High Glucose-suppressed Endothelin-1 Ca2+ Signaling via NADPH Oxidase and Diacylglycerol-sensitive Protein Kinase C Isozymes in Mesangial Cells*

High glucose (HG) is the underlying factor contributing to long term complications of diabetes mellitus. The molecular mechanisms transforming the glomerular mesangial cell phenotype to cause nephropathy including diacylglycerol-sensitive protein kinase C (PKC) are still being defined. Reactive oxygen species (ROS) have been postulated as a unifying mechanism for HG-induced complications. We hypothesized that in HG an interaction between ROS generation, from NADPH oxidase, and PKC suppresses mesangial Ca2+ signaling in response to endothelin-1 (ET-1). In primary rat mesangial cells, growth-arrested (48 h) in 5.6 mm (NG) or 30 mm (HG) glucose, the total cell peak [Ca2+]i response to ET-1 (50 nm) was 630 ± 102 nm in NG and was reduced to 159 ± 15 nm in HG, measured by confocal imaging. Inhibition of PKC with phorbol ester down-regulation in HG normalized the ET-1-stimulated [Ca2+]i response to 541 ± 74 nm. Conversely, an inhibitory peptide specific for PKC-ζ did not alter Ca2+ signaling in HG. Furthermore, overexpression of conventional PKC-β or novel PKC-δ in NG diminished the [Ca2+]i response to ET-1, reflecting the condition observed in HG. Likewise, catalase or p47phox antisense oligonucleotide normalized the [Ca2+]i response to ET-1 in HG to 521 ± 58 nm and 514 ± 48 nm, respectively. Pretreatment with carbonyl cyanide m-chlorophenylhydrazone or rotenone did not restore Ca2+ signaling in HG. Detection of increased intracellular ROS in HG by dichlorofluorescein was inhibited by catalase, diphenyleneiodonium, or p47phox antisense oligonucleotide. HG increased p47phox mRNA by 1.7 ± 0.1-fold as measured by reverse transcriptase-PCR. In NG, H2O2 increased membrane-enriched PKC-β and -δ, suggesting activation of these isozymes. HG-enhanced immunoreactivity of PKC-δ visualized by confocal imaging was attenuated by diphenyleneiodium chloride. Thus, mesangial cell [Ca2+]i signaling in response to ET-1 in HG is attenuated through an interaction mechanism between NADPH oxidase ROS production and diacylglycerol-sensitive PKC.

High glucose (HG) 1 is the key factor contributing to long term complications of diabetes mellitus (1). One of the phenotypic changes observed in mesangial cells exposed to HG is altered Ca 2ϩ signaling. Several groups have shown that the Ca 2ϩ signal induced by vasoactive compounds, including endothelin-1 (ET-1), is markedly reduced in the presence of HG. The mechanism(s) by which HG may depress Ca 2ϩ signaling is unknown. One possible candidate is the activation of protein kinase C (PKC) in HG. Mené et al. (2) have shown that HG inhibits Ca 2ϩ influx through store-operated channels via a PKC-dependent mechanism. An alternative postulate is the involvement of reactive oxygen species (ROS), which have been demonstrated to modify intracellular Ca 2ϩ signaling responsiveness depending on the cell type, the species of ROS, and the magnitude and duration of ROS generation. HG induces dysfunction in mesangial cells and other target cells through enhanced synthesis of autocrine growth factors such as transforming growth factor-␤ 1 , ET-1, and altered signaling via pathways such as PKC (3,4). In the last few years, enhanced production of ROS in response to HG, identified in many target cells including mesangial cells (5,6), has been postulated as a unifying mechanism causing diabetes complications (7)(8)(9).
Although ROS have been implicated in causing cell damage and apoptosis, they also play a physiological role in intracellular signaling pathways (10,11). In particular, several growth factors including ET-1, angiotensin II, plateletderived growth factor, and epidermal growth factor stimulate production of ROS as second messengers (12,13). In several cell types, signaled ROS production is due to activation of NADPH oxidase, a multicomponent enzyme (14). In phagocytic cells, the multiple subunits of NADPH oxidase are localized in subcellular compartments. gp91 phox , the catalytic moiety of the phagocyte oxidase, and p22 phox associate to form a flavocytochrome in the plasma membrane. The cytosol components p47 phox , p67 phox , p40 phox , and the small GTPase, Rac1 (or Rac2), are recruited to the membrane for assembly of a fully active oxidase (15)(16)(17). In nonphagocytic cells, most of the subunits of NADPH oxidase have been identified, although the precise mechanisms of regulation are not completely understood. A functional glomerular mesangial NADPH oxidase has been inferred through the use of diphe-nyleneiodium chloride (DPI), an inhibitor of flavoproteins, in response to cytokine (18,19) and serotonin stimulation (20). An earlier report identified the expression of components of human glomerular mesangial cell NADPH oxidase (21). To date, no report has described the role of NADPH oxidase in HG-induced altered mesangial cell phenotype.
We reasoned that if HG causes enhanced and sustained ROS generation in mesangial cells, Ca 2ϩ signaling responsiveness to ET-1 may be modified through a ROS-dependent mechanism. Since previous reports (22)(23)(24)(25)(26), including work from our laboratory (27)(28)(29)(30), have demonstrated enhanced PKC activity in response to HG, we postulated a mechanistic interaction between diacylglycerol (DAG)-dependent PKC and enhanced NADPH oxidase function.
Cell Culture-Primary rat mesangial cells were isolated from Sprague-Dawley rat kidney glomeruli and characterized as previously described (32). Passages 12-17 were used for all studies. Mesangial cells were grown in Dulbecco's modified Eagle's medium containing 20% fetal bovine serum to confluence and then growth-arrested in 0.5% fetal bovine serum containing either 5.6 mM glucose (NG) or 30 mM glucose (HG) for 48 h.
Confocal Imaging of Intracellular Ca 2ϩ -Mesangial cells were cultured on glass coverslips to subconfluence under the above conditions and then loaded with 2.5 M fluo-3 in Dulbecco's modified Eagle's medium (containing CaCl 2 ) with 0.02% Pluronic F-127 for 60 min at 37°C. The coverslip was mounted in a chamber on the stage of a Zeiss confocal microscope (LSM 410; Dü sseldorf, Germany), and the cells were imaged prior to and during the response to ET-1. Digitized images were captured every 15 s for 120 s. The digitized confocal images from basal condition and peak response of total cell [Ca 2ϩ ] i were analyzed using Scion Image Analysis (Scion Corp., Frederick, MD). For each condition, 30 -35 cells from different coverslips of separate experiments were analyzed. Intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) was calculated as follows, where, K d for fluo-3 was 320 nM (33). The R min and R max represent the fluorescence ratios measured with excitation at 340/380 nm under minimum free Ca 2ϩ (bound with EGTA) and maximum Ca 2ϩ (ionomycin plus 10 mM CaCl 2 ). S b2 represents the fluorescence at saturating Ca 2ϩ levels. S f2 is the fluorescence at 380-nm excitation in the absence of Ca 2ϩ . Autofluorescence, measured by quenching fluo-3 with MnCl 2 , was subtracted from the R max and R min values (32,34). Transient Transfection of GFP-PKC and PKC-Inhibition-Mesangial cells were transiently transfected with either GFP-PKC-␤ or GFP-PKC-␦ using FuGENE6 (Roche Applied Science) as instructed by the manufacturer. To inhibit PKC-activity, mesangial cells were pretreated with the PKC-peptide inhibitor (Zi; 10 M, 24 h) (31) in HG. Cells were loaded with fluo-3 and imaged for Ca 2ϩ response to ionomycin or ET-1.
Reverse Transcriptase-Polymerase Chain Reaction Expression of p47 phox and ␤-Actin-Total RNA was extracted using the Qiagen RNeasy kit (Qiagen) as instructed by the manufacturer. The first strand cDNA was synthesized using RevertAid First strand synthesis kit (Fermentas) according to the manufacturer's instructions. After first-strand synthesis of DNA, 3 l of cDNA was amplified by using specific primers selected on the basis of published sequences (35) including 5Ј-AATGTAGCTGACATCATGGGT-3Ј and 5Ј-TCTAGGAGCT-TATGAATGACC-3Ј (320-bp product). The sequence for the ␤-actin primers is as follows: 5Ј-TTGTAACCAACTGGGACGATATGG-3Ј and 5Ј-GATCTTGATCTTCATGGTGCTAGG-3Ј. PCR products were electrophoresed on a 1.5% agarose gel, and bands corresponding to reverse transcriptase-PCR products were visualized by UV light.
Confocal immunofluorescence of p47 phox and PKC-Mesangial cells were cultured on coverslips to subconfluence and growth-arrested in 0.5% fetal bovine serum NG or HG. Cells were fixed in 3.7% formaldehyde and permeabilized with methanol. To prevent nonspecific binding of antibody, cells were blocked with goat serum containing 0.1% bovine serum albumin and then incubated with anti-p47 phox monoclonal antibody or PKC-␦ antibody. The primary antibody was detected with fluorescein isothiocyanate-conjugated secondary antibody using confocal image analysis.
Transient Expression of p47 phox Antisense Oligonucleotide-Mesangial cells were grown to 80% confluence on coverslips and transiently transfected with either antisense or scrambled phosphorothioate-modified oligonucleotide for p47 phox using Effectene or Fugene6 transfection reagents according to the manufacturer's instructions. The sequence to the p47 phox antisense oligonucleotide was derived from a published report by Bey and Cathcart (36). The sequence is as follows: TTCACCTGGCT-GTCATTGG. Cells were transfected with a 5 M concentration of the oligonucleotide for 48 h, and decreased expression of p47 phox protein was ascertained by immunoreactivity of p47 phox using confocal imaging.
Measurement of ROS-Generation of intracellular ROS was measured with the fluoroprobe carboxymethyl-H 2 -dichlorofluorescein diacetate (CM-H 2 DCFDA; Molecular Probes). CM-H 2 DCF-DA is a nonpolar compound that is converted into a nonfluorescent polar derivative (CM-H 2 DCF-DA) by cellular esterases after incorporation into cells. CM-H 2 DCF-DA is rapidly oxidized to the fluorescent 2Ј,7Јdichlorofluorescein in the presence of intracellular ROS. Briefly, mesangial cells were incubated for 60 min at 37°C with 10 mol/liter CM-H 2 DCF-DA in Dulbecco's modified Eagle's medium in NG or HG (1, 3, 24, or 48 h). Fluorescence intensity was measured by confocal microscopy (Zeiss; excitation 488 nm, emission 513 nm). Average intensity for each experimental group of cells was determined using Scion Image Analysis software, and values were expressed as above control. As a positive control, cells were stimulated with 100 M H 2 O 2 .
Statistical Analysis-All results are expressed as means Ϯ S.E. Statistical analysis was performed using InStat 2.01 statistics software (Graph Pad, Sacramento, CA). The means of three or more groups were compared by one-way analysis of variance. If significance of p Ͻ 0.05 was obtained in the analysis of variance, the Tukey multiple comparison post-test was applied. Differences described as significant are p Ͻ 0.05 unless otherwise stated.

Ca 2ϩ Signaling in NG and HG in Response to Ionomycin-To
study Ca 2ϩ signaling, mesangial cells were grown on glass coverslips and growth-arrested in 5.6 mM (NG) glucose or 30 mM (HG) glucose for 48 h and stimulated with the calcium ionophore, ionomycin. Ca 2ϩ release was determined by converting the mean pixel intensity of the entire cell into values [Ca 2ϩ ] i expressed in nM using the calibration formula (32).
There was no difference between the Ca 2ϩ response in NG and HG to ionomycin at 100 nM (data not shown). However, at a low concentration of ionomycin (12.5 nM), the peak total cell [Ca 2ϩ ] i response in NG was 657 Ϯ 86 nM, and in HG, the response was reduced to 346 Ϯ 22 nM (Fig. 1).

DAG-sensitive PKC Inhibition Reverses HG-suppressed Ca 2ϩ
Release in Response to ET-1-To study the Ca 2ϩ response of mesangial cells to the vasoactive peptide ET-1, cells were grown on glass coverslips and growth-arrested in NG or HG glucose for 48 h and stimulated with ET-1 (50 nM). In all experiments, values from total cell [Ca 2ϩ ] i release from basal and peak response are reported. In NG, the total cell peak [Ca 2ϩ ] i response to ET-1 (50 nM) was 630 Ϯ 102 nM, which was attenuated to 159 Ϯ 15 nM in HG (p Ͻ 0.001 versus NG peak) (Fig. 2). To determine whether PKC was involved in decreased Ca 2ϩ signaling in HG, we down-regulated DAG-sensitive PKC isozymes by chronic treatment of PMA (100 nM) in HG for 48 h. Inhibition of PKC reversed the effect seen in HG. The peak total cell [Ca 2ϩ ] i response was elevated to 541 Ϯ 74 nM (p Ͻ 0.01 versus HG peak) (Fig. 2).
Since chronic PMA down-regulates only DAG-sensitive PKC isozymes, the role of PKC-in Ca 2ϩ signaling was ascertained using a cell-permeant myristoylated peptide inhibitor, previously demonstrated to be specific for this PKC isozyme (31). Mesangial cells were growth-arrested in HG with or without the PKCinhibitor (Zi; 10 M, 24 h). Pretreatment with the PKC-inhibitor in HG did not alter Ca 2ϩ signaling in response to ET-1 (268 Ϯ 31 nM, p Ͻ 0.01 versus NG peak) (Fig. 2).
To further support a role of DAG-sensitive PKC isozymes in Ca 2ϩ release in response to ET-1, mesangial cells were transiently transfected with green fluorescent protein (GFP)-PKC-␤ or GFP-PKC-␦ to represent both the conventional and novel class of PKC isozymes. Mesangial cells expressed functional GFP-PKC-␦ and -␦ in these experiments as they responded to PMA (1 M, 10 min) stimulation by translocating to the membrane, the hallmark of PKC activation (Fig. 3). In fluo-3-loaded cells, expression of the GFP-PKC constructs was clearly distinguishable from the surrounding cells (Fig.  4). When stimulated with ET-1, cells transfected with either GFP-PKC-␤ or GFP-PKC-␦ displayed a blunted Ca 2ϩ response (Fig. 4A). This was particularly evident in the nucleus. Conversely, when cells were stimulated with ionomycin (25 M), all cells responded, including the transfected ones, with a robust intranuclear release of Ca 2ϩ (Fig. 4B). Thus, the transfection procedure did not affect the loading of fluo-3.
Protein Expression of p47 phox and Modified Expression by Antisense Oligonucleotide-To investigate a role of NADPH oxidase, particularly the p47 phox subunit, in HG-induced production of ROS, we first determined the expression of p47 phox in mesangial cells. Confocal immunofluorescence imaging showed that mesangial cells indeed express p47 phox , which is localized in a perinuclear pattern (Fig. 7). HG caused increased intensity of the staining. Semiquantitative analysis indicates that the mean pixel intensity per cell in NG was 73 Ϯ 2 S.E., whereas in HG, the value was 110 Ϯ 2 S.E. (100 cells analyzed, p Ͻ 0.01) (Fig. 7).
To address the precise function of p47 phox , mesangial cells were transiently transfected with p47 phox antisense oligonucleotides. Fig. 7 shows that this method was sufficient to reduce expression of p47 phox protein, normally found in a perinuclear distribution.

Inhibition of ROS Formation Reverses the Effect of HG on Ca 2ϩ
Signaling in Response to ET-1-To determine the role of ROS on the observed reduced Ca 2ϩ signaling, mesangial cells were growth-arrested in HG and then treated with the addition of catalase (100 units, 1 h). In these experiments, the peak total cell [Ca 2ϩ ] i response in NG to ET-1 (50 nM) was 446 Ϯ 79 nM. In HG, the effect was reduced to 146 Ϯ 13 nM. Catalase significantly rescued the ET-1-induced Ca 2ϩ re- sponse in HG to 521 Ϯ 58 nM (p Ͻ 0.001 versus HG peak without catalase) (Fig. 9).
To further elucidate the role of ROS in Ca 2ϩ signaling, particularly to determine whether the p47 phox subunit of NADPH oxidase is necessary for the effect of HG on Ca 2ϩ release, mesangial cells were transiently transfected with p47 phox antisense oligonucleotides in HG and stimulated with ET-1. In NG, the peak total cell [Ca 2ϩ ] i response was 426 Ϯ 34 nM. In HG, the peak was reduced to 177 Ϯ 17 nM. Inhibition of p47 phox by down-regulation with antisense oligonucleotides reversed the effect of HG to a level that was not significantly different than the response seen in NG (514 Ϯ 48 nM, p Ͻ 0.001 versus HG peak) (Fig. 10). Transfection of the scrambled (SCR) version of p47 phox antisense oligonucleotide did not reverse the decreased [Ca 2ϩ ] i observed in HG (144 Ϯ 16 nM, p Ͻ 0.001 versus peak NG ϩ ET-1) (Fig. 10). Thus, ROS derived from NADPH oxidase plays a pivotal role in the reduced Ca 2ϩ response in HG.
Since several reports demonstrated the importance of ROS derived from a mitochondrial source in the diabetic milieu, we tested the effects of CCCP, an uncoupler of oxidative phosphorylation, and rotenone on Ca 2ϩ signaling in response to ET-1. These experiments were performed using measangial cells in passages 19 -21. In these experiments, the peak total cell [Ca 2ϩ ] i response was 173 Ϯ 10 nM, which was significantly attenuated to 88 Ϯ 10 nM (p Ͻ 0.05 versus peak NG ϩ ET-1) in HG (Fig. 11). CCCP (100 nM, 1 h) did not significantly alter the peak total cell [Ca 2ϩ ] i in NG (187 Ϯ 14 nM, p Ͼ 0.05 versus peak NG ϩ ET-1) or in HG (103 Ϯ 17 nM, p Ͼ 0.05 versus peak HG ϩ ET-1) (Fig. 11). Pretreatment with rotenone (1 g/ml, 1 h) decreased the basal and peak total cell [Ca 2ϩ ] i in NG and was therefore not used in HG.
Effect of PKC Down-regulation on p47 phox mRNA Expression and ROS on PKC Activation-To determine the effect of HG on p47 phox expression, mesangial cells were growth-arrested in NG or HG, and p47 phox mRNA was measured by reverse transcriptase-PCR. HG increased p47 phox mRNA by 1.7 Ϯ 0.1-fold over control (Fig. 12). To determine whether PKC plays a role in HG-increased p47 phox mRNA expression, mesangial cells were pretreated with chronic PMA exposure (100 nM, 48 h). Fig.   FIG. 6. HG- 12 shows that down-regulation of mesangial cell PKC did not alter the effect of HG regulation on p47 phox mRNA (1.9 Ϯ 0.4-fold over control).
Since PKC inhibition did not affect p47 phox mRNA expression, we postulated that ROS act upstream of PKC. As previously shown in our laboratory, the immunoreactivity of PKC-␦ is enhanced in HG (29), depicted in the confocal micrographs (Fig. 12). Pretreatment of mesangial cells with DPI prevented the increase in immunofluorescence. Mesangial cells were also stimulated with H 2 O 2 , and membraneenriched cell fractions were immunoblotted for various PKC isozymes. Interestingly, a low concentration of H 2 O 2 (10 M, 10 min) increased PKC-␦ in the cell membrane, whereas a slightly longer exposure time (30 min) or higher H 2 O 2 concentration (100 M) was required for PKC-␤ to be isolated in the membrane (Fig. 12). The enrichment of PKCs in the cell membrane fraction is in agreement with our previous report showing activation of PKC in HG (30). DISCUSSION In this study, we have demonstrated that Ca 2ϩ signaling in response to vasoactive peptides such as ET-1 is suppressed in HG through a mechanism requiring both NADPH oxidase ROS and DAG-sensitive PKC isozymes. We have verified that mesangial cell generation of ROS in HG is inhibited by catalase or DPI. Furthermore, transient transfection of p47 phox antisense oligonucleotide, at a concentration that significantly reduced expression of p47 phox protein prevented the increased generation of ROS in HG. Likewise, transient transfection of p47 phox antisense oligonucleotide was sufficient to reverse the effect of HG-induced depressed Ca 2ϩ signaling in response to ET-1. Treatment with catalase also normalized the ET-1-induced Ca 2ϩ response in HG. DAGsensitive PKC isozymes, known to be activated in HG, are necessary for suppression of Ca 2ϩ signaling to ET-1 in HG, since down-regulation of DAG-sensitive PKC isozymes by chronic exposure to PMA in HG restored the Ca 2ϩ response, and a PKC-peptide inhibitor failed to normalize the Ca 2ϩ signaling in response to ET-1 in HG. Moreover, overexpression of a conventional (GFP-PKC-␤) or novel (GFP-PKC-␦) PKC isozyme, mimicking their increased activity observed in HG, attenuated Ca 2ϩ signaling in response to ET-1. These data conclusively reveal the functional relationship between enhanced ROS generation in HG through NADPH oxidase and DAG-sensitive PKC-dependent suppression of Ca 2ϩ signaling.
Several reports indicate that HG attenuates Ca 2ϩ signaling in response to vasoactive peptides in certain cell types, although the hypothesized mechanism(s) for the attenuated Ca 2ϩ signaling appear to be multifaceted. In mesangial cells, Nutt and O'Neil (37) recently reported that both acute and chronic exposure to HG depressed ET-1-stimulated Ca 2ϩ signaling that is probably mediated by receptor-operated calcium influx. In transformed mouse mesangial cells, transforming growth factor-␤ 1 was shown to decrease the expression of inositol 1,4,5-trisphosphate receptors, leading to a reduced Ca 2ϩ response (38). We have observed that ET-1-stimulated Ca 2ϩ signaling measured by confocal fluorescence imaging probably originates predominantly from endoplasmic reticulum stores, since treatment with thapsigargin abolishes the Ca 2ϩ response. 2 Mené et al. (2) showed that HG inhibited store-operated capacitance Ca 2ϩ influx in mesangial cells by a PKC-dependent mechanism. Our findings are similar, since we also observed that reduced Ca 2ϩ signaling in HG is PKC-dependent. Ours is the first study to show that only DAG-sensitive PKC isozymes are involved, since inhibition of PKC-did not affect Ca 2ϩ release in response to ET-1. We are the first to demonstrate that overexpression of individual GFP-PKC isozymes is sufficient to attenuate Ca 2ϩ signaling.
A rapidly emerging area of research suggests that ROS play a major role in several aspects of signaling (39,40). In bovine aortic endothelial cells, acute exposure to HG was shown to abolish Ca 2ϩ oscillation in response to ATP (41). The authors stipulated that glucose-derived superoxide anion diminished Ca 2ϩ oscillation by accelerating Ca 2ϩ leak from intracellular stores and impairing Ca 2ϩ release-activated Ca 2ϩ entry. In human aortic endothelial cells, it was shown that NADPH oxidase-dependent ROS are critical to the generation of histamine-induced Ca 2ϩ oscillations, perhaps by altering the sensitivity of the endoplasmic reticulum to inositol 1,4,5-trisphosphate (42). Here we have established that in HG, NADPH-generated ROS significantly diminished Ca 2ϩ release stimulated by ET-1 in mesangial cells as catalase, and antisense oligonucleotide to p47 phox restored Ca 2ϩ signaling in HG, similar to that seen in NG. In order to address the role of mitochondrial ROS, we used two reagents that inhibit the mitochondrial electron transport chain as reported by others (43). In our experiments, CCCP failed to reverse HG-attenuated Ca 2ϩ signaling. This finding suggests that mitochondria production of ROS is not involved in reducing the Ca 2ϩ response to ET-1 in HG. Rotenone reduced Ca 2ϩ response to ET-1 in NG and therefore could not be used in HG.
Although ROS derived from a mitochondrial source was shown to be predominant in endothelial cells in HG (7,8), certainly other enzymes play pivotal roles. In Otsuka Long-Evans Tokushima Fatty rats, a model of type 2 diabetes, the aorta showed increased superoxide production compared with the control LETO rats, and the effect was inhibited by DPI (44). Furthermore, expression of p22 phox mRNA (44) as well as gp91 phox (45) was up-regulated. Likewise, smooth muscle cell- FIG. 10. p47 phox oligonucleotide normalized the reduced Ca 2؉ signaling in HG. Mesangial cells were growth-arrested in NG, HG, and HG with p47 phox antisense oligonucleotide (oligo) or HG with the scrambled version of the oligonucleotide (SCR) as described under "Experimental Procedures." Cells were loaded with fluo-3 and imaged with confocal microscopy with ET-1 (50 nM) stimulation. A-D, basal conditions; E-H, peak Ca 2ϩ responses. I, the analyzed results are depicted graphically. *, p Ͻ 0.001 versus NG peak; **, p Ͻ 0.001 versus HG peak. Bar, 25 m. or endothelial cell-increased free radical production in HG was inhibited by DPI (46). In streptozotocin-induced diabetic rats, immunoreactivity of p47 phox was increased in the kidney (47). Furthermore, increased generation of ROS in monocytes from diabetic humans was prevented by p47 phox antisense oligonucleotide (48). We showed that HG-induced ROS generation in rat mesangial cells was most likely through enhanced NADPH oxidase activity. Both catalase and DPI significantly reduced DCF fluorescence in HG. In addition, transient transfection of p47 phox antisense oligonucleotide also prevented ROS stimulation by HG in rat mesangial cells.
Although the NADPH oxidase is a multicomponent complex, several reports suggest that down-regulation of a single subunit is sufficient to inhibit enzymatic activity. Inhibition of p22 phox by stable transfection of antisense p22 phox cDNA into vascular smooth muscle cells resulted in significantly reduced angiotensin II-stimulated NADPH-dependent superoxide production (49). Coronary microvascular endothelial cells isolated from p47 phoxϪ/Ϫ mice lost ROS formation in response to tumor necrosis factor-␣, but the effect was restored when full-length p47 phox cDNA was transfected into the cells (50). Smooth muscle cells from p47 phox null mice also showed diminished superoxide production (51) and decreased the proliferative response to growth factors, suggesting that superoxide from NADPH oxidase plays a prerequisite role in atherosclerosis (52). In the same manner, we have shown that inhibition of NADPH oxidase with p47 phox antisense oligonucleotide was sufficient to prevent HG-induced ROS production, leading to an altered Ca 2ϩ response to ET-1.
In neutrophils, PKCs may regulate NADPH oxidase, particularly p47 phox , by direct phosphorylation (53). In nonphagocytic cells, the role of PKC in NADPH oxidase activation is less well known. Inoguchi et al. (46) showed that in aortic smooth muscle cells and endothelial cells, HG and palmitate stimulate ROS through PKC-dependent activation of NADPH oxidase by the use of DPI. In our study, we asked whether HG activation of PKC could regulate p47 phox . We found that down-regulation of PKC isozymes with chronic PMA was insufficient to prevent HG-induced increase of p47 phox mRNA. Conversely, pretreatment with DPI decreased the "activation pattern" visualized by confocal immunofluorescence imaging of PKC-␦ in HG. Likewise, H 2 O 2 caused increased accumulation of PKC-␦ and -␤ in the membrane-enriched cellular fraction, indicating that ROS probably stimulate these mesangial cell DAG-sensitive PKC isozymes.
Activation of PKC is an essential mechanism for HG-induced cellular dysfunction, since it represents a critical downstream event in the pathogenesis of diabetic complications (3,24,25,54). In this study, the importance of DAG-sensitive PKC isozymes is further elucidated in the response of mesangial FIG. 11. Effect of mitochondrial electron transport chain inhibitors on Ca 2؉ signaling. Mesangial cells were growth-arrested in NG or HG with and without CCCP (100 nM, 1 h), loaded with fluo-3, and imaged with confocal microscopy with ET-1 stimulation. A, NG basal; B, NG plus CCCP basal; C, HG basal; D, HG plus CCCP basal; E, NG plus ET-1 peak; F, NG plus ET-1 and CCCP peak; G, HG plus ET-1 peak; H, HG plus ET-1 and CCCP peak; I, graphical analysis. Bar, 25 m. cells to HG. Our data reflect a sequence of events whereby HG suppressed Ca 2ϩ signaling in response to ET-1 is dependent on DAG-sensitive PKC and NADPH oxidase ROS.