Peroxisome proliferator-activated receptor gamma inhibits transforming growth factor beta-induced connective tissue growth factor expression in human aortic smooth muscle cells by interfering with Smad3.

Activation of peroxisome proliferator-activated receptor gamma (PPAR gamma) after balloon injury significantly inhibits VSMC proliferation and neointima formation. However, the precise mechanisms of this inhibition have not been determined. We hypothesized that activation of PPAR gamma in vascular injury could attenuate VSMC growth and matrix production during vascular lesion formation. Since connective tissue growth factor (CTGF) is a key factor regulating extracellular matrix production, abrogation of transforming growth factor beta (TGF-beta)-induced CTGF production by PPAR gamma activation may be one of the mechanisms through which PPAR gamma agonists inhibit neointima formation after vascular injury. In this study, we demonstrate that the PPAR gamma natural ligand (15-deoxyprostaglandin J(2)) and a synthetic ligand (GW7845) significantly inhibit TGF-beta-induced CTGF production in a dose-dependent manner in HASMCs. In addition, suppression of CTGF mRNA expression is relieved by pretreatment with an antagonist of PPAR gamma (GW9662), suggesting that the inhibition of CTGF expression is mediated by PPAR gamma. To elucidate further the molecular mechanism by which PPAR gamma inhibits CTGF expression, an approximately 2-kilobase pair CTGF promoter was cloned. We found that PPAR gamma activation inhibits TGF-beta-induced CTGF promoter activity in a dose-dependent manner, and suppression of CTGF promoter activity by PPAR gamma activation is completely rescued by overexpression of Smad3, but not by Smad4. Furthermore, PPAR gamma physically interacts with Smad3 but not Smad4 in vitro in glutathione S-transferase pull-down experiments. Taken together, the data suggest that PPAR gamma inhibits TGF-beta-induced CTGF expression in HASMCs by directly interfering with the Smad3 signaling pathway.


Activation of peroxisome proliferator-activated receptor ␥ (PPAR␥) after balloon injury significantly inhibits VSMC proliferation and neointima formation.
However, the precise mechanisms of this inhibition have not been determined. We hypothesized that activation of PPAR␥ in vascular injury could attenuate VSMC growth and matrix production during vascular lesion formation. Since connective tissue growth factor (CTGF) is a key factor regulating extracellular matrix production, abrogation of transforming growth factor ␤ (TGF-␤)-induced CTGF production by PPAR␥ activation may be one of the mechanisms through which PPAR␥ agonists inhibit neointima formation after vascular injury. In this study, we demonstrate that the PPAR␥ natural ligand (15-deoxyprostaglandin J 2 ) and a synthetic ligand (GW7845) significantly inhibit TGF-␤-induced CTGF production in a dose-dependent manner in HASMCs. In addition, suppression of CTGF mRNA expression is relieved by pretreatment with an antagonist of PPAR␥ (GW9662), suggesting that the inhibition of CTGF expression is mediated by PPAR␥. To elucidate further the molecular mechanism by which PPAR␥ inhibits CTGF expression, an ϳ2kilobase pair CTGF promoter was cloned. We found that PPAR␥ activation inhibits TGF-␤-induced CTGF promoter activity in a dose-dependent manner, and suppression of CTGF promoter activity by PPAR␥ activation is completely rescued by overexpression of Smad3, but not by Smad4. Furthermore, PPAR␥ physically interacts with Smad3 but not Smad4 in vitro in glutathione S-transferase pull-down experiments. Taken together, the data suggest that PPAR␥ inhibits TGF-␤-induced CTGF expression in HASMCs by directly interfering with the Smad3 signaling pathway.
Early atherosclerotic lesions are characterized by accumula-tion of inflammatory cells and intimal smooth muscle cell proliferation and migration, as well as by extracellular matrix deposition. Cytokines and growth factors such as transforming growth factor ␤ (TGF-␤) 1 participate in these processes as evidenced by TGF-␤-induced overproduction of extracellular matrix proteins in intimal vascular smooth muscle cells (VSMCs) (1,2). Connective tissue growth factor (CTGF) is a novel cysteine-rich secreted peptide that is a key regulator of extracellular matrix production and plays an important role in atherosclerosis and restenosis (3). It was reported that CTGF could mediate the effects of TGF-␤ on stimulation of extracellular matrix production in atherosclerosis (4). In addition, CTGF is implicated in fibrotic disorders such as systemic scleroderma (5,6).
TGF-␤ signaling is initiated upon its binding to two cell membrane receptors termed type I (T␤RI) and type II (T␤RII). Both receptors are serine/threonine kinases, and binding by TGF-␤ results in phosphorylation of T␤RI and T␤RII. Smad proteins are the primary substrates known to date for phosphorylated TGF-␤ receptors. The phosphorylation of Smad2 or Smad3 by TGF-␤ receptor causes their association with Smad4. The Smad complexes translocate into the nucleus and bind to cofactors that determine the choice of the target gene (7). In addition to Smad proteins, c-Jun NH 2 -terminal kinase and mitogen-activated protein kinases are reported to be involved in TGF-␤ signaling (8,9).
Peroxisome proliferator-activated receptors (PPARs) including ␣, ␥, and ␦/␤ are a family of ligand-activated nuclear transcriptional factors that are emerging as important determinants of vascular function and structure (10 -12). Recent studies have documented that PPAR␥ is present in all critical vascular cells as follows: endothelial cells (13), VSMCs (14), and monocytes/macrophages (15). It was reported that thiazolidinediones (TZD), a class of antidiabetic drugs that are specific ligands of PPAR␥, inhibit neointima formation after balloon injury (16) and the development of atherosclerosis in low density lipoprotein receptor-deficient mice (17). Based on the studies above, we hypothesized that the up-regulation of PPAR␥ gene expression induced by cytokines and growth factors in response to vascular injury may function as a countervailing influence that attenuates VSMC growth and matrix production during lesion formation.
To understand the role of PPAR␥ in VSMC, we used commercially available filter-based microarrays from Research Genetics (GF-211, Huntsville, AL) to quantitate changes in expression of mRNAs in human aortic smooth muscle cells (HASMCs). Over 50 genes are either up-or down-regulated by at least 2-fold after treatment with PPAR␥ ligands. Connective tissue growth factor (CTGF) was one of the genes most downregulated by PPAR␥ activators (Ϫ5.6-, Ϫ6.5-, and Ϫ8-fold by ciglitazone, GW7845, and 15-d-PGJ 2 , respectively). Therefore, we postulated that CTGF down-regulation by PPAR␥ activation might be one of the mechanisms by which PPAR␥ agonists inhibit neointima formation.

Materials-Human
Cell Culture-HASMCs were purchased from BioWhittaker (San Diego, CA) and cultured in smooth muscle cell growth medium-2 containing 5% FBS, 2 ng/ml human basic fibroblast growth factor, 0.5 ng/ml human epidermal growth factor, 50 g/ml gentamicin, 50 ng/ml amphotericin B, and 5 g/ml bovine insulin. For all experiments, early passage (5-7) HASMCs were grown to 80 -90% confluence and made quiescent by serum starvation (0.4% FBS) for at least 24 h. Each agonist examined was added 60 min before the addition of human recombinant TGF-␤1. The HepG2 cell line was obtained from the American Type Culture Collection and cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10%(v/v) FBS in a 5% CO 2 humidified atmosphere at 37°C.
Northern Blotting Analysis-Twenty g of total RNA, isolated from each condition by using acid/guanidinium thiocyanate, was subjected to electrophoresis through 1% formaldehyde-agarose gels. After transferring to nylon membranes (Bio-Rad), the RNA was cross-linked to the membrane by a UV cross-linker (Stratagene, La Jolla, CA). 32 P-Labeled cDNA probes were generated by using the random primer labeling system (Life Technologies, Inc.). Blots were pre-hybridized, hybridized, and washed once with 1ϫ SSC at 65°C and once with 0.1ϫ SSC, 1.0% SDS (w/v) at 65°C over 1 h. The lane loading differences were normalized by using a GAPDH cDNA probe.
Western Blotting Analysis-Cell culture medium was removed and concentrated with a Biomax Column (Millipore, Bedford, MA) after 24 h of TGF-␤1 stimulation. The cells were harvested and lysed in solubilization buffer (20 mM Tris-HCl, pH 7.5; 150 mM NaCl; 1 mM EDTA; 1 mM EGTA; 1% Triton X-100; 2.5 mM sodium pyrophosphate; 1 mM sodium vanadate; 10 g/ml each of aprotinin and leupeptin; 2 mM phenylmethylsulfonyl fluoride). The samples were cleared by centrifugation at 13,000 rpm for 10 min. Thirty g of total cell lysate or concentrated cell medium were subjected to SDS-polyacrylamide gel electrophoresis and electrotransferred to nitrocellulose membrane (Bio-Rad). After blocking in 20 mM Tris-HCl, pH 7.6, containing 150 mM NaCl, 0.1% Tween 20, and 5% (w/v) non-fat dry milk, blots were incubated with specific antibodies against PPAR␥ (Santa Cruz Biotechnology) or CTGF (Torrey Pines Biolabs) overnight at 4°C. The blots were then incubated with horseradish peroxidase-conjugated secondary antibody (Santa Cruz Biotechnology). The immunoactivity was visualized by the enhanced chemiluminescence detection system (ECL, Amersham Pharmacia Biotech) according to the manufacturer's instructions.
Cloning of CTGF Promoter-A 661-base pair human CTGF promoter (GenBank TM accession number X92511) was used to search the human genome data base and found to match an ϳ61-kb DNA fragment (Gen-Bank TM accession number AL354866) from chromosome 6, the map position of CTGF. Based on this ϳ61-kb DNA sequence, we designed two primers to amplify an ϳ2-kb CTGF promoter by polymerase chain reaction. This CTGF promoter (nt Ϫ2065 to nt ϩ72) was cloned into a luciferase reporter plasmid (pGL3 basic, Promega). We designated this CTGF promoter/luciferase reporter plasmid as pCTGF-Luc.
Transient Transfection and Luciferase Assays-HepG2 cells, grown to 50 -60% confluence in DMEM/F-12 supplemented with 10% FBS, were transiently transfected by using LipofectAMINE (Life Technologies, Inc.) with reporter and expression plasmids as described. Green fluorescence protein (GFP) expression plasmid was co-transfected as the control for transfection efficiency. The total amount of transfected DNA was kept constant by using a corresponding empty vector mock DNA. Twenty four hours after transfection, cells were cultured for 24 h in serum-free medium and incubated for 24 h in the same medium containing the appropriate reagents for the experiments. The reporter luciferase assay kit (Promega) was used with a luminometer (Victor II, PerkinElmer Life Sciences) to measure luciferase activity in the cells according to the manufacturer's instructions. The luciferase activity was normalized by GFP.
GST Pull-down Experiment-Recombinant PPAR␥1 was transcribed and translated in vitro using a rabbit reticulocyte lysate (Promega) and labeled with [ 35 S]methionine. GST, GST-Smad3, and GST-Smad4 were produced in Escherichia coli DH5␣ and purified using glutathione-Sepharose beads according to the manufacturer's instructions (Amersham Pharmacia Biotech). Ten g of GST-Smad3 or GST-Smad4 fusion protein, loaded on glutathione-Sepharose beads, was incubated with 10 l of [ 35 S]methionine-labeled PPAR␥1 for 2 h in 500 l of binding buffer (20 mM Tris-HCl, pH 7.5, 100 mM KCl, 0.7 mM EDTA, 0.05% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride). The PPAR␥ ligand, GW7845 (at a final concentration of 1 M), was added at the beginning of incubation. After washing four times with binding buffer, the bound proteins were eluted by boiling for 5 min in 20 l of SDS sample buffer. Then the samples were separated by SDS-polyacrylamide gel electrophoresis and analyzed by autoradiography.
Measurement of Collagen Synthesis-We directly measured collagen synthesis as described previously (21). Briefly, 90% confluent HASMCs were rendered quiescent in media supplemented with 1% FBS for 24 h and subsequently stimulated with GW7845 (0.1, 0.5, 1 mol/liter) for 1 h before stimulation with TGF-␤1 (4 ng/ml) and sodium ascorbate (10 g/ml) for 48 h. Control cultures were incubated with ascorbate plus vehicle. HASMCs were pulsed for the last 24 h with 15 Ci/ml L-[ 3 H]proline (100 mCi/mmol) in proline-free DMEM. During this period the medium was supplemented with dialyzed 1% FBS, fresh sodium ascorbate (10 g/ml), and ␤-aminoproprionitrile (80 g/ml). The medium was mixed with buffer containing 0.65 mol/liter NaCl, 100 mmol/liter Tris-Cl, pH 7.4, 4.7 mmol/liter CaCl 2 , 1.25 mg/ml N-ethylmaleimide, and 50 g/ml BSA. Samples were split into two equal portions, and highly specific collagenase (Sigma type VII, 10 units/ml) was added to one portion. The samples were incubated for 90 min at 37°C, and undigested proteins were precipitated with 10% trichloroacetic acid at 4°C. The pellets were washed, air-dried, and dissolved in 0.1 N NaOH. Radioactivity incorporated into trichloroacetic acid-precipitable counts was measured by liquid scintillation counting. The rate of collagen synthesis, relative to that of all proline-containing proteins, was calculated assuming that the number of proline residues in collagen is 5.4-fold higher than that in noncollagen proteins.

RESULTS
Activation of PPAR␥ Inhibits TGF-␤-induced CTGF Expression in HASMCs-To understand the role of PPAR␥ in VSMC, we compared the expression of ϳ4100 genes in human aortic smooth muscle cells (HASMCs) in response to PPAR␥ activators using a human known gene filter (Research Genetics, GF-211, Huntsville, AL). Passage 7 HASMCs were grown to 80 -90% confluence, made quiescent by serum starvation for 24 h, and then stimulated with 10 mol/liter ciglitazone, 1 mol/liter GW7845, or 5 mol/liter 15-d-PGJ 2 for 16 h. Connective tissue growth factor (CTGF), which is a key regulator of extracellular matrix production, was down-regulated about Ϫ5.6-, Ϫ6.5-, and Ϫ8.0-fold by ciglitazone, GW7845, and 15-d-PGJ 2 respectively (data not shown). This led us to postulate that CTGF is a PPAR␥ target gene in VSMCs.
The effects of 15-d-PGJ 2 and GW7845 on CTGF protein levels in the culture medium of TGF-␤-stimulated cells were next examined. Consistent with previous work (4), we found that CTGF protein was undetectable in untreated HASMCs by Western blotting analysis, whereas TGF-␤ significantly induced CTGF protein production and secretion (Fig. 1, C and D). Interestingly, both 15-d-PGJ 2 (5 mol/liter) and GW7845 (1 mol/liter) dramatically inhibited TGF-␤-induced CTGF secretion. Taken together, the results showed that the inhibition of CTGF expression by PPAR␥ ligands is more profound at the protein level than at the mRNA level, suggesting that PPAR␥ ligands might be exerting some translational or post-translational effects on CTGF expression. In addition, the inhibition of CTGF expression by PPAR␥ ligands was not due to cell death because neither 15-d-PGJ 2 (10 mol/liter) nor GW7845 (1 mol/liter) was toxic to HASMCs (viability ϭ 100%, data not shown). Taken together, the data indicate that PPAR␥ activation inhibits TGF-␤-induced CTGF production at both the mRNA and protein levels.
Suppression of CTGF Expression Is Mediated by PPAR␥-If the suppression of CTGF expression is mediated by PPAR␥, we would expect that the PPAR␥-specific antagonist would negate this effect. GW9662 from GlaxoSmithKline has been shown to be a PPAR␥-specific antagonist (18,22). To test this hypothesis, HASMCs were pretreated with or without GW9662 at 1 mol/ liter for 30 min prior to the addition of 15-d-PGJ 2 (5 mol/liter) or GW7845 (1 mol/liter) and were subsequently stimulated with TGF-␤ (4 ng/ml) for 6 h. Northern blotting analysis showed that the suppression of TGF-␤-induced CTGF mRNA by GW7845 is abrogated by GW9662 ( Fig. 2A), whereas the effect of 15-d-PGJ 2 on CTGF mRNA expression is partially reversed by GW9662 (Fig. 2B). These data indicate that the effect of GW7845 on TGF-␤-induced CTGF expression is largely mediated by PPAR␥. However, the inhibition of 15-d-PGJ 2 on TGF-␤-induced CTGF expression is only in part through PPAR␥. The data suggest that 15-d-PGJ 2 activates a PPAR␥independent signaling pathway to repress CTGF expression in addition to the activation of PPAR␥.
Cloning of Human CTGF Promoter-To study the molecular mechanisms by which PPAR␥ activation inhibited CTGF expression, we cloned the human CTGF promoter using the human genome data base as described under "Experimental Procedures." As shown in Fig. 3, the TATA box (TATAAAA) is located at nt Ϫ33 to nt Ϫ27 of the human CTGF promoter. Sequence analysis of this CTGF promoter revealed that there are two putative NF-B sites and two putative AP-1 sites. In addition, a putative Smads-binding site (SBE) was reported at nt Ϫ175 to nt Ϫ167 (CAGACGGAG) (6). Surprisingly, we did not find any putative PPAR␥-like elements. Cloning of this ϳ2-kb human CTGF promoter has provided a powerful tool with which to study the mechanism of the regulation of CTGF gene expression.
PPAR␥ Inhibits the Transcriptional Activity of the CTGF Promoter-To understand the transcriptional regulation of CTGF gene expression, we transiently transfected pCTGF-Luc with Smad3, Smad4, Smad3/Smad4, p65, or c-Jun/c-FosB expression plasmids into the HepG2 cell line, which is a well established cell model for testing TGF-␤ signaling. As shown in Fig. 4, CTGF promoter activity was increased about 1.9-, 3.2-, and 1.8-fold by overexpression of Smad3, Smad3/Smad4, or c-Jun/c-FosB, respectively. However, overexpression of Smad4 or p65 did not affect CTGF promoter activity. Taken together, these results indicate that the Smad3, Smad3-Smad4 complex, and AP1 regulate the transcriptional activity of CTGF promoter.
To examine further the mechanisms by which PPAR␥ activation inhibits TGF-␤-induced CTGF expression, we investigated the regulation of the CTGF promoter in HepG2 cells. Since PPAR␥ is not expressed in the HepG2 cell (Fig. 5C), it is a good model to test the role of PPAR␥ in mediating the inhibitory effects of PPAR␥ ligands on TGF-␤-induced CTGF expression. Without transfecting the PPAR␥ expression plasmid, treatment of HepG2 cells with GW7845 (1 mol/liter) had no effect on CTGF promoter activation induced by TGF-␤ (Fig.  5A). When an increasing concentration of PPAR␥ expression plasmid (50, 100, and 150 ng/well) was transfected into these cells, GW7845 strongly inhibited TGF-␤-induced CTGF promoter activation in a PPAR␥ dose-dependent manner. In addition, we transfected the constant PPAR␥ expression plasmid (150 ng/well) to HepG2 cells to test the effect of GW7845 at 0.01, 0.1, and 1 mol/liter on CTGF promoter activation. We also found that PPAR␥ inhibited TGF-␤-induced CTGF promoter activation in a GW7845 dose-dependent manner (data not shown). These results suggest that GW7845 inhibits TGF-␤-induced CTGF expression at the transcriptional level in a PPAR␥-dependent mechanism.
Because activation of the CTGF promoter depended on a combinatorial interaction between Smad3, Smad4, and AP1, we investigated the minimal promoter containing the binding sites only for Smad3 and Smad4 to determine whether they were targets for negative regulation by PPAR␥ ligands. Cotransfection of a luciferase reporter construct containing three tandem repeats of SBE with an increasing concentration of PPAR␥ expression plasmid into HepG2 cells showed that GW7845 also effectively inhibited SBE activity in a PPAR␥-dependent manner (Fig. 5B). These data suggest that PPAR␥ inhibits CTGF promoter activation by interfering with Smad3/Smad4.
Overexpression of Smad3 or Smad3/Smad4 Abrogates the PPAR␥ Inhibition on TGF-␤-induced CTGF Promoter Activation-To examine whether overexpression of Smad3 and/or Smad4 abrogate the PPAR␥ inhibition on TGF-␤-induced CTGF promoter activation, we overexpressed Smad3 and/or Smad4 in HepG2 cells by transient transfection. As shown in Fig. 6A, overexpression of Smad3 or Smad3/Smad4 abrogated the PPAR␥ inhibition on TGF-␤-induced CTGF promoter activation. However, overexpression of Smad4 alone had no effect. Similar results were observed by using the minimal promoter containing SBE (Fig. 6B). These data suggest that PPAR␥ directly antagonizes the effects of Smad3. The antagonistic relationship between Smad3 and PPAR␥ suggests a direct interaction between Smad3 and PPAR␥.
In Vitro Binding of PPAR␥ to Smad3-To determine whether PPAR␥ directly interacts with Smad3, an in vitro binding assay was performed with bacterially generated GST-Smad3 fusion protein and in vitro transcribed/translated PPAR␥. A matrix-bound GST-Smad3, but not GST, retained [ 35 S]methionine-labeled PPAR␥ in the presence or absence of the PPAR␥ ligand (Fig. 7A). The presence of GW7845, a PPAR␥ ligand in the assay mixture, significantly increased this physical interaction. In addition, we demonstrated that both GST and GST-Smad4 did not retain PPAR␥ (Fig. 7B). These data indicate that PPAR␥ physically interacts with Smad3 but not Smad4.

Effect of PPAR␥ Activation on Collagen Synthesis Rate in
HASMCs-We examined the effect of PPAR␥ activation on the rate of collagen synthesis in HASMCs by metabolically labeling cells with L-[ 3 H]proline. As illustrated in Fig. 8, the relative rate of collagen synthesis was increased ϳ2.5-fold by TGF-␤ (4 ng/ml) stimulation. GW7845 (0.1-1 mol/liter) inhibited TGF-␤-induced collagen synthesis in a dose-dependent manner, whereas GW7845 had no effect on the basal level of collagen synthesis in HASMCs. Similar results were observed when using the PPAR␥ natural ligand, 15-d-PGJ2 (data not shown). Taken together, our results suggest that PPAR␥ activation inhibits TGF-␤-induced collagen synthesis that is likely to be mediated by inhibition of CTGF expression.

DISCUSSION
The results of these studies demonstrate that PPAR␥ activation inhibits TGF-␤-induced CTGF expression in HASMCs by directly interfering with the Smad3 signaling pathway. Because CTGF is a key factor in the regulation of extracellular matrix production, this repression of CTGF expression by PPAR␥ activation may be one of the mechanisms through which PPAR␥ agonists inhibit neointimal formation after vascular lesion.
It is postulated that pathological changes in vessel structure are induced in part by transcription factors that govern cell growth, death, differentiation, inflammation, and matrix production. PPARs are a family of ligand-activated nuclear transcriptional factors that are emerging as important determinants of vascular function and structure (10 -12). Recent studies have documented that the expression of PPAR␥ is upregulated in intimal VSMC (23). Moreover, it has been reported that thiazolidinediones (TZD), a class of anti-diabetic drugs that function as synthetic ligands of PPAR␥, inhibit neointima formation after balloon injury in association with decreased DNA synthesis (23,24). In this study, we have extended these observations by documenting that both synthetic and natural FIG. 5. PPAR␥ inhibits the transcriptional activity of CTGF promoter in a dose-dependent manner. HepG2 cells were co-transfected with the PPAR␥ expression plasmid (50, 100, and 150 ng/well), with pCTGF-Luc (A) or with p3xSBE-Luc (B). The GFP reporter plasmid (50 ng/well) was used as the control for transfection efficiency. The pcDNA3 plasmid was used to ensure the equal amount of DNA in each sample. Twenty four hours after transfection, cells were treated with Opti-MEM for 24 h and then stimulated with or without TGF-␤1 (4 ng/ml) and GW7845 (1 mol/liter), as indicated, for 6 h. The luciferase activities, normalized by GFP activity, were expressed in relative units, with no PPAR␥, no TGF-␤, and no GW7845 as 1 (mean Ϯ S.E., n ϭ 6). C, the PPAR␥ protein level in HASMCs or HepG2 cell lines was determined by Western blotting analysis. An equal amount of protein (50 g) was loaded into each lane.
ligands of PPAR␥ promote a decrease in the expression of CTGF as well as collagen III in VSMC. In addition, previous studies (3,4) have well documented that CTGF plays an important role in atherosclerosis and restenosis by mediating the effects of TGF-␤-induced extracellular matrix production and by stimulating collagen and fibronectin production in fibroblasts (25). Taken together, our data suggest a novel antifibrotic mechanism by which PPAR␥ activation may reduce vascular lesion formation via decreasing CTGF expression.
Recently, it was reported (26 -28) that there are PPAR␥independent effects of PPAR␥ ligands (i.e. TZD) at high doses. In addition, the PPAR␥ natural ligand 15-d-PGJ 2 has many functions other than that of a PPAR␥ activator (29). In this study, we documented that the high affinity PPAR␥ ligand, GW7845, had the effect of inhibiting TGF-␤-induced CTGF expression, beside TZD and 15-d-PGJ 2 , in HASMCs. By using the PPAR␥-specific antagonist, GW9662 (18,22), from Glaxo-SmithKline, we demonstrated that the effect of GW7845 on TGF-␤-induced CTGF expression was essentially mediated by PPAR␥. However, the inhibition of 15-d-PGJ 2 on TGF-␤-in-duced CTGF expression was partly due to PPAR␥ activation, suggesting that 15-d-PGJ 2 can activate other PPAR␥-independent signaling pathways to repress CTGF expression. Interestingly, it was recently reported that 15-d-PGJ 2 inhibits cyclooxygenase 2 expression, not only by activation of PPAR␥ but also by directly inhibiting NF-B as well as by modification of IB (29).
Cloning of this ϳ2-kb human CTGF promoter provided a powerful tool with which to study the mechanisms of regulation of CTGF gene expression. Sequence analysis of this CTGF promoter revealed that there were two putative NF-B sites, two putative AP-1 sites, and a putative SBE. It is interesting to document that Smad3 and Smad4 have a synergistic effect on activating the CTGF promoter. This may represent the fact that overexpression of Smad4 can translocate more Smad3 into the nucleus (7). In this study, our data document that activation of the CTGF promoter depends on Smad3/Smad4 and c-Jun/c-Fos but not on NF-B. At present, we cannot completely rule out the possibility of NF-B in the regulation of CTGF expression. To understand better the transcriptional regulation of CTGF, we are currently performing a systematic deletion mapping analysis of this promoter, which is beyond the scope of this study. FIG. 6. Inhibition of PPAR␥ on TGF-␤-induced CTGF promoter was rescued by overexpression of Smad3 or Smad3/Smad4. HepG2 cells were co-transfected with pCTGF-Luc (A) or p3xSBE-Luc (B) with or without PPAR␥ expression plasmid (100 ng/well). The Smad3 (200 ng/ml), Smad4 (200 ng/ml), or Smad3/Smad4 (100 ng/well of each) expression plasmids were also transfected into cells as indicated. pcDNA3 plasmid was used to ensure an equal amount of DNA in each sample. GFP reporter plasmid (50 ng/well) was used as the control for transfection efficiency. Twenty four hours after transfection, cells were made quiescent in serum-deprived medium for 24 h and then treated with the combination of TGF-␤1 (4 ng/ml) and GW7845 (1 mol/liter), as indicated, for 24 h. The luciferase activities, normalized by GFP activity, were expressed in relative units, with no PPAR␥, no TGF-␤, and no GW7845 as 1 (mean Ϯ S.E., n ϭ 6). Smad proteins are the primary TGF-␤ receptor substrates capable of signal transduction. The Smad complex translocates into the nucleus and binds to cofactors that determine the choice of target genes. Smad proteins consist of two conserved globular domains known as the MH1 (Mad homology 1) and the MH2 domains, coupled by a linker region. The MH1 domain recognizes the Smad-binding site in the target gene promoter, whereas the MH2 domain binds to transcriptional coactivators such as p300 and CREB-binding protein in competition with the corepressors such as Ski and SnoN (7). To test our hypothesis that PPAR␥ directly interacts with Smad proteins, we performed a series of transfection/reporter experiments. Overexpression of Smad3 or Smad3/Smad4 but not Smad4 completely abrogates the PPAR␥ inhibition on TGF-␤-induced CTGF promoter activation. In addition, using an in vitro binding assay, we demonstrate that PPAR␥ physically interacts with Smad3 but not Smad4. Taken together, our data suggest that PPAR␥ can inhibit TGF-␤-induced CTGF expression in HASMCs by directly interacting with Smad3.
Although we have documented that PPAR␥ activation inhibits TGF-␤-induced activation of CTGF through a Smad3-dependent mechanism, it will be interesting to know whether TGF-␤ directly regulates PPAR␥ gene expression in VSMCs. Recently, it was reported (30) that TGF-␤ could decrease the expression of CD36, the macrophage type B receptor that is a well characterized PPAR␥ target gene, by phosphorylation of PPAR␥ through the mitogen-activated protein kinase. Although we have not detected CD36 in HASMCs, using both reverse transcriptase-polymerase chain reaction and Western blotting analyses (data not shown), this provocative paper stimulated us to examine the effect of TGF-␤ on PPAR␥ gene expression. We found that TGF-␤ inhibits PPAR␥ gene expression in a dose-dependent manner in VSMCs. 2 Taken together, our data suggest that the inhibition of CD36 expression by TGF-␤ stimulation may be mediated by PPAR␥ at multiple levels including phosphorylation of PPAR␥, repression of PPAR␥ gene expression, and the interaction of PPAR␥ and Smad3. Studies are underway to determine the mechanisms of TGF-␤-induced PPAR␥ inhibition.
In conclusion, we show for the first time that CTGF, which is a key regulator of extracellular matrix production and plays an important role in atherosclerosis and restenosis, is a PPAR␥regulated gene. We have documented that activation of PPAR␥ abrogates TGF-␤-induced CTGF expression by directly interfering with the Smad3 signaling pathway. Taken together, our data suggest that abrogation of TGF-␤-induced CTGF production by PPAR␥ activation may be one of the mechanisms through which PPAR␥ agonists inhibit neointimal formation after vascular lesion.