Reduction of Diabetes-Induced Oxidative Stress, Fibrotic Cytokine Expression, and Renal Dysfunction in Protein Kinase Cβ–Null Mice

PKC-β inhibitors were provided by Dr. Michael Leitges. The derivation of PKC-β^−/− mice has been described previously (19). Wild-type mice with the same genetic background (129×C57BL6) as the knockout mice were used as control animals. Diabetes was induced in 6-week-old male PKC-β^−/− and wild-type mice fasted for 12 h with intraperitoneal injections of streptozotocin in citrate buffer (70 mg/kg body wt) (Sigma Chemical, St Louis, MO) on 2 consecutive days. The diabetic state was confirmed with blood glucose levels exceeding 250 mg/dl. Control animals received an injection of sterile citrate buffer. All experiments were conducted in accord with the Joslin Diabetes Center’s Animal Care and Use Committee guidelines. Blood pressure was monitored by tail-cuff plethysmography (Ueda Electric Works, Tokyo, Japan). Glucose levels from tail blood samples were measured with a Glucometer Elite XL (Bayer, Elkhart, IL). At 8 and 24 weeks after diabetes, mice were placed in metabolic cages for collection of urine. The samples were stored frozen at −80°C. Albuminuria, urinary isoprostane, and 8-hydroxydeoxyguanosine (8-OHdG) were all measured by enzyme-linked immunosorbent assay kits from Exocell (Philadelphia, PA), Oxis International (Foster City, CA), and Genox (Baltimore, MD), respectively.

Renal clearance studies were performed at 8 weeks of diabetes as previously described, with modification for mice (20). Briefly, mice were anesthetized with an intraperitoneal injection of thiopental sodium. A catheter was inserted into the left jugular vein for continuous infusion of inulin and para-aminohippurate (PAH) solution. The urinary bladder was catheterized for urine collection. After equilibration, blood samples from mice tail and urine samples were obtained for measurement of inulin and PAH at timed intervals. The concentrations of inulin in plasma and urine were measured by cysteine-tryptophan reaction, and PAH was determined by a calorimetric technique as described previously (20). GFR was calculated from inulin clearance and renal plasma flow (RPF) from PAH clearance and filtration fraction from a standard formula, the GFR-to-RPF ratio.

For morphometric analysis of glomeruli, kidneys from animals were fixed in 10% neutral buffered formalin, embedded in paraffin, and sectioned at 4 μm. Sections were stained with hematoxylin and periodic acid Schiff. Sections were coded and read by an observer unaware of the experimental protocol applied. In each animal, glomerular area was measured in 50 consecutive glomeruli encountered in a back and forth scan between the outer and inner cortex. The images were imported into Image-Pro Plus (Media Cybernetics, Silver Spring, MD), and the glomerular tuft area was quantified for each glomerular cross-section.

To evaluate PKC activity in the renal cortex, a modified in situ PKC assay was used (20). To assess activation of different isoforms of PKC, expression of PKC isoforms in the cytosolic and membranous fractions of the renal cortex were assessed by immunoblot analysis. The fractionation of the renal cortex was performed with a method involving ultracentrifugation steps as previously described (10).

The cytosolic and membranous fractions obtained were then used for immunoblotting according to methods described previously (10). Antibodies to the various PKC isoforms were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). p47^phox antibody was supplied from Abcam (Cambridge, MA).

Real-time PCR was performed to evaluate the mRNA expression of transforming growth factor (TGF)-β, connective tissue growth factor (CTGF), type IV and VI collagens, endothelin-1 (ET-1), and vascular endothelial growth factor (VEGF) in the renal cortex of various groups of mice. Total RNA extraction and reverse transcription were performed with the RNeasy Mini kit (Qiagen, Valencia, CA) and the First-Strand cDNA Synthesis Kit (Amersham, Piscataway, NJ) according to the manufacturer’s instructions. Quantitative real-time PCR of TGF-β1, CTGF, and collagen IV and VI was performed with a TaqMan Universal PCR MasterMix and TaqMan Gene Expression assays (TGF-β1, Mm 00441724_m1; CTGF, Mm 00515790_g1; collagen IV, Mm 00802372_m1; and collagen VI, Mm 00487160_m1) (Applied Biosystems, Forest City, CA). Primer and probe sequences of NADPH oxidase subunits are listed as follows: p47^phox (MN 010876.2, forward 5′-ACCTTCATTCGCCATATCGCCCT-3′, reverse 5′-TTCTGTAGACCACCTTCTCCGACA-3′, and probe 5′6FAM-CCCAGCCAGCACTATGTGTACATGTT-TAMRA3′), Nox2 (MN 007807.2, forward 5′-TGCAGCCTGCCTGAATTTCAACTG-3′, reverse 5′-AGATGTGCAATTGTGTGGATGGCG-3′, and probe 5′6FAM-TGCGTGTTGCTCGACAAGGATTCGAA-TAMRA3′), and Nox4 (NM_015760, forward 5′-TGTTGGGCCTAGGATTGTGTT-3′, reverse 5′-AAAAGGATGAGGCTGCAGTTG-3′, and probe 5′-6FAM-AAGCAGAGCATCTGCATCTGTCCTCAACC-TAMRA 3′). Primers and probes for ET-1 and VEGF were previously described (21). 18S ribosomal RNA expression was used for normalization.

All data are presented as means ± SD. Comparisons between groups were performed using one-way ANOVA, with post hoc correction by Tukey’s method. Comparisons of urinary albumin excretion between groups were made with the Mann-Whitney nonparametric test, since the values were not normally distributed. Comparisons between two groups were performed using an unpaired t test. P < 0.05 was considered significantly different.

Both diabetic wild-type and PKC-β^−/− mice exhibited marked hyperglycemia during the experimental periods (8 and 24 weeks) (Table 1). Diabetic wild-type and PKC-β^−/− mice had significantly lower body weight at the end of the experimental period compared with nondiabetic counterparts, regardless of genotype, but higher than at beginning of the study. There was no significant difference between wild-type and PKC-β^−/− mice in body weight, blood glucose, or blood pressure. Similarly, there was no significant difference in these physiological parameters between diabetic wild-type and PKC-β^−/− mice at 8 and 24 weeks.

Induction of diabetes for 8 weeks was associated with elevated PKC activity in the renal cortex in wild-type mice, from 7.35 ± 0.66 to 13.9 ± 2.39 pmol · mg protein⁻¹ · min⁻¹, an increase of 89% (P < 0.001). When diabetes was induced in PKC-β^−/− mice, activated PKC pool was only elevated by 46% to 9.96 ± 0.9 pmol · mg protein⁻¹ · min⁻¹, which was significantly less than the increase in wild-type mice (P < 0.01, Fig. 1A). To compare the relative contribution of different isoforms of PKC to the elevated PKC activity observed in the diabetic state, expression of various isoforms of PKC in membranous and cytosolic fractions of renal cortex were examined (Fig. 1B). In wild-type mice, diabetes increased PKC-β1 expression by 58% (P < 0.01) in the membranous fraction, consistent with activation of this isoform. There was no detectable PKC-β1 in the PKC-β^−/− mice. Diabetes increased the expression of PKC-α by 31% (P < 0.01) in the membranous fractions, which was observed in both wild-type and PKC-β^−/− mice. There was no difference in expression of PKC-α, -δ, and -ε isoforms in the cytosolic fractions between nondiabetic wild-type and PKC-β^−/− mice.

Renal function was evaluated by GFR and filtration fraction (FF) following 8 weeks of diabetes (Fig. 2). GFR and FF increased in diabetic wild-type mice by 55 and 57%, respectively, compared with nondiabetic wild-type mice (P < 0.05). In contrast, increases in GFR and FF were not observed in diabetic PKC-β^−/− mice for the same duration. Furthermore, PKC-β^−/− mice with diabetes had significantly lower GFR (P < 0.05) and FF (P < 0.05) than wild-type mice with diabetes.

After 8 weeks of diabetes, urinary albumin excretion increased 10-fold in wild-type mice (Fig. 3, P < 0.01). Induction of diabetes in PKC-β^−/− mice was also associated with an increase in urinary albumin excretion by 4.4-fold, which was significantly less than that observed in diabetic wild-type animals (P < 0.01). Similarly, following 24 weeks of diabetes, albuminuria remained significantly increased in both diabetic wild-type and PKC-β^−/− mice. However, the amount of albuminuria in diabetic PKC-β^−/− mice was significantly less than in wild-type mice exposed to diabetes of the same duration (Fig. 3, P < 0.01).

Elevation in oxidative stress has been associated with diabetic nephropathy as measured by increases in the levels of several urinary oxidants (22). Thus, we measured the urinary excretions of isoprostane and 8-OHdG. After 8 weeks of diabetes, urinary isoprostane levels were increased 3.6-fold in diabetic wild-type mice as compared with nondiabetic wild-type mice (P < 0.001). This increase in urinary isoprostane was significantly attenuated in the diabetic PKC-β^−/− mice (Fig. 4A, P < 0.001). Similarly, after 24 weeks of diabetes, diabetic PKC-β^−/− mice continued to excrete significantly less isoprostane when compared with diabetic wild-type mice (P < 0.01). Urinary 8-OHdG levels were not significantly increased in diabetic wild-type mice compared with nondiabetic wild-type control mice. Urinary 8-OHdG levels were significantly reduced in diabetic PKC-β^−/− mice compared with diabetic wild-type mice (P < 0.01, Fig. 4B).

Since activation of NADPH oxidase has been reported to be induced by diabetes and is associated with increased oxidant production induced by hyperglycemia, we assessed changes in several components of this enzyme complex (15). As shown in Fig. 5A, there was increased membranous translocation of p47 subunit of NADPH oxidase by 1.73 ± 0.4-fold (P < 0.01) in diabetic wild-type mice compared with wild-type control. This activation was not observed in diabetic PKC-β^−/− mice. Furthermore, diabetes also significantly increased the mRNA expressions of p47 subunit of NADPH oxidase, Nox2, and Nox4 as compared with wild-type control mice. These increases were not seen in diabetic PKC-β^−/− mice (Figs. 5B–D). Expression of Nox1 mRNA level was too low to be assessed credibly (data not shown).

Elevation of ET-1 in glomeruli has been postulated to be involved in diabetes-induced alteration of hemodymics and changes in extracellular matrix (23). As reported previously, expression of ET-1 mRNA in renal cortex of diabetic wild-type mice was significantly increased by twofold (1.90 ± 0.21-fold change versus nondiabetic wild-type mice, P < 0.01) as compared with wild-type control mice. In contrast, the mRNA level of ET-1 was not increased in diabetic PKC-β^−/− mice (fold change compared with nondiabetic wild-type mice: nondiabetic PKC-β^−/− 0.31 ± 0.21-fold versus diabetic PKC-β^−/− 0.37 ± 0.16-fold, P > 0.05), in which expression level was significantly less compared with that observed in diabetic wild-type mice (diabetic PKC-β^−/− 0.37 ± 0.16-fold versus diabetic wild-type 1.90 ± 0.21-fold, P < 0.01). The baseline expression of ET-1 in renal cortex of nondiabetic PKC-β^−/− mice was also significantly less than that of nondiabetic wild-type mice (0.31 ± 0.21-fold of nondiabetic wild-type mice, P < 0.01).

Similiarily, diabetes induced a significant increase in the renal cortical expression of VEGF mRNA in diabetic wild-type mice (1.31 ± 0.16-fold of nondiabetic wild-type mice, P < 0.05) but not in diabetic PKC-β^−/− mice (fold change compared with nondiabetic wild-type: nondiabetic PKC-β^−/− 0.43 ± 0.14-fold versus diabetic PKC-β^−/− 0.67 ± 0.28-fold, P > 0.05). Ablation of the PKC-β gene reduced the basal expression of VEGF mRNA in the renal cortex of nondiabetic PKC-β^−/− mice as compared with nondiabetic wild-type mice (0.43 ± 0.14-fold of nondiabetic wild-type mice, P < 0.01).

Increases in the expressions of fibrotic factors, such as TGF-β and CTGF, and extracellular matrix proteins, such as collagens, are believed to be partly responsible for the glomerular enlargements and fibrosis (24). Wild-type mice with diabetes had increased mRNA expression of TGF-β (191%, P < 0.01), CTGF (63%, P < 0.01), collagen IV (134%, P < 0.01), and collagen VI (102%, P < 0.01), which were observed following 8 weeks of diabetes (Fig. 6). In contrast, there was no significant increase in mRNA expression of CTGF and collagen IV and VI among diabetic PKC-β^−/− mice after 8 weeks (Fig. 6B–,D). There was significant increase in expression of TGF-β after 8 weeks of diabetes in the PKC-β^−/− mice, although this was significantly less than the increase observed in wild-type mice (Fig. 6A). Following 24 weeks of diabetes, there was a sustained increase in the expression of TGF-β, CTGF, and collagen IV and VI in the diabetic wild-type mice (Fig. 6). Long duration of diabetes was associated with increased expression of only collagen VI in the PKC-β^−/− mice, although the increase was less than in the wild-type mice.

Kidney weight was significantly increased (24%, P < 0.01) in wild-type mice following 8 weeks of diabetes (Fig. 7A). This increase in kidney weight was not found in diabetic PKC-β^−/− mice for the same duration. Likewise, while there was significant increase in kidney weight in wild-type mice after 24 weeks of diabetes, there was significantly less renal enlargement in PKC-β^−/− mice with diabetes of the same duration (Fig. 7A).

Morphometric analysis showed that glomerular size was increased by 24.2% (P < 0.05) in wild-type mice following 24 weeks of diabetes. In contrast, there was no glomerular enlargement observed in PKC-β^−/− mice exposed to the same duration of diabetes (Fig. 7B).

Previous studies have used RBX to determine which of the various renal dysfunctions and pathologies are due to the activation of PKC-β isoforms (10,12,13). However, the results are not definitive, since RBX is an isoform-selective inhibitor, which at the dose used may inhibit PKC-α activation by 10–20% (10). We showed that PKC-β^−/− mice could be used to test the specific effects of PKC-β isoforms on renal function, since immunoblot analysis of the renal cortex demonstrated that diabetes can still activate PKC-α isoform in these mice, unlike the wild-type mice where both PKC-α and -β1 were increased. These results show for the first time that PKC-α and -β are activated independently by diabetes or hyperglycemia; several reports have shown that PKC-β activation may increase PKC-α expression (25). This suggestion of PKC-α and -β isoforms being independently activated is more than a biochemical interest because it indicates that a selective PKC-β isoform inhibitor may not completely prevent all renal pathologies of diabetes unless it can also partially inhibit the activation of PKC-α.

One of the most interesting findings in this study is the protection from hyperglycemia-induced oxidative stress in mice lacking PKC-β isoforms, as shown by the significant decreases in urinary isoprostane and 8-OHdG, which were maintained even after 24 weeks of diabetes. This lowering of oxidative stress appears to be partially due to renal origin, since several subunits and isoforms of NADPH oxidase, a major producer of oxidants (26), were found to be decreased in their expression in the renal cortex of PKC-β^−/− mice. The decreases in Nox activities appear to be due to lowering of both activation and expression, since the translocation of p47^phox and the expressions of Nox2 and -4 were significantly reduced in the diabetic PKC-β^−/− mice. These data suggest that PKC-β is a major inducer of oxidative stress in diabetes, possibly via the activation of Nox complex, although confirmation by NADPH oxidase activities would be needed to support the changes in mRNA levels. This is consistent with recent work suggesting that diabetes leads to increased expression of NADPH oxidase subunits Nox4 and p22^phox, which may be responsible for overproduction of ROS and the renal damage seen in diabetes (15,16). Nox4-derived ROS were found to induce renal hypertrophy and increase fibronectin expression (27). In addition, chronic inhibition of NADPH oxidase was found to prevent podocyte apoptosis, albuminuria, and mesangial expansion in a mouse model of diabetes (28). Amelioration of several markers of renal damage in mice lacking PKC-β, along with absent activation of NADPH oxidase, suggests NADPH oxidase is a downstream target of PKC-β. Previous studies have demonstrated that glucose and fatty acids increase ROS production via PKC-dependent activation of NADPH oxidase (29). Recent in vitro studies using inhibitors specific for PKC-β isoforms suggest that PKC-β is the main isoform responsible for activating Nox (30). Furthermore, the increased activation of NADPH oxidase and increased urinary 8-hydroxyguanosine excretion in diabetic animals can be improved by treatment with the PKC-β inhibitor ruboxistaurin mesylate (22). Our in vivo studies confirmed the key role of PKC-β in diabetes-induced NADPH oxidase activation and the protective effect of PKC-β against ROS-induced renal damage. This is consistent with the renoprotective effect observed in both animal and human studies using RBX (10,12,14). However, the results from the present study also showed that the increases in isoprostane were not completely suppressed in diabetic PKC-β^−/− mice, which is different from findings when diabetic animals were treated with RBX. This suggests that PKC-α can also induce oxidant production either via Nox or other pathways. Further studies are needed to determine cellular sites of oxidant production by measuring intracellular oxidative markers such as nitrosylated proteins and 4-hydroxynonenal adducts in renal tissues and cells.

The source of ROS production remains controversial, with evidence suggesting that several pathways, including enhanced formation of AGEs (6), altered polyol pathway activity (31), increased PKC activity (29), and increased superoxide release from the mitochondria (6), may be involved. Recently, there is accumulating evidence that NADPH oxidase is the most important source of ROS production in vascular tissues and may be involved in vascular damage caused by hypercholesterolemia, hypertension, and atherosclerosis (32–34). The kidney is vulnerable to oxidative damage and is known to express NADPH oxidase and generate ROS (26). Increased production of ROS within the glomeruli decreases NO bioactivity on mesangial contraction and arteriolar tone and may contribute directly to the renal hemodynamic and vascular abnormalities observed during the initiation and established phase of diabetic nephropathy (28). The beneficial effects of ACE inhibitors (35) and angiotensin receptor blockers (36) could also be related to reduction of ROS production, since angiotensin has been reported to induce or activate NADPH oxidase (32–34).

PKC-β activation can also affect hemodynamic changes by increasing the expression of ET-1 possibly via excessive oxidant production. ET-1 has been implicated in diabetic kidney injury and may contribute to glomerular injury and interstitial tubular changes in diabetes (23,37). Changes in the expression of ET-1 in PKC-β^−/− mice are consistent with previous observations suggesting that PKC-β activation can regulate both basal and diabetes-induced ET-1 expression (38). A similar pattern of changes is noted for VEGF expression, which was increased in diabetic wild-type but not changed in the diabetic PKC-β^−/− mice. The physiological role of VEGF overexpression in the renal cortex is not clear but will likely have effects on renal plasma flow (39). It is interesting to note that diabetes-induced ET-1 and VEGF expressions were completely suppressed in the diabetic PKC-β^−/− mice, unlike the oxidative parameters and Nox2 and Nox4 expressions, suggesting that PKC-β activation is more important for the induction of ET-1 and VEGF than PKC-α isoform.

Renal enlargement in diabetes results especially from tubular hypertrophy and also from glomerular enlargement and accumulation of extracellular matrix (24). This hypertrophy usually precedes the development of irreversible renal changes including glomerulosclerosis and tubulointerstitial fibrosis. The present studies demonstrate that PKC-β plays a critical role in diabetes-induced tubular hypertrophy and glomerular enlargement, since diabetic PKC-β^−/− mice did not show the expected increase in kidney weight or glomerular area. TGF-β and CTGF, in particular, have been identified as key cytokines mediating the changes that lead to extracellular matrix accumulation and glomerular enlargement (40,41). Whereas induction of diabetes led to marked increases in the expression of TGF-β and CTGF in wild-type mice, this increase in TGF-β expression was attenuated at 8 weeks of diabetes and was absent at 6 months of disease, whereas the increased expression of CTGF was completely absent in PKC-β^−/− mice. This may account for the lack of glomerular enlargement and nephromegaly and marked attenuation in diabetes-induced increase in expression of collagen IV and VI seen in the PKC-β^−/− mice. These findings are consistent with previous observations suggesting that PKC-β isoforms can mediate extracellular matrix accumulation in the diabetic kidney (10,13,42) and in other tissues (43).

In contrast to the complete lack of renal and glomerular hypertrophy in diabetic PKC-β^−/− mice, diabetes in the PKC-β^−/− mice was associated with only partial protection from increased urinary albumin excretion. In addition to extracellular matrix accumulation and mesangial expansion, several other factors are known to be important in determining albuminuria. This includes reduced glomerular nephrin expression, effacement of podocyte foot processes, as well as apoptosis and progressive loss of glomerular podocytes (24). Interestingly, effects of PKC activation on podocyte biology and functions are not known. Other pathways in addition to PKC-β may also mediate some of these processes. Recently, it was shown that hyperglycemia-induced downregulation of the negatively charged basement membrane heparin sulfate proteoglycan perlecan was prevented in mice lacking PKC-α, which also had a reduction in diabetes-associated albuminuria (44). Our data on PKC activity in renal cortex and activation of the α isoform in the PKC-β^−/− mice would be consistent with the view that activation of PKC-α may also contribute to albuminuria observed in diabetes despite it not being involved in causing extracellular matrix accumulation and renal hypertrophy.

In summary, the results derived from using the PKC-β^−/− mice have identified the specific role of PKC-β in mediating diabetes-induced increases in the expression of CTGF and TGF-β with the resultant glomerular and renal enlargement. These effects of PKC-β may be partially due to its actions on enhancing the actions or expressions of the Nox complex of oxidases. In addition, these results have also strongly suggested that PKC-α or other pathways may be partly responsible for causing diabetes-induced increases in oxidative stress and elevated albuminuria. Strategies to target the DAG-PKC pathway using isoform-specific inhibitors of PKC-β may need to take in to consideration that some inhibition of PKC-α may also be necessary to prevent most hyperglycemia-induced injury in the renal glomeruli. Inhibition of PKC-β isoforms does not appear to have significant clinical side effects. However, inhibiting both PKC-β and -α isoforms may have more side effects than inhibiting a single isoform, since PKC activation has been implicated in essential physiological functions, even for insulin secretion (45).

R.C.M. is the recipient of a Croucher Foundation Fellowship; a training fellowship from the Hong Kong Society of Endocrinology, Metabolism and Reproduction; and a William Randolph Hearst Foundation Fellowship. Y.Y. and K.I. were recipients of a mentor-based fellowship from the American Diabetes Association (ADA). This work was supported by National Institutes of Health (DK071359) and ADA (1-05-RA-61) grants.