Targeting of RhoA/ROCK Signaling Ameliorates Progression of Diabetic Nephropathy Independent of Glucose Control

Fasudil was a kind gift of Asahi-Kasei Pharma (Shizuoka, Japan). Simvastatin was obtained from Merck & Co. (West Point, PA). Simvastatin was activated by alkaline hydrolysis as previously described (35). Y27632 was obtained from Calbiochem (San Diego, CA).

Early passaged (passages 3–10) rat mesangial cells were grown in Dulbecco’s modified Eagle’s medium/F12 medium containing 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin in a humidified incubator at 37°C under 5% CO₂. Transfections of RhoA mutants were performed as described previously (35).

All animal studies were carried out with the review and approval of the animal care and use committee of Northwestern University. In two separate experiments, db/db mice received fasudil (10 mg · kg⁻ · day⁻ i.p., n = 8) or simvastatin (40 mg · kg⁻ · day⁻ p.o., n = 8) starting at 8 weeks of age. After 16 weeks of treatment (24 weeks age), the mice were housed in individual metabolic cages for collection of urine. Blood glucose was measured after a 12-h fast using the OneTouch UltraSmart Blood Glucose Meter (Lifescan, Milpitas, CA). Urinary albumin and collagen IV concentrations were measured using the Albuwell M and Collagen IV M assay kits (Exocell, Philadelphia, PA). Serum and urine creatinine were measured using high-performance liquid chromatography. Total cholesterol was determined by the Vanderbilt Mouse Metabolic Phenotyping Center (Nashville, TN) using a microtiter plate assay.

For morphometric studies, the kidneys were fixed in 10% neutral buffered formalin and subsequently embedded in paraffin. The 4-μm sections of paraffin-embedded tissues were stained with periodic acid-Schiff (PAS). Light microscopic views of 40 consecutive glomerular cross-sections per mouse were scanned into a computer. Glomerular and mesangial matrix areas were quantified in a blinded fashion using an image analysis system (MetaMorph version 6.1; Universal Imaging). Mesangial matrix index (MMI) was calculated as the ratio of mesangial area to glomerular area × 100 (% area).

Immunoperoxidase staining was carried out on 4-μm paraffin-embedded kidney tissue sections. The sections were preincubated in boiling sodium citrate buffer (10 mmol/l sodium citrate, 0.05% Tween 20, pH 6.0) for antigen retrieval. Specific primary antibodies, anti-fibronectin antibody and anti-collagen IV antibody (Chemicon, Temecula, CA), were incubated with the sections overnight. For visualization, the sections were developed with 3, 3′-diaminobenzidine to produce a brown color, counterstained with hematoxylin, and mounted in VectaMount mounting medium (Vector Laboratories).

Small pieces of kidney cortex were fixed in 2.5% glutaraldehyde in 0.2 mol/l cacodylate buffer (pH 7.4). They were then dehydrated through graded ethanol and propylene oxide, embedded in epoxy resin, and polymerized at 60°C overnight by standard procedures. Ultrathin sections were stained with uranyl acetate and lead citrate. The specimens were observed using a Jeol 1220 transmission electron microscope. For each specimen, images of the glomerular basement membrane (GBM) in 10 consecutive glomeruli were captured onto a computer.

The 4-μm sections of snap-frozen kidneys were permeabilized for 10 min using cold acetone, rehydrated in 1× Tris-buffered saline (TBS), and preincubated in 5% bovine serum albumin, 0.2% saponin, 0.001% Tween in TBS for 45 min at room temperature. Primary antibodies included anti-RhoA (Abcam, Cambridge, MA), anti-vWF antibody (Sigma, St. Louis, MO), and collagen IV (Abcam). The actin cytoskeleton was stained with rhodamine phalloidin and visualized using scanning confocal fluorescence microscopy (Zeiss LSM510).

RhoA activity was determined by a pull-down assay according to the manufacturer’s instructions (Upstate Biotechnology). In brief, fresh kidney cortex tissues were washed in cold PBS and lysed in ice-cold MLB buffer. Samples were centrifuged and incubated for 60 min at 4°C with 100 μl Rhotekin agarose to precipitate GTP-bound RhoA. Precipitated complexes were washed three times in MLB buffer and resuspended in 30 μl of 2× Laemmli buffer. Total lysates and precipitates were analyzed by Western blot analysis using rabbit polyclonal antibody against RhoA (Santa Cruz, CA).

ROCK phosphorylates the myosin phosphatase target subunit 1 (MYPT1), also known as myosin-binding subunit (MBS), of myosin light-chain phosphatase at Thr853. ROCK activity was measured as previously described (36). Briefly, equal amounts of kidney cortical tissue extracts were analyzed by Western blot analysis by using rabbit phospho-MBS/MYPT1-Thr853 (CycLex) and rabbit anti-MYPT1 (Santa Cruz Biotechnology). ROCK activity was expressed as the ratio of phospho-MYPT1 to total MYPT1.

The values are reported as means ± SE. Statistical significance was determined using one-way ANOVA followed by Tukey’s multiple comparisons test. A value of P < 0.05 (two tailed) was considered statistically significant.

To establish the potential role of RhoA/ROCK in the pathophysiology of diabetic nephropathy, we investigated the effect of either fasudil (10 mg · kg⁻ · day⁻ i.p.) or simvastatin (40 mg · kg⁻ · day⁻ p.o.) in two separate experiments. After 16 weeks of therapy, animals were killed and various biochemical, and histological parameters were evaluated.

ROCK activity in the kidney cortex was directly assessed by the amount of phospho-Thr853 in the MYPT1 of myosin light-chain phosphatase, a downstream target of ROCK (36). As shown in Fig. 1A, ROCK activity was significantly increased in the kidney cortex of diabetic db/db mice, while treatment with fasudil inhibited the increased ROCK activity in diabetic mice. However, there were no significant changes in total MYPT1 in different groups. Fig. 1B depicts a semiquantitative measure of the percentage differences in ROCK activity in db/m, db/db, and fasudil-allocated mice (100 ± 4, 152 ± 10, and 120 ± 3, respectively, *P < 0.05).

Fig. 2 depicts several metabolic and renal parameters of fasudil-treated db/db mice at 24 weeks of age compared with nondiabetic db/m mice. These parameters include blood glucose levels, serum creatinine and cholesterol levels, and urine albumin-to-creatinine ratio. As expected, the mean blood glucose level of the diabetic db/db mice (n = 8) was significantly higher than that of db/m mice (n = 8) (Fig. 2A). Treatment of diabetic animals with fasudil did not significantly lower blood glucose levels (n = 8). However, after 16 weeks of therapy with fasudil, the diabetic mice exhibited a significant reduction in albuminuria (Fig. 2C, 417 ± 120 vs. 0.150 ± 20 μg/mg, P < 0.01). As shown in Fig. 2B and D, there were no significant changes in the serum creatinine and cholesterol levels between fasudil-treated and db/db mice (P > 0.05).

Mesangial matrix expansion is considered a hallmark of diabetic nephropathy. Figure 3A suggests accelerated mesangial matrix expansion in db/db mice as characterized by an increase in PAS-stained area compared with that observed in db/m mice at 24 weeks of age. However, treatment with fasudil prevented mesangial expansion in db/db mice. To assess the effect of fasudil therapy on mesangial matrix accumulation, we evaluated MMI. As shown in Fig. 3B, the MMI in db/db diabetic animals was significantly increased compared with that in nondiabetic db/m mice (23 ± 0.5 vs. 7 ± 0.2%, P < 0.01). Treatment with fasudil markedly reduced MMI in the glomeruli of db/db mice (23 ± 0.5 vs. 14 ± 0.3%, P < 0.01).

Furthermore, to evaluate for the effect of fasudil on extracellular matrix proteins, we performed immunohistochemical studies of type IV collagen and fibronectin (Fig. 3C). Consistent with our previous results for mesangial matrix expansion, there was a strong staining for type IV collagen and fibronectin in the diabetic db/db mice compared with the db/m mice. In contrast, the expression of collagen IV and fibronectin were reduced in the glomeuli of fasudil-treated db/db mice. Figure 3D demonstrates the effect of fasudil on collagen IV protein levels using Western blot analysis. Consistent with previous reports, type IV collagen protein expression was significantly increased in the kidney cortices of db/db mice. Treatment with fasudil significantly reduced type IV collagen protein expression.

A number of previous studies have suggested that changes in collagen IV accumulation may also be accompanied by increased excretion of urinary collagen IV in db/db mice (37). To address the effect of fasudil on urinary collagen IV excretion, we monitored urinary collagen IV excretion as previously described (37). The data presented in Fig. 3E indicate that urinary collagen IV increased in db/db mice but not in nondiabetic db/m mice (1.2 ± 0.25 vs. 0.18 ± 0.03 ng/24 h, P < 0.01). Furthermore, the fasudil-treated animals exhibited a significant reduction in urinary collagen IV excretion after 16 weeks of therapy (0.46 ± 0.081 vs. 1.2 ± 0.25 ng/24 h, P < 0.05).

To assess the effect of fasudil on GBM, we performed transmission electron microscopy. As shown in Fig. 3F and G, the GBM thickness in db/db diabetic animals was increased compared with that in the normal animals (236 ± 4 vs. 150 ± 2 nm, P < 0.001). Interestingly, fasudil-treated animals showed a significant reduction in GBM thickness as compared with untreated db/db mice (171 ± 2 vs. 236 ± 4 nm, P < 0.001).

We and others have previously shown that statins, by inhibiting mevalonate pathway, also inhibit isoprenoids, which are critical for the proper function and activity of RhoA GTPases. Thus, to assess the effect of simvastatin on RhoA activation in vivo, we initially assessed a quantitative measure of activated RhoA in db/db mice. As depicted in Fig. 4A, membrane-bound RhoA (active form) was increased in the diabetic kidney cortex compared with that in the normal control group (n = 8/group). However, total RhoA protein (active and inactive form) was similar among the two groups. Moreover, RhoA activity in the kidney cortex was also directly assessed by an affinity pull-down assay using GST fusion protein rhotekin, which recognizes only the active form of RhoA (GTP-RhoA). As shown in Fig. 4B, RhoA pull-down assay further confirmed that GTP-RhoA (active form) is highly expressed in the kidney cortex of diabetic db/db mice. Treatment with simvastatin suppressed the increased activity of RhoA in diabetic mice, which is demonstrated in the semiquantitative analysis displayed in Fig. 4C.

Figure 5 depicts several metabolic and renal parameters of simvastatin-treated db/db mice at 24 weeks of age compared with nondiabetic db/m mice and PBS-treated diabetic db/db mice. As expected, the mean blood glucose level of the diabetic db/db mice was significantly higher than that of the db/m mice (Fig. 5A). Treatment of diabetic animals with simvastain lowered blood glucose levels, although it did not reach statistical significance (397 ± 23 vs. 452 ± 18 mg/dl, P > 0.05). Urinary albumin excretion was significantly greater in db/db animals than in db/m animals (Fig. 5B). After 16 weeks of therapy with simvastatin, the diabetic mice exhibited a significant reduction in albuminuria (400 ± 37 vs. 717 ± 76. μg/mg, P < 0.001). As shown in Fig. 5C, there were no significant changes in the plasma creatinine in the different groups. And finally, db/db mice allocated to simvastatin exhibited lower cholesterol levels (249 ± 24 vs. 223 ± 14 mg/dl, P < 0.05, n = 8/group) than PBS-treated db/db mice.

The representative glomerular histology of PAS-stained sections is shown in Fig. 6A. The glomerular accumulation of PAS-positive matrix was prominent in PBS-treated diabetic animals. In simvastatin-treated animals, matrix expansion was significantly less than that in PBS-treated animals. Similarly, as shown in Fig. 6A, the increases in fibronectin and type IV collagen protein expression were prominent in the glomeruli of diabetic db/db mice. Treatment with simvastatin significantly reduced the accumulation of these proteins in the glomeruli of diabetic db/db mice. As shown in Fig. 6B and C, simvastatin-treated animals also showed a significant reduction in GBM thickness compared with PBS-treated db/db mice (176 ± 2 vs. 236 ± 4 nm, P < 0.001). Figure 6D depicts that the MMI in simvastatin-treated animals was also significantly lower than that in the vehicle-treated group.

We have previously provided data that high glucose mediates RhoA activation in mesangial cells (32). We also reported that RhoA activation plays an important role in collagen IV accumulation in mesangial cells via cytoskeletal remodeling (33). To decipher whether the RhoA/ROCK pathway mediates high glucose–induced mesangial matrix expansion and collagen IV accumulation, cultured rat mesangial cells were exposed to high glucose (30 mmol/l), and ROCK activity was measured as previously described. As shown in Fig. 7A, high glucose significantly increased ROCK activity, as measured by MYPT1-phospho Thr853, a downstream target of ROCK, after 8 h of stimulation. To test whether RhoA/ROCK activation mediates high glucose–induced increase in type IV collagen protein levels via cytoskeletal remodeling, mesangial cells were transfected with Myc-tagged dominant-negative mutant (N19RhoA) and wild-type RhoA. Cells were then stimulated with high glucose milieu (30 mmol/l). Cytoskeletal remodeling was evaluated by confocal laser microscopy. As shown in Fig. 7B (upper panel), high glucose caused a significant increase in actin stress fiber formation. However, there was a significant reduction in high glucose–induced actin stress fiber formation in cells co-treated with Y-27632 (10 nmol/l), a specific inhibitor of ROCK, as well as in N19RhoA-transfected mesangial cells, suggesting a pivotal role for RhoA/ROCK activation in high glucose–induced cytoskeletal remodeling. The data in Fig. 7B (lower panel) also show that high glucose stimulation led to an increase in cell-associated collagen IV protein levels as evaluated by confocal microscopy. Collagen IV protein levels did not increase in cells co-treated with Y-27632 (10 nmol/l) or in N19RhoA-transfected cells, indicating a critical role for RhoA/ROCK activation in high glucose–induced collagen IV accumulation. Consistent with these results, a Western blot analysis in mesangial cells stimulated with high glucose also indicated that high glucose–induced collagen IV protein accumulation is significantly reduced in dominant-negative RhoA-transfected cells and in Y-27632 co-treated mesangial cells.

The results of this study implicate RhoA/ROCK in the signaling pathways contributory to the development of diabetic nephropathy. A principal finding of this study is that RhoA/ROCK activation is increased in the kidneys of db/db mice and that the inhibition of the RhoA/ROCK pathway can protect against the progression of diabetic nephropathy.

The RhoA/ROCK pathway has recently attracted a great deal of attention in various research fields, particularly diabetes-related research, for several reasons. First, at the cellular level, the RhoA/ROCK pathway has a prominent role in various signaling pathways that are potentially implicated in the pathogenesis of microvascular complications of diabetes. For instance, our laboratory and several others have recently unraveled the role of RhoA/ROCK in vascular endothelial growth factor (VEGF) signaling, an important mediator of microvascular complications of diabetes (34,35). We determined that VEGF-induced glomerular endothelial cell hyperpermeability is mediated by an increase in RhoA/ROCK activation, which leads to phosphorylation of myosin regulatory light chain, a substrate of ROCK. The second reason for the emergence of ROCK as a novel target in several research areas is because it has been suggested that the pleiotropic effects of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors (statins) are mediated, at least in part, by their inhibitory effects on RhoA/ROCK signaling (28,32,38). Indeed, we and others have previously shown that the inhibitory effects of statins on RhoA/ROCK signaling pathways may play an important role in the pleiotropic effects of statins (31,33,34). The final reason for the recent focus on RhoA/ROCK pathway may be based on the involvement of ROCK in extracellular matrix accumulation. For instance, we have recently shown that RhoA activation mediates actin cytoskeleton remodeling, intracellular FAK phosphorylation, and β1 integrin activation, leading to increased type IV collagen synthesis (33). However, despite a growing body of evidence suggesting a pathogenic role for RhoA and ROCK activation in vitro, the role of RhoA/ROCK activation in vivo in diabetic nephropathy is not well defined.

In the current study, we established the effect of pharmacological blockade of RhoA and ROCK pathway in the progression of diabetic nephropathy. We found that RhoA and ROCK activity is increased in the kidney cortices of db/db mice. The pathogenic role of RhoA/ROCK activation was evident when pharmacological inhibition of the pathway resulted in a significant reduction in RhoA/ROCK activity, mesangial matrix expansion, collagen IV accumulation, and albuminuria. Thus, our data indicate that targeting RhoA/ROCK is an effective and novel strategy in preventing the progression of experimental diabetic nephropathy.

The effect of fasudil on albuminuria in this study is consistent with recent reports indicating that fasudil may reduce albuminuria in different animal models of diabetes (39,40). For instance, in a previous report, high-dose fasudil administration significantly attenuated albuminuria and glomerulosclerosis but also led to a significant decrease in blood glucose (39). The authors concluded that fasudil, at least in part, by improving glucose control, ameliorates progression of renal disease. Our results differ from this report in that the beneficial effects of fasudil in our study were independent of its potential effect on glucose control. Indeed, the fasudil-allocated db/db mice in our experiments did not exhibit significant changes in their blood glucose levels compared with untreated db/db mice. This finding suggests that ROCK inhibition may have renoprotective effects independent of its potential effects on glucose levels. We also recognize that differences in the observed effects of fasudil may be explained, at least in part, by the experimental model of diabetes and dose of fasudil used.

Currently, two major families of ROCK inhibitors are used: isoquinoline derivatives and 4-aminopyridine derivatives. Fasudil is one of the most prominent isoquinoline derivatives, which is used clinically to treat cerebral vasospasm in Japan, and several clinical trials are currently examining its anti-anginal effects (41,42). The 4-aminopyridine derivative, Y-27632, is widely utilized in biological and pharmacological experiments. Both Y-27632 and fasudil are selective ROCK inhibitors that target its ATP-dependent kinase domain. Hydroxyfasudil, an active metabolite of fasudil after oral administration, has been shown to have greater selectivity for ROCK than fasudil itself (43). Several efforts have been made to develop more specific and more potent ROCK inhibitors. Recently, the 3-D crystal structure of ROCK and its binding site for fasudil have been determined, which should facilitate the development of more selective ROCK inhibitors in the future (44).

Because of the ability of statins to attenuate several experimental models of inflammatory diseases, considerable interest has arisen in their therapeutic potential for treating a broad spectrum of chronic kidney diseases including diabetic nephropathy. In support of a beneficial role for statins in ameliorating the progression of diabetic nephropathy, several clinical studies have reported findings of a preserved glomerular filtration rate and diminished urine albumin excretion in diabetic patients with proteinuria treated with statins (18,20,24,45,46). However, the molecular mechanism(s) responsible for the renoprotective effects of statins has been the subject of much debate. Several mechanisms may account for the possible beneficial effects of statins in diabetic nephropathy. For instance, because of their direct lipid-lowering properties, statins may diminish Ox-LDL–induced nephrotoxicity (47). It has become increasingly apparent, however, that the effects of statins cannot be ascribed solely to their lipid-lowering properties. For instance, Usui et al. (48) have suggested that the key anti-fibrotic effects of statins lie in their inhibitory effect on nuclear factor-κB activation. Kim et al. (49) have also reported that the anti-fibrotic effect of lovastatin occurs through modulation of transforming growth factor-β1 signaling. And Yokota et al. (50) have proposed that the renoprotective effects of statins in diabetic milieu is via suppression of Ras-MAPK cascade.

Our data also suggest that the therapeutic potential of RhoA and ROCK inhibition in diabetic nephropathy may, at least in part, stem from their inhibitory effect on mesangial matrix accumulation. Indeed, the present study indicates that RhoA/ROCK blockade effectively prevented the increase in collagen IV accumulation in the cortical kidneys of db/db mice. These findings are of significant interest since they provide not only a rationale for the renoprotective effects of fasudil and simvastatin but also offer a novel therapeutic targets in diabetic nephropathy, namely RhoA and its downstream effector ROCK.

Our results provide strong evidence that fasudil and simvastatin protect against the progression of diabetic nephropathy in an experimental model of type 2 diabetes. Further studies elucidating the potential benefits of targeting specifically RhoA/ROCK activation via pharmacological and genetic manipulations are underway in our laboratory. Elucidating signaling pathways by which RhoA/ROCK exert their pathogenic effects may not only enable us to understand the mechanisms of microvascular complications of diabetes but may also lead us to strategies for targeted therapy and protection against diabetic nephropathy.

F.R.D. is supported by a grant from the National Institutes of Health (NIH)/National Institute for Diabetes and Digestive and Kidney Diseases (R01DK067604). This study was supported by the Lipid, Lipoprotein and Atherosclerosis Core of the Vanderbilt Mouse Metabolic Phenotyping Centers (NIH DK59637).

We appreciate the technical contributions of Steve Dunn, Thomas Jefferson University, Philadelphia, Pennsylvania.