PPAR Agonists Amplify iNOS Expression While Inhibiting NF-κB

Implications for Mesangial Cell Activation by Cytokines

Abstract

ABSTRACT. In acute inflammation, the transcription factor NF-κB is activated and increases the expression of multiple pro-inflammatory genes. Agonists of peroxisome proliferator activated receptors (PPAR) have been reported to exert antiinflammatory effects in various systems. In keeping with such an antiinflammatory role, it was found that several PPAR agonists, including Wy14,643, clofibrate, carbaprostacyclin, and ciglitazone inhibited NF-κB activity and increased IκBα levels in cytokine-stimulated mesangial cells (MC). Activation of NF-κB has been found to be crucial to the cytokine-elicited expression of inducible nitric oxide synthase (iNOS). Despite the inhibitory effect of PPAR agonists on NF-κB activity, this study provides experimental data demonstrating that these agonists amplify cytokine-elicited NO generation in MC, potentiating iNOS protein expression approximately threefold. The upregulation of iNOS expression occurred at the mRNA level and apparently did not result from iNOS mRNA stabilization. Clofibrate and ciglitazone amplified the cytokine-elicited stimulation of a 16-Kb human iNOS promoter construct in stably transfected MC, suggesting that PPAR agonists potentiate iNOS induction through transcriptional mechanisms. MC express all three PPAR proteins. However, iNOS potentiation did not correlate with increased PPAR activity. In addition, Wy14,643-induced amplification of cytokine-elicited iNOS levels also occurred in RAW264.7 macrophages and in human epithelial Caco-2 and HT-29 cells. The observation that these epithelial cell lines express an inactive, truncated PPARα variant suggests that a classical PPARα agonist, such as Wy14,643, may act through PPARα-independent mechanisms. In conclusion, these results show that, despite reducing NF-κB activity, PPAR agonists may amplify the expression of certain NF-κB–dependent genes that are relevant to the inflammatory process, like iNOS.

The ligand-activated transcription factors that are known as peroxisome proliferator activated receptors (PPAR), previously accepted to control lipid metabolism and adipocyte differentiation, have received considerable attention due to their recently identified role as modulators of vascular pathophysiology (¹). Three PPAR subtypes, α, β or δ, and γ, exhibit a certain degree of tissue specificity and ligand selectivity. Leukotriene B₄ (LTB₄), polyunsaturated fatty acids and the hypolipidemic agents, Wy14,643, clofibrate, and other fibrates, are considered selective PPARα agonists (^2,3). Carbaprostacyclin (cPGI₂) has been reported to selectively activate PPARδ (⁴). PPARγ can be activated by 15-deoxy-Δ ^12,14-PGJ₂ (15d-PGJ₂) and by the antidiabetic compounds thiazolidinediones, such as troglitazone, ciglitazone, and BRL49653 (^3,5). It has been proposed that many of the above-cited drugs, beside reducing risk factors such as hyperlipidemia, could exert beneficial effects on the cardiovascular system by directly modulating the mechanisms of disease within the vessel wall through their ability to bind PPAR (⁶). In the kidney, a protective role of PPAR in several pathophysiologic kidney processes, including ischemia/reperfusion injury, glomerulosclerosis, and diabetic glomerular disease, has been hypothesized (⁷). Numerous experimental evidences suggest an antiinflammatory role for PPAR. PPARα ligands have been shown to inhibit interleukin-1 (IL-1)–elicited cyclooxygenase-2 (COX-2) expression in human smooth muscle cells (⁸), thrombin-induced endothelin-1 (⁹), and cytokine-induced VCAM-1 expression in endothelial cells (¹⁰). In addition, PPARγ agonists have been shown to reduce the production of inflammatory cytokines in monocytes (¹¹) and matrix metalloprotease expression and migration of vascular smooth muscle cells (¹²). However, there are also examples of exacerbation of the response to inflammatory mediators in the presence of PPAR agonists. PPAR agonists have been reported to increase the production of tumor necrosis factor-α (TNF-α) during endotoxemia in mice (^13,14), to activate Kupffer cells in vivo (¹⁵), to promote the expression of MCP-1 and IL-8 in endothelial cells (¹⁶), and to amplify PLA₂ expression in MC (¹⁷). Some of these contradictory evidences may arise form the use of different cell types, in which the expression of PPAR isoforms is not always fully characterized. The complexity of this field of research increases when it is taken into account that some commonly used PPAR agonists, including Wy14,643, eicosanoids, 15d-PGJ₂, and troglitazone, may display PPAR-independent effects on cell metabolism or activation (^18–21), which renders the interpretation of the reported observations more complex.

PPAR may exert their antiinflammatory effects by several mechanisms. PPARα-induced catabolism of lipidic pro-inflammatory mediators may contribute to a feedback control of inflammation (²). PPAR have also been reported to interfere with the activity of transcription factors, such as AP-1, NF-κB, or STAT-1 (²²), which are critically involved in the signal transduction of cytokines, lipopolysaccharides (LPS), and interferons (IFN), either by competing for limiting amounts of general co-activators (²³) or by direct protein-protein interaction (²⁴). In numerous experimental systems, a correlation between the inhibitory effects of PPAR and a reduction in NF-κB activity has been observed (^10,22,25). This converts NF-κB and NF-κB–dependent pro-inflammatory genes into an important therapeutic target for PPAR agonists.

Accumulating experimental evidence proposes inducible nitric oxide synthase (iNOS) as an important target for NF-κB in the cellular response to inflammatory stimuli (^26,27). Cytokines and products of the bacterial wall elicit the expression of iNOS in a variety of tissues and cell types. NO is a key mediator in inflammation that promotes vasodilatation and participates in the nonspecific immune response. NO also contributes to the regulation of the inflammatory process through the modulation of enzyme activity and gene expression in an autocrine and paracrine fashion (²⁸). An important aspect of NO pathophysiology is its role as a negative modulator of the inflammatory process that contributes to the resolution of inflammation (²⁹). NO can inhibit the activity and expression of iNOS (^30,31), adhesion molecules (³²), and cytokines (³³). In mesangial cells, NO can amplify cytokine-induced IκBα levels, reduce NF-κB activity, and limit the expression of iNOS and COX-2 at late stages of cell activation by cytokines (^34,35). As yet, the available data on the implication of PPAR in the regulation of iNOS expression are conflicting. Certain PPAR agonists have been reported to attenuate iNOS induction in murine macrophages (^22,23,36), whereas others amplify it in macrophages or in smooth muscle cells (^36,37). In addition, the role of PPAR in the regulation of iNOS is not completely elucidated.

In the present study, we have explored the effect of several PPAR agonists on NF-κB activation and the induction of the putative NF-κB target and inflammatory mediator iNOS in cells exposed to pro-inflammatory agents. Our results show that several PPAR agonists potently amplify cytokine-elicited iNOS induction in mesangial cells while inhibiting NF-κB activity.

Materials and Methods

Materials

Recombinant human TNF-α was from Serotec (Oxford, UK). Recombinant human IL-1β (5 × 10⁷ U/mg) was from Roche Molecular Biochemicals (Mannheim, Germany). Interferon-γ was from R&D Systems Europe (Abingdon, UK). Cell culture media and supplements were from Life Technologies SA (Barcelona, Spain). Deoxycytidine 5′-triphosphate α-[³²P] (3000 Ci/mmol) was from Amersham (Buckinghamshire, UK). Polyclonal anti-iNOS (sc-651), anti-IκBα (sc-371), anti-COX-2 (sc-1746), anti-PPARα (sc-1985), anti-PPARβ (sc-7197), anti-PPARγ (sc-7196), and horseradish peroxidase (HRP)–conjugated anti-goat were from Santa Cruz Biotechnology (Santa Cruz, CA). HRP-conjugated anti-rabbit immunoglobulins were from Dako (Glostrup, Denmark). Antiserum to medium chain fatty acyl CoA dehydrogenase (MCAD) was from Alexis Biochemicals (San Diego, CA). Wy14,643 and ciglitazone were from Biomol Research Laboratories Inc. (Plymouth Meeting, PA). cPGI₂ was from Cayman Chemical (Ann Arbor, MI). 15-deoxy-Δ ^12,14-prostaglandin J₂ was from Calbiochem (San Diego, CA). All other reagents used were of the highest purity available from Sigma Chemical Co. (St. Louis, MO).

Cell Culture

Rat mesangial cells (RMC) were obtained as reported earlier (²⁷). Cells were grown in RPMI 1640 supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Throughout this study, passages 5 to 15 were used. For iNOS induction, MC were activated with IL-1β (2 ng/ml) plus TNF-α (25 ng/ml). Human intestinal epithelial HT-29 and Caco-2 cells and the RAW264.7 murine macrophage cell line were obtained from the American Tissue Culture Collection (ATCC, Manassas, VA). HT-29 cells were cultured in DMEM supplemented with 10% FBS. Caco-2 cells were cultured in EMEM with 10% FBS. In HT-29 cells, iNOS was induced with combinations of IL-1β (8 ng/ml) plus IFN-γ (300 U/ml) with or without TNF-α (25 ng/ml) (³⁸). Caco-2 cells were stimulated with combinations of IL-1β (8 ng/ml) plus IFN-γ (100 U/ml) with or without TNF-α (25 ng/ml) (³⁹). RAW264.7 macrophages were cultured in DMEM plus 10% FBS. For iNOS induction, they were stimulated with bacterial LPS (10 μg/ml) with or without IFN-γ (100 U/ml) for 24 h. PPAR agonists were dissolved in DMSO (vehicle). The final DMSO concentration in cell cultures was 0.1%. This concentration affected neither cell viability nor cytokine-elicited iNOS induction. Cells not treated with PPAR agonists always received an equivalent volume of DMSO. Several concentrations of the various PPAR agonists were tested, and those found to be optimal for modulation of iNOS expression were used in subsequent experiments. All treatments were done in serum-free medium without phenol red in confluent cells that had been cultured in medium without FBS for 24 h before stimulation. Cell viability was assessed by trypan blue exclusion as described previously (³⁴). According to these criteria, cell viability was above 90% under all experimental conditions employed.

Transfections

To assess NF-κB activity, MC at 80% confluence were transiently transfected with 1.5 μg of luciferase reporter plasmids containing either three consensus NF-κB response elements (3× NF-κB-pConA-Luc) (²⁷) or the corresponding empty vector (pConA-Luc) plus 2 ng of the Renilla luciferase reporter vector pRL-CMV (Promega, Madison, WI) for 3 h in Optimem (Life Technologies) using the Lipofectamine Reagent (Invitrogen-Life Technologies, Barcelona, Spain). After a 24-h recovery period in serum-free medium, cells were treated as indicated. Firefly and Renilla luminescences were measured in cell lysates according to the instructions of the manufacturer using a dual luciferase reporter system from Promega. Promoter activities were expressed as the ratio between the two luciferase measurements. To estimate PPAR activity, the oligonucleotide 5′-CCCGAACGTGACCTTTGTCCTGGTCCCGAGCT-3′, containing the PPRE from the rat acyl-CoA oxidase gene (⁴⁰), was cloned into the KpnI/SacI sites of the pGL3promoter vector (Promega) to give the construct pGL3p-ACO-PPRE-Luc. MC were transfected as detailed above. In all cases, luciferase activity of the empty vectors was negligible.

For assessment of iNOS promoter activity, a 16-Kb fragment of the human iNOS promoter (nucleotides −16 Kb to +33 bp from the transcriptional start site in the reported sequence [(⁴¹)]), was cloned in front of the luciferase reporter gene pXP₂ (pNOS2(16)Luc). This construct has been previously characterized (⁴²). To generate stably transfected MC, a previously published procedure was used with minor modifications (⁴²). Briefly, cells were transfected with 4.5 μg of pNOS2(16)Luc or with the empty pXP₂ vector and 0.5 μg of pRc-CMV containing a neomycin resistance gene, using Lipofectamine. Transfected cells were selected by culture with G-418 (1 mg/ml). Several cell clones were analyzed for luciferase activity in the absence and presence of cytokines as well as for integration of the transfected DNA by PCR. Luciferase activity was measured in lysates from transfected cells and corrected by protein content. Luciferase activity and inducibility in the pXP₂-transfected cells was negligible compared with cells transfected with the iNOS promoter. The activity of the iNOS promoter construct was further characterized in clones that were positive for inducibility and integration. Cytokine-elicited iNOS promoter activity was time-dependent, showing maximal activity after 6 h of treatment. In addition, the activity of the promoter responded to a battery of agents known to modulate iNOS transcription, including dexamethasone, IL-13, pyrrolidine dithiocarbamate (PDTC), and tyrphostin, as expected from the published evidences. Nitrite and/or iNOS protein expression were routinely monitored in stably transfected MC to confirm that their response to the different agents employed throughout the study did not differ from that of untransfected cells.

SDS-PAGE and Immunoblotting

Cells were lysed by treatment with 50 mM Tris-HCl (pH 7.4), 0.1 mM EDTA, 0.1 mM EGTA, 0.1 mM β-mercaptoethanol, and 0.5% SDS, containing 2 μg/ml of each of the following protease inhibitors: leupeptin, pepstatin A, and aprotinin. Protein concentration was determined by the BCA protein assay from Pierce (Rockford, IL). Fifteen micrograms of protein from each experimental condition were electrophoresed on 10% polyacrylamide gels and transferred to Immobilon-P membranes (Millipore, Bedford, MA). As a control for even protein loading and transfer, membranes were stained with Ponceau S. Blots were probed with primary antibodies at 1:1000 dilution followed by secondary antibodies at 1:2000 dilution. Immunocomplexes were visualized using an enhanced chemiluminescence (ECL) detection system from Amersham. The intensity of the bands was estimated by image scanning and corrected by the values obtained from the Coomassie staining of membranes or by the signal given by an antibody against a nonrelated protein.

Nitrite Determination

The accumulation of nitrite in the cell culture supernatant was taken as an index of iNOS activity. After treatment with the various agents for 16 h, nitrite was determined in cell supernatants by the Griess reaction as described previously (²⁷), using sodium nitrite as a standard.

RNA Isolation and Northern Blot Analyses

RNA isolation and Northern blot were performed essentially as described (³⁴), using the guanidinium thiocyanate-phenol-chloroform method. For analysis of iNOS mRNA expression, a 699-bp fragment of rat iNOS cDNA (²⁷), was labeled with α-[³²P]dCTP using the Rediprime kit for random primer labeling from Amersham. To ensure even loading of the samples, blots were stripped and rehybridized with a probe for human 28S rRNA gene (pTRI RNA 28S) from Ambion (Austin, TX). Densitometric analysis was performed using an Agfa StudioStar TPO scanner and the public domain software NIH IMAGE 1.60b5. Results were calculated as the ratio of iNOS mRNA versus 28S rRNA expression.

Statistical Analyses

Results are expressed as mean ± SEM. Statistical analyses were performed by using the unpaired two-tailed t test or ANOVA where applicable. Comparisons were considered statistically significant at the P < 0.05 level.

Results

Effect of PPAR Agonists on NF-κB Activation and IκBα Levels in MC

To assess the overall consequences of treatment with PPAR agonists on the response of MC to pro-inflammatory stimuli, we explored the effect of PPAR agonists on the activation of NF-κB by IL-1β/TNF-α in MC. Wy14,649 and clofibrate, considered classical PPARα activators, and cPGI₂, considered a selective agonist for PPARδ (⁴), moderately reduced the activity of NF-κB elicited by cytokines as evaluated by transient transfection of MC with an NF-κB reporter plasmid (Figure 1A). Ciglitazone, considered a selective PPARγ agonist (³), markedly reduced the activity of the reporter in both the presence and absence of cytokine stimulation (50% inhibition; Figure 1A). In addition, PPAR agonists increased the protein levels of the inhibitory subunit of NF-κB, IκBα, in cytokine-stimulated MC to variable extents (Figure 1B). In cells supplemented with Wy14,643 during an 8-h activation with cytokines, the levels of IκBα protein were 1.7-fold higher than in cells treated with cytokines alone (n = 4), although no significant effect on IκBα expression was observed in cells treated with Wy14,643 alone (1.1 fold versus control, n = 4). Clofibrate, cPGI₂ and ciglitazone increased IκBα protein levels by 1.6-, 1.3- and 1.3-fold, respectively, in cytokine-treated cells (n = 2), although the variations in IκBα levels in the absence of cytokines were below 20%. These effects could contribute to the inhibition of NF-κB activity by PPAR agonists in MC.

Figure 1.: Effect of peroxisome proliferator activated receptors (PPAR) agonists on NF-κB activity and IκBα protein levels in mesangial cells (MC). (A) The activity of NF-κB in MC was measured by transient transfection of a 3× NF-κB reporter construct as described in Materials and Methods. MC were treated with a combination of interleukin-1β (IL-1β; 2 ng/ml) plus tumor necrosis factor–α (TNF-α; 25 ng/ml) (Ck) for 6 h in the absence or presence of 100 μM Wy14,643, 50 μM clofibrate, 5 μM ciglitazone, or 10 μM cPGI₂ as indicated. Results are average values ± SEM of at least three experiments performed in triplicate (*P < 0.05 versus the same condition without agonist). (B) Levels of IκBα protein in MC treated as above were assessed by immunoblot. Results shown are representative of at least two different experiments.

Effect of Several PPAR Agonists on IL-1β/TNF-α–Elicited NO Synthesis

NF-κB activity has been found to be key for the induction of the pro-inflammatory iNOS gene (²⁶). To assess the effect of PPAR agonists on iNOS induction in MC, cells were incubated with a combination of IL-1β and TNF-α, as described previously (³⁵). Treatment of MC with several PPAR agonists during stimulation with IL-1β/TNF-α resulted in a modulation of NO generation, as estimated from the accumulation of nitrite in the cell culture medium (Figure 2). Of the agonists employed, Wy14,643, clofibrate, fenofibrate, and bezafibrate amplified cytokine-induced nitrite accumulation to a variable extent, Wy14,643 and clofibrate being the most effective. Oleic and linoleic acid had no detectable effect. cPGI₂ and ciglitazone augmented cytokine-elicited nitrite generation by 2.5- to 3-fold. None of the compounds tested induced nitrite accumulation per se (not shown).

Figure 2.: Effect of PPAR agonists on inducible nitric oxide synthase (iNOS) activity. MC were treated with IL-1β/TNF-α (Ck) in the absence or presence of the following PPAR agonists: Wy14,643 (100 μM), clofibrate (100 μM), fenofibrate (5μM), bezafibrate (100μM), oleic acid (5μM), linoleic acid (5μM), cPGI₂ (10μM), and ciglitazone (2μM). Generation of NO was estimated from the accumulation of nitrite in the cell culture supernatant over a 16-h period as determined by the Griess reaction. Results are average values ± SEM of at least three experiments. (*P < 0.05 versus Ck by t test).

Effect of PPAR Agonists on iNOS Expression

To assess whether the observed effects on nitrite generation were related to changes in iNOS expression, we explored the levels of iNOS protein attained in MC treated with IL-1β/TNF-α in the presence of various concentrations of PPAR agonists. As it can be observed in Figure 3A, Wy14,643, clofibrate, cPGI₂, and ciglitazone potentiated iNOS protein expression in a dose-dependent manner. In the case of Wy14,643 and clofibrate, the maximal potentiating effect (3 and 2.7-fold amplification, respectively; n = 3) was achieved with 100 μM agonist. The extent of the amplification of iNOS protein levels by 100 μM Wy14,643 was similar at all time points studied from 4 to 24 h (not shown). cPGI₂ and ciglitazone also amplified iNOS protein levels (3.5- and 2.4-fold amplification in the presence of 10 μM cPGI₂ and 2 μM ciglitazone, respectively; n = 3). None of the tested agonists elicited iNOS protein expression by themselves. As a control for this experiment, we evaluated the abundance of the PPAR target gene MCAD (⁴³) under the various experimental conditions used. As is shown in Figure 3B, MCAD protein levels did not increase after treatment with PPAR agonists.

Figure 3.: Effect of PPAR agonists on iNOS protein expression. MC were treated with a combination of IL-1β (2 ng/ml) plus TNF-α (25 ng/ml) (Ck) in the absence or presence of the indicated PPAR agonists. (A) iNOS protein expression was assessed by Western blot after a 6-h treatment with the indicated agents. Results shown are representative of three experiments. (B) The expression of MCAD was estimated in MC treated in the absence or presence of IL-1β plus TNF-α and/or 100 μM Wy14,643, 100 μM clofibrate, 5 μM ciglitazone, or 5 μM cPGI₂ for 6 h. Results are representative of two experiments.

Effect of Wy14,643 on Cytokine-Elicited iNOS Induction in Various Cell Types

To rule out that the potentiating effect of fibrates on iNOS induction was a singularity of MC, we explored the effect of Wy14,643 on the expression of iNOS in several cell types activated with different stimuli (Figure 4). In accordance to previous reports, a minimal combination of IFN-γ and IL-1β was needed for iNOS induction in the human intestinal epithelial cell lines Caco-2 and HT-29 (^38,39). Supplementation with Wy14,643 clearly amplified iNOS expression induced by combinations of IL-1β/IFN-γ (twofold potentiation) or IL-1β/TNF-α/IFN-γ (1.6-fold potentiation) in Caco-2 cells. In HT-29 cells stimulated with IL-1β/IFN-γ, Wy14,643 amplified iNOS induction when used at concentrations up to 100 μM, although the potentiating effect was no longer observed at 200 μM. In contrast, when iNOS was induced in HT-29 cells by treatment with IL-1β/TNF-α/IFN-γ, the effect of Wy14,643 supplementation was inhibitory when used at concentrations of 100 μM or higher. The amplification of iNOS expression by Wy14,643 could also be observed in LPS and LPS/IFN-γ–stimulated murine macrophages (RAW264.7 cells), a cell type in which negative effects of PPAR agonists have been frequently reported after overexpression of PPAR. Wy14,643 induced iNOS expression per se in none of these cell types. These results show that the potentiating effect of fibrates on iNOS induction can take place in several cell types of human, rat, and murine origin and indicate that the nature of the effects encountered may be cell- and stimulus-dependent.

Figure 4.: Effect of Wy14,643 on iNOS protein levels in several cell types. The different cell lines were stimulated with the indicated agents as specified in the experimental section in the absence or presence of Wy14,643 for 24 h as indicated.

Effect of PPAR Agonists on Cytokine-Elicited iNOS mRNA Levels

To gain insight into the mechanism involved in the potentiation of iNOS protein and activity by PPAR agonists, we first assessed the levels of iNOS mRNA in cytokine-stimulated MC in the absence and presence of the various agonists. Wy14,643 and clofibrate amplified cytokine-elicited iNOS mRNA expression (Figure 5). Quantitation of these results by densitometric analysis (Figure 5, lower panels) showed a 2.5-fold potentiation in the presence of either agonist, indicating a close correlation between iNOS protein and mRNA levels. We next explored whether the amplification of iNOS mRNA levels brought about by fibrates could be related to an increase in the stability of the iNOS transcript. For this we assessed the half-life of the iNOS transcript in mesangial cells stimulated with IL-1β/TNF-α in the absence or presence of Wy14,643 by measuring the decay of iNOS mRNA after inhibiting transcription with actinomycin D as previously reported by us (³⁵). Wy14,643 did not prolong but actually reduced iNOS mRNA half-life from 8.8 to 3.4 h (average of two experiments, not shown). cPGI₂ and ciglitazone were also effective at amplifying cytokine-induced iNOS mRNA levels (Figure 6). The extent of potentiation was similar to that attained with fibrates (Figure 6, lower panel).

Figure 5.: Effect of Wy14,643 and clofibrate on iNOS mRNA. MC were treated with IL-1β/TNF-α (Ck) and/or 100 μM Wy14,643 (A) or 50 μM clofibrate (B) for 4 h. iNOS mRNA and 28S rRNA levels were determined by Northern blot. Results are representative of three experiments. Densitometric quantitation of iNOS mRNA expressed in arbitrary units as the ratio of iNOS mRNA to 28S rRNA levels is depicted in the lower panels. Results are average values from three experiments ± SEM (*P < 0.05 with respect to Ck).

Figure 6.: Effect of cPGI₂ and ciglitazone on iNOS mRNA. MC were treated with IL-1β/TNF-α (Ck) and/or 10 μM cPGI₂ or 5 μM ciglitazone for 4 h. iNOS mRNA and 28S rRNA levels were determined by Northern blot. Results are representative of two experiments. Densitometric quantitation of iNOS mRNA expressed in arbitrary units as the ratio of iNOS mRNA to 28S rRNA levels is depicted in the lower panels. Results are average values from two experiments.

Effect of PPAR Agonists on iNOS Promoter Activity

To explore the potential involvement of transcriptional mechanisms in the amplification of iNOS induction by PPAR agonists, we generated MC stably transfected with a 16-Kb fragment of the human iNOS promoter cloned into a luciferase reporter vector (pNOS2(16)Luc). In accordance with previous observations (⁴²), this construct showed measurable activity in unstimulated cells, in which iNOS expression is not detectable (Figure 7). It has been previously proposed that this is due to the occurrence of significant iNOS transcription in unstimulated cells that would not give rise to detectable iNOS transcript levels because of insufficient iNOS mRNA stabilizing mechanisms (⁴²). Cytokine stimulation of stably transfected MC induced a twofold increase in iNOS promoter activity. Supplementation with various PPAR agonists during activation with cytokines resulted in a modulation of the activity of the iNOS promoter construct to variable extents. Clofibrate induced a moderate (40%) increase in promoter activity in both the absence and presence of cytokine stimulation. In this experimental setting, Wy14,643 did not elicit significant changes in the activity of the promoter either under basal conditions or in stimulated cells at the concentrations tested (25 and 100 μM Wy14,643). Ciglitazone was the PPAR agonist that proved most potent at activating iNOS promoter, giving rise to a 2.6-fold increase in unstimulated cells and a 3.7-fold increase over basal activity in the presence of cytokines. These results suggest that the amplification of iNOS induction by certain PPAR agonists may occur by transcriptional mechanisms.

Figure 7.: Effect of PPAR agonists on the activity of the iNOS promoter in stably transfected MC. Confluent MC, stably transfected with a 16-Kb reporter construct of the human iNOS promoter, were treated with IL-1β/TNF-α (Ck) in the absence or presence of PPAR agonists as indicated: (1) vehicle; (2) 25 μM Wy14,643; (3) 100 μM clofibrate; (4) 5 μM ciglitazone. After 6-h treatment, luciferase activity in the cell lysates was measured and corrected for the total protein content. Results shown are average values ± SE of three experiments performed in triplicate (*P < 0.05). Essentially the same results were obtained using a single cell clone or the pool of transfected cells.

Expression and Activity of PPAR in MC

To explore the role of PPAR in the effects of agonists on iNOS expression, we first assessed the presence of PPAR in MC and other cell types by Western blot. As is shown in Figure 8A, MC express PPARα, β, and γ proteins. Interestingly, Caco-2 and HT-29 cells predominantly express a shorter form of PPARα that coincides with the molecular weight of the truncated dominant negative variant of PPARα, which has been identified in human hepatocytes and HepG2 cells (⁴⁴). The antibodies against PPARβ and PPARγ detected protein bands of the expected molecular weights in all cell types employed. In addition, we estimated the activity of PPAR by transiently transfecting MC with a PPRE-luciferase reporter construct (Figure 8B). PPRE activity could be detected under basal conditions, suggesting that MC contain significant amounts of endogenous PPAR agonists, as has been observed in other systems (⁴⁵). Treatment of MC with Wy14,643 under the conditions resulting in iNOS potentiation did not increase PPRE activity, instead, a decrease in PPRE activity was observed in cytokine-treated cells, both in the presence or absence of agonist.

Figure 8.: Expression and activity of PPAR in MC. (A) The presence of PPAR isoforms in MC and other cell types was assessed by Western blot. (B) The activity of PPAR in MC treated in the absence or presence of cytokines and/or 100 μM Wy14,643 for 16 h was explored by transient transfection with a PPRE-luciferase reporter construct as detailed in the experimental section. Results are average values ± SE of two experiments performed in duplicate.

Discussion

The ligand-activated transcription factors known as PPAR have been shown to play important roles in vascular pathophysiology by acting as negative modulators of inflammatory processes. However, the results herein reported suggest that the effects of PPAR agonists on cellular activation may be more complex than originally thought, because, despite of their ability to inhibit NF-κB activity, several PPAR agonists potently amplify iNOS induction in response to cytokine treatment in various cell types. The amplification of iNOS expression by PPAR agonists may have positive or negative effects on the response of cells to pro-inflammatory stimuli. NO is a dual regulator of inflammation that promotes vasodilation and cell activation but also triggers mechanisms that contribute to the resolution of inflammation both in animal and cellular models (^29,46). These NO-mediated mechanisms include the negative regulation of NF-κB and AP-1 activity and lead to the downregulation of cytokines, adhesion molecules, iNOS, and COX-2 expression (^28,32,34,35). PPAR ligands have been reported to exert antiinflammatory activities in different cell types by antagonizing the transcriptional activity of NF-κB through mechanisms that may involve the induction of IκBα expression (²⁵). The observations that PPAR agonists increase IκBα expression and reduce the activity of NF-κB in MC suggest that they may have antiinflammatory effects. Consistent with this, we have observed that treatment with Wy14,643 anticipates the decay of COX-2 expression in cytokine-stimulated MC (results not shown). Interestingly, IκBα induction and the subsequent inhibition of NF-κB have been reported to stand at the basis of the inhibition of adhesion molecules and COX-2 expression by NO (^32,34). Thus, the NF-κB pathway appears as an important point of interaction common to peroxisome proliferator– and NO–mediated modulation of the inflammatory response in MC.

The observation that various PPAR agonists amplify iNOS induction is in apparent contradiction with several reports that have proposed a negative role for PPAR activation on iNOS expression (^22,47). These differences may arise from the use of different cell types or experimental approaches. Overexpression of PPAR was required in several studies to observe the inhibitory effects of PPAR agonists on the activity of iNOS promoter reporter constructs, although either no effect or an increase in the activity of the iNOS promoter could be observed in the absence of PPAR overexpression (^22,47). Moreover, as stated above, several PPAR agonists, including 15d-PGJ₂, troglitazone, and cPGI₂, have been shown to display PPAR-independent effects (^19–21) that could contribute to the modulation of iNOS expression. The increasing identifications of PPAR-independent effects of PPAR ligands may bring up the need to reevaluate the involvement of PPAR in some of the reported effects (^18,48).

We have observed that iNOS amplification by PPAR agonists occurs at the protein and mRNA levels. In a previous report, the PPARγ agonist troglitazone has been reported to increase the stability of iNOS mRNA in vascular smooth muscle cells (³⁷). However, our results indicate that the half-life of iNOS mRNA is reduced in cytokine-stimulated MC supplemented with Wy14,643 and that the amplifying effect of some PPAR agonists, including ciglitazone, on iNOS induction occurs by transcriptional mechanisms. As stated above, the effect of PPAR agonists on MC iNOS expression could be mediated by PPAR-dependent or independent mechanisms. On one hand, it could result from the activation of PPAR response elements (PPRE) present in the iNOS promoter that have not yet been characterized. As an alternative mechanism, activation of PPAR could enhance cytokine-elicited iNOS induction by promoting the degradation of lipidic negative modulators, among which could be some COX products, including PG of the J series, like 15d-PGJ₂. In fact, inhibition of COX activity has been reported to lead to enhanced iNOS expression in a model of immune glomerulonephritis (⁴⁹). At this point, however, our results do not support the involvement of PPAR in the effects herein described. We have observed that mesangial cells in culture express all three PPAR proteins. However, we have not detected an increase in PPAR activity, as measured by a PPRE reporter assay, nor increases in the levels of the PPAR target gene MCAD, under the experimental conditions associated with enhanced iNOS expression. Moreover, amplification of iNOS expression also took place in cell lines expressing a dominant negative variant of PPARα. In addition, the attempt to block PPAR activity in MC by treatment with a PPRE decoy oligonucleotide using a previously published procedure (⁵⁰) did not significantly affect iNOS induction or potentiation by PPAR agonists (results not shown). These observations raise the possibility that the effect of PPAR agonists on iNOS expression can occur through receptor-independent mechanisms.

Taken together, the findings reported here show that certain PPAR agonists may enhance iNOS expression and NO generation in cells exposed to pro-inflammatory stimuli. NO itself has potent antiinflammatory effects. Whether an increased generation of NO may contribute to the antiinflammatory properties of some PPAR agonists will be the subject of future studies.

We thank Estrella Soria and María Jesús Carrasco for technical assistance, Dr. Blanca Pérez-Maceda for help with cell lines, and Dr. Ana Aranda (Instituto de Investigaciones Biomédicas, C.S.I.C.) for generously sharing ideas and reagents. This work was supported by grants from Ministerio de Ciencia y Tecnología SAF2000–0149, Comunidad Autónoma de Madrid 08.4/0031/2000, and Fundación Ramón Areces, Madrid, Spain. E. Cernuda-Morollón is the recipient of fellowships from Fundación Ramón Areces and Residencia de Estudiantes, Ayuntamiento de Madrid, Spain.