Hyperoside pre-treatment prevents glomerular basement membrane damage in diabetic nephropathy by inhibiting podocyte heparanase expression

Glomerular basement membrane (GBM) damage plays a pivotal role in pathogenesis of albuminuria in diabetic nephropathy (DN). Heparan sulfate (HS) degradation induced by podocyte heparanase is the major cause of GBM thickening and abnormal perm-selectivity. In the present study, we aimed to examine the prophylactic effect of hyperoside on proteinuria development and GBM damage in DN mouse model and the cultured mouse podocytes. Pre-treatment with hyperoside (30 mg/kg/d) for four weeks could significantly decrease albuminuria, prevent GBM damage and oxidative stress in diabetes mellitus (DM) mice. Immunofluorescence staining, Real time PCR and Western blot analysis showed that decreased HS contents and increased heparanase expression in DN mice were also significantly improved by hyperoside pre-treatment. Meanwhile, transmission electron microscope imaging showed that hyperoside significantly alleviated GBM thickening in DN mice. In addition, hyperoside pre-treatment inhibited the increased heparanase gene (HPR1) promoter activity and heparanase expression induced by high glucose or reactive oxidative species (ROS) in cultured podocytes. Our data suggested that hyperoside has a prophylactic effect on proteinuria development and GBM damage in DM mice by decreasing podocyte heparanase expression.

Heparanase is an identified endo-β(1-4)-D-glucuronidase which can specifically cleave carbohydrate chains of HS in extracellular matrix such as GBM and glomerular endothelial glycocalyx [17][18][19] . Loss of HS contents degraded by heparanase in the GBM has been confirmed in experimental DN models and DN patients [19][20][21] . Heparanase-driven inflammatory cascade also plays an important role in pathogenesis of DN 22 . Renal heparanase, mainly in podocytes is abnormally up-regulated by high glucose, ROS and renin-angiotensin system (RAS) in diabetes condition [23][24][25] . The recent studies suggest that heparanase plays a crucial role in pathogenesis in DN 19,26 .
Hyperoside is a flavone glycoside showing anti-oxidant, anti-cancer, and anti-inflammatory properties [27][28][29] . Hyperoside inhibited high glucose-induced vascular inflammations 27 and mitigated cultured podocyte apoptosis induced by the advanced glycation end-products (AGEs) 30 . The recent study suggested that hyperoside could ameliorate glomerulosclerosis in DN by downregulating miR-21 and increasing matrix metalloproteinases-9 (MMP-9) expression 31 . Our previous study showed that hyperoside could decrease albuminuria at the early stage of DN through alleviating podocyte injury 32 . Based on the pathogenesis and prevalence of DN, it is clinically more important and meaningful to prevent the onset of albuminuria and GBM damage in DN than to cure them. Whether hyperoside has a prophylactic effect on development of albuminuria and can postpone the onset of DN in DM remains to be investigated.
The purpose of this study was to explore the possible prophylactic effect of hyperoside on development of albuminuria and GBM damage in STZ-induced DM mouse model. Furthermore, its effect on glomerular podocyte heparanase up-regulation in vivo and in vitro was also observed to clarify its molecular and pharmacologic mechanisms.

Results
Hyperoside pre-treatment alleviated albuminuria and ROS in DN mice. To explore the prophylactic effect of hyperoside on albuminuria at the early stage DN mice, urinary mAlb/Cr was measured in the four groups. As compared with the normal mice, DN mice exhibited significant increase in urinary mAlb/Cr (P < 0.05). Pre-treatment with lower dose of hyperoside (10 mg/kg/d) had a decreasing effect on urinary mAlb/ Cr but did not reach the statistically difference. Higher dose of hyperoside (30 mg/kg/d) for four weeks could significantly decrease the albuminuria in DN mice (P < 0.05). Both doses of hyperoside did not significantly affect blood pressure and glucose metabolism in DN mice ( Table 1). The activities of ROS-related enzymes and ROS-related end product MDA in kidney reflect the status of renal oxidative stress. The MDA and ROS producing enzyme XO were significantly elevated in kidney cortex of DN mice by about three fold. Hyperoside pre-treatment (30 mg/kg/d) could significantly decrease the elevated MDA and XO levels (Fig. 1A,D). It could also increase the ROS preventing enzyme SOD and CAT levels which were found to be down-regulated in DN mice (Fig. 1B,C). Meanwhile, as the indicatives of ROS, increased MDA and decreased SOD levels in serum of DN mice were improved significantly by hyperoside pre-treatment (Table 1), indicating hyperoside displayed a decreasing effect on renal and serum ROS production in DN mice.
Hyperoside pre-treatment decreased renal heparanase up-regulation in DN mice. The immune-reactivity of heparanase was mainly distributed in mouse glomeruli and semi-quantitative analysis was adopted to measure the amount of glomerular heparanase positive area. As shown in immunofluorescence staining with two different heparanase primary antibodies, hyperoside pre-treatment (30 mg/kg/d) for four weeks could significantly inhibit the increased glomerular heparanase immune-reactivity in DN mice ( Fig. 2A-D). Furthermore, quantitative PCR and western blot analysis showed that renal cortex heparanse expression was abnormally up-regulated in DN mice as compared with the normal mice (P < 0.05). Consistently, hyperoside pre-treatment could significantly decrease the up-regulated heparanase protein (Fig. 1E,F) and mRNA expression (Fig. 2E) in DN mice.  (Fig. 3A,C). Accordingly, western blot analysis showed that hyperoside pre-treatment could markedly improve the decreased HS protein levels in renal cortex of DN mice (Fig. 3E). GBM thickness was assessed by transmission electron microscopy and semi-quantitative analysis 33 . The results indicated that hyperoside pre-treatment (30 mg/kg/d) for four weeks could significantly alleviate the degree of GBM thickening in DN mice (Fig. 3B,D).
Hyperoside pre-treatment mitigated heparanase expression induced by high glucose. As compared with the 5 mmol/L glucose (as the Control), incubation of high concentration of glucose (10, 15 and 20 mmol/L) could up-regulate heparanase mRNA levels in cultured podocytes in a dose-related manner (Fig. 4A). The mannitol of the same osmolity with 20 mmol/L glucose had no significant effect on heparanase expression, which excluded the effect of high osmolity on heparanase expression. Pre-incubation with hyperoside (30 or 100 µg/ml) could significantly decrease the up-regulated heparanase mRNA levels induced by 20 mmol/L glucose (Fig. 4B). Accordingly, hyperoside pre-treatment could significantly decrease heparanase protein expression in cultured podocytes by western blot analysis (Fig. 4C).

Hyperoside pre-treatment inhibited heparanase promoter activity induced by high glucose.
To determine whether hyperoside could regulate HPR1 transcription induced by high glucose, a luciferase reporter assay was performed. The cultured podocytes was transfected with the luciferase reporter gene driven by a 0.7 or 3.5 kb HPR1 promoter. 10, 20 and 25 mmol/L glucose could induce luciferase activity by about four fold as compared with cells treated with 5 mmol/L glucose as the control (Fig. 4D). Pre-treatment with hyperoside (100 µg/ml) could significantly decrease the heparanase promoter activity induced by 20 mmol/L glucose (P < 0.05), whereas luciferase activity in cells with the empty vector was very low and not affected by glucose or hyperoside treatment (Fig. 4E).
Hyperoside pre-treatment attenuated heparanase up-regulation induced by ROS. Incubation with xanthine and xanthine oxidase (X/XO) generates extracellular radicals such as superoxide and hydrogen peroxide while exposure to NADH induces intracellular generation of ROS 25 . Exposure to X/XO at concentration of 3, 10, 30 and 100 (µM/mU) or NADH (0.3, 1, 3 and 10 mM) could significantly increase podocyte heparanase mRNA levels in a dose-related manner (Fig. 5A

ROS inhibition decreased heparanase elevation induced by high glucose.
The dose-response effect of ROS inhibitor N-acetyl-L-cysteine (NAC) and YCG063 on heparanase expression induced by high glucose was explored. Pre-incubation with NAC (3 mM) or YCG063 (30 µM) could partly antagonize the up-regulated heparanase mRNA expression induced by 20 mmol/L glucose (P < 0.05). Incubation with NAC or YCG063 alone did not influence heparanase expression (Fig. 6A,B).

Discussion
DN is becoming one of the life-threatening diabetic micro-vascular complications 4,34 . The epidemical evidence showed that the incidence of DN has increased dramatically among diabetes population worldwide in recent year 4,35 . The data from evidence-based medicine suggested that proteinuria occurred inevitably and kept increasing at the early stage of DN, despite intensive blood glucose control and RAS inhibition 36,37 . DN often develops insidiously over many years before clinical manifestations including proteinuria and declining renal function are evident 3 . Therefore, taking precaution against the appearance of albuminuria and onset of DN is more meaningful and important than curing it in clinical practice 38 . There is a great need for alternative drugs to prevent the development of albuminuria and postpone the onset of DN in DM patients 39 .
Alteration of the permeability characteristics of GFB is an early cause for proteinuria 38,40 . GBM damage plays an important role in GFB dysfunction and pathogenesis of DN 5,9,41 . The presence of glycosaminoglycans, primarily HS in GBM can prevent passage of negatively charged macromolecules cross to urinary space 7, 12 . The decreased HS content in GBM results in loss of negative selectivity at the early stage of DN 12,16 and then impairs the perm-selectivity to negatively charged albumin 10,42 . Substantive albumin is leaked from plasma into urinary space and causes the formation of albuminuria in DN. The recent study showed that heparanase plays a pivotal role in the development of proteinuria in DN, which suggests that loss of HS contributes to the development of proteinuria 17,26 .
Heparanase is the only enzyme capable of degrading HS specifically in extracellular matrix 18 . HS chains are important constituents and organizers of GBM and have a key role in maintaining the integrity and function of the glomerular filtration barrier 43 . The HPR1 gene first synthesizes a 68 kD pre-proheparanase and it is processed to 65 kD pro-heparanase in the endoplasmatic reticulum by glycosylation and cleaving off the signal peptide. Then it is transported to the Golgi apparatus, packaged into vesicles and secreted. After internalization, the 65 kD heparanase protein is then transferred to endosomes and lysosomes, where it is proteolytically into the active heterodimer consisting of an 8 kD and 50 kD subunits 18 . Heparanase is majorly up-regulated in podocytes induced by high glucose in DM status 23,24 . Its expression could also be increased by ROS, angiotensin II and aldosterone in cultured podocytes 25,44 and the increased podocyte heparanase expression in kidneys has been demonstrated in DN 19,23 , which is essential for the development of albuminuria DN in both animal model and human 26,45 . Inhibition of heparanase expression and restoring HS contents proved to be effective to decrease proteinuria of glomerular diseases 26,[46][47][48][49] .
As an active flavone glycoside, hyperoside is documented to have therapeutic effect on ischemic reperfusion injury in brain and heart, hepatitis and cancer [50][51][52][53] . It can protect cells against oxidative stress and oxidation-related apoptosis 29,54,55 . Hyperoside displayed potential anti-inflammatory on high glucose induced inflammation in vitro and in vivo 27,28 . Our previous study showed that hyperoside has therapeutic effect on albuminuria at the early stage of DN 32 . More meaningfully, the present study offered the further experimental evidence that hyperoside has a prophylactic effect on albuminuria and GBM damage in DM mice, which could postpone the onset of DN. The results showed that orally pre-treatment with hyperoside (30 mg/kg/d) for four weeks could significantly ameliorate the albuminuria (urinary mAlb/Cr) in DM mice and slow the progression of DN. There was no significant difference in blood pressure, glucose and lipids parameters, plasma angiotensin II and aldosterone levels between DN and hyperoside pre-treatment groups. The prophylactic effect of hyperoside might not be due to its remedial effect on metabolic and hemodynamic factors. We could not completely exclude the possibility of involvement of the intra-renal RAS system which is independent of systemic RAS system because it was difficult to measure the renal tissue angiotensin II and aldosterone levels. Notably, Hyperoside could decrease ROS end product (MDA) and ROS producing enzyme (XO) and increase ROS preventing enzymes (SOD and CAT) both in kidney and circulation of DM mice, showing that it could alleviate renal oxidative stress in DM mice. The anti-oxidant property of hyperoside might partly explain its prophylactic effect on albuminuria and GBM damage in DM mice.
Consistently, transmission electron microscopy observation demonstrated that GBM thickening in DN was also alleviated by hyperoside pre-treatment. GBM thickening precedes clinically evident albuminuria and is regarded as the early pre-disposing factor of albuminuria in DN patients 5,6 . This morphological change suggested that hyperoside had a protective effect on GBM damage in DM mice. Our results showed that hyperoside pre-treatment could restore the glomerular HS contents decreased in DN mice. Due to the major role of HS in GBM composition, its reduction was regarded as a determinant of glomerular hyper-permeability in DN 12,16,42 . We presumed that the preservation of HS by hyperoside might be associated with decreased heparanase function which can specifically degrade HS in GBM. With sensitive heparanase antibody that specifically interacts with the 65-kDa heparanase, we were able to detect small quantity of heparanase protein. The data manifested that renal heparanase mRNA and protein up-regulation in DM mice could be significantly improved by hyperoside pre-treatment. Our results indicated that the normalization of glomerular heparanase expression and HS content in GBM are closely associated with the protective effect of hyperoside on albuminuria development and GBM damage in DN. Admittedly, we cannot exclude the possible involvement of the other ingredients of GBM besides HS. Taken together, hyperoside pre-treatment could decrease glomerular heparanase expression and restore HS contents in GBM, displaying a protective effect on GBM damage and albuminuria in DM animal.
Because renal heparanase is mainly distributed in podocytes 23 , we speculated that hyperoside has a negative effect on podocyte heparanase expression. The in vitro studies showed that heparanase expression in cultured podocytes could be induced by high glucose, which is consistent with the previous study 23 . Hyperoside pre-treatment could significantly inhibit heparanase mRNA and protein up-regulation induced by high glucose. To further test whether this effect is caused by regulating heparanase promoter activation, we conducted HPR1 gene luciferase reporter gene assays in cultured podocytes. Hyperoside was found to have a direct negative effect on increased HPR1 promoter activation and heparanase transcription induced by high glucose. Certainly, we could not exclude the possibility that hyperoside could directly bind to heparanase and affect the enzyme catalytic activity of heparanase. Furthermore, whether hyperoside directly binds to the transcription regulating factor or promoters of HRP1 also remains to be deciphered.
The previous studies suggested ROS act as an important risk factor for heparanase up-regulation 25,44 . Our data have also demonstrated that both extracellular and intracellular ROS stimulus could induce heparanase mRNA expression in podocytes. ROS-induced heparanase mRNA and protein expression in cultured podocytes could be significantly attenuated by hyperoside pre-treatment, which was coherent with our in vivo data. Hyperoside pre-treatment could markedly inhibit heparanase up-regualtion induced by ROS.
Oxidative damage including both free radical and non-radical oxygen species contributes to the pathogenesis of onset and progression of DN 56 . Both experimental and clinical studies have documented a close link among hyperglycemia, oxidative stress and DN 57 . High-glucose treatment increased ROS generation in podocytes 58 and the recent study has implicated that high glucose-induced podocyte injury and apoptosis are mediated by ROS 59 . It was observed that ROS inhibitor NAC and YCG063 could partly antagonize podocyte heparanase up-regualtion induced by high glucose. Based on that, we deduced that high glucose induced heparanase up-regulation might also be partly mediated by ROS generated in podocytes. The inhibitory effect of hyperoside on podocyte heparanase expression was likely to be partly mediated by its anti-oxidative effect but needs more supportive evidence.
Although the detailed underlying molecular mechanism remains to be investigated, our current in vitro and in vivo data suggested that hyperoside pre-treatment could inhibit podocyte heparanase expression induced by high glucose and ROS and then restore HS contents in GBM in DN mice. This effect contributed to maintaining GBM composition integrity and preventing the development of albuminuria in DN. Taken together, hyperoside could be taken as a potential and promising candidate chemical compound in preventing and treatment of DN. It would be necessary to perform the translational medicine research or clinical trial upon it in the future.

Conclusions
Our study indicated that hyperoside pre-treatment can prevent development of proteinuria and postpone the onset of DN in DM mice. The prophylactic effect of hyperoside on GBM damage in DM mice attributed to inhibiting podocyte heparanase expression induced by high glucose and oxidative stress.

Methods
Reagents. Hyperoside with the purity higher than 98% was purchased from Zelang Biological Technology Company (Nanjing, China) and suspended in 1% carboxymethyl cellulose (CMC) solution for animal experiments. Hyperoside (CAS#482360), N-acetyl-L-cysteine (NAC), Xanthine and xantine oxidase used for cell experiments were from Sigma Corporation (Hong Kong, China) and dissolved in 0.1% DMSO. YCG063 was bought from Calbiochem (San Diego, CA, USA) and NADH from Roche (Roche Diagnostics, Indianapolis, IN, USA). Diabetic nephropathy model. Type 1 diabetes was induced by streptozotocin (STZ) injection according to the Animal Models of Diabetic Complications Consortium (AMDCC) protocol with minor modification 60,61 . Seven to eight weeks old female C57BL6 mice were given three intra-peritoneal injections of STZ (Sigma, 100 mg/ kg body weight) dissolved in 10 mM sodium citrate every 48 hours. The normal mice only receiving the same volume of sodium citrate served as the control group. The mice with blood glucose between 13.9 and 22.2 mmol/L were regarded as DM mice two weeks after STZ injection. Two weeks after the DM establishment, DM mice were randomly divided into three groups: DM mice orally treated with the vehicle CMC solution alone (DN group), low or high dose of hyperoside (10 or 30 mg/kg/d, once every day) for four weeks (LHPS and HHPS groups). Each group had 15 mice and their ratio of urinary micro-albumin to creatinine (urinary mAlb/Cr) was assayed every two weeks. Those with elevated urinary mAlb/Cr (30~300 mg/g) were regarded as DN mice 60 . All animals were fed with standard laboratory diet and provided with water ad libitum.

Experimental protocols. The experimental protocol was approved by the Animal Ethics Committee of
Fudan University and all experiments were performed in accordance with relevant guidelines and regulations. Blood glucose levels of DM mice were monitored in tail vein blood every week and urine was tested for ketone bodies. When necessary, DM mice were given supportive insulin treatment (Ultratard, Novo Nordisk, Denmark) at dose of 1 IU/kg BW (twice a week) to prevent apparent exhaustion and ketosis. When experiment terminated, individual urine sample was collected and the urinary mAlb/Cr was measured and calculated according to our previous study 32 . Mean arterial pressure (MAP) was measured in conscious, trained mice at room temperature using a tail-cuff monitor (BP-2000 Blood Pressure Analysis system, Visitech Systems). Fresh kidney cortices were excised right after animals were sacrificed. Blood samples were collected and centrifuged to obtain serum. All samples were stored at −80 °C until further study. Histological analysis. For ultrastructural evaluation, the renal cortex was cut into small pieces (1 mm 3 ) and pre-fixed with 2.5% glutaraldehyde (pH = 7.4) for 48 h, then washed by 0.1 M PBS (10 min × three times). The samples were post-fixed in 1% buffered potassium osmate for 1 h, dehydrated by gradient acetone and embedded in Epon812 overnight. Ultra-thin sections were stained with 3% lead citrate/uranyl acetate and examined under transmission electron microscope (CENTRA100, Carl Zeiss, Germany). Three samples were selected from every renal cortex and each with five electron micrographs at ×25000 magnification. Five glomeruli in every electron micrograph were randomly selected for further measurement. Image-Pro Plus software was utilized to measure the mean GBM thickness and all images were analyzed by an investigator who was blinded to the identity of the samples.

Enzyme
Immunofluorescence. Mice kidney cortices were fixed in 4% formalin and embedded in paraffin. Antigen retrieval and deparaffinization were performed using the antigen retrieval solution for 20 minutes at 95 °C. All sections (5 μm) were blocked with 10% bovine serum albumin (BSA) and then incubated with monoclonal anti-mouse HS primary antibody (AMS Biotechnology, Milton Park, Abingdon, UK) in the dilution of 1:100 at 4 °C overnight. The rabbit anti-mouse heparanase primary antibody (Ab85543, Abcam, Cambridge, MA) which recognizes 65 kD precursor as well as the 50 kD and 8 kD subunits of heparanase and the heparanase primary antibody (MBS711555, My Biosource) reacting with 65 kD heparanase heterodimer were adopted to display glomerular heparanase expression. The sections were washed in 0.1 M PBS and stained with FITC (or rhodamine) conjugated secondary antibodies for 60 min at room temperature. DAPI was adopted to label nuclear DNA. After washing in 0.1 M PBS for three times, the sections were imaged by a Zeiss microscope (LSM-510, Carl Zeiss, Germany). Fluorescence intensity was quantified and analyzed by National Institute of Health (NIH) Image software.
Cell culture. Conditionally immortalized mouse podocytes were cultivated in RPMI 1640 cell medium containing 10% fetal bovine serum (FBS), 100 U/ml penicillin, 100 mg/ml streptomycin, and 2 mM L-glutamine 62 . To permit immortalized growth, the culture medium was supplemented with 10 U/ml of recombinant mouse γ-interferon to induce the expression of T antigen and cells were cultured at 33 °C (permissive conditions). To induce differentiation, cells were cultured on collagen IV under non-permissive condition (37 °C) without γ-interferon for 10-14 days 63,64 . The differentiated podocytes were used in our cell experiments. Glucose, ROS or ROS-related inhibitors pre-treatment were performed as previous studies with minor modification 23,25 .
Western blot. The mouse kidney cortices or cultured cells were homogenized in radioimmunoprecipitation assay (RIPA) buffer at 4 °C. After centrifugation (21000 × g) for 30 min at 4 °C, supernatant was collected. Total protein concentration was quantified by using the DC-protein determination system (Bio-Rad, Shanghai, China). Samples were processed for SDS-PAGE, and 30 μg protein extract were loaded and electro-transferred onto nitrocellulose membrane (Hybond-ECL, Amersham). The membranes were blocked with 5% non-fat dry milk in 1 × TBS (0.1% Tween-20) for 30 min and incubated overnight with polyclonal rabbit anti-mouse heparanase-1 antibody (MBS711555, My Biosource) which can recognize the 65 kD precursor of heparanase-1 or the monoclonal anti-mouse HS primary antibody (AMS Biotechnology, Milton Park, Abingdon, UK) in the dilution of 1:1000 at 4 °C. After incubation with horseradish peroxidase-labeled secondary antibody for 60 min, the signals were detected with an enhanced chemiluminescence system (Amersham). Membranes were incubated with rabbit polyclonal anti-β-actin antibodies (Cell Signaling Technology, USA) to serve as the internal controls for equal loading. The density of heparanase in each band was determined using NIH Image software and expressed as a relative value to the density of the corresponding β-actin band.
Transfection and luciferase activity assay. A pGL3 luciferase reporter vector containing the 0.7 or 3.5 kb promoter region of the mouse HPR1 (heparanase) gene was constructed as the previous study 23 . The cultured podocytes in 12-well plates were treated with vehicle, Hyperoside (Sigma-Aldrich) and (or) 20 mmol/L glucose before they were transfected with either pGL3-HPR1 promoter construct or pGL3-basic empty vector. The pRL-CMV (Promega, WI, USA) containing renilla luciferase gene was adopted as an internal control for transfection efficiency using 1.5 µl Lipofectamine in 100 µl of Opti-MEM medium (Invitrogen, CA, USA). Cultured podocytes were harvested 24 h after transfection and the luciferase activity was measured by Dual-Luciferase reporter assay kit (Promega, Madison, WI). Statistical analysis. Data were presented as mean ± standard deviation (SD) and analyzed by Graphpad Prism 6.0 software (San Diego, CA, USA). Differences between multiple groups were analyzed by one-way ANOVA in combination with Student's t test. The post hoc comparisons using the Student Newman-Keuls test were used for inter-group comparisons of multiple variables. P < 0.05 was considered significant.

Ethics approval. Our experimental protocol was approved by the Animal Ethics Committee of Fudan
University and all experiments were performed in accordance with relevant guidelines and regulations. Data availability. All the data in our manuscript are available in Jinshan Hospital of Fudan University.