Renal Fibrosis and Glomerulosclerosis in a New Mouse Model of Diabetic Nephropathy and Its Regression by Bone Morphogenic Protein-7 and Advanced Glycation End Product Inhibitors

We purchased 7-week-old C57BL/6, 129Sv, and CD1 mice from Charles River (Wilmington, MA). These mice received standard chow and water. Blood samples were collected between 12 and 1 p.m., and plasma glucose was measured by the glucose oxidase method.

Diabetes was induced in 8-week-old C57BL/6, 129Sv, and CD1 mice by an injection of STZ. CD1 and C57BL/6 mice were made diabetic by single intraperitoneal injection of STZ at 200 mg/kg in 10 mmol/l citrate buffer (pH 4.5). For 129Sv mice, 50 mg/kg of STZ was injected intraperitoneally for 5 consecutive days. Citrate buffer was injected as a control arm of the experiments. At 3 days after the STZ injection, diabetes was confirmed by a urine dipstick method, and mice were divided into experimental study groups. In each of the mouse strains, mortality was ∼30% with a peak at 3.5 months. Diabetic and nondiabetic CD1 mice were killed at 3.5 and 6 months after the injection of STZ. Diabetic and nondiabetic C57BL/6 and 129Sv mice were killed at 6 months after the STZ injection. Diabetic CD1 mice were randomly divided into four groups and treated with the following three compounds, starting 1 month after STZ injection for 5 months thereafter. Aminoguanidine (n = 8; Sigma, St. Louis, MO) and pyridoxamine dihydrochloride (n = 6; Sigma) were dissolved in drinking water at a concentration of 1 and 2g/l, respectively (7,8). rhBMP-7 was injected into four diabetic mice at a dose of 300 μg/kg every other day (9). Seven diabetic mice injected with vehicle buffer every other day served as a control. On death at 6 months, renal tissues were evaluated by morphometric analysis for mesangial expansion, interstitial fibrosis, tubular atrophy, and interstitial fibroblast accumulation. In all groups (treated and untreated), the mortality was ∼30% between 2 and 4 months after the single STZ injection. Deaths before 2 months were not noted in any groups. All mouse studies were reviewed and approved by the institutional animal care and use committee.

Spot urine samples were collected and urinary albumin and creatinine concentration estimated using a bromocresol green–based colorimetric assay according to the manufacturer's recommendation (Sigma). Urine albumin excretion was estimated as the quotient of urine albumin and urine creatinine as described in our previous publications (10,11). Clinical chemistry analysis of plasma was performed by Antech Diagnostics (Boston, MA).

Renal cross sections were fixed in 4% paraformaldehyde, embedded in paraffin, and deparaffinized in xylene, and then 4-μm sections were stained with hematoxylin-eosin, periodic acid Schiff (PAS), and Masson trichrome. The extent of renal injury was assessed by morphometric analysis of the glomerular disease, tubular damage, and interstitial fibrosis as previously described (12). For glomerular damage, we evaluated mesangial expansion and enlargement of the glomeruli. A point-counting method was used to quantify mesangial matrix deposition according to a previous method with some modifications (13). We analyzed 20 PAS-stained glomeruli from each mouse on a digital microscope screen grid containing 667 (29 × 23) points. The number of grid points that hit pink or red mesangial matrix deposition were divided by the total number of points in the glomerulus to obtain the percentage of mesangial matrix deposition (mesangial matrix index) in a given glomerulus. The relative interstitial volume was evaluated by morphometric analysis using a 10-μm graticule fitted into the eyepiece of the microscope. We evaluated 10 randomly selected cortical areas under 200× magnification for each mouse. Tubules were evaluated for their widened lumen, atrophy, or thickened basement membranes to estimate the percentage of damaged tubules (% tubular damage) (9).

Indirect immunofluorescence studies were performed as described previously (14). Briefly, 4-μm cryosections were fixed in acetone for 5 min at 4°C. To block the nonspecific antibody bindings, 1% BSA in 10 mmol/l PBS were applied for 20 min at room temperature. As primary antibodies, we used a rabbit antibody to S100-A4 as a fibroblast marker (Dako, Carpinteria, CA), goat anti–collagen III antibody (Southern Biotechnology Associates, Birmingham, AL), and goat anti–transforming growth factor (TGF)-β1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). As secondary antibodies, fluorescein isothiocyanate–conjugated anti-goat IgG and anti-rabbit IgG were obtained from Jackson ImmunoResearch (West Grove, PA). The primary antibodies were applied for 1 h at room temperature followed by incubation with secondary antibodies. For AGE staining, 4% paraformaldehyde–fixed paraffin-embedded sections were stained with a CSAII biotin–free tyramide signal amplification system according to the manufacturer's recommendations (Dako). Primary antibodies against carboxymethyllysine (CML; Wako Chemical, Richmond, CA) were used.

After frozen renal sections were labeled using antibodies specific to TGF-β1, collagen III, or CML, photomicrographs from a mouse kidney cortex were taken using 100× magnification under an Axioscope 2 Plus fluorescence microscope (Carl Zeiss Micro Imaging, Thornwood, NY). The cortical TGF-β1–positive tubular areas were measured by Carl Zeiss Axiovision digital imaging software. The TGF-β1–positive area was divided by the total cortical area to obtain the percentage of TGF-β1–positive area. We evaluated ∼300–600 tubules in the cortex for each mouse kidney. For quantification of collagen III and CML, five 100× magnification pictures from each mouse were assessed with National Institutes of Health Image J software.

Values are the means ± SE. The significance of the differences between the two groups was analyzed by Student's t test. Comparisons among three groups were performed by two-way ANOVA followed by Sheffe's test to evaluate the significance of the differences between any two groups. A level of P < 0.05 was defined as statistically significant.

To establish a model of STZ-induced type 1 diabetes, a single injection of STZ is used to elicit diabetes secondary to its toxicity to the pancreatic β-cells (4,15). Eight-week-old C57BL/6, 129Sv, and CD1 mice were made diabetic by intraperitoneal injection of STZ. All diabetic mice were severely hyperglycemic at the time of death, and plasma glucose levels were elevated in all mouse strains (Fig. 1A). Urine protein excretion (estimated as the urine albumin–to–urine creatinine ratio) was evaluated at the time of death, and all three strains of mice had significant proteinuria (Fig. 1B), similar to humans with diabetic nephropathy.

In human diabetic nephropathy, renal histopathologic findings include diffuse thickening of the glomerular basement membrane, prominent mesangial expansion, nodular lesions in the periphery of glomerular tufts, and tubulointerstitial fibrosis (16–18). Glomerular morphometric analysis demonstrates that diabetic C57BL/6, 129Sv, and CD1 mice develop glomerular hypertrophy, confirming previous studies (19,20) (Fig. 1C). CD1 mice exhibited the most prominent glomerular lesions and mesangial matrix deposition compared with C57BL/6 and 129Sv mice with diabetes. Control experiments in which kidney sections were labeled with antibodies specific to CML revealed that in each mouse strain AGE deposition was significantly increased in the STZ-injected diabetic mice. AGE levels did not differ significantly between the diabetic mouse strains (Fig. 1K–Q).

Tubular injury associated with diabetic nephropathy in humans is characterized by widened lumina, flattened tubular cells, and thickened tubular basement membranes. Tubular injury was present in all diabetic mouse strains (Fig. 2A–G). C57BL/6 mice and 129Sv mice did not develop interstitial fibrosis, again confirming previous studies (4). Diabetic CD1 mice, however, exhibited significant interstitial fibrosis at 6 months of diabetes (Figs. 2H and N). Analysis of kidneys from diabetic CD1 mice collected at earlier time points (3.5 months) revealed mild tubulointerstitial fibrosis (data not shown) (21).

After 6 months of diabetes caused by STZ injection, significant widening of the interstitial space associated with collagen accumulation was observed (Fig. 3A). Tubular damage included a loss of eosinophilic pink staining in the tubular cytoplasm (reminiscent of the Armanni-Ebstein anomaly observed in human diabetic nephropathy patients) (Fig. 3B) and atrophy of renal tubules, preferentially restricted to the cortical regions (Fig. 3C). The most frequently seen tubular damage was cyst-like enlargement of tubular lumen (Fig. 3C). Diabetic CD1 mice presented with a diffuse glomerulosclerosis with minimal focal glomerulosclerosis (Fig. 3D). Increased interstitial volume in diabetic CD1 mice was associated with increased accumulation of fibroblasts, as evaluated by labeling with antibodies to fibroblast-specific protein 1 (FSP1) (Fig. 3E and F). TGF-β1 is a mediator of progressive renal disease in experimental diabetic mice (22,23). Immunolabeling using specific antibodies to TGF-β1 revealed a significant increase of TGF-β1 in the tubulointerstitium of diabetic CD1 mice compared with nondiabetic control mice (Fig. 3H and I). Tubulointerstitial fibrosis in CD1 mice correlated with decreased renal excretory function in diabetic CD1 mice (Table 1). Plasma creatinine levels increased from 0.58 mg/dl (in nondiabetic control mice) to 1.6 ± 0.17 mg/dl (in diabetic CD1 mice) after 6 months of STZ-induced diabetes (Table 1). Increased plasma creatinine levels were associated with increased potassium and phosphate levels (Table 1).

Diabetic CD1 mice develop chronic renal disease associated with tubulointerstitial fibrosis associated with decreased excretory renal function. Therefore, we next tested the efficacy of three different experimental therapeutics to inhibit the progression of renal disease in these mice. AGEs (products of the nonenzymatic reaction of amino groups in proteins and lipids with reducing sugars such as glucose) are considered to be important for the progression of diabetic nephropathy. Here, we used two different inhibitors of AGE accumulation, aminoguanidine and pyridoxamine, to address their therapeutic benefit. Pyridoxamine is shown to inhibit the accumulation of AGE by inhibiting conversion of Amadori intermediate to AGEs through binding of redox metal ions (24,25). Aminoguanidine has been suggested to facilitate cross-link cleavage of AGEs, leaving uncomplexed AGE, and facilitating excretion by the kidney (22).

BMP-7 (rhBMP-7) is a growth factor of the TGF-β superfamily, and several independent studies demonstrate that administration of rhBMP-7 inhibits progression of renal fibrosis in animal models for chronic renal disease (9,12,26,27). Antifibrotic action of BMP-7 is mediated via direct antagonism of TGF-β1 signaling (12,28,29). Previous studies suggest a therapeutic benefit for rhBMP-7 in a rat model of STZ-induced diabetic glomerulosclerosis (which do not develop interstitial fibrosis) (30).

In our study, aminoguanidine and pyridoxamine inhibited glomerular hypertrophy and mesangial matrix expansion (Fig. 4), confirming that AGEs are involved in the progression of glomerular lesions associated with diabetic nephropathy. Efficacy of aminoguanidine and pyridoxamine in these studies further confirms that renal fibrosis observed in diabetic CD1 mice is not caused by nonspecific toxicity of STZ (4). Administration of rhBMP-7 significantly inhibited glomerular hypertrophy but had an insignificant effect on mesangial matrix expansion and proteinuria (Fig. 4). However, administration of rhBMP-7 was most effective in inhibiting tubulointerstitial fibrosis (Figs. 5 and 6). Mechanisms of action for AGE inhibitors and rhBMP-7 were confirmed by immunolabeling using antibodies specific for the AGE CML. Although both pyridoxamine and aminoguanidine significantly decreased CML accumulation, administration of rhBMP-7 had no significant impact (Fig. 7). Body weight at the conclusion of the studies (STZ 30.0 ± 0.9 g, STZ-aminoguanidine 28.9 ± 0.7 g, STZ-pyridoxamine 28.4 ± 1.0 g, and STZ–rhBMP-7 29.5 ± 0.5 g) did not differ, confirming that none of the treatment modalities affected the diabetic metabolism in mice per se. In all groups (treated and untreated), the mortality was ∼30% between 2 and 4 months after the single STZ injection. Deaths before 2 months were not noted in any groups, ruling out any acute toxicity. Some previous studies have suggested that the STZ dose administered in our study might be toxic. In this regard, we believe that the outbred CD1 mice might be more resilient than the inbred strain of mice. We in fact found more toxicity when C57BL/6 mice were used. The outbred CD1 mice are larger than the inbred mice at the same age and thus might be able to withstand the effects of STZ much better than the inbred C57BL/6 mice.

Diabetic nephropathy is a leading cause of ESRD worldwide. Progress in evaluating the underlying pathogenic pathways and development of new therapeutic strategies has been impaired because of the lack of representative animal models mimicking human diabetic nephropathy. Here, we report that CD1 mice developed chronic renal disease associated with tubulointerstitial fibrosis and decreased renal excretory function within 6 months after a single injection of STZ. Interestingly, C57BL/6 and 129Sv mouse strains did not develop interstitial fibrosis, but they exhibited prominent glomerular lesions, increased blood glucose levels, and albuminuria, comparable to STZ-induced diabetic CD1 mice. Although our study does not address the mechanistic reason behind such differences, they reiterate the importance of genetic background in diabetic nephropathy progression. Among the advantages of this new mouse model are that it recapitulates several pathologies associated with human diabetic nephropathy.

We further provide evidence that each of the experimental therapeutics—aminoguanidine, pyridoxamine, and rhBMP-7—are effective in inhibiting the progression of renal disease in STZ-induced diabetic CD1 mice. Aminoguanidine and pyridoxamine inhibit glomerular disease in STZ-induced diabetic CD1 mice, whereas rhBMP-7 inhibits predominantly tubulointerstitial fibrosis. Our results demonstrate that pyridoxamine does not offer any advantage over aminoguanidine in inhibiting diabetic nephropathy. In addition to inhibition of AGE formation, aminoguanidine is known as an effective free radical scavenger (31), and pyridoxamine has the capacity to scavenge toxic carbonyl products (32). Thus, whereas both aminoguanidine and pyridoxamine inhibit AGE accumulation, additional beneficial properties cannot be excluded. Our findings suggest that combination of an AGE inhibitor and BMP-7 can provide synergistic effects in the clinic and argue for combining them in human clinical trials.

This study was supported in part by research grants DK62987 and DK55001 from the National Institutes of Health, a research fund from the Beth Israel Deaconess Medical Center for the Division of Matrix Biology, and a Stop and Shop Pediatric Tumor Foundation fellowship from Dana-Farber Cancer Institute (DFCI) (to H.S.). M.Z. was funded by a grant from the National Institutes of Health (5K08DK074558-01) and a American Society of Nephrology Carl W. Gottschalk Award.