Ciglitazone, a PPARγ agonist, ameliorates diabetic nephropathy in part through homocysteine clearance

Abstract

Diabetes and hyperhomocysteinemia (HHcy) are two independent risk factors for glomeruloslerosis and renal insufficiency. Although PPARγ agonists such as ciglitazone (CZ) are known to modulate diabetic nephropathy, the role of CZ in diabetes-associated HHcy and renopathy is incompletely defined. We tested the hypothesis that induction of PPARγ by CZ decreases tissue Hcy level; this provides a protective role against diabetic nephropathy. C57BL/6J mice were administered alloxan to create diabetes. Mice were grouped to 0, 1, 10, 12, and 16 wk of treatment; only 12- and 16-wk animals received CZ in drinking water after a 10-wk alloxan treatment. In diabetes, PPARγ cDNA, mRNA, and protein expression were repressed, whereas an increase in plasma and glomerular Hcy levels was observed. CZ normalized PPARγ mRNA and protein expression and glomerular level of Hcy, whereas plasma level of Hcy remained unchanged. GFR was dramatically increased at 1-wk diabetic induction, followed by hypofiltration at 10 wk, and was normalized by CZ treatment. This result corroborated with glomerular and preglomerular arteriole histology. A steady-state increase of RVR in diabetic mice became normal with CZ treatment. CZ ameliorated decrease bioavailability of NO in the diabetic animal. Glomerular MMP-2 and MMP-9 activities as well as TIMP-1 expression were increased robustly in diabetic mice and normalized with CZ treatment. Interestingly, TIMP-4 expression was opposite to that of TIMP-1 in diabetic and CZ-treated groups. These results suggested that diabetic nephropathy exacerbated glomerular tissue level of Hcy, and this caused further deterioration of glomerulus. CZ, however, protected diabetic nephropathy in part by activating PPARγ and clearing glomerular tissue Hcy.

peroxisome proliferator-activated receptor-γ (PPARγ) is a member of nuclear hormone receptor superfamily of ligand-activated transcription factors. It plays important roles in maintaining glucose and lipid homeostasis (39). The PPARγ agonists, such as ciglitazone (CZ), are agents that improve glycemic control by increasing insulin sensitivity (5). These agonists also mediate direct antiatherogenic effects in the diabetic vasculature that are independent of their metabolic actions (4). In the pathogenesis of diabetic vasculopathies, such as in glomerulosclerosis, downregulated PPARγ expression is associated with matrix accumulation and glomerulonephritis (23, 30, 39). Several studies have shown the efficacy of PPARγ agonists to ameliorate the progression of glomerulosclerosis (15) and have suggested that PPARγ ligands have a direct beneficial renal effect. However, the role of PPARγ in diabetic nephropathy and tissue levels of homocysteine (Hcy) in glomerulus remains unknown.

Hcy is a sulfur-containing amino acid, and hyperhomocysteinemia (HHcy) is an independent vascular risk factor; HHcy is also associated with kidney insufficiency (3, 38). Studies have shown that, in diabetic nephropathy, Hcy disposal and clearance are impaired, thereby increasing plasma and tissue levels of Hcy (32). Moreover, HHcy leads to renal microvascular impairment and vasoconstriction (26). This further leads to volume retention and accumulation of Hcy, causing a vicious cycle and chronic renal failure. This mechanism involves, at least in part, imbalances of the matrix metalloproteinase (MMP)-tissue inhibitor of matrix metalloproteinase (TIMP) axis that causes glomerular matrix remodeling.

MMPs are a class of enzymes that reside in the latent form and are activated by various physiological threats (35). MMPs are regulated by naturally occurring TIMPs. We have shown previously that increases in glomerular Hcy and activation of MMP-2 are associated with glomerulosclerosis (26). It is, however, unclear how MMPs and TIMPs are involved in glomerulosclerosis and whether PPARγ in part regulates these enzymes that modulate glomerular dysfunction in diabetic nephropathy. Therefore, we sought to determine the role of PPARγ in the glomerular tissue level of Hcy and renal dysfunction in alloxan-induced diabetic kidney in this study. Also, we determined the effect of the PPARγ agonist CZ in mediating structural and functional changes of glomerulus in this animal model.

MATERIALS AND METHODS

Animals and treatments.

Eight-week-old C57BL/6J male mice were obtained from Jackson Laboratory (Bar Harbor, ME) and maintained in the animal facility center at the University of Louisville. The mice were fed regular mice chow (category no. 5015; PMI LabDiet, St. Louis, MO) and treated with a single dose of alloxan (65 mg/kg body wt) intraperitoneally. Mice were grouped into five different groups, which are 0, 1, 10, 12, and 16 wk. All mice received alloxan treatment. The duration of treatment varied between the five groups. The 0-, 1-, and 10-wk mice received alloxan treatment for the respective time period. However, in the 12- and 16-wk groups, after 10 wk of alloxan treatment, mice were given CZ for another 2 and 6 wk and were killed at 12 (10 wk alloxan + 2 wk CZ) and 16 (10 wk alloxan + 6 wk CZ) wk, respectively.

CZ (EMD Chemicals, Gibbstown, NJ) administration consisted of 8 μg/ml in water until the end of the experiment. Because CZ is not stable in aqueous phase, the manufacturer does not recommend use of aqueous solution >24 h. Therefore, every 24 h we replaced the drinking water with freshly prepared CZ solution in the appropriate groups. It was estimated that adult mice have a blood volume of ∼2 ml and drink ∼5 ml of water/day. Therefore, each mouse ingested ∼1.6 mg·kg⁻¹·day⁻¹ CZ, considering the fact that each mouse weighed ∼25 g. This amount of ingested CZ was enough to saturate most binding sites on PPARγ (25). At the end of each experiment, mice were killed; blood and tissues were harvested immediately for analysis. All animal procedures were performed in accordance with National Institute of Health Guidelines for animal research and were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Louisville.

Antibodies and reagents.

Anti-TIMP-1, anti-β-actin antibodies, and other analytical reagents were from Sigma-Aldrich (St. Louis, MO). Rabbit polyclonal anti-PPARγ antibody and horseradish peroxidase-conjugated secondary antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Polyvinylidene difluoride (PVDF) membrane was purchased from Bio-Rad (Hercules, CA).

Gene array.

The cDNA gene expression array analyses were done as described previously (34).

RT-PCR.

For RT-PCR, total RNA from kidney samples was isolated using Trizol (Invitrogen), following the manufacturer's instructions, and reverse transcribed using oligo(dT) primers, with a total reaction volume of 20 μl. The reverse transcription program was 25°C for 10 min, 42°C for 50 min, and then 70°C for 15 min. PCR was performed using 2 μl of each RT product (cDNA), with a total reaction volume of 20 μl. The PPARγ primers for PCR used were forward primer (5′-CACCAGTGTGAATTACAGCA-3′) and reverse primer (5′-GGTGGAGATGCAGGTTCTAC-3′). The PCR thermal cycle was 94°C for 6 min, then 35 cycles at 94°C for 50 s, 60°C for 1 min, and 72°C for 1 min, and finally 72°C for 10 min, which gives a product of ∼320 bp. Primers for GAPDH (116-bp product) were sense (5-GATGCAGGGATGATGTTCTG-3) and antisense (5-ACAACTTTGGCATTGTGGAA-3). The PCR thermal cycle for GAPDH was 94°C for 2 min, then 30 cycles at 94°C for 30 s, 57°C for 30 s, and 72°C for 1 min, and finally 72°C for 2 min. All of the primers were obtained from Invitrogen (Carlsbad, CA).

Western blot.

For immunoblot analysis, glomeruli were isolated (26) and glomerular protein was extracted as described in Gelatin zymography, containing proteinase inhibitors, PMSF (1 mM phenylmethylsulphonyl fluoride), aprotinin, leupeptin and pepstatin (1 μg/ml) each. Equal amounts of protein for each of the samples were separated by SDS-PAGE, transferred to PVDF membrane, and detected immunochemically as described earlier (29).

Glucose and Hcy measurement.

At the end of each experiment, animals were anesthetized with 2,2,2-tribromoethanol (0.5 ml/25 g body wt), and blood was collected from inferior vena cava in a heparinized syringe (1:9, heparin/blood). Plasma, separated from blood and glucose, was measured using a Bio-Rad glucose measurement kit. Plasma total Hcy was measured, following a previously published procedure (35).

Glomerular fitration rate measurement.

To determine the functional status of glomerulus, glomerular fitration rate (GFR) was measured as described previously (26). Briefly, a bolus dose (0.3 mg/25 g body wt ip) of FITC-inulin (Sigma) was injected, and mice were kept in metabolic cages individually. After 24 h, urine from metabolic cages and blood were collected. The levels of inulin in blood and urine were measured from a standard curve generated by known concentrations of FITC-inulin. In the spectrofluorometric measurement, excitation and emission were measured with the following wavelengths: excitation 488 nm and emission 530 nm, respectively.

Renovascular resistance.

A previously adopted method for renovascular resistance (RVR) measurement was used (26). Briefly, in 2,2,2-tribromoethanol, (100 mg/kg bw ip), anesthetized mouse renal artery blood pressure was measured with an indwelling PE catheter and analyzed in customized Micro-Med software. Renal artery blood flow was measured by a transonic probe. RVR was measured by renal artery pressure (mmHg) divided by blood flow (ml/min) and expressed as mmHg·ml⁻¹·min⁻¹.

Renal histology.

Frozen kidney tissue was cryosectioned into 6-μm-thick sections and stained with trichrome. Glomerulosclerosis was identified by thickened basement membrane, as described earlier (26).

Glomerular nitric oxide and Hcy.

Glomerular tissue was homogenized in Tris·HCl (50 mM, pH 7.4) buffer, and homogenate extract was used to measure total tissue level of Hcy, as described previously (35). Nitric oxide (NO) was measured by commercially available kit (R & D Systems, Minneapolis, MN).

Gelatin zymography.

Protein from glomerular tissue was extracted for gelatin zymography, as described previously (20). In brief, glomerular tissues were minced into small pieces in ice-cold extraction buffer (1:3 wt/vol) containing (in mmol/l) 10 cacodylic acid, 20 ZnCl, 1.5 NaN₃, and 0.01% Triton X-100 (pH 5.0) and incubated overnight at 4°C with gentle shaking. The homogenate was then centrifuged for 10 min at 800 g, and supernatant was collected. Protein concentration in the sample was measured using Bradford method, and 100 μg of the protein was electrophoretically resolved for each sample in 8% SDS-PAGE containing 1.5% gelatin as MMP substrate. Gels were washed in 2.5% Triton X-100 for 30 min to remove SDS, rinsed in water, and incubated for ≥24 h in activation buffer (50 mmol/l Tris·HCl, 5 mmol/l CaCl₂, and 0.02% NaN₃, pH 7.5) at 37°C in a water bath with gentle shaking. Gels were then transferred to staining solution (acetic acid-methanol-water, 10:50:40) containing 0.5% Coomassie blue for 1 h at room temperature. MMP activity in the gel was detected in a dark blue background with white bands.

Preglomerular arteriole structure and renal artery function.

Preglomerular arteriole structure was identified with trichrome stain as, described in Glomerular histology. Renal artery ring was prepared according to the previously developed method in our laboratory, and arterial function was measured as described (26).

Statistical analysis.

Values are given as means ± SE; n = 6/group. Differences between groups were tested by two-way ANOVA followed by the Bonferroni post hoc test (31). Significance was accepted at P < 0.05.

RESULTS

Diabetes reduces PPARγ cDNA, mRNA, and protein expression.

Induction of PRARγ has been shown to reverse the renal damage and restore glomerular function (21). Moreover, diabetes is associated with impaired renal function (26). Therefore, we sought to determine whether diabetes modulates the expression of PPARγ gene by cDNA gene array, mRNA by RT-PCR, and protein expression by Western blot. We found a reduced expression of PPARγ gene in diabetic kidney following 10 wk of alloxan treatment (Fig. 1A). Expressions of PPARγ, both at mRNA and protein level, were attenuated in alloxan-induced diabetic kidney following 1 and 10 wk of treatment (Fig. 1, B and C). Antidiabetic drug and PPARγ agonist CZ, however, reversed and normalized alloxan-induced attenuation of PPARγ mRNA and protein expression (Fig. 1, B and C). These results suggested that reduction in PPARγ gene, mRNA, and protein were directly associated with diabetic-induced renal failure.

Fig. 1.

A: expression of peroxisome proliferator-activated receptor-γ (PPARγ) gene in diabetic kidney tissues. Mice were treated with alloxan for 10 wk, as stated in materials and methods. RNA was isolated from kidney tissues, and equal amounts of poly (A)⁺ RNA from control and diabetic tissue were hybridized to cDNA array membranes. B: total mRNA was isolated from the kidney samples, and expression of PPARγ mRNA was measured by agarose gel eclectrophoresis; bottom: densitometric analysis of PPARγ mRNA expression. C: PPARγ protein expression was measured by Western blot, as described in the materials and methods; bottom: densitometric analysis of PPARγ expression. Note that PPARγ cDNA, mRNA, and protein show attenuated expression in diabetic kidney tissue, whereas ciglitazone (CZ) normalized PPARγ mRNA and protein expression (n = 6). #P < 0.01 compared with 0 wk.

Plasma glucose and Hcy.

Plasma glucose level was increased significantly within 1 wk of alloxan treatment compared with 0 wk and remained elevated until 10 wk (Fig. 2A). However, CZ attenuated this increased plasma level of glucose within 2 wk of treatment and brought the level to almost normal after 6 wk of treatment (Fig. 2A). Plasma Hcy level did not increase after 1 wk of diabetic induction; however, after 10 wk, a robust increase in the plasma Hcy level was observed in diabetic mice (Fig. 2B). Contrary to glucose level after CZ treatment, plasma Hcy level did not change and remained elevated until 16 wk (Fig. 2B). These results suggested that PPARγ agonist reduced glucose level in alloxan-induced diabetic model but had no role in plasma Hcy level.

Fig. 2.

Plasma levels of glucose and homocysteine (Hcy). A: mice were treated with alloxan (65 mg/kg body wt) as shown. Separate groups of alloxan-treated mice received CZ after 10 wk, and blood was collected at 12 and 16 wk. Plasma glucose levels were measured and compared with 0 wk (n = 6 in each group). *P < 0.01 compared with 0 wk. **P < 0.02 compared with 10 wk. B: plasma Hcy was separated with HPLC and measured by a spectrophotometer (n = 6 in each group). *P < 0.05 when compared with 0 wk. Note that CZ ameliorated hyperglycemia; however, CZ did not change the levels of plasma Hcy.

GFR and RVR.

We determined the functional status of glomerulus by measuring GFR. There was hyperfiltration in diabetic mice after 1 wk of alloxan treatment followed by hypofiltration at 10 wk (Fig. 3A). CZ normalized GFR after 2 and 6 wk of treatment. Although not significant, there was an increasing trend of RVR at 1 wk of diabetic induction. However, at 10 wk, a dramatic increase in RVR was observed. This increased RVR was normalized with CZ treatment at 12 and 16 wk, as shown in Fig. 3B. These results suggested that acute diabetes induced hyperfiltration followed by hypofiltration, whereas the PPARγ agonist CZ normalized glomerular function. In addition to that, acute diabetes-induced RVR was reversed toward normal by CZ.

Fig. 3.

Glomerular filtration rate (GFR). A: mice were treated with alloxan (65 mg/kg alloxan body wt), and GFR was measured at different time periods as shown (n = 6 in each group). Mice were acclimatized in metabolic cages for 3 days, and 0.3 mg of inulin-FITC/25 g body wt was given intraperitoneally; urine was collected for 24 h. Separate groups of alloxan-treated mice that received CZ were placed in metabolic cages, and urine was collected at 12 and 16 wk. Plasma and urine inulin-FITC was measured, and GFR was expressed as μl·min⁻¹·g kidney wt⁻¹ (n = 6 in each group). *P < 0.02 compared with 0 wk; **P < 0.05 compared with 1 wk. B: renal vascular resistance (RVR) was measured by renal artery blood flow (ml/min) and pressure (mmHg) and expressed as mmHg·ml⁻¹·min⁻¹. *P < 0.01 compared with 0 wk; **P < 0.02 compared with 10 wk. Note that in this model of diabetes, initially there was hyperfiltration followed by hypofiltration.

Glomerular histology.

Histological staining with trichrome showed glomerular hypertrophy of 1-wk diabetic kidney (Fig. 4). However, there was a reduction in glomerular size in 10-wk diabetic kidney compared with 1 wk. This reduction was possibly due to thickened parietal layer of Bowman's capsule (Fig. 4). Interestingly, these data corroborated with the GFR data (Fig. 3A). CZ treatment, however, reversed glomerular hypertrophy and thickened parietal layer toward normal level at 16 wk (Fig. 3A). These results suggested the involvement of PPARγ in diabetic glomerulopathy and that CZ ameliorated these changes.

Fig. 4.

CZ attenuates glomerular hypertrophy. Six-micrometer sections of kidney tissue samples from different treated groups as shown were stained with Masson-Trichrome stain and visualized under dissecting microscope; 10 × 20 magnification (n = 6 in each group).

Glomerular Hcy and NO.

Glomerular tissue level of Hcy was increased significantly, and maximum level was observed in 10-wk diabetic mice (Fig. 5A). Treatment of diabetic mice with CZ attenuated this increased level of Hcy significantly at 12 and 16 wk (Fig. 5A). Glomerular tissue level of NO, as determined by total NO₂/NO₃ assay, was increased significantly in diabetic mice at 1 wk. These changes, however, abruptly and significantly decreased in 10-wk diabetic kidney. Although CZ ameliorated NO production of the diabetic glomerular tissue at both 12 and 16 wk compared with 10 wk, the significant improvement was observed only at 16 wk (Fig. 5B). These results demonstrated the differential role of CZ in diabetic glomerular Hcy and NO production.

Fig. 5.

Glomerular tissue levels of Hcy and nitric oxide (NO). A: kidney was removed from anesthetized mice, and cortical tissues were separated from medullary mass. Cortical tissue homogenates were prepared. Total Hcy was extracted, separated by HPLC, and quantitated by a spectrophotometer. Hcy was expressed as ng/mg of protein. *P < 0.01 compared with 0 wk; **P < 0.01 compared with 10 wk. B: total nitrate/nitrite was measured by Griess method. NO levels were expressed as nM/l of cortical tissue homogenate. Equal amounts of total protein were used. *P < 0.02 compared with 0 wk; **P < 0.05 compared with 0 wk. ***P < 0.05 compared with 10 wk. Note that tissue levels of Hcy were normalized by CZ treatment.

MMP-2 and -9 activities.

Control (0 wk) diabetic mice showed a constitutive expression of MMP-2 and -9 in glomerular tissue; however, tissue levels of MMP-2 and -9 were increased significantly in diabetic mice, with a robust increase of both of the MMPs in 1-wk diabetic mice (Fig. 6A). These increases were observed for both latent and active MMPs. CZ treatment significantly attenuated these increased changes in MMP levels.

Fig. 6.

A: matrix metalloproteinase (MMP)-2 and -9 analysis. Glomeruli were isolated under a dissecting microscope from different groups of animals as indicated, and glomeruli-extracted protein was analyzed by 1.5% in-gel gelatin zymography (zymography image of MMP-2 and -9). To verify equal amount of loading, homogenates of the gelatin zymography for MMP-2 and -9 groups containing 25 μg of protein from each of the samples were separated by 10% SDS-PAGE gel and transferred to PVDF membrane, and Western blot was performed using anti-β-actin antibody (Western blot image of β-actin). B: densitometric analyses of MMP activity that were shown in A. *P < 0.01 compared with 0 wk (means ± SE; n = 6 in each group); **P < 0.05 compared with 10 wk (means ± SE; n = 6 in each group). Note that both MMP activities were attenuated by CZ at 12 and 16 wk.

TIMP-1 and -4 in glomeruli.

There was a constitutive expression of both TIMP-1 and -4, with a higher level of TIMP-4 in control mice. TIMP-4 expressions were attenuated in diabetic mice; however, TIMP-1 showed increased expression (Fig. 7A). These results suggested a differential role of TIMP-1 and -4 in control and diabetic mice. CZ treatment normalized these changes in diabetic mice (Fig. 7, A and B).

Fig. 7.

A: tissue inhibitor of metalloproteinase (TIMP)-1 and -4 analysis. Glomeruli were isolated under a dissecting microscope from different groups of animals as indicated, and glomeruli homogenates were subjected for Western blot analyses using anti-TIMP-1 and anti-TIMP-4 antibody. Membranes were reprobed with anti-β-actin antibody to verify equal amount of loading. A representative β-actin loading control is shown at A, bottom. B: densitometric analyses of TIMP-1 and -4 protein that were shown in A. *P < 0.01 compared with 0 wk (means ± SE; n = 6 in each group); **P < 0.05 compared with 10 wk (means ± SE; n = 6 in each group). Note that TIMP-1 expression was increased at 1 and 10 wk, and this expression was attenuated by CZ at 12 and 16 wk, whereas TIMP-4 expression was reversed.

Arterial structure and function.

There was a robust increase in media content and narrowing of lumen; an increase in media/lumen ratio was observed at 10 wk of alloxan treatment compared with 0-wk control (Fig. 8, A, B, and D). However, treatment with CZ reversed this effect and normalized the vessel architecture at 16 wk (Fig. 8, C and D).

Fig. 8.

Preglomerular arteriole and tubule histology. Part of kidney from 0 wk (A), 10 wk of alloxan treatment (B), and 10 wk of alloxan plus another 6 wk of CZ treatment (C) was frozen using tissue-freezing medium for cryosection and histological analysis. Six-micrometer sections were stained with Masson-Trichrome. D: preglomerular arterioles were identified under a microscope, and medial/lumen ratio was measured by a digital micrometer and plotted. Bar graph represents mean ± SE of 6 animals/group. *P < 0.01 compared with 0 wk; **P < 0.05 compared with 10 wk. Note that medial/lumen ratio was increased dramatically after 10 wk of alloxan treatment and was ameliorated significantly by CZ treatment at 16 wk.

Figure 9A shows a typical dose-response curve of renal artery with phenylepinephrine. One and ten weeks of diabetic mice showed significant decreases in renal artery contractile function compared with 0-wk control (Fig. 9B). This contractile dysfunction was almost recovered after CZ treatment, as shown by improved contractile function (Fig. 9B). Similarly, attenuated relaxation of renal artery by acetylcholine in diabetic mice was normalized with CZ (Fig. 9C). These results suggested the impaired renal artery function in diabetic mice and that CZ improved this function by activating, in part, PPARγ.

Fig. 9.

A: typical dose-response curve of phenylepinephrine (PE) in renal artery isolated from 0-wk animal. B: dose-response curve of PE in renal artery isolated from different groups of animal as shown. C: the vessels were precontracted with PE and treated with different doses of acetylcholine (Ach) as shown. *P < 0.01 compared with 0 wk (means ± SE; n = 6 in each group). Note that CZ ameliorated diabetes-induced renal artery contraction and endothelial-mediated renovascular relaxation.

DISCUSSION

This study demonstrated that PPARγ cDNA, mRNA, and protein expressions were attenuated in diabetes animals with elevated level of plasma Hcy. Diabetic nephropathy was observed with hyperfiltration followed by hypofiltration and increased in renovascular resistance. Initial hyperfiltration was due to glomerular hypertrophy, whereas at late stages of diabetes, glomerular atrophy exhibited reduced GFR. This atrophy and glomerular collapse was in part mediated through imbalances in the MMP-TIMP axis that probably modulated matrix protein. Moreover, structural changes in preglomerular arteriole were reflected in acetylcholine-induced vessel relaxation since diabetic arteries were less responsive. This study also showed that treatment with CZ, a PPARγ inducer, reversed attenuated PPARγ mRNA and protein expressions, normalized increased plasma glucose level, and improved RVR and GFR with normalized preglomerular arteriole and glomeruli. Although CZ did not change plasma Hcy level in diabetes model, there was a dramatic reduction of glomerular tissue level of Hcy with an increased bioavailability of NO that was observed. Increases in MMP-2 and -9 activity and TIMP-1 expression were associated with diabetic-induced deterioration of renal structure and function, whereas CZ reversed deleterious effects on glomerular structure and restored functional changes. In contrast to TIMP-1, TIMP-4 expression was opposite; this suggested that these two TIMPs may have differential roles in diabetic nephropathy and glomerulosclerosis.

Diabetic nephropathy is the most common single condition found in end-stage renal disease. The majority of diabetic patients with renal failure suffer from glomeruloslerosis; however, diabetes and acute renal failure coexist as well (13). Impairment of renal function due to vasoconstriction of glomerular arteriole was related to volume retention that accumulated and further increased plasma Hcy levels (26). Elevated plasma Hcy, in turn, caused chronic and impaired renal filtration and was also reported as a risk factor for diabetic nephropathy (17, 36). Activation of PPARγ, a nuclear receptor protein that functions as transcription factor, induced insulin sensitivity in type 2 diabetes and promoted tissue uptake of Hcy; these resulted in lowering of plasma Hcy levels (26, 35). However, in type 1 diabetes, the mechanism could be different. In the present study, although CZ induced PPARγ and reduced plasma glucose, it did not change plasma Hcy level, although glomerular tissue level of Hcy became normal with CZ treatment. This is probably due to improvement of diabetic nephropathy that reduces renal volume retention and clearance of tissue Hcy. This result supported clinical trials where PPAR agonists ameliorated endothelial dysfunction in HHcy with no effect on plasma Hcy level (1).

Remodeling of extracellular matrix (ECM) is an important physiological phenomenon of normal and diseased condition. The regulation of glomerular ECM turnover engages a number of MMPs and their natural inhibitor, TIMPs. These MMPs and TIMPs include MMP-2 and -9 and TIMP-1 and -4. Diabetes affects ECM degradation, and activities of MMPs and TIMPs mostly regulate this process (27). Type IV collagenases, MMP-2 and -9, have been studied extensively in various glomerular diseases with conflicting results (10, 14, 26, 27). However, in the present study, we found that both MMP-2 and -9 activities were increased significantly in diabetic kidney. Similar results have been reported by independent laboratories, including our own (6, 8, 26, 40). We also showed that expression of TIMP-1, a regulator of matrix remodeling, corroborated with the MMP-2 and -9 activities. This is in agreement with the previously reported study by Eddy et al. (9) where progressive renal fibrosis was characterized by upregulation of TIMP-1 expression. Interestingly, TIMP-4 expression was found to be opposite of TIMP-1 expression in our study. PPARγ agonist CZ ameliorated these matrix protein changes in diabetic kidney through activating PPARγ and Hcy clearance. These resulted in restoration of renal damage and normal glomerular function.

GFR has been used to measure level of kidney function and to determine the stage of kidney disease (16). It is reported that, at the onset of diabetes, the kidney grows large along with glomerular hypertrophy, and the GFR becomes supranormal (19, 33). Hypertrophy also contributes to cellular oxidative stress, and this may precede the reactive oxygen species (ROS) perturbation (28). This eventually leads to gradual deterioration of kidney with reduced GFR and subsequently results in sclerosis and kidney failure. Moreover, renal arterial resistance has been shown to play a nontrivial role in deteriorating renal function in diabetic patients (22). In our present study, we found a robust increase in GFR with very little change in RVR at the onset of diabetes. This is in agreement with previous reports (19, 33). However, at latter stage, subnormal GFR and a more than twofold increase in RVR were observed. This subnormal GFR occurs, however, due to glomerular collapse and eventually increases RVR. Hypertrophy-induced ROS generation and increases in tissue Hcy in the glomerulus at later stages further exacerbated oxidative damage and malfunction. ROS causes endotheial injury (24), and therefore, we found that acetylchonine-induced endothelial-dependent renal artery function was severely diminished in diabetes. Decreased production and bioavailability of NO supported this and mitigated endothelial-dependent renal artery function. CZ treatment, however, ameliorated both structural and functional changes through at least two different mechanisms in our study. First, changes were ameliorated by activating PPARγ, which normalized glucose in the plasma; therefore, this activation modulated diabetes-induced kidney damage. Second, changes were ameliorated by Hcy clearance from the tissue that reduced oxidative damage and glomerulosclerosis.

We considered 1.6 mg·kg⁻¹·day⁻¹ of CZ sufficient to effectively stimulate PPARγ based on the fact that 100 mg/day PPAR agonist in humans has previously been shown to be potentially effective (11). If we assume the average weight of a person to be 70 kg, then each individual was given 1.42 mg·kg⁻¹·day⁻¹ of PPAR agonist in the above reference report. However, in our experiment, we gave higher doses of CZ to maximize its effect on PPARγ in alloxan-induced diabetic mice. Additionally, we have reported elsewhere that 1.6 mg·kg⁻¹·day⁻¹ CZ has a strong effect on PPARγ induction in alloxan-induced diabetic mouse heart (25). Because CZ is not a very stable compound in aqueous solution for the long term, we considered that measurement of CZ in collected plasma might not accurately determine the actual CZ content. To ensure that we were providing experimental animals with a sufficient amount of active CZ to agonize PPARγ, every 24 h we replaced drinking water supplemented with freshly prepared CZ.

It is true that our experimental design resulted in the lack of a control population of mice, such as glycemic control, that extended through the experiment that also had their blood glucose levels controlled to the same extent as the CZ groups. Also, this study utilized a paired statistical evaluation. There is evidence that glycemic control can correct homocysteinemia; for example, both insulin and sulfonylurea have been shown to do this, probably through pathways other than PPAR stimulation (12, 37). There are also independent effects of hyperglycemia on GFR and renal plasma flow (18). The experimental design utilizing a control group throughout the experimental period would have substantially strengthened the conclusions of the study. The research outcome of future studies might settle the issue of whether or not PPARγ has a regulatory role in Hcy metabolism in diabetic-induced renal failure. Additionally, it may also be possible that CZ fosters hertodimerization between PPAR and another nuclear receptor, such as retinoid X receptor (7). Nevertheless, data from our present study showed that PPARγ is one of the regulatory pathways, if not the only mechanism, that modulates tissue Hcy metabolism and HHcy-associated glomerular remodeling in diabetic nephropathy.

In summary, we have shown that PPARγ agonist CZ had renoprotective effects in experimental diabetes. This effect was in part through reduction in glomerular tissue Hcy level over and above the effects on plasma Hcy level. This finding also linked to the reversal of glomerular damage and to the restoration of function following PPARγ induction by CZ. This is consistent with the protective role of PPARγ against diabetic nephropathy (2).

GRANTS

A part of this study was supported by National Heart, Lung, and Blood Institute Grants HL-71010, HL-74185, and HL-88012.