Fig. 3b) were reduced by NXT. Fibronectin is a major Figure 1. NXT inhibits diabetes-induced hyperglycemia in db/db mice. Male db/db mice (~6-week old) in two groups (10/group) received following treatment for 14 weeks: Control group, mice were fed normal chow; Naoxintong (NXT) group, mice were fed normal chow containing NXT (620 mpk). Wild type mice fed normal chow were used as the normal control. During the treatment, mouse bodyweight (a) and fasting blood glucose levels (b) were determined, at the indicated time points of treatment. At the end of study, fasting serum insulin levels (c) were also determined. * p < 0.05 vs. db/db control group; # p < 0.05 vs. wild type group (n = 10). molecule responsible for accumulation of carbohydrate macromolecules. Correspondingly, fibronectin expression was increased in db/db control mouse kidneys, but the increase was substantially reduced by NXT (lower panel, Fig. 3a; right panel, Fig. 3b). AGE is an important risk factor for DN development. We found that AGE levels in db/db control mouse kidney were increased with the majority in tubules, which is in the line that diabetes-induced AGE accumulation is in a tissue or cell type-dependent manner 31 . Interestingly, NXT blocked AGE accumulation in tubules of db/db c) kidney frozen cross sections were used to detect morphological changes of glomerulus by HE staining (b) followed by quantitation of the glomerular area (c). * p < 0.05 vs. wild type group; # p < 0.05 vs. db/db control group (n ≥ 5); (d,e) VEGFA expression was determined by immunohistochemical staining with kidney frozen cross sections (d), and by Western blot with total cellular proteins extracted from a piece of kidney (e), respectively. *p < 0.05 vs. db/db control group (n ≥ 5). mouse kidney (Fig. 3c). In addition, expression of the receptor for AGE (named as RAGE or AGER) also plays an important role in pathogenesis of DN 32 . Figure 3d shows that NXT significantly decreased diabetes-induced AGER mRNA expression, indicating that NXT can clearly correct the dysfunctions of AGE-RAGE pathway.

Control
Podocytes cover the outer aspect of the GBM and form the final barrier to prevent protein loss. The podocyte injury (e.g., reduction of podocyte number and density per glomerulus) is linked to development of proteinuria and progression of DN in patients. Wilm's tumor 1 (WT1) protein is a marker of podocytes and plays an important role in maintenance of podocyte function 33,34 . The results of immunohistochemical staining and qRT-PCR analysis show that NXT was able to partially restore diabetes-reduced WT1 protein and mRNA expression (Fig. 3e,f). Nephrin is a key slit diaphragm protein and expressed by podocytes. Nephrin can directly affect insulin signaling via modulation of glucose transporters vesicle trafficking at the plasma membrane 35 . Similar to WT1, the decreased nephrin expression in db/db mouse kidney was also partially recovered by NXT (Fig. 3g). NXT inhibits glomerular mesangial expansion, expression of fibronectin, AGEs and AGER, but activates expression of WT1 and nephrin in db/db mouse kidneys. After treatment, kidney cross sections and total RNA were used to complete following assays: (a) accumulation of carbohydrate macromolecules and fibronectin expression were determined by PAS staining and immunofluorescent staining, respectively; (b) the glomerulosclerosis scores were obtained by calculating the percent of sclerosis area in total area of each glomerulus based on the images of PAS staining (left panel) and the method described in the "Methods". *, # p < 0.05 (20 glomeruli in each sample were counted). The density of immunofluorescence (mean fluorescence intensity, MFI) of images for fibronectin expression was quantified using segmentation color-threshold analysis method (right panel); (c,d) expression of AGEs protein or mRNA was determined by immunofluorescent staining (c) or qRT-PCR (d); (e-g) expression of WT1 protein (e) was determined by immunohistochemical staining; expression of WT1 (f) and nephrin (g) mRNA was determined by qRT-PCR. * p < 0.05 vs. wild type group; # p < 0.05 vs. db/db control group (n ≥ 5).
NXT inhibits ECM accumulation in db/db mouse kidney by inactivating TGFβ/Smad pathway and inhibiting CTGF expression. The accumulation of ECM proteins including collagen type I/IV is a main hallmark for DN development 36 . Collagen type IV is a typical collagen of the basement membrane while collagen type I is an important composition of ECM of renal interstitial fibrosis. The results of immunofluorescent  Fig. 4a,b indicate that diabetes-increased collagen levels were substantially reduced by NXT. Matrix metalloprotein 2/9 (MMP2/9) catalyze the degradation of ECM components including collagen type IV 36 . By completing immunofluorescent staining and qRT-PCR, we determined that MMP2/9 expression was decreased in db/db mouse kidneys, but the decrease was alleviated by NXT ( Fig. 4c-f).
NXT activates insulin signaling pathway and improves glucose metabolism. To disclose the mechanisms by which NXT inhibits DN, we investigated the effects of NXT on insulin signaling pathway in mouse liver and other tissues. Compared with wild type mice, expression of insulin receptor (INSR) was reduced in db/db control mouse liver. However, the reduction was clearly recovered by NXT (Fig. 6a). The results of Western blot analysis confirm that induction of INSR by NXT is mainly contributed by increased INSRα (Fig. 6b). Consequently, the reduced insulin receptor substrate 1/2 (IRS1/2) and phosphorylated IRS1 (pi-IRS1) levels in db/db control mouse liver were increased by NXT (Fig. 6c,d). Furthermore, we found that NXT induced expression of both regulatory and catalytic subunit of PI3K, p85 and p110 (Fig. 6e), and consequently the activated PI3K increased both total Akt and phosphorylated Akt (pi-Akt) levels (Fig. 6f).
In mouse kidney, compared with wild type mice, either diabetes or NXT had little effect on INSR expression (upper panel, Fig. 6g; left panel, Fig. 6h). However, the reduced phosphorylated INSR in db/db control mice was restored to normal by NXT treatment (lower panel, Fig. 6g; right panel, Fig. 6h). Therefore, Fig. 6 suggests that NXT activates insulin signaling pathway in db/db mouse liver and kidney.
Activation of glucokinase (GCK) expression can reduce diabetes by enhancing glycogen synthesis and glycolysis. In contrast, phosphoenolpyruvate carboxykinase 1 (PCK1) and glucose-6-phosphatase (G6Pase) have pro-diabetic functions since they are key enzymes for gluconeogenesis. In the liver, NXT had little effect on PCK1 or G6Pase expression indicating that gluconeogenesis is not affected. However, GCK expression was substantially activated (Fig. 7a) suggesting that glycogen synthesis and glycolysis is activated by NXT.
AMPKα can regulate energy metabolism in mammalian cells and be activated by increased ratio of AMP to ATP (AMP/ATP). The activated AMPKα (pi-AMPKα) inhibits fatty acid/cholesterol synthesis and gluconeogenesis, while stimulating fatty acid uptake/oxidation, glucose uptake and mitochondrial biogenesis 37 . Interestingly, we determined that NXT increased both AMPKα and pi-AMPKα levels (Fig. 7b), which suggests that the energy metabolism in db/db mice is enhanced. Correspondingly, in the skeletal muscle, NXT increased glucose transporter 4 (GLUT4) expression (Fig. 7c), a molecule responsible for glucose uptake/energy metabolism in the tissue.
Fibroblast growth factor 21 (FGF21) enhances insulin sensitivity and glucose/energy metabolism 38 . Figure 7d shows that NXT increased FGF21 expression in both liver and white adipose tissue of db/db mice. Taken together, the results above suggest that NXT protects db/db mice against DN by increasing insulin sensitivity and improving glucose and energy metabolism through multiple actions.

Discussion
DN is one of the diabetic complications with a high morbidity and mortality in patients. Although the current therapeutic interventions can delay the onset and progression of DN, the effects on DN mortality are still limited. Therefore, development of alternative therapeutic approaches is urgent. Traditional Chinese medicine, at least as an adjunctive therapy, has been demonstrated various benefits to patients with different types of diseases. Clinically, NXT is prescribed to patients with cardiovascular and cerebrovascular diseases [19][20][21][22][23] . However, the anti-diabetic effects of NXT have been observed in patients with hyperlipidemia and hyperglycemia. In this study, we treated db/db mice with NXT for a long-term and found that NXT clearly inhibited DN development by inhibiting diabetes-induced abnormal kidneys, mesangial expansion, renal accumulation of lipids, AGEs and collagens (Figs 2 and 3). The amelioration of renal functions by NXT is associated with reduction of serum lipid profiles, urea nitrogen and creatinine (Table 1), and excretion of urinary microalbumin ( Table 2). The inhibitory effect of NXT on DN is mainly attributed to amelioration of glucose metabolism through activation of insulin signaling pathway in multiple tissues (Figs 1, 6 and 7). Mechanistically, we determined that NXT inhibited TGFβ/Smad signaling pathway and decreased CTGF expression in the kidney which resulted in restoration of diabetes-inhibited MMP2/9 expression (Figs 4 and 5).
Binding of insulin to INSR can activate IRS1 and consequent PI3K/Akt. In db/db mice, deficiency of leptin receptor expression results in dysfunction of leptin and inactivation of IRS1 39,40 . Therefore, the animals exhibit severe insulin resistance and hyperglycemia from a very young age (~5-week old). In this study, we determined that NXT increased INSR expression (Fig. 6a,b) which was associated with restoration of IRS1 expression/ phosphorylation in db/db mouse liver to that in wild type mice (Fig. 6c). NXT also activated IRS1/2 expression (Fig. 6c,d). Consequently, expression of PI3K molecules (p85 and p110) and Akt expression/phosphorylation were increased by NXT (Fig. 6e,f). Although NXT had no effect on INSR expression, it activated INSR by enhancing its phosphorylation in the kidney (Fig. 6g,h). The activation of insulin signaling pathway by NXT results in induction of GCK expression (Fig. 7a). Moreover, both AMPKα expression/phosphorylation in the liver, and FGF21 expression in liver and adipose tissue were activated by NXT (Fig. 7b,d). Taken together, our results demonstrate that NXT controls glycemia by multiple mechanisms in different tissues, mainly by the activation of insulin signaling pathway. Activation of VEGFA expression and AGE-RAGE pathway as well as their interaction greatly influence DN development. Hyperglycemia induces formation and accumulation of AGE. The consequently activated AGE-RAGE pathway activates reactive oxygen species generation and PKC pathway 41 . Meanwhile, activation of RAGE in podocytes increases VEGFA expression and enhances recruitment/activation of inflammatory cells in diabetic glomeruli which can further accelerate albuminuria and glomerulosclerosis in diabetic kidneys 42 . In animal model, podocyte-specific VEGFA overexpression leads to proteinuria, glomerulomegaly, GBM thickening, mesangial expansion and decreased nephrin 43 . In this study, we observed that administration of NXT inactivated AGE-RAGE pathway, decreased VEGFA protein expression and prevented the podocyte injury by restoring nephrin and WT1 expression in db/db mouse kidneys (Figs 2 and 3).
TGFβ plays an important role in tubule glomerular sclerosis in diabetic kidneys by activating matrix synthesis and inhibiting matrix degradation in glomerular mesangial cells, which results in cell proliferation and ECM expansion 44 . In this study, our results show that NXT regulated TGFβ signaling pathway since it decreased TGFβ1 and TGFβR2 protein expression, and correspondingly enhanced MMP2/9 expression in db/db mouse kidney. Therefore, inhibition of DN by NXT should be attributed to the blockage of ECM accumulation through inhibition of TGFβ signaling and activation of MMP2/9. Associated with hyperinsulinemia and hyperglycemia, cholesterol metabolism in db/db mice is also exacerbated 45 . Similarly, the impaired lipoprotein metabolism, such as increased VLDL-CHO and LDL-CHO, can be observed in diabetic patients. The diabetes associated with dyslipidemia might be an independent risk factor for DN since the dyslipidemia enhances macrophage infiltration and ECM production in glomeruli 46 . Meanwhile, clinical studies indicate the renoprotective effects of lipid-lowering therapy on diabetic patients 47 . In this study, we determined that NXT restored T-CHO and LDL-CHO levels (upper panel, Table 1), suggesting the lipid-lowering effects of NXT may also make contribution to inhibition of DN. Interestingly, we previously reported that NXT has little effect on lipid profiles in apoE deficient mice 48 which suggests that improvement of lipid metabolism by NXT in db/db mice might be completed by different signaling pathways, such as activation of AMPKα to enhance energy metabolism. In the context of that the effects of current interventions on diabetes are still limited, the alternative strategies including Chinese medicine might offer additional benefits to the patients. Several herbal medicines, such as Tangningtongluo formula, Cistanche tubulosa, Swetia punicea Hemsl and tuberous root of Liriope spicat var., have been demonstrated anti-diabetic effects on patients or animal models [49][50][51][52] . In this study, based on the clinical observations, we determined that administration of db/db mice with NXT inhibited the development of DN. Furthermore, we presented the results demonstrating that the inhibitory effects of NXT on DN should be attributed to its multiple anti-diabetic actions including improving glucose and lipid metabolism, activating insulin signaling pathway to reduce accumulation of ECM and AGE, and inactivating TGFβ/Smad signal pathway in the kidney (Fig. 8). Our study suggests an important and potential application of NXT for DN treatment.

Animals.
The in vivo studies with mice were conducted according to the protocol which was granted by the Ethics Committee of Nankai University (Tianjin, China) and conformed to the Guide for the Care and Use of Laboratory Animals published by NIH. Both male type 2 diabetic (BKS.C g-m+/+ Lepr db /J, db/db, 6-week-old, ~33 g average bodyweight) and C57BLKS/J wild type mice (6-week-old, ~20 g average bodyweight) were purchased from the Animal Center of Nanjing University (Nanjing, China). The animals were maintained at the Animal Center of Nankai University with free access to food and drinking water.
Based on the clinical usage, the dose of NXT to mice could be converted to ~620 mg/day/kg body weight (mpk) or 620 mg/100 g food (mice eat food at ~10% of their body weight daily). Male db/db mice at an age of 6-week old were randomly divided into two groups (10 mice/group) and received following treatment: Control group, mice were fed normal chow; NaoXinTong (NXT) group, mice were fed normal chow containing NXT (620 mpk). Male C57BLKS/J wild type mice also at an age of 6-week old were used as a non-diabetic or normal control. The treatment was continued for ~14 weeks. During the treatment, we routinely checked bodyweight, food intake, water drinking and exterior appearance, and did not find difference caused by NXT treatment except that the bodyweight gain was reduced by NXT. At the end of the experiment, all the mice were anesthetized and euthanized by i.p injection of 2,2,2-tribromoethanol (640 mg/kg bodyweight) followed by collection of blood and tissue samples.  Determination of renal functions. At the indicated durations, mice were housed in the metabolic chambers (Nalgene) with free access to food and drinking water for 24 h to collect urine samples. Urinary microalbumin levels were determined with the ELISA assay kit purchased from Elabscience Biotechnology (Wuhan, China). Nitrogen and creatinine levels in urine samples were determined with the assay kits purchased from BioSino Bio-technology and Science Inc. (Beijing, China).

HE, PAS and Oil Red O staining, and determination of glomerulosclerosis scores.
After treatment, the 5 μm cross sections of kidney were prepared and used to conduct the following staining: HE staining for determination of glomerular area; PAS staining for determination of carbohydrate macromolecules; and Oil Red O staining for lipid accumulation 53 .
The images of PAS staining were used to determine the sclerosis area and total area in each glomerulus. The glomerulosclerosis scores were obtained using the method as described 54 with the following modifications: the percent of sclerosis area in total area of each glomerulus was timed a sclerosis grade factor of, (0) normal glomerulus; (1) sclerotic area ≤25% of the total glomerular area; (2) sclerosis of 25-50% of the total glomerular area; (3) sclerosis of 50-75% of the total glomerular area; and (4) sclerosis ≥75% of the glomerulus. For example, if the percent of sclerosis area in total area of glomerulus in one sample is 26.8%, the glomerulosclerosis for this sample is 26.8% × 2 = 0.54; if the percent of sclerosis area in total area of glomerulus in another sample is 78.9%, the glomerulosclerosis for it is 78.9% × 4 = 3.16.
Immunofluorescent and immunohistochemical staining. The sections of kidney and liver were subjected to immunofluorescent/ immunohistochemical staining to determine protein expression of fibronectin, AGE, COL1A2, COL4A1/3, MMP2, MMP9, pi-Smad2/3, IRS1, pi-IRS1, INSR and pi-INSR as described 53 . The liver or kidney 5 μm frozen sections were also used to conduct immunohistochemical staining as follows: the slides were rinsed with PBS and incubated in 0.3% H 2 O 2 /PBS solution at room temperature for 10 min. After rinsing with PBS, the sections were blocked with goat serum for 1 h followed by incubation with primary antibody in a humidified chamber for 1 h at room temperature or overnight at 4 °C. After removal of primary antibody by washing with PBS, the sections were incubated with biotin-conjugated goat anti-rabbit IgG for 15 min at room temperature. The sections were then washed with PBS followed by incubation in a HRP-conjugated avidin solution for 15 min before adding the developing solution.
After development, sections were stained with hematoxylin solution for nucleus and then mounted under cover slides with Permount. After adequate drying, the slides were viewed and photographed using a Leica microscope. The density of images was quantified by segmentation color-threshold analysis using morphometry software (IP Lab, Scanalytics, Rockville, MD) as described 55,56 . Quantitative real time RT-PCR (qRT-PCR). After treatment, total RNA was extracted from a piece of kidney followed by cDNA synthesis using a reverse transcription kit (Promega, Madison, WI) and real time PCR with SYBR Green Master Mix (Bio-Rad, Los Angeles, CA) as described 58 . The sequences of primers are listed in Table 3. Expression of AGER, Nephrin, WT1, MMP2, and MMP9 mRNA was normalized with β-actin mRNA in the corresponding samples.

Data analysis.
All experiments were repeated at least three times, and the representative results are presented. All data were initially subjected to a normal distribution analysis with SPSS software (1-sample K-S of non-parametic test), and the data in normal distribution were then analyzed by the parametric statistics, post hoc test of one-way analysis of variance. A difference was considered to be statistically significant at p < 0.05. In addition, all the raw data are available with authors and can be submitted upon request.  Table 3. Sequences of primers for qRT-PCR. AGER: advanced glycosylation end product-specific receptor; MMP2/9: matrix metallopeptidase 2/9; WT1: Wilm's tumor 1.