Diminished Loss of Proteoglycans and Lack of Albuminuria in Protein Kinase C-α—Deficient Diabetic Mice

Experiments were performed with male 129/SV PKC-α^−/− mice (12) and 129/SV wild-type (WT) animals from the strain that was used to generate the 129/SV PKC-α^−/− mice. The animals received a standard diet with free access to tap water. All procedures were carried out according to guidelines from the American Physiological Society and were approved by local authorities. Seven-week-old weight-matched mice received either 125 mg/kg body wt streptozotocin (Sigma-Aldrich) in 50 mmol/l sodium citrate (pH 4.5; n = 20 per group) or sodium citrate buffer (n = 8 per group) intraperitoneally on days 1 and 4. Glucose levels from tail blood were measured with the Glucometer Elite (Bayer, Leverkusen, Germany) every other day. Animals with glucose levels >16 mmol/l on two consecutive measurements were regarded as hyperglycemic, and glucose measurements were extended to once weekly. Animals that were not hyperglycemic within 14 days after the first injection were excluded. The mice received no insulin within the complete study period. Ketonuria did not occur (data not shown). After 2 or 8 weeks of hyperglycemia, the animals were killed according to the following protocol. After anesthesia with Avertin (2.5%), a laparotomy was performed and urine was collected by puncturing the bladder with a 23-gauge needle. Then, the abdominal aorta was cannulated with a 23-gauge needle, and the organs were perfused with lactated Ringer solution. After ligation of the left renal artery, the left kidney was removed, weighed, and snap frozen in isopentane (−40°C). The right kidney was perfused with 3% paraformaldehyde (PFA) in 0.1 mol/l Soerensen’s phosphate buffer. The right kidney was fixed for an additional 20 h in 3% PFA in Soerensen’s phosphate buffer and embedded in paraffin.

Albumin concentration in spot urine samples was measured with a commercially available competitive enzyme-linked immunosorbent assay following the instructions of the manufacturer (Exocell, Philadelphia, PA) and was normalized to urine creatinine.

Histological and morphometric analysis was carried out on paraffin sections (3-μm thickness) cut on a rotation microtome (Microm) and stained with trichrome stain after Masson-Goldner. Glomerular tuft volume was estimated as described before (13,14). In each animal, 50 random cross-sectional profiles of superficial to midcortical glomeruli (first two rows of glomeruli beneath the kidney capsule) were recorded with a digital video camera (Axiocam; Zeiss, Jena, Germany) connected to a light microscope (Axioplan-2; Zeiss), and the glomerular tuft area (A_T) was measured using an image analysis system (Axiovision; Zeiss). Average tuft area (Ā_T) was used to calculate an average glomerular tuft volume (V_T) for each animal by the formula V_T = β/k × (Ā_T)3/2, where β = 1.38 (shape coefficient for spherical particles) and k = 1.1 (size distribution coefficient) (15). Average V_T was corrected for the effects of shrinkage (roughly 48%) during paraffin embedding (16).

Immunohistochemistry was performed on cryostat sections of the frozen kidneys or on paraffin sections using the following primary antibodies: anti–PKC-α (catalog no. sc-208; Santa Cruz Biotechnologies, Santa Cruz, CA) anti–TGF-β₁ (catalog no. sc-146; Santa Cruz), anti–vascular endothelial growth factor (VEGF; catalog no. sc-152; Santa Cruz), VEGF RII (catalog no. sc-504; Santa Cruz), anti-fibronectin (catalog no. 14-109-0568; Paesel+Lorei, Frankfurt, Germany), anti–type IV collagen (catalog no. 1340-01; Southern Biotechnology, Birmingham, AL), and anti–type III collagen (catalog no. 234189; Calbiochem). For indirect immunofluorescence, nonspecific binding sites were blocked with 10% normal donkey serum (Jackson ImmunoResearch Laboratory, West Grove, PA) for 30 min. Then sections were incubated with the primary antibody for 1 h. For fluorescent visualization of bound primary antibodies, sections were further incubated with Cy3-conjugated secondary antibodies (Jackson ImmunoResearch Laboratory) for 1 h. Specimens were analyzed using a Zeiss Axioplan-2 imaging microscope with the computer program AxioVision 3.0 (Zeiss). Semiquantitative analysis of VEGF, VEGF-RII, and TGF-β₁ expression was done by counting the numbers of glomeruli with high, moderate, and weak expression. A total of 40 glomeruli/animal were counted. The scoring was done without knowledge of the identity of the animal group by two independent observers.

For Western blotting, the frozen kidneys were pulverized in liquid nitrogen and resuspended in 2 ml of lysis buffer (20 mmol/l Tris buffer [pH 7.5] containing 10 mmol/l glycerophosphate, 2 mmol/l pyrophosphate, 1 mmol/l sodium fluoride, 1 mmol/l phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 mmol/l dithiothreitol, and 1 mmol/l EDTA). Homogenates were sonicated for three 20-s bursts on ice and centrifuged at 500g for 1 min to remove cell debris. Aliquots of the supernatants were stored at −80°C. The protein amount was measured using the Lowry assay. A total of 70 μg of protein of each sample was resuspended in loading buffer and run on a 10% polyacrylamide gel and electrophoretically transferred to nitrocellulose membrane. Membranes were blocked in 5% skim milk and 1% BSA for 1 h at room temperature. Primary antibody against factor TGF-β₁ (sc-146; Santa Cruz) was applied with gentle rocking overnight at 4°C. After three 10-min washing steps with TBST buffer (50 mmol/l Tris HCl [pH 7.5], 150 mmol/l NaCl, 0.01% Tween 20), incubation with horseradish peroxidase–conjugated goat anti-rabbit secondary antibody (Dianova, Hamburg, Germany) was performed for 1 h at room temperature. After three additional TBST washes, the membrane was incubated with Renaissance reagent (NEN Life Science, Zaventem, Belgium) according to the manufacturer’s instructions and exposed to X-ray film (Kodak). Quantification was done by measuring relative density (Scion Image).

For electron microscopic investigation, additional animals were anesthetized with 2.5% avertin (150 μl/10 g body wt) and the kidneys were fixed by perfusion via the abdominal aorta. After flushing of the vasculature with Ringer’s saline, the animals were perfused with fixative containing 1.5% PFA and 1.5% glutaraldehyde in 0.1 cacodylate buffer (pH 7.4) for 5 min. Tissue slices were postfixed in the same solution for 2 h and with 1% OsO₄ in cacodylate buffer for 1 h and embedded in Epon. Thin sections (70 nm) were stained with uranyl acetate and lead citrate and examined in a Zeiss EM 10 electron microscope.

For real-time qPCR, 2 μg of DNase-treated total RNA was reverse transcribed using a mix of random hexamers and oligo(dT)12–15 oligonucleotides (Stratagene, Amsterdam, Netherlands) and Superscript II Reverse Transcriptase (Invitrogen, Carlsbad, CA). qPCR was performed on an SDS 7700 system (Applied Biosystems, Darmstadt, Germany) using 10 ng of transcribed RNA, Rox dye as internal control (Invitrogen), FastStart Taq Polymerase (Roche Diagnostics, Mannheim, Germany), and gene-specific primers in combination with SYBR-Green chemistry (Molecular Probes, Eugene, OR). PCR amplification was carried out for 10 min at 95°C, 40 cycles for 10 s at 95°C, and 1 min at 60°C. Specificity of the amplification product was verified by melting curve analysis. For each group, three RNA samples were used. For normalization of the samples, distribution of 18S ribosomal RNA was measured. Quantification was carried out using qgene software. Primers were designed using Primer Express software (Applied Biosystems) based on Unigene clusters and GenBank accession numbers, respectively (given in parentheses). The primer sequences are the following: 5′-3′ direction, r18S (NC_001665) ACATCCAAGGAAGGCAGCAG (primer 1) and TTTTCGTCACTACCTCCCCG (primer 2); and for Perlecan (Mm.273662) GGGAGGCCCGTCTTGTCT (primer 1) and GTGTTGACCGCCACATTAGGA (primer 2).

Data are shown as mean ± SD. The data were compared by ANOVA, and the Bonferroni multiple comparison test was used as posttest. The Mann-Whitney U test was used as posttest when analyzing the albuminuria. Significant differences were accepted when P < 0.05. Data analysis was performed using InStat.

Hyperglycemia was induced in 7-week-old mice by intraperitoneal injection of streptozotocin on days 1 and 4. Approximately 80% of the animals were diabetic 14 days after the first injection. The hyperglycemia persisted in the diabetic animals during the 8-week study period (Fig. 1). We then analyzed PKC-α expression in the kidney by immunohistochemistry. In WT animals, PKC-α was predominantly expressed in the glomeruli. Under diabetic conditions, an increased expression of PKC-α in glomeruli was observed (data not shown). No glomerular PKC-α expression was found in control or diabetic PKC-α^−/− mice (data not shown).

During the study period, a significant increase in body weight was observed (Table 1). After 8 weeks of diabetes, the animals were killed and the kidneys removed and analyzed further. A significant increase in the kidney weight and the kidney–to–body weight ratio during hyperglycemia was observed in WT and PKC-α^−/− mice. The kidney–to–body weight ratio increased significantly from 8.8 ± 0.4 to 11.3 ± 1.4% (P < 0.01) in the WT animals and from 7.8 ± 0.4 to 10.8 ± 1.5% (P < 0.01) in the PKC-α^−/− mice (Table 1).

In contrast to our observations on renal hypertrophy, we found a significant effect of PKC-α deficiency on albumin excretion. After 2 weeks of hyperglycemia, a slight increase in albumin excretion was present in the WT animals, whereas no albuminuria was observed in PKC-α^−/− mice (Fig. 2). After 8 weeks of diabetes, a further increase was observed in the WT mice, whereas only a few of the PKC-α^−/− animals showed slight albuminuria (Fig. 2). The albumin–to–creatinine ratio increased from a median of 7.38 to 21.43 g/mol in the diabetic WT animals (P < 0.01). In contrast, the albumin–to–creatinine ratio remained stable, with a median of 8.03 g/mol in control and 10.79 g/mol in hyperglycemic PKC-α^−/− mice.

The lack of albuminuria in hyperglycemic PKC-α^−/− mice indicates that PKC-α is involved in the diabetes-induced changes of the glomerular basement membrane (GBM) or the slit membrane. We therefore analyzed the glomeruli using histology and electron microscopy (Fig. 3). No apparent difference in glomerular volume and mesangial expansion was observed between the two diabetic groups (Fig. 3A and C). Hyperglycemia induced a significant increase of the mean glomerular V_T from 226 ± 27 to 452 ± 79 μm³ × 10³ (P < 0.01) in WT mice and from 205 ± 27 to 387 ± 115 μm³ × 10³ (P < 0.01) in PKC-α^−/− mice (Table 1). The electron micrographs showed no signs of endothelial cell swelling or podocyte damage. Podocytes displayed well-developed foot processes without an obvious difference between both diabetic groups. Inspection of the GBM revealed a wide variation of GBM thickening in the diabetic animals. Thickened GBM segments with formation of subepithelial “humps” seemed to be less frequent in diabetic PKC-α^−/− mice as compared with diabetic WT animals (Fig. 3B and D).

We next investigated the molecular composition of the GBM using immunohistochemistry. We first assessed the expression of perlecan, a major heparan sulfate proteoglycan of the GBM. As shown in Fig. 4A and C, there was a strong expression of perlecan in nondiabetic animals. The expression of perlecan was greatly reduced in the glomeruli of diabetic WT animals (Fig. 4B). This diabetes-induced loss of glomerular perlecan expression was completely prevented in diabetic PKC-α^−/− mice (Fig. 4D). In addition, we analyzed the perlecan mRNA expression. We found a significant reduction of perlecan mRNA levels in whole-kidney extracts from diabetic versus control WT mice (0.08 ± 0.14 vs. 1.0 ± 0.24 relative units; P < 0.05) (Fig. 4E). No reduction of perlecan mRNA levels was observed in diabetic versus control PKC-α^−/− mice (1.43 ± 0.59 vs. 0.76 ± 0.26 relative units). Further analysis of the extracellular matrix molecules type III collagen, type IV collagen, and fibronectin showed an increase under diabetic conditions in both groups (data not shown).

To elucidate the possible mediators between PKC-α and perlecan expression in the GBM, we analyzed the role of TGF-β₁ and VEGF and VEGF receptor II (VEGFR-II). The results for TGF-β₁ are shown in Fig. 5. Immunohistochemistry showed a comparable increase of the glomerular TGF-β₁ expression in hyperglycemia with no significant difference between WT and PKC-α^−/− mice (Fig. 5A–E). Furthermore, no significant difference in the TGF-β₁ expression between WT and PKC-α^−/− diabetic mice was detected by Western blot (Fig. 5F).

To elucidate the possible role of VEGF, we analyzed the glomerular expression of VEGF165 by immunohistochemistry. These results are shown in Fig. 6. Only a weak expression of VEGF165, mainly located in podocytes, was observed in nondiabetic control animals of both groups. Under hyperglycemic conditions, a significant increase of VEGF was observed in the WT animals. In contrast, this increase was significantly reduced but not completely abolished in PKC-α^−/− mice (Fig. 6). These results with VEGF165 prompted us to investigate the influence of PKC-α expression on the diabetes-induced expression of VEGFR-II. These results are shown in Fig. 7. The VEGFR-II expression was weak in the nondiabetic animals, as shown by immunohistochemistry. The fluorescence signal was mainly located in the endothelial and mesangial cells. Under hyperglycemic conditions, a significant increase of the VEGFR-II expression was observed in WT animals compared with nondiabetic mice (Fig. 7). This increase was significantly reduced but not abolished in PKC-α^−/− mice (P < 0.01). These results indicate that the expression of VEGF and VEGFR-II is regulated by PKC-α.

The most important finding of our study is that diabetic PKC-α^−/− animals were protected from albuminuria, whereas WT mice were not. By electron microscopy, we found no convincing evidence that a different basement membrane thickness or deterioration of podocyte structure explained the finding. However, the hyperglycemia-induced downregulation of the negatively charged heparan sulfate proteoglycan perlecan was completely prevented in the PKC-α^−/− mice. The glucose-induced albuminuria seems to be mediated by the PKC isoform α via downregulation of heparan sulfate proteoglycans in the GBM. It is interesting that the lack of albuminuria in the diabetic PKC-α^−/− animals is dissociated from the diabetes-induced renal and glomerular hypertrophy. We then asked whether expression of TGF-β₁ and/or VEGF is implicated in the PKC-α–mediated changes of the basement membrane. The hyperglycemia-induced glomerular expression of VEGF165 and its receptor flk-1 was ameliorated in PKC-α^−/− mice, whereas expression of TGF-β₁ was not affected by the lack of PKC-α. Our findings indicate that two important features of diabetic nephropathy—glomerular hypertrophy and albuminuria—are differentially regulated. In addition, glucose-induced glomerular expression of VEGF in diabetes may be mediated by PKC-α.

It is assumed that the loss of negative charges in the basement membrane might be causative for the development of albuminuria (17,18). According to the Steno hypothesis, diabetic complications result from the loss of heparan sulfate proteoglycans in the basement membrane (19). This hypothesis is based on immunohistochemical findings from diabetic patients with early nephropathy. These patients have a reduced amount of heparan sulfate in the GBM (20,21). Perlecan is the most common heparan sulfate proteoglycan in the body (22,23). Perlecan was not detectable by immunohistochemistry in WT diabetic animals; however, downregulation was prevented in the PKC-α^−/− diabetic mice. Therefore, we assume that the missing loss of negative charges in the PKC-α^−/− diabetic mice might be responsible for the lack of albuminuria in these mice. It has been demonstrated that perlecan core protein is downregulated in the liver of diabetic patients (24). In the retinal basement membrane of diabetic patients, a diminished or unaltered perlecan expression has been reported (25,26). Thus far, no report exists about the expression of perlecan in the kidney of humans or animals with diabetes. However, incubation of glomerular epithelial cells with 30 mmol/l glucose resulted in a reduction of the heparan sulfate proteoglycan synthesis (27). In diabetes, heparan sulfate proteoglycan side chains are diminished or altered (28,29). It has been suggested that PKC isoforms may be involved in the regulation of heparan sulfate production (30). Our data suggest that PKC-α plays a key role in the synthesis of heparan sulfate proteoglycan perlecan mRNA expression. Further studies will be needed to elucidate the molecular pathway of this interaction. It will also be necessary to analyze the expression of heparan sulfate synthesis and other heparan sulfate proteoglycans such as agrin (31). It has been suggested that perlecan plays a role in the development of arteriosclerotic diseases. Therefore, PKC-α might be involved in the development of other diabetes-related complications. Further studies are necessary to clarify this topic.

Our study suggests a role for PKC-α in the regulation of VEGF165 and its receptor in diabetes. VEGF is a cytokine that potentially induces angiogenesis, endothelial permeability, and endothelium-dependent vasodilation and represents a key factor for the development of (proliferative) diabetic retinopathy (32). However, the role of VEGF in the development of diabetic nephropathy has been less thoroughly investigated. VEGF and its receptor are expressed in the healthy human kidney (33). VEGF expression is increased in patients with diabetic nephropathy (34,35), and patients with type 1 diabetes and diabetic nephropathy have increased plasma levels of VEGF (34). Furthermore, blockade of VEGF with systemic antibody administration decreased albuminuria in rats with streptozotocin-induced hyperglycemia or db/db mice with type 2 diabetes (36,37). Various cell culture studies have proved that high-glucose–induced stimulation of VEGF expression is regulated via a PKC-dependent pathway (35,38,39). However, the specific PKC isoform involved in this cascade has not been identified yet (38,40,41). Our data suggest that the PKC-α isotype is involved in the regulation of VEGF and its receptor (42). Therefore, our finding of a reduced glomerular expression of VEGF and VEGFR-II in diabetic PKC-α^−/− compared with WT mice could be one possible explanation for the observed difference in the development of albuminuria.

We could not find a role for PKC-α in the diabetes-induced activation of TGF-β₁. TGF-β₁ plays a key role in the development of renal hypertrophy and accumulation of extracellular matrix in diabetic nephropathy (43). We have recently shown that high glucose leads to a PKC-α–dependent expression of TGF-β₁ in vascular smooth muscle cells under in vitro conditions (11). Surprisingly, our present findings do not support a role for PKC-α in the regulation of TGF-β₁ expression in vivo. This discrepancy might be explained by the fact that glucose stimulates different PKC isoforms in different cell types. It could also be that short-term activation of a PKC isoform by high- glucose concentrations in vitro has a different effect on TGF-β₁ expression than the long-term effects of hyperglycemia in vivo.

Ziyadeh et al. (44) showed that long-term inhibition of TGF-β₁ with a specific antibody almost completely prevented the mesangial expansion and partially prevented renal hypertrophy in db/db mice. However, the breakdown of the glomerular filtration barrier with development of albuminuria was not prevented by inhibition of TGF-β₁ in this study. Our data confirm these previous results and extend the findings by the fact that PKC-α signaling cascade is not involved in the regulation of TGF-β₁ in the diabetic milieu. Furthermore, our findings suggest that hyperglycemia induces the pathological changes in the kidney by activation of different mediator systems whereby VEGF is associated with endothelial and podocyte function, whereas TGF-β₁ plays a role in mesangial expansion.

Inoguchi et al. (45) observed an increased expression of the PKC isoform βII in tissues from diabetic animals and hypothesized that this PKC isoform is primarily responsible for the glucose-induced effects in diabetes. Further evidence for this hypothesis stems from reports that a specific PKC-β inhibitor ameliorates hyperglycemia-induced changes in the kidney. In rats with streptozotocin-induced type 1 diabetes and db/db mice with type 2 diabetes, the development of albuminuria and mesangial expansion could be prevented by the administration of the aforementioned inhibitor (46–48). However, it is interesting that only PKC-βI but not PKC-βII is expressed in glomerular cells (49,50). From the existing data, it can be concluded that PKC-α and -β are both important in the development of diabetic complications. We assume that PKC-β is more important in the upregulation of TGF-β₁ and for the development of glomerular hypertrophy under diabetic conditions, whereas PKC-α seems to play a critical role in the development of albuminuria by perpetuating the loss of the negatively charged heparan sulfate and upregulation of glomerular VEGF expression. As the structure of PKC-α and -β are very similar, it is conceivable that inhibition of one of the two isoforms will lead to a compensatory rise in the activity of the other isoform. Therefore, blockade of both PKC isoforms may be necessary to achieve a therapeutic effect.

This work was supported by a grant-in-aid from the Deutsche Forschungsgemeinschaft to H.H. (Ha 1388-7/1) and from the EFSD (European Foundation for the Study of Diabetes)-Servier grant for vascular complications of type 2 diabetes to M.M.