PPARγ Inhibition of Cyclooxygenase-2, PGE₂ Synthase, and Inducible Nitric Oxide Synthase in Cardiac Myocytes

Peroxisome proliferator-activated receptors (PPARs) are a family of 3 nuclear hormone receptors, PPARα, PPARβ (also named PPARδ), and PPARγ, which are members of the steroid receptor superfamily. PPARs bind to cognate DNA elements called peroxisome proliferator response elements (PPRE) as obligate heterodimers with the retinoid X receptor (RXR). After ligand activation, they work as transcription factors.^1,2 There are two PPARγ subtypes, PPARγ₁ and PPARγ₂, which are derived from alternative splicing and promoter usage.^3,4 PPARγ₂ is highly expressed in adipose tissue, whereas PPARγ₁ has been found in the kidney, heart, liver, and activated monocytes.^5,6 PPARγ can be activated by docosahexanoic acid and certain prostaglandins.⁷ Other PPARγ ligands include the natural prostaglandin metabolite 15-deoxy-Δ^12,14-prostaglandin J₂ (15dPGJ₂),^8,9 polyunsaturated fatty acids,¹⁰ nonsteroidal anti-inflammatory drugs,¹¹ and members of the thiazolinedione family.¹²

PPARγ has been associated with control of inflammation by inhibiting cytokine stimulation of COX-2 and inducible nitric oxide synthase (iNOS) expression in different cell types,^13–15 decreasing release of proinflammatory cytokines¹⁶ and inhibiting vascular smooth cell migration.^17–19 PPARγ has also been shown to have antineoplastic and antigrowth properties.¹⁸ However, the precise mechanisms defining these effects are unknown.

We recently found that in neonatal ventricular myocytes (NVM), induction of cyclooxygenase-2 (COX-2), and prostaglandin E₂ synthase (PGES) by the proinflammatory cytokine interleukin-1β (IL-1β) results in preferential production of PGE₂.²⁰ In addition, we found that PGE₂ induces growth in NVM, suggesting that it plays a detrimental role in the heart. Also, it is known that iNOS is induced on stimulation with IL-1β²¹ and that it contributes to myocyte damage after infarction in the rat²² and mouse.²³ Thus, we hypothesized that PPARγ would exert an anti-inflammatory effect in NVM, decreasing COX-2 and PGES expression and leading to decreased production of the hypertrophic prostanoid PGE₂, as well as decreasing iNOS.

Methods

Cell Culture

Primary cultures of NVM were derived from digestion of 1- to 2-day-old neonatal Sprague-Dawley rat hearts (Charles River).²⁴ This protocol was approved by the Henry Ford Hospital IACUC. Myocytes were plated at a density of 1×10⁵/cm² (1×10⁶ cells per well of a 6-well dish or 5.8×10⁶ cells/10-cm dish) in DMEM (Gibco-BRL) plus 10% FBS (Gibco) and 0.1 mmol/L bromo- deoxyuridine for 40 hours. Serum-free medium supplemented with glutamine, insulin, selenium, and transferrin (SF-DMEM) was then added for 24 hours. Unless otherwise specified, all treatments lasted 24 hours. Cells were pretreated for 1 hour with 15dPGJ₂, troglitazone, and GW9662 before treatment with 5 ng/mL IL-1β.

Isolation of Protein and Western Blot of COX-2, PGES, and iNOS

Protein was isolated from NVM through the use of lysis buffer and protease inhibitors.²¹ Lysate protein (50 μg per lane) was separated out by electrophoresis on an 8% SDS polyacrylamide gel (for COX-2 and iNOS detection) and transferred to an Immobilon-P PVDF membrane (Millipore); 72-kDa COX-2 and 130-kDa iNOS proteins were detected as described previously.^25,26 Microsome-enriched preparations were used for Western blot of PGES. For this, 5.8×10⁶ cells/10-cm dish were washed twice with PBS and scraped from the plate in the same solution. Cells were then centrifuged at 16 000g for 5 minutes, and the pellet was resuspended in a solution composed of 60 mmol/L PIPES, 25 mmol/L HEPES, 0.75% Triton-X 100, 10 mmol/L EGTA, 2 mmol/L MgCl₂, 10 μg/mL leupeptin, 5 μg/mL aprotinin, 1 μg/mL pepstatin, 156 μg/mL benzamidine, 1 mmol/L iodoacetamidine, and 0.5 mmol/L PMSF (pH 7.4). After 15 minutes of incubation on ice, the homogenate was placed on top of a 6% sucrose cushion and ultracentrifuged at 230 000g with the use of a Beckman ultracentrifuge and a 70.1-Ti rotor for 40 minutes. The pellet was resuspended in lysis buffer containing protease inhibitors. The 17-kDa PGE₂ synthase protein was separated out on a 15% SDS polyacrylamide gel and detected with an anti-rabbit polyclonal antibody (Cayman) diluted 1:500. The appropriate secondary antibody linked to horseradish peroxidase was used for chemiluminescent detection with ECL Western blot reagents (Amersham Pharmacia Biotech). The signal was detected by exposure to Fuji RX film, quantified by laser densitometry, and expressed as optical density (OD) units.

Enzyme Immunoassay for Measurement of PGE₂, PGF_2α, and the Stable PGI₂ Metabolite 6-Keto PGF_1α

1×10⁶ cells per well of a 6-well plate were treated with vehicle, IL-1β, 15dPGJ₂, or 15dPGJ₂+IL-1β in 1 mL SF-DMEM. A 1-mL aliquot of medium from each well was dried down and resuspended in 0.15 mL ultrapure H₂O. Aliquots were diluted if necessary and assayed for specific prostaglandins with the use of EIA kits from Cayman. Values from triplicate wells were averaged, and mean±SEM values are expressed as nanograms per milligram of protein.

Isolation of RNA and RT-PCR of COX-2 and PGES

Total RNA was isolated from control and IL-1β–treated NVM by means of Tri Reagent (MRC), according to the manufacturer’s instructions; 2 μg total RNA was reverse-transcribed in a total volume of 25 μL, using 1 μg random primer (Gibco/BRL) and 200 U MMLV reverse transcriptase (Promega). Five-microliter aliquots of the resulting cDNA were subjected to PCR in a 50-μL reaction volume for amplification of COX-2 (462 bp), PGES (473 bp), and GAPDH (554 bp). Oligonucleotide primers (5′-3′) for rat COX-2 were AGT ACG AAG ACC CTG CCT ACG (sense) and TAA GTT GGT GGG CTG TCA AT (antisense). Primers for the complete PGES cDNA were ATG ACT TCC CTG GGT TTG GTG (sense) and TCA GCT GCT GGT CAC AGA TGG (antisense). Primers for rat GAPDH were AAT GCA TCC TGC ACC ACC TGC (sense) and GGA GGC CAT GTA GGC CAT GAG GTC (antisense). A PTC-100 thermal controller (MJ Research) was used for amplification, and the program settings were denaturation for 30 seconds at 95°C, annealing for 60 seconds at 55°C for COX-2, and 60°C for PGES and GAPDH, and extension for 90 seconds at 72°C for 35 (COX-2 and PGES) or 25 cycles (GAPDH). PCR products were separated out on a 2% agarose gel, stained with ethidium bromide, and photographed. Densitometry was used for quantification. COX-2 and PGES were normalized to GAPDH.

Transient Transfection and Luciferase Assay

Transfection was performed by electroporation, and luciferase activity was assayed as described previously^27,28;1 μg PPRE-TK-luciferase [containing 3 copies of PPRE, coupled to a minimal thymidine kinase (TK) promoter, kindly provided by Dr Ronald Evans, Salk Institute] was transfected per 3×10⁶ NVM, and the cells were aliquoted into 3 wells of a 12-well plate. After 40 hours, the medium was changed to SF-DMEM, and 24 hours later the cells were treated with 15dPGJ₂. After a 24-hour treatment period, NVM were lysed and assayed for luciferase activity using Promega reagents in an OptoComp 1 luminometer. Relative light units (RLU) from triplicate wells were averaged. RLU values in control (untreated) lysates were normalized, and the values of 15dPGJ₂-treated lysates were compared with control to give the fold increase. Data are expressed as mean±SEM.

Chemicals

15dPGJ₂, GW9662, PGES antibody, and EIA kits for PGE₂, PGF_2α, and 6-keto PGF_1α were obtained from Cayman. Troglitazone was obtained from Biomol. PPARγ₁, COX-2, and iNOS antibodies were obtained from Santa Cruz, and IL-1β was from Collaborative Biomedical Products.

Statistics

Values are represented as mean±SEM. Differences in mean values were analyzed by Student t test or 1-way ANOVA, with pairwise multiple comparisons made by the Student-Newman-Keuls method. A value of P<0.05 was considered significant.

Results

Detection and Activation of PPARγ in NVM

To determine whether PPARγ receptors are present in NVM, protein from cell lysates was analyzed by Western blot. PPARγ₁ was detected as a 52-kDa protein (Figure 1A). We next tested whether endogenous PPARγ is transcriptionally active in cardiac myocytes. For this, NVM were transiently transfected with the PPRE-TK-luciferase reporter plasmid and treated with the PPARγ agonist 15dPGJ₂. Figure 1B shows that 15dPGJ₂ increased luciferase activity 12.0±2.2-fold compared with untreated NVM. Finally, we cotransfected NVM with the PPRE-TK-Luc reporter gene and dominant negative PPARγ cDNA (dnPPARγ). As shown in Figure 1B, 15dPGJ₂ activation of the PPRE-TK-Luc reporter gene was significantly decreased.

• Download figure
• Download PowerPoint

Effect of 15dPGJ₂ on IL-1β Stimulation of COX-2, PGES, and Prostaglandin Production

High PGE₂ resulting from induction of COX-2 participates in inflammation. We tested whether PPARγ activation is anti-inflammatory by examining the effect of the natural PPARγ agonist 15dPGJ₂ on PGE₂ production and COX-2 mRNA. 15dPGJ₂ reduced IL-1β–stimulated PGE₂ release by 73% (Figure 2A); however, IL-1β–induced PGF_2α and PGI₂ levels were not affected (data not shown). In addition, 15dPGJ₂ decreased IL-1β–induced COX-2 mRNA levels by 40% (Figure 2B). IL-1β induction of COX-2 protein (72 kDa) was similarly inhibited (data not shown).

• Download figure
• Download PowerPoint

Since PGE₂ production was inhibited to a greater extent than COX-2 mRNA, we tested whether 15dPGJ₂ was affecting the inducible form of PGE₂ synthase (mPGES). Figure 3A shows that 15dPGJ₂ completely inhibited IL-1β induction of PGES mRNA. Western blots of microsome-enriched preparations revealed that 15dPGJ₂ also blocked IL-1β induction of PGES protein in NVM (Figure 3B).

• Download figure
• Download PowerPoint

Effect of the PPARγ Agonist Troglitazone on COX-2 and PGE₂

We used a second PPARγ agonist, troglitazone (TRO, 100 μmol/L) to verify our findings with 15dPGJ₂. TRO transactivated the PPRE-TK-LUC reporter plasmid (data not shown). TRO also blocked IL-1β–induced COX-2 (Figure 4A) and PGE₂ production (Figure 4B).

• Download figure
• Download PowerPoint

Effect of a PPARγ Antagonist, GW9662, on COX-2 and PGES

To further verify that the effects of 15dPGJ₂ are PPARγ-dependent, we studied the effect of the PPARγ antagonist GW9662 (GW, 2 μmol/L). GW blocked the effects of 15dPGJ₂ on IL-1β–induced COX-2 (Figure 5A) and PGES mRNA (Figure 5B). GW alone did not modify either COX-2 or PGES mRNA (data not shown).

• Download figure
• Download PowerPoint

Effect of the PPARγ Activator 15dPGJ2 on iNOS

We have previously shown that IL-1β induces iNOS expression and nitrite production in NVM.²⁵ Since iNOS has been implicated in inflammation, we examined the effect of 15dPGJ₂ on IL-1β induction of iNOS in NVM. Although treatment of NVM with IL-1β significantly increased iNOS protein, cotreatment with IL-1β and 15dPGJ₂ inhibited iNOS expression (Figure 6A). Troglitazone mimicked the effect of 15dPGJ₂ on IL-1β–induced iNOS expression (Figure 6B). GW was able to partially reverse the inhibitory effect of 15dPGJ₂ on IL-1β–induced iNOS protein (Figure 6C).

• Download figure
• Download PowerPoint

Discussion

We found that PPARγ is present and functional in cardiac myocytes. We also demonstrated that its natural agonist 15dPGJ₂ exerts anti-inflammatory effects by modulating IL-1β–stimulated COX-2, PGES, and iNOS in a PPAR-dependent manner.

The PGD₂ dehydration product 15dPGJ₂ is the most potent endogenous ligand for PPARγ yet discovered.^8,9 Recently, 15dPGJ₂ production was measured in tryptase-stimulated mast cells in culture²⁹ as well as human breast tissue.³⁰ Moreover, prostaglandin D synthase (PGDS) mRNA is expressed in the heart.³¹ Thus, locally produced PGD₂ may result in 15dPGJ₂ in myocytes or surrounding cells. Although it is not known whether 15dPGJ₂ is produced in myocytes and has autocrine effects to activate PPARγ, it is known that the 15dPGJ₂ precursor Δ¹²-PGJ₂ is transported into cells and accumulates in the nucleus.³²

We found that 15dPGJ₂ decreased IL-β induction of COX-2 mRNA. This inhibition probably occurred at the transcriptional level. The anti-inflammatory effects of PPARγ are thought to be mediated by antagonizing transcription factors such as NF-κB, AP-1, STAT,² and NFAT,³³ some of which are necessary for COX-2 transcription. In fact, Subbaramaiah et al¹³ showed that stimulation of a breast epithelial cell line with phorbol ester induced COX-2 and that two different PPARγ agonists interfered with AP-1–dependent activation of the COX-2 promoter. A PPARγ agonist also inhibited IL-β–induced COX-2 expression in mesangial cells, although the authors suggested that the effects seen are PPAR-independent.³⁴

Previous studies have demonstrated that COX-2 and microsomal PGES are colocalized in the perinuclear envelope, suggesting coordinated biosynthetic activity between enzymes.³⁵ In fact, induction of both COX-2 and PGES results in high level PGE₂ production by cardiac myocytes.³⁶ Our data show that 15dPGJ₂ blocked IL-1β stimulation of PGE₂ production but did not modify IL-1β stimulation of PGI₂ or PGF_2α, indicating that PPARγ ligand effects are specific for PGES. Naraba et al³⁷ reported that the mouse PGES gene contains several cis-acting elements including C/EBPα, AP-1, C/EBPβ, and CACCC-binding factor. In contrast to COX-2, no binding sites for NF-κB or CREB were found in the PGES promoter region. This may indicate that the mechanisms for PPAR regulation of COX-2 and PGES are distinct. In the same report, the authors showed that induction of PGES is exclusively regulated by the transcription factor early growth response-1 (Egr-1), which can be rapidly and transiently upregulated by various growth factors including IL-1β.^37,38 Also, Egr-1 interacts with CBP/p300 to modulate gene transcription.³⁹ Interestingly, ligand-induced transactivation of PPARγ is achieved by receptor recruitment of the coactivator CBP/p300.⁴⁰ Hence, we speculate that PPAR-mediated inhibition of PGES could result from competition for limited amounts of CBP/p300.

It has been shown that the PPAR ligand 15dPGJ₂ exerts anti-inflammatory functions independent of PPAR activation.⁴¹ Some of the independent effects of 15dPGJ₂ have been attributed to an electrophilic carbon that can react with nucleophils.⁴⁰ To validate our results, we tested a different PPAR agonist. Troglitazone, an antidiabetic drug belonging to the thiazolinedione family, mimicked the effect of 15dPGJ₂ on IL-1β–induced COX-2, iNOS, and PGE₂ production. We found that troglitazone was more effective than 15dPGJ₂ in inhibiting IL-1β induction of COX-2. This may result from the fact that these two selective PPARγ agonists bind PPAR and alter its structure in specific ways, resulting in distinct PPARγ-coactivator interactions.⁴² It is also possible that 15dPGJ₂ and troglitazone may differ in their stability and affinity binding constant for PPARγ in NVM.

Recently, a PPAR-selective antagonist, GW9662, was identified.⁴³ GW9662 has been demonstrated to covalently modify PPAR in a selective and irreversible manner and was found to block the effects of 15dPGJ₂ in bone marrow,⁴⁴ hepatocytes,⁴⁵ and human smooth muscle cells.⁴⁶ By using this PPAR-specific antagonist, we demonstrated that the effects of 15dPGJ₂ on IL-1β–induced COX-2 and PGES are essentially mediated by PPAR.

As expected, activation of PPARγ inhibited IL-1β stimulation of the proinflammatory gene iNOS. Previous reports demonstrated that 15dPGJ₂ inhibited IL-1β regulation of iNOS by blocking NF-κB activation.⁴⁷ A similar mechanism is probably responsible for regulation of iNOS in cardiac myocytes in our studies, since we know that IL-1β activates NF-κB in cardiac myocytes, and 15dPGJ₂ can partially abrogate this effect (Mendez and LaPointe, unpublished data). However, PPAR-independent effects have been reported.^40,48 Our results show that GW only partly reverses the effect of 15dPGJ₂ on IL-induced iNOS protein. Thus, we cannot rule out the possibility of some PPAR-independent effect of the natural ligand 15dPGJ₂ on iNOS regulation. At this time, we know of no studies that definitively address the PPAR-independent mechanisms that regulate iNOS expression in cardiac myocytes.

iNOS and inducible COX-2 have been implicated in pathophysiological events in the heart.^22,49,50 We reported that PGE₂ is produced by COX-2 in cardiac myocytes and fibroblasts and that PGE₂ analogues stimulate myocyte hypertrophy²⁰ and fibroblast proliferation in vitro (Mendez and LaPointe, unpublished observations). In addition, COX-2 inhibition reduces hypertrophy and fibrosis in vivo in a mouse model of myocardial infarction (LaPointe, unpublished data). PPAR inhibitors have been shown to reduce protein synthesis in myocytes^51,52 and vascular smooth muscle cell proliferation.¹⁸ It is interesting to speculate that 15dPGJ₂ exerts its antigrowth effects in NVM by blocking the release of the hypertrophic prostaglandin PGE₂.

Perspectives

PPARγ has been shown to participate in cell growth, differentiation, and inhibition of inflammatory responses in multiple cell types important to cardiovascular function, such as vascular smooth muscle cells and cardiac myocytes. In this study, we provide compelling evidence that PPARγ activation blunts the release of the hypertrophic and proinflammatory prostanoid PGE₂, along with blockade of the inflammatory genes COX-2 and PGES, in a PPAR-dependent manner. In addition, blockade of iNOS is both PPAR-dependent and PPAR-independent. Knowledge of key components of the transcriptional regulation of inflammatory genes by PPARγ may provide a viable therapeutic target in the control of cardiac diseases resulting from inflammatory stimuli.

Footnotes