Transcription Suppression of Thromboxane Receptor Gene by Peroxisome Proliferator-activated Receptor-γ via an Interaction with Sp1 in Vascular Smooth Muscle Cells*

Thromboxane (TX) A2exerts contraction and proliferation of vascular smooth muscle cells (VSMCs) via its specific membrane TX receptor (TXR), possibly leading to the progression of atherosclerosis. A nuclear hormone receptor, peroxisome proliferator-activated receptor (PPAR)-γ, has recently been reported to be expressed in VSMCs. Here we examined a role of PPAR-γ in TXR gene expression in VSMCs. PPAR-γ ligands 15-deoxy-Δ12,14-prostaglandin J2 and troglitazone reduced TXR mRNA expression levels as well as cell growth as assessed by [3H]thymidine incorporation. Transcriptional activity of the TXR gene promoter was suppressed with PPAR-γ ligands, and the suppression was augmented further by PPAR-γ overexpression. By deletion and mutation analyses, the transcription suppression was shown to be the result of a −22/−7 GC box-related sequence (upstream of transcription start site). Electrophoretic mobility shift assays also showed that the sequence was bound by Sp1 but not by PPAR-γ, and the formation of a Sp1·DNA complex was inhibited either by coincubation with PPAR-γ or PPAR-γ ligand treatment of VSMCs. Moreover, glutathione S-transferase pull-down assays demonstrated a direct interaction between PPAR-γ and Sp1. In conclusion, PPAR-γ suppresses TXR gene transcription via an interaction with Sp1. PPAR-γ may possibly have an antiatherosclerotic action by inhibiting TXR gene expression in VSMCs.

Thromboxane (TX) 1 A 2 is a labile metabolite of arachidonic acid synthesized by TX synthase (1). TX is known to exert many biological effects such as platelet aggregation, contraction and growth of vascular smooth muscle cells (VSMCs) (2,3), and renal electrolyte metabolism (4). TX is also known to play a pathophysiological role in the inflammatory diseases such as atherosclerosis (5) and glomerulonephritis (6). The biological action of TX is mediated via its specific membrane TX receptor (TXR). We previously isolated a cDNA for rat TXR (7), localized it in either the kidney (8) or testis (9), and identified its chromosomal localization (10). Moreover, we have isolated 5Ј-flanking region (FL) of the rat TXR gene and studied its transcription regulation in VSMCs (11).
Peroxisome proliferator-activated receptor (PPAR)-␥ is a nuclear hormone receptor that was shown to transactivate adipocyte-specific genes and induce adipocyte differentiation (12). Either insulin-sensitizing thiazolidinediones including troglitazone (TRO) or 15-deoxy-⌬ 12,14 -prostaglandin J 2 (PGJ 2 ) (13,14) has been identified as a ligand of PPAR-␥. Recently, PPAR-␥ has been shown to be present not only in adipocytes but also in vascular tissues including VSMCs (15), and an inhibitory effect of PPAR-␥ on gene expression in atherosclerosis has been studied. Activation of PPAR-␥ with its ligands suppressed expression of plasminogen activator inhibitor type 1 (16) in vascular endothelial cells and that of matrix metalloproteinase-9 in VSMCs (15). Moreover, we have observed that PPAR-␥ can suppress TX synthase gene transcription in macrophages (17).
In the present study, we examined the role of PPAR-␥ in TXR gene expression in VSMCs. We observed that PPAR-␥ inhibited the TX-mediated cell growth of VSMCs and TXR mRNA expression. Suppression of TXR gene transcription was confirmed, and the suppression was shown to be dependent on a GC box-related sequence present at the Ϫ22/Ϫ7 region of TXR gene promoter (upstream of transcription start site), which was bound by Sp1 but not by PPAR-␥. PPAR-␥ was shown to interact physically with Sp1 by glutathione S-transferase (GST) pull-down assays. Taken together, PPAR-␥ was suggested to suppress TXR gene transcription via a protein-protein interaction with Sp1. An antiatherosclerotic action of PPAR-␥ by inhibiting TXR gene expression in VSMCs may be suggested.
Cell Culture-Rat VSMCs that were isolated from male Sprague-Dawley rat thoracic aortas were gifts from Dr. K. K. Griendling (Emory University, Atlanta, GA) and maintained as described previously (22). Passages between 7 and 15 were used for the following experiments.
[ 3 H]Thymidine Incorporation-When rat VSMCs grown in 3.5-cm plates became 70% confluent, media were changed to serum-free minimum Eagle's medium and incubated for 2 days. Then the cells were incubated in the absence or presence of 1 M U-46619 (2) (a stable TXA 2 mimetic) (Cayman Chemical), 0.5 or 1 M PGJ 2 (Cayman Chemical), and 1 or 10 M TRO (kindly provided by Sankyo Company, Ltd., Tokyo) for 12 h with 1 Ci/ml [ 3 H]thymidine (Amersham Biosciences, Inc.) in serum-free minimum Eagle's medium. The cells were washed twice with ice-cold phosphate-buffered saline, incubated with 500 l 5% trichloroacetic acid for 30 min at 4°C, washed twice with 1 ml 5% trichloroacetic acid, and incubated with 400 l of 1 N NaOH for 20 min at room temperature. After neutralization with 400 l of 1 N HCl, the cell extracts were put into vials with 5 ml of scintillation solution and counted in a ␤-counter.
Northern Blot Analysis-When rat VSMCs became 70% confluent, media were changed to Dulbecco's modified Eagle's medium with 1% resin and charcoal-treated calf serum (stripped medium) (23) and incubated for 5-6 h. The cells then were incubated either with or without 2.5 M PGJ 2 or 50 M TRO for an additional 12 h. Their total RNAs were then extracted using an RNeasy mini kit (Qiagen). 10 g of isolated total RNA was subjected to electrophoresis in 1% agaroseformaldehyde gels, transferred to nylon membrane (Hybond-N, Amersham Biosciences, Inc., and the blot was hybridized with 32 P-labeled TXR cDNA probe as described previously (7). Intensity of the blot was calculated using Luminous Imager (AI-C), and all values were normalized to the densities of ethidium bromide staining of 28 S ribosomal RNA (rRNA).
Luciferase Reporter Gene Assay-When rat VSMCs became 70% confluent, media were changed to stripped medium and were incubated for 5-6 h. Then the transfection using Lipofectin was performed according to the manufacturer's instructions (Invitrogen). Briefly, 1.2 g of reporter plasmid and 0.8 g of ␤-galactosidase control plasmid in pCMV (Clontech) were mixed with 6 l of Lipofectin/3.5-cm plate. In some experiments, 1 g of PPAR-␥1, RXR-␣, and/or Sp1 expression plasmids were cotransfected. 12 h after transfection, media were changed to stripped medium, and the cells were incubated for an additional 12 h. The cells were then incubated either with or without several concentrations of PGJ 2 or TRO for 12 h. In some experiments, the cells were incubated with 1 M leukotriene (LT) B 4 (24) (Cayman Chemical) for 12 h. After harvesting, the cell extracts were analyzed for both luciferase and ␤-galactosidase activities (23). Transfection efficiency was normalized by the ␤-galactosidase expression.
GST Pull-down Assay-Full-length GST-PPAR-␥ fusion protein was synthesized from pGEX-PPAR-␥, and full-length GST-Sp1 fusion protein was synthesized from pGEX-Sp1, using the GST Gene Fusion system (Amersham Biosciences, Inc.). The proteins were loaded onto glutathione-Sepharose beads, which were washed and resuspended in binding buffer (20 mM HEPES (pH 7.7), 75 mM KCl, 0.1 mM EDTA, 2.5 mM MgCl 2 , 0.05% Nonidet P-40, 2 mM dithiothreitol, and 10% glycerol) in the presence or absence of 10 or 50 M TRO. The beads were incubated with 5 l of in vitro translated 35 S-labeled Sp1 or 35 S-labeled PPAR-␥ for 3 h at 4°C in the presence or absence of 10 or 50 M TRO, followed by washing six times with binding buffer in the presence or absence of 10 or 50 M TRO. They were then resuspended in 30 l of SDS sample buffer and analyzed by SDS-PAGE.
Statistical Analysis-Statistical significance was calculated by onefactor analysis of variance using StatView 4.0 (ABACUS Concepts).

Effect of PPAR-␥ Ligands on TX-induced [ 3 H]Thymidine
Incorporation-We first examined the effect of PPAR-␥ ligands PGJ 2 and TRO on DNA synthesis in VSMCs. As shown in Fig into VSMCs was induced by 1 M TX mimetic U-46619 (lines 1 and 2) as reported previously (2). When VSMCs were coincubated with either 0.5 or 1 M PGJ 2 or 1 or 10 M TRO, the U-46619-stimulated increase in [ 3 H]thymidine incorporation was inhibited in a dose-dependent manner (lines 3 and 4 for PGJ 2 and lines 5 and 6 for TRO). The results indicate that both PPAR-␥ ligands can significantly inhibit the TX-stimulated DNA synthesis in VSMCs.
Effect of PPAR-␥ Ligands on TXR mRNA Expression-We next performed Northern blot analysis to examine TXR mRNA expression regulation in VSMCs. TXR mRNA expression (3.7 kb, indicated by an arrow) was fully observed in VSMCs in the absence of PPAR-␥ ligands (Fig. 2, lane 1). TXR mRNA expression levels were decreased significantly by PPAR-␥ ligands such as PGJ 2 (2.5 M) and TRO (50 M) (Fig. 2, lanes 2 and 3,  respectively). The results indicate that PPAR-␥ ligands negatively regulate TXR mRNA expression, which may cause inhibition of TX-stimulated DNA synthesis, in VSMCs.
Effect of PPAR-␥ Ligands on TXR Gene Promoter Activity-We then examined the effect of PPAR-␥ ligands on transcription of the TXR gene. As shown in Fig. 3, PPAR-␥ ligand PGJ 2 (lines 2-4) and TRO (lines 5-7) significantly decreased transcription of Ϫ989/ϩ184-luc in a dose-dependent manner. In contrast, PPAR-␣ ligand LTB 4 did not affect the transcription (line 8). These results suggest that PPAR-␥ ligands specifically suppress TXR gene transcription.
Involvement of PPAR-␥ in TXR Gene Suppression Independent of RXR-␣-PPAR-␥ overexpression in VSMCs was performed. As shown in Fig. 4, PPAR-␥ overexpression decreased transcription of Ϫ989/ϩ184-luc significantly in the absence (lines 1 and 2) or the presence of PPAR-␥ ligands (lines 5 and 6 for PGJ 2 and lines 9 and 10 for TRO), whereas RXR-␣ overexpression did not affect it (lines 3 and 7). Moreover, overexpression of both factors did not affect the transcription suppression caused by that of PPAR-␥ alone (lines 4 and 8). It is indicated that RXR-␣ is not involved in the TXR gene suppression.
Sp1 Binding to the Ϫ22/Ϫ7 GC Box-related sequence-We next examined protein-DNA interactions on the Ϫ47/Ϫ7 region of the TXR gene promoter. When 32 P-labeled Ϫ47/Ϫ7 region oligonucleotides were incubated with in vitro translated PPAR-␥, formation of a protein⅐DNA complex was not observed (Fig. 6B, lane 3). Moreover, even by coincubation with in vitro translated PPAR-␥ and RXR-␣, no formation of a protein⅐DNA complex was observed (data not shown). Generation of PPAR-␥1 protein at the expected molecular size (52-kDa) (28) was confirmed by SDS-PAGE of 35 S-labeled PPAR-␥1 (Fig. 6A,  lane 2). It is therefore suggested that PPAR-␥ does not bind to the Ϫ47/Ϫ7 region either as a monomer/homodimer or as a PPAR-␥⅐RXR-␣ heterodimer. On the other hand, incubation with in vitro translated Sp1 and 32 P-labeled Ϫ47/Ϫ7 induced a Sp1⅐DNA complex (described Sp1 in Fig. 6B, lane 2, indicated by an arrow). The complex formation was attenuated with 32 P-labeled ⌬(Ϫ22/Ϫ7)/(Ϫ47/Ϫ7) oligonucleotides (containing a mutation at the Ϫ22/Ϫ7 GC box-related sequence in the Ϫ47/ Ϫ7) (Fig. 6B, lane 5). In contrast, formation of the complex was not affected with 32 P-labeled ⌬(Ϫ39/Ϫ29)/(Ϫ47/Ϫ7) oligonucleotides (containing a mutation at the Ϫ39/Ϫ29 GC box-related sequence in the Ϫ47/Ϫ7) (Fig. 6B, lane 4). Generation of Sp1 protein at the expected molecular size (95 kDa) (29) was confirmed by SDS-PAGE of 35 S-labeled Sp1 (Fig. 6A, lane 3). Incubation with anti-Sp1 antibody completely abolished formation of the Sp1⅐DNA complex (described as Sp1 in Fig. 6C, lane 1, indicated by an arrow) and newly induced complexes (indicated as Supershifted Sp1 in Fig. 6C, lane 2). These data suggest that Sp1 binds to the Ϫ22/Ϫ7 GC box-related sequence of the TXR gene promoter.
Binding of VSMCs Endogenous Sp1 to the Ϫ22/Ϫ7 GC Boxrelated Sequence-To identify a possible factor that binds to the Ϫ22/Ϫ7 GC box-related sequence, we performed EMSA using nuclear extracts from VSMCs. When 32 P-labeled Ϫ47/Ϫ7 region oligonucleotides were incubated with nuclear extracts, two protein⅐DNA complexes were observed (indicated with arrows, Fig. 6D, lane 2). Coincubation with a 100-fold excess of the unlabeled Ϫ47/Ϫ7 region oligonucleotides completely abolished formation of both complexes (Fig. 6D, lane 3), suggesting that both complexes were specific to the Ϫ47/Ϫ7. Coincubation with a 100-fold excess of unlabeled oligonucleotides containing the hydroxymethylglutaryl-CoA synthase gene PPRE (27) did not affect the complex formation (Fig. 6D, lane 5). On the other hand, coincubation with a 100-fold excess of the unlabeled oligonucleotides containing the SV40 early promoter Sp1 site (26) abolished the complex formation (Fig. 6D, lane 4), suggesting that Sp1 family proteins are involved in the complex formation. To identify the Sp1 family, we then performed antibody supershift experiments. Incubation with anti-Sp1 antibody abolished formation of most of both complexes and newly induced a complex (indicated as Supershifted Sp1 with an arrow ,  Fig. 6E, lane 2). Incubation with anti-Sp3 antibody did not affect both complexes, and supershifted bands were hardly observed (Fig. 6E, lane 3). Moreover, incubation with anti-Sp2 and anti-Sp4 antibodies also did not affect both complexes, and no new bands were observed (data not shown). These data suggest that most of the DNA-binding protein is composed of Sp1. We further tried to identify the sequence responsible for the Sp1 binding in the Ϫ47/Ϫ7. With the 32 P-labeled ⌬(Ϫ22/ Ϫ7)/(Ϫ47/Ϫ7) oligonuceotides, formation of Sp1⅐DNA complexes was inhibited (Fig. 6D, lane 7). In contrast, formation of both complexes was not affected with ⌬(Ϫ39/Ϫ29)/(Ϫ47/Ϫ7) oligonucleotides (Fig. 6D, lane 6). It is thus suggested that Sp1 binds to the Ϫ22/Ϫ7 GC box-related sequence in VSMCs.
Inhibition of Sp1 Binding to DNA by PPAR-␥ and Its Ligands-We next examined the effect of PPAR-␥ on Sp1 binding to DNA. By incubation with increasing amounts of in vitro translated PPAR-␥, the Sp1⅐DNA complexes were decreased comparably (Fig. 6D, lanes 8 -10). Moreover, when we treated rat VSMCs with either 2.5 M PGJ 2 or 50 M TRO for 12 h before preparation of nuclear extracts, formation of the Sp1⅐DNA complexes decreased significantly (Fig. 6F, lanes  1-3). These data suggest that PPAR-␥ and its ligands inhibit formation of the Sp1⅐DNA complex on the Ϫ22/Ϫ7 region.
Physical Interaction between PPAR-␥ and Sp1-A GST pulldown assay was performed to study a possible protein-protein interaction between PPAR-␥ and Sp1. As shown in Fig. 8A, no significant band was identified by incubation with GST alone with in vitro translated 35  These data suggest that PPAR-␥ physically interacts with Sp1, and the interaction is enhanced by PPAR-␥ ligand in a dosedependent manner. DISCUSSION TX-stimulated DNA synthesis in VSMCs was inhibited by PGJ 2 and an insulin-sensitizing drug TRO, both of which have been shown to be ligands for PPAR-␥ (13, 14). Because TXR mRNA expression was also decreased by these PPAR-␥ ligands, the inhibition of TX-stimulated DNA synthesis by PPAR-␥ ligands was suggested to be caused at least in part by the repression of TXR expression. Moreover, the TXR expression suppression was dependent on a gene transcription mechanism involving PPAR-␥ because overexpression of PPAR-␥ suppressed transcription of the TXR gene, which was augmented by the addition of PPAR-␥ ligands. The TXR gene transcription suppression appeared to be PPAR-␥ ligand-specific because PPAR-␣ ligand LTB 4 (24) showed little effect. It is suggested that PPAR-␥ specifically suppresses the TXR transcription possibly leading to a decrease in TXR expression as well as its function in VSMCs.
PPAR-␥ has been shown to heterodimerize with RXR-␣, bind to PPRE, and induce ligand-dependent transactivation (27). Deletion mutant analyses have demonstrated that the region between Ϫ47 and Ϫ7, most likely the Ϫ22/Ϫ7 GC box-related sequence, is responsible for the transcription suppression by PPAR-␥. However, we did not observe any PPAR-␥ binding to the Ϫ47/Ϫ7 region by EMSA either as a monomer/homodimer or as a PPAR-␥⅐RXR-␣ heterodimer. Additionally, no PPRE consensus sequence was identified within the Ϫ47/Ϫ7 region. The Ϫ22/Ϫ7 GC box-related sequence was actually bound by Sp1, and the binding was decreased either by incubation with PPAR-␥ or PPAR-␥ ligand treatment. It was thus suggested that PPAR-␥ interacts with Sp1 causing the transcription suppression. PPARs have been reported to regulate transcription by interacting with other transcription factors or cofactors. For instance, PPAR-␣ was shown previously to activate prolactin gene transcription via an interaction with growth hormone factor (GHF)-1 or coactivators such as cAMP response elementbinding protein-binding protein (CBP) and SRC-1 (30). It also suppressed interleukin-6 gene transcription via an interaction with nuclear factor-B or AP-1 (31). Regarding PPAR-␥, it suppressed interleukin-12 gene transcription via an interaction with nuclear factor-B (32) and inducible nitric oxide synthase gene transcription via a direct interaction with CBP (33). Moreover, we have observed recently that PPAR-␥ suppresses TX synthase gene transcription via a direct interaction with nuclear factor E2-related factor 2 (17). On the other hand, Sp1 was shown to interact physically with other nuclear hormone receptors such as retinoic acid receptor (34), estrogen receptor (35), and progesterone receptor (36), but all of these interactions enhanced their Sp1-induced gene transcriptions. In the present study, we have observed a direct physical interaction between PPAR-␥ and Sp1 by GST-pull down assays. The interaction is specific because it has been augmented in the presence of PPAR-␥ ligand. We thus propose a mechanism by which PPAR-␥ inhibits the action of Sp1 on the GC box-related sequence. This is the first report describing an interaction between PPAR-␥ and Sp1 causing the transcription inhibition by PPAR-␥.
The effects of PPAR-␥ and its ligands on the cardiovascular system have been studied. PPAR-␥ ligands suppress expression of plasminogen activator inhibitor type 1 (16) and inhibit neointimal formation after balloon injury (37) in vascular endothelial cells. In VSMCs, expression of matrix metalloproteinase-9 was suppressed by PPAR-␥ ligands (15). We have reported recently the suppression of type-1 angiotensin II receptor gene transcription by PPAR-␥ in VSMCs (38) and that of TX synthase gene transcription by PPAR-␥ in macrophages (17). Moreover, in terms of atherogenesis, PPAR-␥ inhibited macrophage activation (39,40) and stimulated cholesterol efflux from macrophages by inducing the ABCA1 gene (41,42). Furthermore, in clinical studies, TRO was reported to lower blood pressure (43), ameliorate microalbuminuria (44), and prevent cardiac mass increase and cardiac function impairment (45) in diabetic patients. An inhibitory effect of TRO on carotid arterial wall thickening in diabetic patients was also reported (46). It thus appears that PPAR-␥ has an antiatherosclerotic effect.
An increase in TX production is widely reported in patients with diabetes mellitus (DM) (47,48) as well as in animal DM models (49,50), which may induce the hypercoagulable state leading to cardiovascular complications including atherosclerosis. Moreover, TX-induced aortic contraction is augmented in diabetic rats (51). Therefore, both synthesis and action of TX are suggested to be up-regulated in DM. We have shown previously that PPAR-␥ and its ligands also suppressed expression of the TX synthase gene (17), which is supported by the observation that TRO reduced TX production in human platelets and erythroleukemia cells (52). PPAR-␥ and its ligands may thus attenuate both synthesis and action of TX, exerting a beneficial effect on cardiovascular complications in DM patients.
In conclusion, PPAR-␥ suppresses TXR gene expression at the transcription level via an interaction with Sp1 and inhibits TX-induced cell proliferation in VSMCs. PPAR-␥ can inhibit TX system possibly causing antiatherosclerotic and anticoagulative effects on the vasculature.