Modulation of Notch-1 Signaling Alleviates Vascular Endothelial Growth Factor–Mediated Diabetic Nephropathy

Conditionally immortalized human podocytes (22) were routinely cultured in RPMI-1640 medium supplemented with 10% FBS and 1% insulin transferrin disodium selenite (Sigma, St. Louis, MO) at 33°C. To stimulate cell differentiation, the culture temperature was changed from 33 to 37°C for 14 days, and the FBS was replaced with human plasma (23).

HEK 293 cells were maintained in Dulbecco's Modified Eagle's Medium with 10% FBS. Additional incubations with 10 units/ml bovine erythrocyte SOD-PEG (Sigma), 1 μmol/l N-[N-(3,5-difluorophenacetyl)-l-alanyl]-S-phenylglycine t-butyl ester (DAPT) (Calbiochem, Gibbstown, NJ), or specific monoclonal antibodies (R&D Systems, Minneapolis, MN) were also performed in cultured cells.

Total RNA from 1 × 10⁶ cells or from rat kidneys was prepared with QIAzol reagent (Qiagen, Valencia, CA) or Tri reagent (Sigma) according to the manufacturers' instructions. RT-qPCR was performed according to the MIQE guidelines (24). Total RNA (1 μg) was reverse transcribed to cDNA with the Moloney murine leukemia virus reverse transcriptase (Fermentas, Glen Burnie, MD). The PCR mixture (25 μl) containing the cDNA template equivalent to 20 ng total RNA, 2.5 μmol/l forward and reverse primers, and 2X iQ SYBR Green Supermix (Bio-Rad, Hercules, CA) was prepared and subjected to PCR amplification. The specific primer pairs targeting the Jagged-1, Notch-1, Notch-2, MAML-1, VEGF, hairy and enhancer of split (Hes)-1, nephrin, and β-actin genes were generated using the MIT Primer3 software (http://fokker.wi.mit.edu/primer3/). The individual PCR primer sequences and cycle number used in this study included the following: 5′-GACTCATCAGCCGTGTCTCA-3′ and 5′-TGGGGAA CACTCACACTCAA-3′ for amplifying Jagged-1, 25–28 cycles; 5′-CAATGTG GATGCCGCAGTTGTG-3′ and 5′-CAGCACCTTGGCGGTCTCGTA-3′ for Notch-1, 25–30 cycles; 5′-GGACTTCGAGCAAGAGATGG-3′ and 5′-AGCACT GTGTTGGCGTACAG-3′ for Notch-2, 25–30 cycles; 5′-CCCTGATGAGATC GAGTACATCTT-3′ and 5′-AGCAAGGCCCACAGGGATTT-3′ for VEGF, 25–28 cycles; 5′-CAGCATCAGTTGCTTTTGGA-3′ and 5′-CCCTGTGAACTGTC CAACCT-3′ for MAML-1, 25–30 cycles; 5′-CAACACGACACCGGATAAAC-3 and 5′-TTCAGCTGGCTCAGACTTTC-3′ for HES-1, 25–30 cycles; 5′-CACGGTCAG CACAACAGAGG-3′ and 5′- GAAACCTCGGGAATAAGACACCT-3′ for nephrin, 30 cycles; 5′-TCCTTCCTGGGCATGGAGTC-3′ and 5′-TTCTGCATCCTGTCG GCAATG-3′ for β-actin, 25 cycles; and 5′-GAGTCAACGGATTTGGTCGT-3′ and 5′-TTGATTTTGGAGGGATCTCG-3′ for glyceraldehyde-3-phosphate dehydrogenase, 25 cycles. All real-time PCR experiments were performed in duplicate from at least three independent treatments. The relative gene expression was calculated as previously described (25) and as described in the MIQE guidelines (24).

Membrane, cytosolic, and nuclear extracts from cultured cells were prepared as previously described (26). Antibodies specific for Jagged-1 (Rockland, Gilbertsville, PA), full-length Notch-1 (Santa Cruz, CA), NICD (Abcam, Cambridge, U.K.), MAML-1, HES-1 (Abcam), hypoxia-inducible factor (HIF)-1α (Cell Signaling, Beverly, MA), VEGF (Santa Cruz), nephrin (kindly provided by Prof. Karl Tryggvason, Karolinska Institute, Stockholm, Sweden), poly(ADP-ribose) polymerase (PARP)-1 (Cell Signaling), cleaved and full-length caspase-3 (Cell Signaling), and β-actin (Santa Cruz) were used.

γ-Secretase activity was quantified by fluorescent spectroscopy using a γ-secretase assay kit according to the manufacturer's instructions (R&D). Briefly, the γ-secretase activity was determined using the amyloid precursor protein peptide (GVVIATVIV) conjugated with 5-[(2-aminoethyl)amino] naphthalene-1-sulfonic acid (EDANS) and 4-(4-dimethylaminophenylazo) benzoic acid (DABCYL) reporter molecules as a substrate. The cleavage-dependent release was measured using a fluorescent microplate reader (Anthos Zenyth 3100; Anthos Labtec, Salzburg, Austria) at 495 nm.

Plasmids expressing human Notch-1 shRNA (5′-CCAC CAGTTTGAATGGTCATT-3′), MAML-1 shRNA (5′-CGAGCAACTCCCTGTT TCT-3′), and a control shRNA were purchased from SuperArray (Frederick, MD). Transfection of podocytes or HEK293 cells (1 × 10⁶ cells/well in a six-well plate) with the indicated plasmids was performed using lipofectamine reagent (Invitrogen).

The NICD expression plasmid has been described previously (27); it was transfected into HEK293 cells using lipofectamine reagent (Invitrogen), and the transfected cells were cultured in medium with 300 μg/ml G418 (Invitrogen) for screening transient transfectants as previously described (27).

For determination of cell viability and proliferation, podocytes under different stimulations were cultured and were evaluated with tetrazolium-based colorimetric assay (XTT assay; Roche, Mannheim, Germany). Colorimetric changes were measured with a microplate reader at a wavelength of 450–500 nm (Molecular Devices, Sunnyvale, CA).

Conditioned medium isolated from podocytes treated with various modulators was collected. Angiogenesis induced by the conditioned medium was determined by coculturing the conditioned medium with human umbilical vein endothelial cells (HUVECs) on a Matrigel-coated (Becton Dickinson, Bedford, MA) 48-well culture plate (1 × 10⁴ cells/well). Formation of capillary/tube-like structures was visualized and counted 2 h after coculture (28).

Male Wistar rats 4 months of age and weighing between 220 and 250 g (National Experimental Animals Production Center, Tapei, Taiwan) were caged and subjected to streptozotocin (STZ) injection as previously described (29); rats with blood glucose levels >350 mg/dl were defined as diabetic. Diabetic rats were treated with DAPT (n = 6; 5–10 mg/kg) or vehicle (n = 6; 200 μl corn oil) via subcutaneous injection once a week for 35 consecutive days. Then, urine samples were collected from all rats for measurement of total protein and creatinine concentrations using assay kits (Sigma). The animals were killed, and the tissues were collected as described previously (25). All in vivo experimental protocols were approved by the Institutional Animal Care and Use Committee of the Chang Gung Memorial Hospital and carried out according to the National Institutes of Health guidelines.

Paraffin-embedded tissues were sliced longitudinally into 5-μm-thick sections and subjected to immunohistochemical staining. Antibodies specific for Jagged-1, NICD, HIF-1, VEGF, nephrin, podocin (Santa Cruz), and CD2AP (Santa Cruz) were used. Six regions within the renal glomeruli from three separate sections were obtained from each rat; images were randomly analyzed (25,29).

Paraffin-embedded rat tissues were evaluated for apoptosis using an in situ cell death detection kit according to the manufacturer's instructions (Roche, Mannheim, Germany) and as previously described (29).

All values were expressed as means ± SE from at least three independent experiments. Wilcoxon's test was used to evaluate differences between the sample of interest and its respective control. To analyze the time course studies, multiple-range ANOVA and post hoc tests were used. P values <0.05 were considered significant.

Dysregulation of VEGF and nephrin expression is commonly observed in diabetic nephropathy (7,10). To determine the effect of HG on human podocyte gene expression, cells were cultured with increasing amounts of d-glucose (15, 25, and 35 mmol/l) for 48 h. In normal control cells, 5 mmol/l d-glucose was maintained in the culture medium to provide the basal requirement for cell growth. Additionally, mannitol served as an osmotic control for HG. As illustrated in Fig. 1 A, VEGF mRNA gradually increased with HG conditions in human podocytes; downregulation of nephrin mRNA was also observed. Neither VEGF nor nephrin expression was significantly altered in the mannitol-treated cells compared with normal control cells, indicating that HG modulated VEGF and nephrin expression in cultured podocytes.

In addition to VEGF and nephrin, we also examined the effect of HG on the Notch signaling pathway in human podocytes. Cells cultured in medium with or without HG (35 mmol/l) were harvested at different time points (24, 48, and 72 h). mRNA expression of Notch signaling components, including Jagged-1, Notch-1, Notch-2, and HES-1, was determined by RT-PCR (Fig. 1,B) or RT-qPCR (Fig. 1,B). Because the expression of β-actin or glyceraldehyde-3-phosphate dehydrogenase genes in both untreated and glucose-treated podocytes at 24, 48, and 72 h remained invariant under the experimental conditions (supplementary Fig. 1 in the online appendix, which is available at http://diabetes.diabetesjournals.org/cgi/content/full/db09-0663), we subsequently used β-actin as an internal reference gene for expression studies. As shown in Fig. 1,B, HG induced the Jagged-1, Notch-1, and HES-1 mRNA expression in a time-dependent manner; however, Notch-2 expression was unaltered. As expected, increased VEGF and decreased nephrin were also detected concomitantly in HG-treated cells. By contrast, no changes in gene expression were observed upon treatment with high mannitol (35 mmol/l) (Fig. 1 B). These results raised the question whether signaling induced by HG may be linked to dysregulation of VEGF and nephrin expression in podocytes.

To examine the relationship between Notch-1 signaling and dysregulation of VEGF and nephrin expression in HG-stressed podocytes, a γ-secretase inhibitor, DAPT, was used to inhibit Notch signaling. Human podocytes were pretreated with DAPT for 2 h, and then the continual DAPT treatment proceeded in the presence or absence of HG. Treatment with DAPT markedly abolished the upregulation of HES-1 and VEGF and downregulation of nephrin in HG-treated human podocytes (P < 0.05) (Fig. 2,A). Notably, DAPT did not alter upregulation of Jagged-1 or Notch-1 mRNAs in HG-treated podocytes (Fig. 2,A). Similarly, Western blot analyses (Fig. 2,B) revealed that DAPT significantly inhibited NICD expression alone; Jagged-1 and full-length Notch-1 expression induced by HG remained unchanged. Dysregulation of VEGF and nephrin induced by HG was also alleviated by DAPT. Superoxide dismutase (SOD), an enzyme that can scavenge superoxide, was also analyzed. Under normal culture conditions, treatment of human podocytes with exogenous SOD did not affect basal Notch-1 signaling (supplementary Fig. 2). However, exogenous SOD inhibited HG-induced Jagged-1, Notch-1, and HES-1 expression (Fig. 2,A and B). Inhibition of Notch-1 signaling by SOD also relieved the dysregulation of VEGF and nephrin in HG-treated podocytes (Fig. 2 A and B).

To further determine the angiogenic activity of VEGF secretion from podocytes, a tube-like forming assay was performed using HUVECs cocultured with conditioned medium from cultured human podocytes treated with various reagents. Compared with the normal control conditions, larger and more abundant tube-like structures were observed upon coculture of HUVECs with conditioned medium isolated from HG-treated cells. However, reduced tube formation was observed using conditioned medium from the cells pretreated with SOD or DAPT and cultured in an HG environment compared with that from cells treated with HG alone (Fig. 2,C). γ-Secretase activity was also examined in response to various stimuli. A significant increase (P < 0.05) in γ-secretase activity was detected in HG-stressed cells compared with cells cultured in normal conditions. Enhanced γ-secretase activity was inhibited by SOD and DAPT (Fig. 2 D).

In addition to human podocytes, we also determined whether HG affected Notch-1 signaling and the expression of VEGF and nephrin in the human embryonic kidney cell line, HEK293. HEK293 cells are known to express nephrin (30,31), and the cell type-specific expression of nephrin protein in HEK293 cells was also confirmed by Western blot analysis (supplementary Fig. 3). Increased Jagged-1, Notch-1, HES-1, and VEGF expression and nephrin repression were detected in HG-treated HEK293 cells by RT-qPCR (Fig. 3,A), indicating that similar regulatory signaling may exist in both podocytes and HEK293 cells in response to HG. Because of the pleiotropy of VEGF (7,10), we examined the causative relationship between VEGF and nephrin expression or cell apoptosis in both human podocytes and HEK293 cells cultured in HG conditions. In HG-stressed podocytes, nephrin repression and apoptotic cellular markers, cleaved caspase-3 and PARP-1, were consistently detected (Fig. 3,B). Addition of a VEGF monoclonal antibody (10 ng/ml), but not a control antibody, efficiently blocked both nephrin repression and apoptosis induced by HG (Fig. 3 B). Similar attenuation of nephrin repression and apoptosis by anti-VEGF antibodies was also observed in HG-stressed HEK293 cells (supplementary Fig. 4).

To further determine the effect of VEGF on nephrin expression and apoptosis, podocytes were treated with exogenous VEGF recombinant protein (1 and 10 ng/ml). As illustrated in Fig. 3 C and D as well as in supplementary Fig. 5A, exogenous VEGF protein increased cleaved caspase-3 and PARP-1 levels but reduced nephrin expression and cell viability. However, reduced expression of nephrin was not due to reduced cell viability, as demonstrated in supplementary Fig. 5B. These results indicated that downregulation of nephrin and induction of the apoptotic cascade in HG conditions were mediated through Notch-1/VEGF signaling.

To further delineate the role of the Notch-1 signaling on dysregulation of VEGF and nephrin in HEK293, a NICD expression plasmid was transfected into HEK293 cells. NICD overexpression significantly induced HES-1, HIF-1α, and VEGF protein expression while reducing nephrin expression in HEK293 cells (Fig. 3 E and F). Taken together, these results suggest that Notch-1 signaling stimulated by HG induces VEGF expression, which subsequently mediates nephrin repression and the apoptotic cascade in human podocytes and HEK293 cells.

Cells expressing Notch-1 shRNA were created, and the expression of VEGF and nephrin were evaluated in these cells. To efficiently knock down the Notch-1 pathway, we generated a pool of transfected podocytes that expressed Notch-1 or control shRNAs. Endogenous full-length Notch-1, NICD, and HIF-1α were substantially reduced in cells expressing Notch-1 shRNA but not in those expressing control shRNA (Fig. 4,A). Knockdown of Notch-1 in cells maintained in HG conditions resulted in reduction of HES-1 and VEGF expression and restoration of nephrin expression (Fig. 4,B). Intriguingly, enhanced levels of MAML-1, a key coactivator of Notch signaling, were repeatedly detected in HG-treated cells (Fig. 4,C and D and data not shown). Transfection of MAML-1 shRNA in HEK293 cells not only attenuated HG-induced MAML-1 but also abrogated HG-induced activation of HIF-1α, HES-1, and VEGF expression. Nephrin downregulation induced by HG was also restored in MAML-1 knockdown cells (Fig. 4 C and D).

The effects of inhibition of Notch signaling by DAPT on diabetic nephropathy were also assessed. Compared with the normal group, diabetic rats had increased blood glucose and urinary protein excretion (Table 1). DAPT treatment significantly reduced urinary protein secretion in diabetic rats; blood glucose and A1C levels remained unchanged (Table 1). No differences in overall body weight, organ/body weight ratios as well as BUN, creatinine, AST, ALT, and CPK levels were observed between the untreated and DAPT-treated diabetic rats (Supplementary Table 1).

Involvement of Notch signaling in diabetic rats was examined by immunostaining and immunoblot analyses. Increased Jagged-1, NICD, HIF-1α, and VEGF expression were observed in podocytes of glomeruli in the diabetic group compared with the normal group (Figs. 5,A and 6,A and Table 1). In contrast, fewer nephrin-positive cells were observed in diabetic rats compared with normal rats (Fig. 6,A; Table 1). In addition, elevated Jagged-1, full-length Notch-1, NICD, HIF-1α, and VEGF expression, as well as reduced nephrin, was detected in tissue lysates from the diabetic kidneys upon Western blot (Figs. 5,B and 6,B) and RT-qPCR (supplementary Fig. 6A) analyses. No significant differences in podocin or CD2AP expression levels were observed between the diabetic and normal rats (Table 1 and supplementary Fig. 6B). Using terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) analysis, elevated podocyte and mesangial cell apoptosis was detected in diabetic rats compared with normal rats (Fig. 5,A and Table 1).

Inhibition of Notch-1 signaling in diabetic rats by DAPT resulted in amelioration of enhanced NICD, HIF-1α, and VEGF (Figs. 5 and 6). DAPT treatment also restored nephrin downregulation in diabetic rats to the normal level observed in control rats (Fig. 6). Furthermore, DAPT treatment resulted in suppression of apoptosis in diabetic rats (Fig. 5,A and Table 1). Although expression of several Notch-1 signaling molecules was blocked by DAPT, enhanced Jagged-1 in diabetic rats was not affected by DAPT treatment. The ability of DAPT to inhibit γ-secretase activity was confirmed by examining γ-secretase activity in DAPT-treated rats (Fig. 5,C). As expected, DAPT treatment significantly repressed γ-secretase activity in diabetic rats. As shown above, SOD inhibited Notch-1 signaling and VEGF expression in HG-treated human podocytes (Fig. 2). The effects of SOD on Notch-1 signaling and expression of VEGF and nephrin in diabetic kidney were analyzed. Exogenous SOD significantly suppressed Jagged-1, Notch-1, HES-1, and VEGF expression and restored nephrin expression in the kidney of diabetic rats (supplementary Fig. 7A). γ-Secretase activity in diabetic rats was also inhibited by exogenous SOD (supplementary Fig. 7B). These results indicate the involvement of oxidative stress, at least in part, in regulating the Notch signaling pathway in diabetic rats.

Podocyte dysfunction is closely associated with marked proteinuria/albuminuria in diabetic nephropathy (12). In this study, we show that the Notch-1 signaling is activated in HG-treated human podocytes and HEK293 cells and within the glomeruli of diabetic rats (Figs. 1, 2, 3, and 5). Activation of Notch-1 signaling coincides with increased VEGF expression, nephrin downregulation, and increased apoptosis (Figs. 1, 3, 5, and 6). Importantly, overexpression of NICD in HEK293 cells and treatment of podocytes with recombinant VEGF are sufficient to induce nephrin repression and apoptosis (Fig. 3 and supplementary Fig. 4). Furthermore, repression of Notch-1 signaling by pharmacologic reagents, including DAPT and SOD, and through genetic knockdown of Notch-1 signaling components alleviated VEGF and nephrin dysregulation in HG-stressed cells and ameliorated urinary protein excretion in diabetic animal model (Figs. 2, 4, 5 and 6). These findings underscore the importance of the Notch/VEGF signaling pathway in diabetic proteinuria (Fig. 7).

Several studies have suggested that overactive Notch signaling in mature podocytes may result in loss of podocytes, glomerular failure, and the development of glomerulosclerosis and focal segmental glomerulosclerosis (21,32,33). Consistent with the importance of Notch signaling in the development of glomerular disease, Notch-1 signaling was detected in diabetic animal model (Figs. 5 and 6) along with marked proteinuria and reduction of podocyte SD. Administration of the γ-secretase inhibitor, DAPT, alleviated proteinuria found within the glomeruli (Table 1), suggesting that Notch-1 signaling may directly contribute to the pathogenesis of proteinuria. The strong correlation between overactive Notch-1 signaling and proteinuria indicates that components of the Notch signaling cascade could possibly be genetic biomarkers for prediction of the progression of diabetic nephropathy.

Elucidation of the molecular mechanism by which Notch-1 signaling is modulated in diabetic rats remains to be addressed. In cultured podocytes or HEK293 cells, an HG environment induced Notch-1 signaling through many distinct manners, including Jagged-1 (a ligand of Notch-1), Notch-1, and MAML-1 (a coactivator of Notch signaling) expression as well as enhancement of γ-secretase activity (Figs. 1, 2, and 4). Interestingly, exogenous SOD dramatically blocked Notch-1 signaling in HG-treated cells (Fig. 2). These results imply that superoxide overproduction under HG stress may act as a trigger in stimulating the Notch-1 pathway.

At the early stages of diabetic nephropathy, the expression of VEGF and VEGF receptor was upregulated within the glomerulus—specifically within podocytes (34). Several reports have shown that inhibition of VEGF in experimental models of diabetic nephropathy restored the slit pore density of podocytes and alleviated albuminuria (5,7,35). Based on our results, we observed elevated VEGF expression as well as reduced nephrin expression and podocyte number in diabetic rats compared with normal rats (Figs. 5 and 6 and Table 1), further supporting the relationship between VEGF signaling and proteinuria. Neutralization of VEGF with monoclonal antibodies reversed the repression of nephrin and attenuated the cleavage of PARP-1 and caspase-3 in HG-treated podocytes and HEK293 cells (Fig. 3 and supplementary Fig. 4). Although VEGF is generally regarded as a prosurvival factor (36), our results revealed that recombinant VEGF stimulates the apoptotic cascade in human podocytes (Fig. 3 B–D). A growing number of studies have suggested that VEGF promotes cell death mediated by or in concert with other apoptotic signaling factors (37,–39).

In addition to nephrin, other SD proteins that maintain the normal function of the glomerular filtration barrier may be controlled by Notch-1 and VEGF signaling in diabetic rats. However, no significant changes in CD2AP or podocin expression were observed in diabetic rats compared with normal controls (supplementary Fig. 6B and Table 1). Given these results, it can be concluded that overactive Notch/VEGF signaling specifically modulates the expression of nephrin.

Previous studies have shown that an HG environment activated oxidative stress through Ras/Rac/extracellular signal–regulated kinase–dependent pathways in renal mesangial cells and ameliorated the deleterious effects of proteinuria in diabetic nephropathy by SOD (29,40). Crosstalk between reactive oxygen radicals and Wnt/ß-catenin signaling in modulating the apoptosis of HG-stressed mesangial cells was also observed (40). In the present study, superoxide may also serve as a Notch1 signaling modulator, thereby regulating VEGF-mediated nephrin expression in diabetic nephropathy (Fig. 2 and supplementary Fig. 7). Therefore, at least two pathways, including Wnt/ß-catenin and Notch-1, that are controlled by redox reactions contribute to diabetic nephropathy.

Taken together, we provide evidence that HG augments Notch/VEGF signaling components. Modulation of the Notch-1 signaling pathway may provide a therapeutic strategy for controlling the deleterious effects of diabetic nephropathy.

This work was supported in part by grants from the National Science Council of Taiwan (NMRPG180661 to C.-L.L.) and Chang Gung Memorial Hospital at Chiayi, Chiayi, Taiwan (CMRPG680351 to C.-L.L.).

We thank the Prevision Medical Corporation of Taiwan for technical assistance. No other potential conflicts of interest relevant to this article were reported.

C.-L.L. researched data and wrote, reviewed, and edited the manuscript. F.-S.W. researched data and contributed to discussion. Y.-C.H. researched data and contributed to discussion. C.-N.C. contributed to discussion. M.-J.T. researched data and contributed to discussion. M.S. researched data and contributed to discussion. P.-J.C. researched data and wrote, reviewed, and edited the manuscript. J.-Y.W. contributed to discussion.

We thank Professor Karl Tryggvason of the Karolinska Institute, Stockholm, Sweden, for kindly providing the nephrin monoclonal antibody.