HB-EGF expression levels were higher in both eNOS^−/− and eNOS^−/−/Lepr^db/db mouse kidneys at 8 weeks of age, indicated by the 25 kDa band corresponding to pro-HB-EGF (Figure 1a and b), while minimal HB-EGF expression was detected in Lepr^db/db or wild type (wt) mouse kidneys of the same age. Further increases in HB-EGF expression level were seen in eNOS^−/−/Lepr^db/db mice at 16 weeks of age (Figure 1c and d) and later (data not shown). By 16 weeks of age, a 29 kDa higher molecular weight band was also detectable in some of the eNOS^−/−/Lepr^db/db mice, consistent with expression of increased glycosylated forms of membrane-associated pro-HB-EGF.²³ Meanwhile, increased EGFR activation were detected in kidneys of diabetic, eNOS^−/− and eNOS^−/−/Lepr^db/db mice compared to wild type mice (Figure 1e and f).

Since pro-HB-EGF can be cleaved to release soluble HB-EGF, we determined whether renal HB-EGF was released into urine or serum. We failed to detect increased HB-EGF levels in the serum in either group (data not shown). However, higher amounts of HB-EGF were found in eNOS^−/−/Lepr^db/db urine compared to other groups. HB-EGF excretion increased as early as 8 weeks of age (Figure 2a and c), and further increased at 16 weeks of age, with the 16 to 22 kDa molecular weight species corresponding to secreted HB-EGF (Figure 2b and d). In Lepr^db/db or eNOS^−/− non-diabetic mice, only trace or low levels of HB-EGF were detected in urine at 16 weeks of age (Figure 2b and d).

Glomeruli from control mice had minimal staining for HB-EGF. Lepr^db/db, eNOS^−/− and eNOS^−/−/Lepr^db/db all had detectable HB-EGF staining in the glomeruli, with the highest levels seen in eNOS^−/−/Lepr^db/db mice. In arteries, HB-EGF staining was not detected in normal endothelial cells or smooth muscle cells (SMC). Although only weak staining was observed in the endothelial cells and SMC of Lepr^db/db mice, strong HB-EGF staining was detected in endothelial cells of both eNOS^−/− and eNOS^−/−/Lepr^db/db mice, and eNOS^−/−/Lepr^db/db mice also had strong HB-EGF immunoreactivity in SMC (Figure 3a). In situ hybridization indicated strong HB-EGF mRNA expression in both glomeruli and tubular epithelial cells of eNOS^−/−/Lepr^db/db mouse kidneys compared to other groups (Figure 3b).

In glomeruli of Lepr^db/db, eNOS^−/−, and eNOS^−/−/Lepr^db/db mice, HB-EGF was detected primarily in podocytes, with lower expression levels in endothelial cells as indicated by double labeling of anti-HB-EGF with anti-podocin (a marker for podocyte) and anti-CD31 (a marker for endothelial cells) antibodies (Figure 3c and d).

In medulla, strong HB-EGF immunoreactivity was detected in tubules of both eNOS^−/− and eNOS^−/−/Lepr^db/db mouse kidneys, with less staining in Lepr^db/db or wild type mouse kidneys (Figure 4a). Tubular-segment specific marker proteins indicated that HB-EGF localized to the apical membrane of the thick ascending limb and collecting duct epithelial cells without expression in proximal tubules (Figure 4b). Similar localization was observed by in situ hybridization (Figure 4c). There was minimal HB-EGF mRNA staining in the tubules of the wild type mouse kidneys.

The increased HB-EGF expression seen in eNOS^−/−/Lepr^db/db mice kidneys suggested that decreased NO bioavailability may contribute to the increased HB-EGF levels. GENC were cultured for 48 hours in medium containing normal glucose (NG, 5.5 mM glucose), high glucose (HG, 30 mM glucose), mannitol (30 mM, used as a high osmolarity control for HG group), or the NOS inhibitors L-NAME (500 μM) or L-NIO (100 μM) with or without HG. Both L-NAME and L-NIO increased HB-EGF protein expression and EGFR phosphorylation after 48 hours, without any further augmented by administration of high glucose (Figure 5a, b, d and e). Since in vivo studies indicated that the highest HB-EGF expression was seen in eNOS^−/−/Lepr^db/db mouse kidneys compared to either eNOS^−/− or Lepr^db/db mouse kidneys, these in vitro studies suggested that in eNOS^−/−/Lepr^db/db mice, high glucose-related metabolic products rather than high glucose itself may contribute to the increased HB-EGF levels.

Up-regulation of HB-EGF expression by NO deficiency was also seen in another endothelial cell type. We detected much higher HB-EGF levels in cultured pulmonary endothelial cells (PEC) without eNOS expression compared to PEC with eNOS expression (cells provided by Mark de Caestecker, Vanderbilt University) (Figure 5c).

L-NAME was administered to 12-week old Lepr^db/db and wild type mice for 2 weeks. Systolic blood pressure (SBP) began to increase by 7 days of L-NAME injection (Figure 6a). There were no significant differences in blood pressure between Lepr^db/db and wild type mice after L-NAME treatment. Albuminuria significantly increased in Lepr^db/db mice treated with L-NAME, with increases after 1 week treatment, and further increases at 2 weeks treatment, while in the non-diabetic wild type mice, ACR remained unchanged after L-NAME treatment (Figure 6b). In Lepr^db/db mice, L-NAME administration increased renal HB-EGF expression (Figure 6c and e). HB-EGF expression was not increased in the L-NAME treated non-diabetic wild type group in the observed period (data not shown). In addition, after L-NAME treatment, urinary HB-EGF excretion from the Lepr^db/db mice was much higher compared to vehicle group, or either L-NAME or vehicle-treated wild type mice (Figure 6d and f).

In further studies, we determined the effects of NO replenishment in eNOS^−/−/Lepr^db/db mice. For this purpose, NaNO3 was added to the drinking water, and renal pathology and HB-EGF expression were evaluated. Studies were begun at 8 weeks of age, at a time when eNOS^−/−/Lepr^db/db mice already have increased albuminuria (24 h ACR: 487 ± 120 vs. 44 ± 3.3 in wild type mice). Although the degree of hyperglycemia was significantly lower in the NaNO3-treated group, there were no statistically significant differences in body weight or SBP (Table 1). The total NOx levels in both plasma and urine were significantly increased in NaNO3-treated mice versus control (Figure 7a and b), while the renal cGMP content did not show significant difference between NaNO3-treated mice and controls (2.67 ± 0.35 and 2.52 ± 0.33 pmol/mg protein, n = 8 in each group).

Compared to control, NaNO3 treatment markedly inhibited albuminuria. After 4 weeks of NaNO3 treatment, albuminuria was essentially stable for the subsequent 12 weeks (Figure 7c). After 16 weeks of NaNO3 administration, significant differences in renal morphology were observed. Compared to control mice of the same age, the NaNO3 treatment group had less glomerulosclerosis as shown by periodic acid Schiff (PAS) staining and assessed by sclerosis index (SI, 0 to 4 score, Figure 7d and e). The SI of the NaNO3 treatment group (1.40 ± 0.31) was significantly lower than the control group (2.32 ± 0.16, P < 0.05) (Figure 7e). Interstitial fibrosis was assessed by sirius red staining and found to be significantly reduced after NaNO3 treatment compared to control group (Figure 7d and f). Those results suggested that NO replenishment to the eNOS^−/−/Lepr^db/db mice impeded overall renal deterioration.

To investigate the effect of NO replenishment on HB-EGF expression, we performed in situ hybridization. In 24-week old eNOS^−/−/Lepr^db/db mice (control), HB-EGF mRNA was extensively expressed in the glomeruli and tubules, and this expression was significantly reduced after 16 weeks NaNO3 treatment (Figure 8a). Kidney HB-EGF protein expression levels were significantly less in the NaNO3 treatment group than in the control group (Figure 8b and d). Urinary HB-EGF excretion levels gradually decreased after NaNO3 treatment and were significantly lower after 12 weeks treatment (Figure 8c and e).

Specific HB-EGF deletion in endothelium was confirmed by X-gal and HB-EGF staining at both 2 weeks and 20 weeks after tamoxifen injection (Figure 9a and b). In kidneys of 28-week old mice, glomerular sclerosis can be easily seen in diabetic eNOS^−/− mice compared to eNOS^−/−/HB^endo−/− mice (Figure 9c). Renal injury was significantly reduced in diabetic eNOS^−/−/HB^endo−/− mice compared to diabetic eNOS^−/−/HB^lox/lox mice as assessed by 24-hour urine albumin/creatinine ratio and glomerular injury score (Figure 9d and e).

D-(+)-glucose, D-mannitol, SIGMAFAST^™ 3,3′-diaminobenzidine tablets, paraformaldehyde, streptozotocin, tamoxifen, rabbit anti-human podocin antibody, goat anti-mouse EGFR antibody, and anti-β-actin antibody were purchased from Sigma-Aldrich (St. Louis, MO); N^G-Nitro-L-arginine methyl Ester (L-NAME) and L-N5-(1-Iminoethyl) ornithine (L-NIO) were from EMD4Biosciences (Gibbstown, NJ). Heparin sepharose (HiTrap Heparin HP) was from GE Healthcare (Piscataway, NJ). Rabbit anti-rat HB-EGF polyclonal antibody was a gift from Lili Feng (Dept. of Immunology, The Scripps Research Institute, La Jolla, CA). Mouse anti-human HB-EGF and phospho-EGF receptor (Tyr1068) antibody were from Santa Cruz Biotechnology (Santa Cruz, CA).

Other antibodies used in this study were rat anti-mouse CD31 (BD Pharmingen, Bedford, MA) and rabbit anti-human Tamm-Horsfall glycoprotein (THP) (Biomedical Technologies Inc. Stoughton, MA). Lotus tetragonolobus lectin (LTL), dolichos biflorus agglutinin (DBA), biotinylated goat anti-rabbit IgG, and avidin-biotin complex (ABC) kits were from VectorLabs (Burlingame, CA).

eNOS^−/− mice on the C57BL/6J background, db heterozygous (Lepr^db/wt) mice on the C57BLKS/J (BKS) background and wild type BKS mice were purchased from The Jackson Laboratory (Bar Harbor, ME). eNOS^−/− mice were backcrossed for 10 generations to the BKS background and then crossed with Lepr^db/wt mice to generate eNOS^−/−/Lepr^db/db mice as previously described.⁴ Genotypes were confirmed by PCR.

Mice with the HB-EGF gene flanked by loxP sites (HB^lox/lox) were kindly provided by Dr. Eisuke Mekada (Department of Cell Biology, Research Institute for Microbial diseases, Osaka University, Japan).²⁸ The targeting vector contained the lacZ gene as a reporter for the expression of HB-EGF. When HB-EGF cDNA is deleted by cre-recombinse, HB-EGF expression can be identified by X-gal staining in HB^+/− or HB^−/− mice. Homozygous HB^lox/lox mice were bred with eNOS^−/− to generate eNOS^−/−/HB^lox/lox mice, which were further bred with mice that carry a tamoxifen-inducible Cre-ER(T) recombinase driven by the 5′ endothelial enhancer of the stem cell leukemia (SCL) locus (endothelial-SCL-Cre-ER(T), provided by Jimmy Hao, Vanderbilt University). 8-week old male littermates of eNOS^−/−/HB^lox/lox and eNOS^−/−/HB^lox/lox/Cre(+) (eNOS^−/−/HB^endo−/−) mice were treated with STZ 50 mg/kg/day for five consecutive days to induce diabetes, and 2 mg tamoxifen was administered daily for five consecutive days to activate cre recombinase. Development of diabetes was evaluated by fasting blood glucose two weeks after the first injection, and those with blood glucose levels higher than 300 mg/dl were used for study. HB-EGF deletion was confirmed by X-gal staining and HB-EGF immunostaining. All the mice used in this part of the study were on C57BL/6 background. All animal experiments were performed under approval of the Institutional Animal Care and Use Committee at Vanderbilt University.

SBP was measured in conscious trained mice at room temperature using a tail-cuff monitor (BP-2000 BP Analysis system, Visitech System, NC). Mice were subjected to 10 minutes per day of acclimation to the system under restrainers with tail cuffs for 4 days. On the fifth day, systolic blood pressure measurements were performed.

Urine from individual mice was collected in metabolic cages (Braintree Scientific, Braintree, MA) over 24 hours. Urine samples were centrifuged at 3,000 g for 5 minutes, and supernatants were stored in aliquots at −80°C until assay. Urinary albumin was determined using the Albuwell M Murine Microalbuminuria enzyme-linked immunosorbent assay kit (Exocell Inc., Philadelphia, PA). Urinary creatinine was measured using the Creatinine Companion kit (Exocell Inc.).

Excreted urinary HB-EGF was measured by a modification of previously described heparin affinity chromatography. Briefly, equal volumes of urine from 24-hour collections were added to heparin sepharose, incubated at 4°C overnight, washed 3 times with 10 mM phosphate buffer (pH 7.4), re-suspended, boiled in protein sample buffer before separation on 18% SDS-polyacrylamide gels, and followed by immunoblotting with an anti-HB-EGF antibody.

GENC were provided by Dr. Michael Madaio, University of Pennsylvania, and condition. Cells were were cultured at 37°C with 5% CO ₂ in a humidified airgrown to 80% confluence, starved for 24 hours in serum-free medium containing 0.1% bovine serum albumin (BSA), followed by incubation with mannitol (30 mM), D-Glucose (30 mM), L-NAME (500 μM), or L-NIO (100 μM) for 48 hours before harvesting for immunoblot analysis as described below.

12-week old male Lepr^db/db mice and non-diabetic BKS mice were given L-NAME (50 mg/kg) subcutaneously every day for two weeks. Sham groups were give equivalent amounts of normal saline. SBP was measured to monitor the effects of L-NAME.

NaNO3 was added to the drinking water of eNOS^−/−/Lepr^db/db mice at a concentration of 85 mg/L (1 mM) beginning at 8 weeks of age for the treatment group and regular water for the control group. Urinary HB-EGF excretion levels and 24-hour urine albumin/creatinine ratios (ACR) were monitored at 4, 8, 12 and 16 weeks of treatment. Mice were sacrificed after 16 weeks of treatment and renal HB-EGF expression and pathological changes were examined. Plasma and urine nitrate/nitrite concentrations were determined using the Greiss reaction by measuring combined oxidation products of NO, nitrate (NO ₃⁻) and nitrite (NO ₂ ⁻), after reduction with nitrate reductase using a colorimetric assay (Cayman Inc., Ann Arbor, Michigan, USA). Determination of cGMP content was performed in acetylated samples using an immunoassay kit (Cayman Chemical Company, Ann Arbor, MI). cGMP levels were normalized to kidney protein content.

Mice were euthanized at indicated ages and were perfused with ice cold normal saline, followed by 4% paraformaldehyde. Kidneys were removed and fixed overnight in 4% paraformaldehyde at 4°C, and 4 Lm thick sections were stained with PAS, sirius red or used for immunostaining. Alternatively, tissues were fixed in 4% paraformaldehyde for 2 hours at 4°C, washed with PBS, and put into 30% sucrose overnight at 4°C and then immersed into OCT for cryosectioning. Immunostaining was performed as previously described.²⁹

Semiquantitative glomerulular injury and sclerosis index (0 to 4 score) was used to evaluate the degree of glomerulosclerosis, as previously prescribed.^{30, 31} Tubular interstitial fibrosis was shown by sirius red staining and assessed by sirius red positive area. All sections were examined without knowledge of the treatment protocol.

A quarter of the kidney was snap-frozen and stored at −80°C until use. Frozen samples were homogenized and lysed in 0.5 ml ice-cold TGH buffer (1% Triton X-100, 10% glycerol, 20 mM HEPES, pH 7.2, 100 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 1 mM Na ₃VO ₄). Western blots were conducted as previously described. ²⁹ Membranes were blotted using antibodies against HB-EGF, EGFR or phospho-EGFR.

Probes for HB-EGF were obtained by subcloning a 661-bp coding region of the mouse HB-EGF cDNA (NM 010415.2) into pCRII-TOPO vector (Invitrogen) and transcribing in vitro to produce antisense and sense riboprobes labeled with digoxigenin using DIG RNA labeling kit (Roche). In situ hybridization was performed as previously described.³²

Data were evaluated using unpaired t-test for two-group comparison or one way ANOVA followed by Tukey-Krymer test for multiple comparisons and expressed as means ± SEM. Values of P < 0.05 were considered statistically significant. The effects of NaNO3 on metabolic and physiologic levels were examined using multivariable generalized least square regression analysis with adjustment for corresponding baseline levels (Table 1). Statistical analyses were performed using R version 2.15.1 (http://www.r-project.org). A two-sided significance level of 5% was required for consideration as statistical significant.