Nitric Oxide and cGMP-dependent Protein Kinase Regulation of Glucose-mediated Thrombospondin 1-dependent Transforming Growth Factor-β Activation in Mesangial Cells*

Excessive transforming growth factor-β (TGF-β) activity in hyperglycemia contributes to the development of diabetic nephropathy. Glucose stimulation of TGF-β activity and matrix synthesis are dependent on autocrine thrombospondin 1 (TSP1) to convert latent TGF-β to its biologically active form. The mechanisms by which glucose regulates TSP1 are not known. High glucose inhibits nitric oxide (NO) bioavailability and decreased NO increases TGF-β activity and extracellular matrix accumulation. Yet, the impact of NO signaling on TSP1 activation of TGF-β is unknown. We tested the role of NO signaling in the regulation of TSP1 expression and TSP1-dependent TGF-β activity in rat mesangial cells exposed to high glucose. On exposure to 30 mmglucose, NO accumulation in the conditioned media and intracellular cGMP levels were significantly decreased. The addition of an NO donor prevented the glucose-dependent increase in TSP1 mRNA, protein, and TGF-β bioactivity. The effects of the NO donor were blocked by ODQ (a soluble guanylate cyclase inhibitor) or Rp-8-pCPT-cGMPS (an inhibitor of cGMP-dependent protein kinase). These effects of high glucose were also reversed by the nitric-oxide synthase cofactor tetrahyrobiopterin (BH4). These results show that high glucose mediates increases in TSP1 expression and TSP1-dependent TGF-β bioactivity through down-modulation of NO-cGMP-dependent protein kinase signaling.

TGF-␤ is synthesized and secreted as latent complex (latent TGF-␤). It must be converted to the active state before binding to its receptors and eliciting cellular functions. Latent TGF-␤ can be activated by a number of factors, including heat, extreme pH, plasmin, altered glycosylation, integrin binding, and reactive oxygen species (7)(8)(9). However, there is increasing evidence that the extracellular matrix protein, thrombospondin 1 (TSP1), is a major physiologic regulator of TGF-␤ activation (10 -12). TSP1, a disulfide-linked 180-kDa trimer, is a multifunctional protein that is produced by a variety of cells, including mesangial cells. In vitro and in vivo studies show that high glucose up-regulates TSP1 expression (13)(14)(15)(16)31) and that TSP1 protein is increased in the kidneys of diabetic patients (17). Recently, we and others showed that TSP1 is responsible for the activation of TGF-␤ in mesangial cells when exposed to high glucose for either short (2 days) or long term (3 weeks) treatment (18,19). However, the mechanisms involved in regulating TSP1 expression by high glucose concentrations are unclear.
Nitric oxide has been implicated in the pathophysiology of diabetic nephropathy (20 -22). In glomerular mesangial cells, decreased nitric oxide (NO) bioavailability is observed at high glucose concentrations. Several glucose-dependent mechanisms have been shown to modulate NO levels. These include impaired nitric oxide synthesis due to a decreased availability of L-arginine or tetrahydrobiopterin (BH 4 ), a cofactor of the nitric-oxide synthases (NOS) (23,24). In addition, high glucose increases breakdown of NO through reactions with reactive oxygen species (26,50). Accumulation of advanced glycation end products with chronic high glucose exposure is similarly associated with scavenging of NO (25,26).
Decreased NO production has been shown to induce TSP1 expression in cultured endothelial cells (27). Moreover, the promoter of TSP1 has both simian virus 40 promoter factor 1 (SP-1) and activator protein 1 (AP-1) sites, which are regulated by the cellular redox state (28). It is likely that NO plays an important role in glucose-dependent regulation of TSP1 expression, although this has not been tested.
NO regulates cellular functions, in part through the activation of soluble guanylate cyclase and formation of cyclic GMP. Cyclic GMP binds to cGMP-dependent protein kinase (PKG), resulting in the activation of PKG and phosphorylation of intracellular proteins (29). Transfection of rat aortic smooth muscle cells with the catalytic domain of PKG-I inhibits TSP1 production (30). Although glucose has been shown to decrease NO-mediated cGMP production in neuroblastoma cells (49), direct evidence for glucose regulation of PKG activity has been lacking.
Based on these findings, we performed the present studies to determine the role of NO and PKG in glucose-mediated upregulation of TSP1 and TSP1-dependent TGF-␤ activity. Data from our studies now show that high glucose significantly decreases NO production and intracellular cGMP levels. At high glucose concentrations, but not at 5 mM glucose, exogenous NO restores basal levels of TSP1 mRNA and protein and TGF-␤ bioactivity in rat mesangial cells. A decrease in endogenous NOS activity appears to play a role since basal levels of TSP1 and TGF-␤ activity can be restored by the NOS co-factor BH4. These effects of exogenous NO in the high glucose media are blocked by both an inhibitor of soluble guanylate cyclase and a selective inhibitor of PKG. These data provide the first evidence that NO-dependent PKG activity is involved in the regulation of TSP1 expression and TGF-␤ bioactivity in response to high glucose concentrations.

EXPERIMENTAL PROCEDURES
Materials-RPMI 1640 medium with L-glutamine without glucose was purchased from Invitrogen. Insulin-transferrin-sodium selenite liquid media supplement, minimal essential medium nonessential amino acid solution, sodium pyruvate solution, and BH 4 were purchased from Sigma. Monoclonal antibody 133 raised against human platelet TSP1 stripped of TGF-␤ activity was purified by our lab in a joint effort with the University of Alabama at Birmingham Hybridoma Core Facility (43). Anti-PKG I antibody was generously provided by Dr. Thomas M. Lincoln, Department of Pathology, University of Alabama at Birmingham. Secondary antibody was purchased from Jackson Im-munoResearch Laboratories, Inc. (West Grove, PA). DetaNONOate was purchased from Alexis Biochemicals (San Diego, CA). The decomposed compound of DetaNONOate was made by leaving the DetaNONOate in 0.1% NaOH at room temperature for at least 1 week, and complete decomposition assessed from the UV spectra. The cGMP analogue 8-pCPT-cGMP was purchased from Sigma. Rp-8-pCPT-cGMPS was purchased from Biolog. (La Jolla, CA). Luciferase assay reagent and passive lysis buffer were purchased from Promega (Madison, WI). Human recombinant TGF-␤1, monoclonal anti-TGF-␤1, -␤2, -␤3 antibody, and the Quantikine human TGF-␤1 immunoassay kit were purchased from R&D Systems, Inc. (Minneapolis, MN). Neomycin Sulfate (G418) was obtained from ICN Biomedicals Inc. (Aurora, Ohio). Anti-inducible NOS antibody was obtained from BD Transduction Laboratory (San Diego, CA).
Cell Culture-Rat mesangial cells (RMCs) were the generous gift from Dr. Anne Woods, University of Alabama at Birmingham. Cells were cultured in RPMI 1640 medium supplemented with 20% heatactivated fetal bovine serum, 5 mM D-glucose, 2 mM L-glutamine, 1% (v/v) nonessential amino acids, 2 mM sodium pyruvate, 10 g/ml transferrin, 5 ng/ml sodium selenite, and 0.6 international units/ml insulin. RMCs were passaged at 80% confluency. Experiments in this study were performed on cells between the 5th and 10th passages.
Mink lung epithelial cells (clone 32) stably transfected with the TGF-␤ response element of the human plasminogen activator inhibitor-1 (PAI-1) gene promoter fused to firefly luciferase reporter gene were a generous gift from Dr. D. B. Rifkin (New York University Medical Center). Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% calf serum, L-glutamine, and 200 g/ml G418.
Measurement of Nitric Oxide Production-NO release from NO donor was measured using an ISONO-PMC microchip electrode (WPI, Sarasota, FL) in a sealed 4-ml chamber that was magnetically stirred at 37°C. DetaNONOate (5 M) was added to the chamber, which contained RPMI media with either 5 or 30 mM glucose (without phenol red) to measure NO release in the media. NO measurements in the presence of cells were made by suspending rat mesangial cells (ϳ1.5 ϫ 10 5 /ml) in RPMI media containing either high (30 mM) or normal (5 mM) glucose. Data were collected by a digital recording device (Dataq, Akron OH) connected to a PC. The NO electrode was calibrated by the addition of known concentrations (measured spectrophotometrically) of S-nitroso-N-acetylpenicillamine to the chamber to establish the [NO]:electrode current relationship. The rate of NO production over the initial 60 s of DetaNONOate decomposition and the concentration of NO released at steady state were determined for each system.
Total nitrite and nitrate levels in RMC culture media were measured utilizing a fluorometric nitric oxide assay kit (Calbiochem). Briefly, nitrate reductase and enzyme co-factors were added to 20 l of conditioned media. Samples were incubated for 2 h at room temperature and then incubated with 20 l of 2,3-diaminonaphthalene for 10 min. Finally, 10 l of NaOH was added to stop the reaction, and fluorescence was read using a CytoFluor plate reader (Perseptive Biosystems, Framingham, MA) using the excitation wavelength of 365 nm and an emission wavelength of 450 nm. The mean values of fluorescence intensity (relative units) were converted into concentrations of total nitrite and nitrate (M) according to the standard curve.
Measurement of cGMP Production-RMCs were cultured in complete medium for 2-3 days to reach 85-90% confluency and were made quiescent for 2 days in serum-free medium. Isobutylmethylxanthine, a nonspecific cellular phosphodiesterase inhibitor, was used to prevent breakdown of cGMP during the exposure period. RMCs were incubated with 5 M DetaNONOate in the presence of 0.1 mM isobutylmethylxanthine for the indicated time. After treatment, culture medium was aspirated, and cells were washed with ice-cold phosphate-buffered saline. To extract cGMP, 200 l of lysis buffer was added to the cells and incubated for 10 min at room temperature. cGMP production was determined by using a commercially available cGMP EIA kit (Amersham Biosciences, Inc.). After the acetylation EIA procedure of the manual in the kit, briefly, rabbit anti-cGMP antibody was added to 96-well plate precoated with donkey anti-rabbit IgG. Acetylated standards or samples were added to the wells and incubated for 2 h at 4°C. cGMPperoxidase conjugate was added and incubated for an additional 1 h at 4°C. After washing, 3,3Ј,5Ј,5Ј-tetramethylbenzidine substrate was added to all wells and incubated for 30 min at room temperature, and absorbance was measured in an enzyme-linked immunosorbent assay reader at 450 nm. Values were expressed at pmol of cGMP/mg of protein according to the standard curve. Protein concentrations did not differ with glucose treatment (data not shown).
Preparation of Condition Medium and Cell Lysate from Cultured RMCs-RMCs were seeded in six-well plates at a density of 3 ϫ 10 4 cells/ml. Cells were grown in the growth medium containing 20% fetal bovine serum for 2-3 days until cells reached 80% confluency and then rendered quiescent by culturing in serum and insulin-free RPMI media containing 5 mmol/liter glucose, 10 g/ml transferrin, and 5 ng/ml sodium selenite for 48 h. Cells were treated for the next 24 h with serum-free media containing either 5 mmol/liter or 30 mmol/liter glucose in the presence or absence of reagents. When using PKG inhibitor (Rp-8-pCPT-cGMPS), cells were pretreated with the inhibitor for 20 min before the addition of other reagents. After 24 h of incubation, conditioned media were harvested and analyzed for thrombospondin 1 expression by immunoblotting. Cells were harvested in ice-cold PEM buffer (20 mmol/L sodium phosphate (pH 6.8), 2 mmol/L EDTA, 15 mmol/L ␤-mercaptoethanol, 0.15 mol/L NaCl, 10 g/ml leupeptin, and 5 g/ml aprotinin) (30), homogenized using a sonicator, and briefly centrifuged. The supernatants were analyzed for PKG expression by immunoblotting.
Immunoblot Analysis-Conditioned media or supernatants of cell lysate were harvested, and equal amounts of protein were subjected to SDS-PAGE gel (8%) under reducing conditions. After electrophoretic transfer to nitrocellulose membranes and blocking, the membranes were incubated either with mouse monoclonal anti-TSP1 antibody 133 (1:5000) or anti-PKG I polyclonal antibody (1: 5000) for 2 h at room temperature. After intensive washing, secondary antibody (1:10,000 dilution) was used for the detection of immunoreactive bands with the enhanced chemiluminescence detection system (Pierce). Equal loading and transfer of protein samples were assayed by staining the blots with Ponceau S.
TGF-␤ Assay-Total and active TGF-␤ levels in the condition media were assayed using the PAI-1/luciferase assay. Briefly, Mink lung epithelial cells were plated into 24-well tissue culture plate at 1.6 ϫ 10 5 cells/ml and incubated for 3-4 h for optimal attachment. After aspiration of the growth medium from the attached cells, 500 l of condition media were added. To measure total TGF-␤, conditioned media samples were heat-activated for 3 min at 100°C. Samples and TGF-␤ standards were incubated overnight at 37°C. After incubation, cells were lysed with lysis buffer at room temperature for 20 min. Lysates were analyzed for luciferase activity using a luminometer after the injection of 100 l of substrate solution and recorded as relative light units. The mean values of triplicate samples were converted into concentrations of TGF-␤ (pM) using a standard curve obtained with human recombinant TGF-␤1. Specificity of the assay was proven by either neutralization of the TGF-␤ activity in conditioned media with an anti-TGF-␤ antibody. Alternately, aliquots of the identical conditioned media used in the PAI-1/luciferase assay were assayed for active TGF-␤ using an enzymelinked immunosorbent assay for TGF-␤1 (Quantikine, R&D System; data not shown) (54).
RNA Isolation and Northern Blot Assay-Quiescent cells were treated for 24 h with serum-free media containing either 5 or 30 mM glucose in the presence of increasing concentrations of DetaNONOate. After treatment, total RNA was extracted using RNeasy mini kit (Qiagen, Valencia, CA). Equal amounts of total RNA (10 g) were denatured, electrophoresed, and transferred to nylon membranes. RNA was fixed by UV-cross-linking. The TSP1 probe (1.1-kilobase fragment) was  Statistical Analysis-Data are expressed as the mean Ϯ S.D. Statistical evaluation of the data was performed using the unpaired t test, considering the p value of Ͻ0.05 as significant.

High Glucose Inhibits NO Production and Intracellular
cGMP Levels-To test the direct effect of high glucose on NO production, RMCs were cultured in RMPI medium (without phenol red and nitrite) with either 5 mM or 30 mM glucose for 24 h. The conditioned media were collected and analyzed for nitrate/nitrite levels (an index of accumulated NO produced by cells). As shown in Table I, 24-h NO production decreased a small but significant amount (ϳ10%) in 30 mM glucose media as compared with 5 mM glucose media (p Ͻ 0.05), which is consistent with previous observations (23).
NO regulates cellular functions in part through the activation of soluble guanylate cyclase and the formation of cGMP (29). To determine whether cGMP levels are decreased after acute exposure of the cells to 30 mM glucose, intracellular cGMP levels were measured in cells in which the media were changed to either 5 or 30 mM glucose. Although there was only a modest stimulation of endogenous NO after the media change, the cGMP level in media with 30 mM glucose is nonetheless only 2% that in 5 mM glucose media. To determine whether cells cultured under high glucose conditions still retain the capability of producing cGMP in response to NO, experiments with the NO donor DetaNONOate were performed. The capacity of soluble guanylate cyclase to generate cGMP was essentially equivalent in cells exposed to either 5 or 30 mM glucose (Table I). These data suggest that the NOsignaling pathway at the level of either NO production or bioavailability is decreased on acute exposure of mesangial cells to media containing high glucose.
NO Inhibits Glucose-induced TSP1 Expression and TGF-␤ Activity-It has previously been shown that NO levels regulate TSP1 expression by endothelial cells (27). Because glucose decreases bioavailable NO, it is possible the decreased NO after exposure to high glucose media is involved in regulation of TSP1. To determine whether NO is involved in up-regulation of TSP1 and TSP1-dependent TGF-␤ bioactivity by high glucose concentrations, RMCs were treated with increasing concentrations of the NO donor (DetaNONOate) for 24 h. The compound DetaNONOate was selected because of its low rate of release, its independence from metabolic activation for NO release, and its lack of reactive side products formed during decomposition (52).
In cells cultured with 30 mM glucose, DetaNONOate downregulates both TSP1 mRNA and protein levels in a concentrationdependent manner (Figs. 1, A and B). Glucose up-regulates TSP1 mRNA ϳ2.5-fold as compared with basal levels; both TSP1 mRNA and protein are reduced to basal levels, equivalent to those found in cells cultured in 5 mM glucose, when RMCs were treated with 2-5 M DetaNONOate. In contrast, when RMCs were cultured in 5 mM glucose, TSP1 protein levels were decreased only slightly to 80 -85% that of basal levels even at higher concentrations of the NO donor. The concentrations of DetaNONOate that were used in these experiments had no effect on cell number and were not toxic, as judged by 95% trypan blue exclusion after trypsinization of cells. We also observed the effect of this NO donor on TGF-␤ activity and protein levels by using the PAI-1/luciferase assay. Incubation of mesangial cells with increasing concentrations of DetaNONOate markedly reduced the stimulatory effect of 30  mM glucose on TGF-␤ activity to the level of TGF-␤ activity observed in mesangial cells cultured with 5 mM glucose (Fig.  1C). Results from the PAI-1/luciferase assay were confirmed using a receptor-based enzyme-linked immunosorbent assay to measure TGF-␤ activity (data not shown). Levels of total TGF-␤ were not significantly affected by DetaNONOate treatment (Fig. 1D). The decomposed compound of DetaNONOate had no effect either on TSP1 expression or active or total TGF-␤ production in cells cultured with either 5 or 30 mM glucose (data not shown). One possible explanation for the inhibitory effect of NO on TSP1 expression and TGF-␤ activity in response to high glucose exposure might be that the culture conditions modify the exposure of the cells to the exogenous NO. To test for this possibility, the release of NO by DetaNONOate was measured in both 5 and 30 mM glucose containing media polarographically. The results show that over a wide range of DetaNONOate concentrations (5-50 M), there was no difference between the rate of NO release or steady state NO concentration in the cell-free media at either glucose concentration. Data for the 5 M concentration of DetaNONOate are shown in Table II. This suggests that NO scavenging or modification of the decomposition rate of the NO donor by glucose is not responsible for the altered responses (33). In contrast, the release of NO by 5 M DetaNONOate when added in the presence of cells differed between the 5 and 30 mM glucose conditions (Table II). The NO concentration is decreased (p Ͻ 0.05) in mesangial cell media conditioned with 30 mM glucose as compared with the NO levels in media conditioned by the cells under basal glucose conditions (5 mM). Together these data suggest that cellular factors are important in regulation of NO levels by glucose. This is consistent with reports of decreased NOS activity under high glucose conditions (24, 44 -47).
A number of cofactors for NOS are necessary for the catalytic activity of the enzyme (55). Among these, BH 4 is potentially important, since it has been shown that glucose can decrease the availability of BH 4 , resulting in inhibition of NO formation from NOS (24). We therefore examined the effects of BH 4 supplementation on NO production and TSP1-dependent TGF-␤ activation in RMCs exposed to high glucose. BH 4 (1 M) prevented high glucose-induced decreases in NO levels (Table  III) and restored TSP1 and TGF-␤ activity in 30 mM glucosecontaining cultures to basal levels found in cells cultured in 5 mM glucose (Fig. 2, A and B). These data further show that NO and the activity of endogenous NOS plays a pivotal role in regulation of glucose-mediated TSP1 expression and TGF-␤ activation.

Effects of cGMP on TSP1 Expression and TGF-␤ Activity in the Media of RMC Cultured in 5 mM or 30 mM Glucose-To
evaluate whether the effect of NO on TSP1 expression and TGF-␤ activity under 30 mM glucose conditions occurs through the cGMP-dependent pathway, an inhibitor of soluble guanylate cyclase (5 M ODQ) was used to treat RMCs before exposure to DetaNONOate (5 M). Pretreatment with ODQ essentially blocked the ability of DetaNONOate to decrease TSP1 expression in cells exposed to 30 mM glucose (Fig. 3A, lane 7). Similarly, ODQ also abolished the ability of DetaNONOate to reduce TGF-␤ bioactivity (Fig. 3B). As expected, if glucose down-regulates the NO available to activate soluble guanylate cyclase, neither TSP1 nor TGF-␤ production was affected by ODQ treatment of cells cultured with 30 mM glucose (Figs. 3).
In further experiments, RMCs were exposed to increasing concentrations of the non-hydrolyzable cGMP analogue, 8-pCPT-cGMP, for 24 h. As anticipated, if NO signaling is mediated via soluble guanylate cyclase, 8-pCPT-cGMP mimicked the effect of the NO donor on TSP1 expression and TGF-␤ activity (Fig. 4, A and B). Only TGF-␤ activity and not protein expression are modulated by cGMP, as total TGF-␤ protein levels were not affected (Fig. 4C). Under conditions (5 mM glucose) in which there are basal levels of cGMP, 8-pCPT-cGMP had no effect on TSP1, TGF-␤ activity, or total TGF-␤ production in 5 mM glucose. These results show that up-regulation of TSP1 and TGF-␤ activity by exposure of cells to high glucose can be blocked by stimulating a cGMP-dependent pathway, consistent with a role for cGMP signaling in glucose regulation of TSP1 and TGF-␤ activity.
Effects of PKG on TSP1 Expression and TGF-␤ Activity in RMC-conditioned Media Cultured in 5 mM Glucose Versus 30 mM Glucose-Cyclic GMP can signal through activation of PKG (32), and it has been shown that PKG-I inhibits TSP1 protein expression in aortic smooth muscle cells (30). To examine the role of PKG in NO-dependent inhibition of TSP1 expression and TGF-␤ activity mediated by high glucose, RMCs were exposed to increasing concentrations of a selective PKG inhibitor (Rp-8-pCPT-cGMPS) for 20 min before the addition of DetaNONOate (5 M). After incubation for 24 h with these compounds, the conditioned media were harvested, and TSP1 expression and TGF-␤ activity were determined. Rp-8-pCPT-cGMPS pretreatment significantly diminished the ability of 5 M DetaNONOate to decrease TSP1 expression in a concentration-dependent manner (Fig. 5B). Similarly, the inhibitory effect of DetaNONOate on the TGF-␤ bioactivity under 30 mM glucose conditions was prevented by Rp-8-pCPT-cGMPS pretreatment in a concentration-dependent manner (Fig. 5C). Total TGF-␤ production was not affected (Fig. 5D). As expected, in the absence of exogenous NO, Rp-8-pCPT-cGMPS (50 M) had no effect on TSP1 expression or TGF-␤ activity (Figs. 5, A and C). These data show that PKG is a key mediator of NO-dependent TSP1 regulation.
To determine whether the responses of the PAI-1/luciferase assay are due to TGF-␤ signaling or due to the possible direct effects of alterations in glucose, NO, or PKG signaling on the PAI-1 promoter, anti-TGF-␤ antibody was added to replicate aliquots of the conditioned media and tested in the PAI-1/ luciferase assay. The anti-TGF-␤ antibody neutralized more than 90% of the luciferase activity in this assay, confirming that these responses are indeed due to active TGF-␤ signaling (Fig. 6). Furthermore, basically identical responses were obtained using a TGF-␤1-specific enzyme-linked immunosorbent assay (data not shown).
To assess whether glucose regulates PKG expression, RMCs cultured in 5 or 30 mM glucose media in the presence of Det-aNONOate were examined for PKG-I protein levels by immunoblotting analysis. Neither glucose nor NO levels altered PKG protein levels in mesangial cells (Fig. 7). This suggests that glucose or exogenous NO regulates the activity of PKG-I other than expression of PKG-I. Taken together, these results demonstrate that under high glucose conditions, NO regulates TSP1 expression and TGF-␤ activity by modulation of PKG-I activity. DISCUSSION The results in this study provide the first direct evidence linking NO and PKG signaling to glucose-induced TSP1 tran- scription and protein expression and TSP1-dependent TGF-␤ activity in mesangial cells. Data presented in this study also show decreased NO and cGMP levels in RMCs in response to high glucose. It is proposed that this suppression of endogenous NO-regulated PKG signaling contributes to glucose-dependent increases in TSP1 expression and TSP1-dependent activation of TGF-␤.
In our study, NO concentrations achieved in cultures exposed to high glucose and treated with DetaNONOate were decreased by ϳ25% as compared with DetaNONOate treatment of cultures in basal glucose. A smaller but significant change in the endogenous NO formed on changing media to high glucose was also evident. This level of change due to glucose is consistent with data from Trachtman et al. (23) in which high glucose concentrations inhibited interferon-␥-and lipopolysaccharide-induced NO production to a similar extent (23). NO is synthesized from L-arginine by the enzyme NOS. Although inducible NOS is reported to be present in kidney mesangial cells (53), its expression (mRNA and protein levels) is not affected by high glucose exposure in mesangial cells (Ref. 24 and data not shown). Rather, under high glucose conditions, decreased availability of BH 4 (a cofactor of NOS) has been implicated in the impaired NOS activity and reduced NO production (24). Our data are consistent with these observations, since supplementation of BH 4 prevented high glucose-induced inhibition of NO production and subsequent TSP1 expression and TGF-␤ activity. The mechanisms of BH 4 action are potentially complex and potentially involve a combination of decreased NO formation and enhanced reactive oxygen species formation from NOS (56,57).
The soluble guanylate cyclase/cGMP-signaling pathway appears to be involved in glucose regulation of TSP1 and TGF-␤ activity, since high glucose causes a substantial decrease in intracellular cGMP levels in mesangial cells. In addition, cGMP levels after 24 h of stimulation by exogenously added NO are also significantly decreased in RMCs exposed to high glucose. Although the NO levels under high glucose conditions are not decreased to the same extent, it is possible that modest changes in NO levels can have amplified effects on downstream mediators. This might occur through modulation of soluble guanylate cyclase activity. Consistent with this concept, guanylate cyclase activity is decreased in streptozotocin-induced diabetic rats (48,51).
A primary downstream signaling molecular of cGMP is PKG (32). PKG has been shown to suppress the production of extracellular matrix proteins including osteopontin and TSP1 in aortic smooth muscle cells (30). In mammalian cells, two genes encoding PKG have been identified, type I and type II. In mesangial cells, only type I PKG has been identified (38,39). Expression of PKG-I in cultured vascular smooth muscle cells is undetectable after several passages (36,37). However, we were able to detect expression of PKG-I in RMCs by immunoblotting even in the 10th passage. It has not previously been shown whether glucose modulates the expression or/and activity of PKG. These present data suggest that glucose regulates PKG at the level of activity as there was no difference in PKG-I protein levels in RMCs cultured in either 5 or 30 mM glucose media. Because the decline in NO synthesis in RMC-conditioned media cultured in high glucose is paralleled by decreases in cGMP (23), PKG activity is predicted to be low under high glucose conditions. As anticipated, under conditions in which PKG activity is expected to be down-regulated (high glucose), inhibition of PKG activity did not affect glucose-mediated TSP1 expression and TGF-␤ bioactivity. In contrast, activation of PKG by NO donors under high glucose conditions inhibited TSP1 expression and resultant TGF-␤ activity. These data demonstrate that PKG activity plays a major role in high glucose-mediated TSP1 expression.
The cross-activation of cyclic nucleotide-dependent protein kinases has been reported; highly elevated cGMP levels can activate PKA (35,40). PKA has been reported to increase TSP1 deposition in cultured human dermal microvascular endothelial cells (41), which might occur through the activation of the cAMP response element (CRE) at positions Ϫ792 and Ϫ1192 of the human TSP1 promoter (42). Protein kinase A does not appear to be involved in the regulation of TSP1 in mesangial cells by pathological glucose levels, since a selective protein kinase A inhibitor (KT5720) did not modulate TSP1 expression (data not shown). However, our current studies do not eliminate the possibility that PKA is involved in basal TSP1 regulation.
At this time, the molecular basis of the regulation of TSP1 expression by PKG-I remains unclear. These studies and others show that glucose increases TSP1 transcription as well as protein expression (18,19). Furthermore, we now show that NO donors in a concentration-dependent manner prevent increases in steady state TSP1 mRNA levels in RMCs upon high glucose exposure (24 h), demonstrating that NO signaling regulates high glucose-mediated TSP1 gene expression at the transcriptional level. However, these data are inconsistent with those of Dey et al. (30) in which steady state TSP1 mRNA levels were only slightly decreased in aortic smooth muscle cells 48 h after transfection with the constitutively active catalytic domain of PKG-I, leading to the conclusion that posttranscriptional mechanisms are involved (30). This difference in our findings might be due to the different nature of these experimental systems; regulated PKG signaling versus constitutively active PKG. The possibility of additional mechanisms of TSP1 regulation by PKG-I warrants further investigation.
In this study, TSP1 expression and TGF-␤ activity under normal glucose conditions are not regulated by NO levels; treatment of cultures in 5 mM glucose with the NO donor DetaNONOate did not show reductions in TSP/TGF-␤ levels. This is at variance with the work of Craven et al. (34), in which NO induced a marked reduction in TGF-␤ activity in mesangial cells cultured in 5.6 mM glucose as well as 30 mM glucose (34). This difference might be due to the high concentration of Snitroso-N-acetylpenicillamine (0.5 mM) used in their studies and the fact that this NO donor is also capable of mediating modification of proteins through S-nitrosylation. This raises the interesting possibility that S-nitrosylation pathways may have effects on TSP1 distinct from those of cGMP-dependent signaling. In contrast, it is evident from our data that the rates of NO release and steady state concentrations achieved in the presence of 5 M DetaNONOate are in the low nM range. These concentrations are readily achieved by the activation of consti- FIG. 8. Model for the role of NO and PKG in regulation of TSP1 and TGF-␤ levels by glucose. High glucose concentrations up-regulate TSP1 production, and TGF-␤ bioactivity occurs through a glucosedependent reduction in nitric oxide availability, apparently due to a reduced NOS co-factor (BH 4 ). Other mechanisms altering NO bioavailability might also be involved. The reduced NO levels result in decreased cGMP generation and, thus, decreased PKG activity. The decreased PKG activity down-regulates an inhibitory signal, allowing for increased TSP1 mRNA transcription and protein expression. The increased TSP1 protein then acts to activate latent TGF-␤.
tutive NOS under physiological conditions such as shear stress.
In summary, we have demonstrated that high glucose upregulates TSP1 expression and TSP1-dependent TGF-␤ bioactivity through down-modulation of the NO/cGMP/PKG-signaling pathway (Fig. 8). These findings suggest that the altered NO-signaling pathway plays an important role in the pathogenesis of diabetic nephropathy and suggest novel means of modulating glucose-dependent renal fibrosis.