Peroxisome Proliferator-activated Receptor γ Ligands Inhibit Retinoblastoma Phosphorylation and G1 → S Transition in Vascular Smooth Muscle Cells*

Peroxisome proliferator-activated receptor γ (PPARγ) is a member of the nuclear receptor superfamily that is activated by binding certain fatty acids, eicosanoids, and insulin-sensitizing thiazolidinediones (TZD). The TZD troglitazone (TRO) inhibits vascular smooth muscle cell proliferation and migration both in vitro and in vivo. The precise mechanism of its antiproliferative activity, however, has not been elucidated. We report here that PPARγ ligands inhibit rat aortic vascular smooth muscle cell proliferation by blocking the events critical for G1 → S progression. Flow cytometry demonstrated that both TRO and another TZD, rosiglitazone, prevented G1 → S progression induced by platelet-derived growth factor and insulin. Movement of cells from G1 → S was also inhibited by the non-TZD, natural PPARγ ligand 15-deoxy-12,14Δ prostaglandin J2(15d-PGJ2), and the mitogen-activated protein kinase pathway inhibitor PD98059. Inhibition of G1 → S exit by these compounds was accompanied by a substantial blockade of retinoblastoma protein phosphorylation. TRO and rosiglitazone attenuated both the mitogen-induced degradation of p27 kip1 and the mitogenic induction of p21 cip1 . 15d-PGJ2 and PD98059 inhibited both the degradation of p27 kip1 and the induction of cyclin D1 in response to mitogens. These effects resulted in the inhibition of mitogenic stimulation of cyclin-dependent kinases activated by cyclins D1 and E. These data demonstrate that PPARγ ligands are antiproliferative drugs that act by modulating cyclin-dependent kinase inhibitors; they may provide a new therapeutic approach for proliferative vascular diseases.


Peroxisome proliferator-activated receptor ␥ (PPAR␥) is a member of the nuclear receptor superfamily that is activated by binding certain fatty acids, eicosanoids, and insulin-sensitizing thiazolidinediones (TZD). The TZD troglitazone (TRO) inhibits vascular smooth muscle cell proliferation and migration both in vitro and in vivo.
The precise mechanism of its antiproliferative activity, however, has not been elucidated. We report here that PPAR␥ ligands inhibit rat aortic vascular smooth muscle cell proliferation by blocking the events critical for G 1 3 S progression. Flow cytometry demonstrated that both TRO and another TZD, rosiglitazone, prevented G 1 3 S progression induced by platelet-derived growth factor and insulin. Movement of cells from G 1 3 S was also inhibited by the non-TZD, natural PPAR␥ ligand 15-deoxy-12,14 ⌬ prostaglandin J 2 (15d-PGJ 2 ), and the mitogen-activated protein kinase pathway inhibitor PD98059. Inhibition of G 1 3 S exit by these compounds was accompanied by a substantial blockade of retinoblastoma protein phosphorylation. TRO and rosiglitazone attenuated both the mitogen-induced degradation of p27 kip1 and the mitogenic induction of p21 cip1 . 15d-PGJ 2 and PD98059 inhibited both the degradation of p27 kip1 and the induction of cyclin D1 in response to mitogens. These effects resulted in the inhibition of mitogenic stimulation of cyclin-dependent kinases activated by cyclins D1 and E. These data demonstrate that PPAR␥ ligands are antiproliferative drugs that act by modulating cyclin-dependent kinase inhibitors; they may provide a new therapeutic approach for proliferative vascular diseases.

Proliferation of vascular smooth muscle cells (VSMC) 1 plays
a key role in the development of restenosis and in the progression of atherosclerosis (1,2). Injury to the endothelium results in the migration of underlying VSMC into the intimal layer of the arterial wall, where they proliferate and synthesize extracellular matrix components. Although many growth factors induce the proliferation of VSMC (3), platelet-derived growth factor (PDGF) is an important mitogenic and chemotactic regulator of VSMC, since blocking antibodies to PDGF inhibits neointimal formation in rat models of arterial injury (4). Insulin is also a potent mitogen for VSMC, and physiologic hyperinsulinemia enhances the mitogenic effects of PDGF (5). In response to vascular injury, quiescent VSMC (G 0 ) must transit through the G 1 phase of the cell cycle and enter into the S phase to undergo DNA replication.
Progression through the cell cycle requires the formation and activation of cyclin and cyclin-dependent kinase (CDK) complexes (6). Progression through the G 1 phase requires cyclin D/CDK4, cyclin D/CDK6, and cyclin E/CDK2 holoenzymes. Functional cyclin A/CDK2 complexes are required for DNA synthesis (S phase), and subsequently, cyclin A/CDC2 and cyclin B/CDC2 pairs are assembled and activated during G 2 phase and mitosis (M phase), respectively. Activation of the G 1 phase cyclin⅐CDK complexes results in the phosphorylation of retinoblastoma gene products (Rb) (7). Rb proteins are critical negative regulators of cell cycle progression by controlling gene expression mediated by E2F transcription factors (7). E2Fregulated genes encode proteins required for S phase DNA synthesis (8). In the absence of its being phosphorylated by CDKs, Rb binds and sequesters E2F, thereby preventing transcriptional activation of target genes. CDK-phosphorylated Rb releases E2F that permits the induction of E2F-dependent genes. CDK inhibitors (CDKIs), p21 cip1 (9, 10), p27 kip1 (11)(12)(13), and p15/p16 ink4 (14), regulate this process by inhibiting cyclin/ CDK activity and phosphorylation of Rb, resulting in G 1 arrest (15). Progression through the mammalian cell cycle is regulated by the balance between the levels and activities of cyclin⅐CDK complexes, the growth-promoting transcriptional factors they regulate, CDKIs and other growth suppressor proteins.
Thiazolidinediones (TZDs) are high-affinity ligands for peroxisome proliferator-activated receptor ␥ (PPAR␥). PPAR␥ is expressed in VSMC (16,17), and ligands for this nuclear receptor inhibit VSMC proliferation induced by several different FIG. 1. PPAR␥ ligands and PD98059 prevent mitogen-induced G 1 3 S progression in RASMC. A, quiescent RASMC (0.4% FBS for 24 h) were stimulated by treatment with PDGF (20 ng/ml) and insulin (1 M). Cells were preincubated with troglitazone (10 M), rosiglitazone (10 M), 15d-PGJ 2 (5 M), and PD98059 (30 M) for 30 min prior to addition of mitogens. 24 h after stimulation, DNA was stained with propidium iodide (PϩI), and 1 ϫ 10 6 cells were analyzed by flow cytometry. A shows representative DNA histograms for quiescent RASMC (0.4% FBS for 24 h) (a), RASMC stimulated with PDGF and insulin (P ϩ I) (b), RASMC stimulated in the presence of TRO (c), RSG (d), 15d-PGJ 2 (e), and PD98059 (f), respectively. The x and y axes represent the intensity of propidium iodide fluorescence and cell number, respectively. The data are representative of three separate experiments. B, PPAR␥ ligands and PD98059 inhibit mitogen-induced G 1 3 S progression in a dose-dependent manner. Results are the mean of three independent experiments. Mean Ϯ S.E. is expressed as percentage of S phase transition. mitogens in vitro (18,19) and intimal hyperplasia in vivo (18). These observations suggest that activation of PPAR␥ interferes with the function of a fundamental component of the cell cycle machinery. The specific mechanism by which PPAR␥ inhibits VSMC proliferation, however, remains to be determined. We have previously shown that PPAR␥ ligands inhibit ERK MAPKdependent mitogenic signaling pathways in VSMC at a step downstream of ERK activation (18). Induction of cyclin D and G 1 3 S progression of nonvascular cells has also been shown to require activation of ERK and MAPK. The purpose of this study was to examine the effect of PPAR␥ ligands on cell cycle regulators in rat aortic smooth muscle cells and to delineate the mechanism of their antiproliferative activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Treatment with Growth Factor and Reagents-Rat aortic smooth muscle cells (RASMC) were prepared from thoracic aorta of 2-3-month-old Harlan Sprague-Dawley rats by using the explant technique. The cells were cultured in Dulbecco's modified Eagle's medium containing 10% FBS (Irvine Scientific, Santa Ana, CA), 100 units/ml penicillin, 100 g/ml streptomycin, and 200 mM L-glutamine. Human coronary artery smooth muscle cells (CASMC, purchased from Clonetics, San Diego, CA) were 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 (all purchased from Clonetics). For all experiments, early passaged (5-8) RASMC or CASMC were grown to 60 -70% confluency and made quiescent by serum starvation (0.4% FBS) for at least 24 h. Each reagent examined was added 30 min before the addition of human recombinant PDGF-BB (Sigma) and insulin (Lilly) at the final concentration of 20 ng/ml and 1 M, respectively. For all data shown, each individual experiment was performed using an independent preparation of RASMC or CASMC. Troglitazone (TRO) was kindly provided by Parke-Davis; rosiglitazone (RSG, formerly BRL 49653) was a generous gift from Smith Kline Beecham Laboratories. 15-Deoxy-12,14 ⌬ prostaglandin J 2 (15d-PGJ 2 ) was obtained from Biomol (Plymouth Meeting, PA). The MEK inhibitor PD98059 was purchased from New England Biolabs (Beverly, MA).
Cell Cycle Distribution-Flow cytometry was performed to analyze cell cycle distribution. Quiescent RASMC were pretreated for 30 min with each compound or vehicle (Me 2 SO), followed by the addition of growth factors (PDGF-BB 20 ng/ml ϩ insulin 1 M). After 24 h cells were trypsinized, centrifuged at 1500 rpm for 3 min, washed with PBS, and then treated with 20 g/ml RNase A (Calbiochem). DNA was stained with 100 g/ml propidium iodide for 30 min at 4°C and protected from light, and 1 ϫ 10 6 cells were then analyzed with a FACScan (Becton Dickinson). DNA histogram analysis was performed using the ModFitLT software (Becton Dickinson). Experiments were repeated at least 3 times.
Immunocomplex Kinase Assay-Cyclin D1⅐CDK complex activity and cyclin E-CDK activity were measured as described previously (20). Briefly, after appropriate treatments, cells were washed with cold PBS and solubilized on ice in lysis buffer (50 mM Tris, pH 8.0; 250 mM NaCl; 0.5% Nonidet P-40; 1 g/ml leupeptin; and 1 mM phenylmethylsulfonyl fluoride). Insoluble materials were cleared through centrifugation at 4°C for 10 min at 12,000 rpm. Protein concentrations were determined, and protein was suspended in 1 ml of lysis buffer and immunoprecipitated by incubating with agarose-conjugated anti-cyclin D1 (sc-450AC, Santa Cruz Biotechnology) or cyclin E rabbit IgG (sc-481AC, Santa Cruz Biotechnology) overnight. Immunoprecipitants were washed three times with kinase buffer (150 mM NaCl; 1 mM EDTA; 50 mM Tris-HCl, pH 7.5; and 0.1% Tween 20). CDK activities present in the immunoprecipitants were determined by resuspension in kinase buffer (50 mM Tris-HCl, pH 7.5; 10 mM MgCl 2 ; 10 mM dithiothreitol; 1 mM ATP; and 1 mM EGTA). Resuspended complexes were incubated for 15 min at 37°C with 0.5 g of soluble Rb (sc-4112, Santa Cruz Biotechnology) or histone-H1 (Upstate Biotechnology, Inc.) and with 3 Ci of [␥-32 P]ATP. Samples were analyzed by SDS-polyacrylamide gel electrophoresis, and the dried gel was exposed on film with an intensifying screen at Ϫ80°C overnight and quantitated by densitometry.
Statistics-Analysis of variance with paired or unpaired t tests was performed for statistical analysis, as appropriate. Values of p Ͻ 0.05 were considered to be statistically significant. Data are expressed as mean Ϯ S.E.

RESULTS
PPAR␥ Ligands and PD98059 Block the Progression of VSMC into S Phase-TZD PPAR␥ ligands TRO and RSG, a non-TZD PPAR␥ ligand 15d-PGJ 2 , and a MEK inhibitor PD98059 all inhibited cell cycle progression, as determined by flow cytometry. Subconfluent RASMC accumulated in G 1 after serum starvation for 24 h (70.74% in G 0 /G 1 phase and 16.14% in S phase; Fig. 1A). Quiescent RASMC were induced to enter S phase by stimulation with the competence factor PDGF (20 ng/ml) and a progression factor insulin (1 M). The population of G 0 /G 1 cells decreased substantially (43.91%; Fig. 1b) with a concomitant increase in RASMC in S phase (48.45%; Fig. 1b). TRO and RSG inhibited G 1 3 S progression as reflected by the higher percentage of G 0 /G 1 cells (65.17% in Fig. 1c and 62.14% in Fig. 1d, respectively) and by the lower percentage of S phase cells (19.59% in Fig. 1c and 21.68% in Fig. 1d, respectively). Movement of cells from G 1 3 S was also inhibited by 15d-PGJ 2 and the MAPK pathway inhibitor PD98059, with an increase in the population of G 0 /G 1 cells (66.29% in Fig. 1e and 66.34% in Fig. 1f, respectively) and with a concomitant decrease in S phase cells (19.25% in Fig. 1e and 20.66% in Fig. 1f, respectively). All PPAR␥ ligands tested, as well as PD98059, prevented mitogen-induced G 1 3 S progression in a dose-dependent manner (Fig. 1B). Inhibition of ϳ80% was observed at 10 M for TRO and RSG, 5 M for 15d-PGJ 2 , and 30 M for PD98059.
PPAR␥ Ligands Inhibit Mitogen-induced Rb Phosphoryla- tion-To elucidate the mechanism by which PPAR␥ ligands inhibit G 1 3 S progression, we examined their effect on Rb phosphorylation. Rb migrates in an SDS-polyacrylamide gel as multiple, closely spaced bands reflecting varying degrees of phosphorylation. After 24 h mitogenic stimulation with PDGF ϩ insulin, a mobility shift of Rb was observed indicative of increased phosphorylation in RASMC. All PPAR␥ ligands tested, as well as PD98059, inhibited the mobility shift ( Fig. 2A).
In some experiments, Rb from RASMC treated with PPAR␥ ligands appeared to migrate through gels with a slightly faster mobility than that observed for Rb in G 0 /G 1 -arrested cells. Enhanced mobility of Rb after PPAR␥ activation could result from dephosphorylation of hypophosphorylated Rb present in G 0 /G 1 cells. To explore this finding further, we performed similar experiments using human CASMC. An advantage of using CASMC is the availability of antibodies that recognize sitespecific phosphorylations on Rb. Several of these antibodies were tried, unsuccessfully, on RASMC. In CASMC, a phosphospecific antibody was used to assess the phosphorylation status of Ser-807/Ser-811 in Rb, which mediates CDK-dependent regulation of Rb function (21,22). PPAR␥ ligands, as well as PD98059 at highest concentration tested, inhibited the mitogen-induced phosphorylation at Ser-807/Ser-811 (Fig. 2B, b). Importantly, even at 20 M, TRO and RSG did not reduce Ser-807/Ser-811 phosphorylation to levels lower than that detected in G 0 /G 1 cells (data not shown). No evidence of PPAR␥ ligand-induced dephosphorylation of hypophosphorylated Rb (i.e. band migrating faster than those detected in G 0 /G 1 cells) was observed when an antibody recognizing total human Rb was used to analyze CASMC (Fig. 2B, a). In combination, these experiments strongly suggest that activation of PPAR␥, or inhibition of ERK-MAPK activity, blocks only mitogen-induced Rb phosphorylation and has no effect on its basal phosphorylation in G 0 /G 1 .
Effects of PPAR␥ Ligands on Expression of Early G 1 Cyclins and CDKs in VSMC-To understand the mechanism by which PPAR␥ ligands inhibit Rb phosphorylation, we examined their effect on the expression of CDKs and their cyclin partners for which Rb is a major physiological substrate. CDK2 levels were low in quiescent cells, increased after 24 h mitogenic stimulation, and did not change with any of these compounds (Fig. 3A).
Quiescent RASMC expressed both CDK4 and CDK6 which did not change after either mitogenic stimulation or treatment with any of these compounds (Fig. 3A). We next examined the effect of these compounds on protein expression of G 1 phase cyclins D1 and E. Both cyclins D1 and E were expressed at low levels in quiescent RASMC and increased after 24 h stimulation with PDGF ϩ insulin. Treatment with TRO and RSG had no effect on the induction of cyclin D1 by mitogens, whereas addition of 15d-PGJ 2 and PD98059 to mitogen-stimulated VSMC attenuated induction of cyclin D1 levels by 89 Ϯ 3.7 and 68 Ϯ 7.7% at the maximum concentration tested, respectively (p Ͻ 0.05 versus PDGF/insulin alone, Fig. 3, A and B). Mitogenic induction of cyclin E was not affected by any agent (Fig. 3A).

Effects of PPAR␥ Ligands on CDKI Expression in VSMC-
The CDKI p27 kip1 inhibits the activities of cyclin E⅐CDK2 and cyclin D1⅐CDK4 complexes (11,12). Down-regulation of p27 kip1 during G 1 in response to mitogens is important for maximal activation of G 1 cyclin/CDK holoenzymes (23). We therefore investigated the effect of PPAR␥ ligands and PD98059 on p27 kip1 expression after mitogenic stimulation. Western analysis of quiescent RASMC revealed substantial p27 kip1 protein.
Effects of PPAR␥ Ligands and PD98059 on Cyclin D1-associated CDK and Cyclin E-associated CDK Activities-To determine whether various PPAR␥ ligands and PD98059 can regulate CDK activity, we measured the effects of these compounds on mitogen-stimulated cyclin D1-associated and cyclin E-associated CDK activities. By using glutathione S-transferase-Rb fusion protein and purified histone H1 proteins as substrates for cyclin D1-associated CDK and cyclin E-associated CDK, respectively, we found that stimulation with PDGF (20 ng/ml) ϩ insulin (1 M) increased activity of both CDKs (Fig. 5, A and  B). All these PPAR␥ ligands and PD98059 inhibited the induction of cyclin D1-dependent kinase activity (PDGF ϩ insulin ϩ 10 M TRO, 58 Ϯ 3.7% inhibition, p Ͻ 0.01 versus PDGF ϩ insulin alone; 10 M RSG, 53 Ϯ 4.5% inhibition, p Ͻ 0.05 versus PDGF ϩ insulin alone; 5 M 15d-PGJ 2 , 79 Ϯ 4.6% inhibition, p Ͻ 0.01 versus PDGF ϩ insulin alone; 30 M PD98059, 77 Ϯ 7.4% inhibition, p Ͻ 0.01 versus PDGF ϩ insulin alone) (Fig. 5A). Similarly, all the compounds inhibited the induction of cyclin E-dependent kinase activity (10 M TRO, 73 Ϯ 7.3% inhibition; 10 M RSG, 67 Ϯ 7.8% inhibition; 5 M 15d-PGJ 2 , 82 Ϯ 8.0% inhibition; 30 M PD98059, 87 Ϯ 9.4% inhibition; all p Ͻ 0.01 versus PDGF ϩ insulin alone) (Fig. 5B). DISCUSSION The principal finding of this study is that PPAR␥ ligands inhibit VSMC proliferation by attenuating the activity of several key cell cycle regulators that control G 1 3 S progression. All PPAR␥ ligands prevented mitogen-induced phosphorylation of Rb by inhibiting cyclin D1-and cyclin E-dependent kinase activity. Attenuation of mitogen-induced p27 kip1 degra- dation by PPAR␥ ligands is likely the major mechanism ultimately resulting in the inhibition of Rb phosphorylation. Indeed, depletion of p27 kip1 by antisense oligodeoxynucleotides promotes cell growth (28), and mice with targeted disruption of the p27 kip1 gene have enhanced growth with enlargement of the pituitary, adrenals, and gonads (29,30). A recent study demonstrated that overexpression of p27 kip1 inhibited serumstimulated DNA synthesis in VSMC (28). Furthermore, in a porcine balloon-injury model, p27 kip1 expression was markedly reduced in the intima and media early after angioplasty, consistent with an injury-induced proliferative response (31,32). Thus, p27 kip1 down-regulation is a necessary event for cell cycle progression. The prostanoid PPAR␥ ligand 15d-PGJ 2 and the ERK/MAPK pathway inhibitor PD98059 exhibited a broader profile of activity toward G 1 cell cycle regulatory proteins, since they also blocked mitogenic induction of cyclin D1.
Stimulation of the Ras-ERK/MAPK pathway triggers proliferation for a variety of different cell types (33). Activation of Ras 3 Raf 3 MEK 3 ERK/MAPK pathway is critical for both mitogen-dependent cyclin D1 induction and p27 kip1 degradation (34,35). In NIH-3T3 fibroblasts, constitutive expression of an activated form of MEK1, the upstream dual-specificity kinase that phosphorylates and activates ERKs, down-regulated p27 kip1 (35). In CCL39 fibroblasts and IEC-6 intestinal epithelial cells, the blockade of the MAPK cascade with the MEK inhibitor PD98059 prevented both S phase entry and p27 kip1 down-regulation (36). These observations are in concordance with our finding that blocking ERK/MAPK signaling with PD98059 attenuated mitogen-induced down-regulation of p27 kip1 . Degradation of several cell cycle regulatory proteins including cyclins, p53, Rb, and p27 kip1 results from ubiquitinmediated proteolysis in proteosomes (37)(38)(39). The mechanism by which the MAPK pathway and proteosomes interact, however, is unclear. Nevertheless, in combination, these studies support an important role for ERK/MAPKs in regulating p27 kip1 levels after mitogenic stimulation.
Similar to other cell types, the ERK/MAPK pathway is required for the growth in VSMC (40). We previously demonstrated in VSMC that the PPAR␥ ligand TRO inhibited mitogen-induced MAPK signaling downstream of ERK phosphorylation and activation by MEK, which is associated with an inhibition of growth (18). In the present study, we find that PPAR␥ ligands and the MAPK pathway inhibitor PD98059 inhibit mitogenic degradation of p27 kip1 , thereby preventing Rb hyperphosphorylation and the exit of quiescent VSMC from G 1 3 S. Consistent with previous studies in nonvascular cells (34,35), we found that blockade of MAPK signaling with PD98059 potently inhibited cyclin D1 up-regulation by mitogens. In contrast, the TZD PPAR␥ ligands TRO and RSG had no effect on cyclin D1 expression. Thus, PPAR␥ ligands do not globally suppress MAPK-dependent processes, but they may inhibit some transcription factors regulated by the MAPK pathway. We have reported that TRO inhibited activation of the transcription factor Elk-1, which is regulated by MAPK (18). Inhibition of MAPK-dependent Elk-1 activation was associated with an inhibition of c-Fos expression and cell proliferation (18). These data also suggest that inhibition of cyclin D1 expression by 15d-PGJ 2 may not be mediated by its activation of PPAR␥ but rather through its binding to prostaglandin receptors.
Ras can also activate the Rho pathway, which is also important for cell proliferation. Rho signaling has been implicated in the regulation of p27 kip1 degradation (41,42). In IIC9 hamster embryo fibroblasts, expression of dominant-negative Rho prevented PDGF-induced degradation of p27 kip1 (42). In marked contrast to results obtained in mouse fibroblasts linking ERK/ MAPK activation to p27 kip1 degradation, overexpression of dominant-negative ERK in IIC9 cells had no effect on p27 kip1 levels. This study concluded that Ras activation of Rho, but not ERK, regulates p27 kip1 degradation. Rho-dependent degradation of p27 kip1 in PDGF-stimulated fibroblasts requires its phosphorylation by cyclin E-CDK2 (41). We observed that all PPAR␥ ligands inhibited cyclin E-dependent kinase activity, which may be the mechanism by which they attenuate p27 kip1 degradation. This effect could result from an inhibition of mitogen-induced ERK signaling by PPAR␥ as we previously identified. Alternatively, PPAR␥ may regulate p27 kip1 turnover through a novel action to block Rho signaling. However, we found that PPAR␥ activation did not decrease mitogen-enhanced Rho protein levels. 2 In addition, it is unlikely that a nuclear receptor like PPAR␥ would affect Rho movement from the cytosol to the plasma membrane, which is important for activation of the Rho pathway (43). Additional experiments are required to establish whether the Rho pathway is targeted by PPAR␥.
Quiescent VSMC expressed high levels of p27 kip1 , but p21 cip1 was not detectable. Mitogenic stimulation with the combination of PDGF and insulin increased expression of p21 cip1 during G 1 3 S transition. Up-regulation of p21 cip1 during G 1 at first glance is paradoxical, given that it and other CDKIs function to regulate negatively cyclin-CDK activity. Several recent reports, however, have revealed that CDKI modulation of the cell cycle is complex and involves both positive and negative regulation by CDKIs (23)(24)(25)(26)44). Threshold levels of p21 cip1 have been shown to be required for the formation of functional cyclin D1⅐CDK4 complexes (44). Higher concentrations of p21 cip1 , however, inhibited cyclin D1-CDK4 activity consistent with its more traditionally recognized CDKI function (44). The potential for p21 cip1 to regulate CDKs positively is supported by a recent genetic study showing that primary mouse embryonic fibroblasts from p27 kip1 /p21 cip1 double knockout animals failed to assemble detectable amounts of cyclin D1⅐CDK complexes (27). At physiological concentration (10 M), the PPAR␥ ligands TRO and RSG inhibited mitogen-induced p21 cip1 , but 15d-PGJ 2 and the MAPK pathway inhibitor had no effect. Since we saw similar effects of TRO and RSG to inhibit p21 cip1 induction, it is possible that this also plays a role in the inhibition of G 1 3 S transition by PPAR␥. PPAR␥ blockade of p21 cip1 induction by mitogens may result from the ability of these nuclear receptors to inhibit the function of transcription factors, such as NFB, STAT, and AP-1, via the mechanism of transrepression (45,46).
Recently, Morrison and Farmer (47) showed that activation of ectopically expressed PPAR␥ in 3T3-L1 fibroblasts inhibited their growth and promoted their differentiation into adipocytes, which was associated with a concomitant increase in mRNA and protein for CDKIs, p18 ink4c and p21 cip1 . There was no effect of PPAR␥ activation on p27 kip1 in these cells. Inhibition of cell cycle progression frequently is a prerequisite to terminal differentiation. Regulatory mechanisms causing cell cycle arrest during differentiation, however, appear to differ from those governing the exit of quiescent cells from G 1 3 S. For example, during adipocyte differentiation PPAR␥-mediated growth arrest did not require a functional pRB (48). In contrast, our data strongly implicate PPAR␥-dependent inhibition of Rb phosphorylation as the mechanism by which PPAR␥ ligands prevent quiescent VSMC from exiting G 1 . PPAR␥ ligands also inhibit growth of tumor cells (49 -51) and growth of endothelial cells to prevent angiogenesis (52). The impact of PPAR␥ ligands on the cell cycle in these cell types remains to be determined.
Recent studies have illustrated the feasibility of targeting specific cell cycle regulators in cardiovascular cells as an alternative antiproliferative therapy (53). A wide range of antiproliferative drugs has been tested as means to prevent restenosis and vein graft neointimal formation. One alternative to drug therapy is the use of modified viruses designed to carry a cell cycle regulatory gene directly into the arterial wall. Infection of porcine femoral or rat carotid arteries with an adenoviral vector designed to express a nonphosphorylatable, constitutively active form of Rb inhibited neointima formation in animal balloon-injury models (54). Our recent studies have shown that TRO inhibited intimal hyperplasia after balloon injury of aortae in the rat (18). The present study suggests that the prevention of the reduction of p27 kip1 levels by TRO in vivo may contribute, at least in part, to its activity to inhibit the vascular injury response. The observation that PPAR␥ ligands inhibit important cell cycle processes activated by growth factors produced in response to vascular damage may provide a new oral therapeutic approach for proliferative vascular disease such as restenosis and atherosclerosis.