Forkhead Transcription Factors Inhibit Vascular Smooth Muscle Cell Proliferation and Neointimal Hyperplasia*[boxs]

Atherosclerosis, a leading cause of mortality and morbidity in the Western world, involves a multitude of pathophysiological processes including endothelial dysfunction, inflammation, vascular smooth muscle cell (VSMC)¹ proliferation, and alteration in the extracellular matrix (1, 2). Percutaneous transluminal coronary angioplasty is a well established procedure for revascularization of patients with arterial occlusive disease. The occurrence of restenosis at the site of angioplasty remains a major limitation of this procedure. An important goal is to elucidate the molecular basis of restenosis and to use this information to identify novel therapeutic targets.

Neointimal hyperplasia, a hallmark of VSMC proliferation, contributes significantly to the restenotic process after percutaneous transluminal coronary angioplasty. The complex response involves initial accumulation of inflammatory cells and the release of growth factors and chemotactic cytokines (3). Various studies have implicated a role for platelet-derived growth factor (PDGF), tumor necrosis factor (TNF)-α, and insulin-like growth factor (IGF)-1 in this process (4-15).

PDGF-BB, which is expressed in multiple cell types, including VSMC, is a potent mitogen for VSMC in vitro and in vivo (16-21). PDGF has been shown to be secreted from platelets at sites of intimal injury and/or thrombus (22, 23). Several recent studies have reported that inhibition of PDGF or PDGF receptor attenuates neointimal hyperplasia (24-26). Others demonstrated a direct association between PDGF expression and neointimal thickening (27-29).

TNF-α is a pluripotent mediator of inflammation (30) and is believed to play a pathophysiological role in vascular remodeling. TNF-α promotes migration of cultured VSMC (31, 32) and induces the expression of growth factors and cytokines (33), adhesion molecules (34), and extracellular matrix-degrading metalloproteinase (35). Under in vivo conditions, TNF-α is expressed by macrophages and VSMC in atherosclerotic plaques but not in normal vessels (36-38). In various animal models, TNF-α levels are increased following balloon injury, following wire carotid injury, following common carotid artery ligation, and during acute rejection of cardiac allografts (32, 39-42). Blockade of TNF-α with a soluble TNF-α receptor inhibited acute coronary neointimal formation in an animal model of coronary graft atherosclerosis (43). Finally, injury- or ligation-induced intimal hyperplasia of the carotid artery is markedly attenuated in mice that are null for TNF-α (40, 42, 44).

IGF-1 is a weak mitogen for VSMC in vitro. Previous studies have demonstrated increased levels of IGF-1 in the region of intimal hyperplasia in rabbit aorta (45) and in the neointima of arteriovenous fistulas (46). In addition, IGF-1 receptor is highly expressed in the VSMC of intact arteries and in cultured VSMC (7, 47).

The forkhead winged helix transcription factor, DAF-16, was shown to regulate longevity and lipid metabolism in Caenorhabditis elegans (48). Mammalian homologues of DAF-16 (namely members of the FoxO family (FKHR, FKHRL1, and AFX)) have been implicated in cell growth, cell cycle, and differentiation (49, 50). In various cell types, insulin and IGF-1 induce phosphorylation of FoxO proteins, which results in their exclusion from the nucleus and decreased expression of downstream target genes involved in programmed cell death and cell cycle arrest (51, 52). Recently, we demonstrated a role for forkhead proteins in mediating endothelial cell cycle arrest and apoptosis (53, 54). In the current study, we show that FoxO proteins lie downstream of PDGF-BB, TNF-α, and IGF-1 signaling pathways in VSMC in vitro, and we provide evidence for an inhibitory role of forkhead proteins in injury-induced neointimal hyperplasia in vivo.

EXPERIMENTAL PROCEDURES

Materials—PD98059, LY294002, and wortmannin were purchased from Calbiochem. Recombinant human PDGF-BB, TNF-α, and IGF-1 were purchased from R&D Systems (Minneapolis, MN). Antibodies to AKT, phospho-AKT-Ser-473, FKHR, FKHRL1, phospho-FKHR-Ser-256 (which also recognizes phospho-AFX-Ser-193), β-actin, and hemagglutinin (HA) were from Cell Signaling (Beverly, MA). Cy3-conjugated anti-smooth muscle actin antibody was purchased from Sigma. Anti-PDGF-B antibody was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Anti-p27^kip1 antibody, annexin V staining kit, and propidium iodide were from BD Biosciences. Anti-HA-fluorescein antibody was purchased from Roche Applied Science. Biotinylated anti-mouse and anti-rabbit IgGs, streptavidin Texas Red and fluorescein isothiocyanate, and Vectastain Elite ABC kit were from Vector Laboratories (Burlingame, CA). siRNAs were purchased from Qiagen Inc. (Valencia, CA). The XTT assay kit was from Roche Applied Science.

Cell Culture—Human coronary artery smooth muscle cells (CASMC) (Cambrex Bio Science, Baltimore, MD) were cultured in smooth muscle basal medium (SmBM) supplemented with SingleQuots and 10% FBS (Cambrex Bio Science). All studies were carried out with CASMC at passages 5-8. CASMC were grown for 16-72 h (as indicated) in SmBM in the absence of serum before treating with PDGF-BB, TNF-α, or IGF-1. CASMC from two independent donors were tested for the in vitro experiments.

Western and Northern Blot Analyses—Immunoblot assays were performed by modification of the procedure previously described (55). Briefly, CASMC were washed in phosphate-buffered saline twice and harvested by scraping in 20 mm Tris-HCl (pH 7.5), 150 mm NaCl, 1 mm EDTA, 1 mm EGTA, 1% Triton X-100, 20 mm β-glycerophosphate, 10 mm sodium pyrophosphate, 10 mm sodium vanadate, 1 μg/ml leupeptin, 1 mm NaF, and 1 mm phenylmethylsulfonyl fluoride, and protease inhibitor mixture (Roche Applied Science). Protein concentration was determined using a Bio-Rad protein assay kit. Protein (25 μg) was separated on 10% of Tris-HCl SDS-polyacrylamide electrophoresis gel and transferred to a nitrocellulose membrane (Bio-Rad Laboratories). The membrane was blocked with TBST (10 mm Tris, 0.2% Tween 20) containing 5% nonfat milk for 1 h and then incubated with primary antibody for 2 h at room temperature. After three washes with TBST, the membrane was incubated with secondary antibody for 1 h at room temperature. The enhanced chemiluminescence kit (Pierce) was used for detection. Membranes were stripped with Restore Western blot stripping buffer (Pierce) for 15 min at room temperature followed by washing with TBST. Primary antibodies were used as indicated. Secondary antibodies included anti-rabbit IgG/horseradish peroxidase conjugate or anti-mouse IgG/horseradish peroxidase conjugate (1:1000 dilution; Amersham Biosciences). RNA was harvested with the Trizol reagent, and hybridization was performed with p27^kip1 cDNA probe as previously described (53). Northern and Western blots were performed in triplicate, and representative blots are shown.

Adenovirus Vector Constructs and Transduction—CASMC were transduced with replication-defective adenovirus encoding the cDNA for β-galactosidase (Adv), wild type FKHRL1 (WT-FKHRL1), and the phosphorylation-resistant triple mutant FKHRL1 (TM-FKHRL1), which contains T32A, S253A, and S319A (53). For virus transductions, CASMC were plated at ∼70-80% confluence and cultured for 5 h at 37 °C, 5% CO₂ in smooth muscle growth medium (SmGM, Cambrex Bio Science). The cells were washed once with and incubated in Opti-MEM I (without phenol red) for 30 min and then transduced at a multiplicity of infection of 20 in SmGM medium for 12 h, grown in fresh medium for another 10 h, and then serum-starved in SmBM for 16-72 h as indicated.

FKHR and FKHRL1 Immunofluorescence—CASMC were plated onto 4-well chamber slides (Lab-Tek, Christchurch, New Zealand) at a density of 50,000 cells/well. The cells were fixed in 3.7% paraformaldehyde for 20 min at room temperature, washed with phosphate-buffered saline, and incubated in blocking buffer (4 mg/ml bovine serum albumin, 0.1% saponin in phosphate-buffered saline) for 10 min. The cells were incubated with anti-FKHR, anti-FKHRL1, or anti-HA antibody (1:100) in 200 μl of blocking buffer for 1 h at room temperature, followed by fluorescein isothiocyanate-conjugated secondary anti-rabbit IgG (1:100) in 200 μl of blocking buffer for 1 h. After extensive washing, the cells were incubated with DAPI-containing mounting medium Vectashield (Vector Laboratories, Burlingame, CA) for 10 min at room temperature, and representative images were captured with a Nikon Eclipse E800 microscope and a Spot digital camera. The images were merged using Adobe Photoshop software.

Proliferation Assays—CASMC were grown to a density of 50,000 cells/well in 12-well plates. At 80% confluence, the cells were starved for 48 h in SmBM in the absence of serum. The cells were treated with TNF-α, PDGF-BB, or 10% fetal calf serum for 20 h to induce cell cycle reentry and pulsed with 1 mCi/liter [³H]thymidine (Amersham Biosciences) during the last 4 h of incubation. After washes with cold phosphate-buffered saline, cells were trypsinized and solubilized with 0.2 m NaOH. Radioactivity incorporated into DNA was measured in a scintillation counter. In addition, the XTT assay was employed to determine cell proliferation. This assay is based on the conversion of the yellow tetrazolium salt XTT into an orange, water-soluble dye, formazan, by metabolically active cells. Cells were serum-starved for 48 h, treated with agonists for 48 h, and then incubated with XTT for 4 h. Formation of formazan was directly quantitated in 96-well plates with an enzyme-linked immunosorbent assay reader at A_{490-655 nm} (model 680; Bio-Rad). Trypan blue exclusion method was also used to count the number of viable cells where indicated (55).

Flow Cytometry for Cell Cycle and Apoptosis Assay—FACS analyses were carried out to determine apoptosis as assayed by annexin V-propidium iodide staining of intact cells and to determine the cell cycle distribution of CASMC by quantifying propidium iodide-labeled DNA content of ethanol-permeabilized cells, as previously described (53). A total of 10,000 cells were counted (gated) by FACS.

In Vitro Knockdown of p27^kip1 by siRNA—p27^kip1 siRNA was synthesized using the DNA target sequence: AAGTACGAGTGGCAAGAGGTG (56). 400,000 CASMC were plated onto a 60-mm plate, and 24 h later, the cells were transfected with 5 μg of siRNA p27^kip1. In brief, 25 μl of Lipofectin was added to 225 μl of Opti-MEM I in a polystyrene tube. Following a 30-min incubation at room temperature, the mixture was combined with 250 μl of Opti-MEM I containing 5 μg of control nonsilencing Alex Fluor 488 (NS) (Qiagen) or p27^kip1 siRNA and incubated for an additional 30 min. 2 ml of fresh Opti-MEM I was added to each polystyrene tube containing 500 μl of the mixture of siRNA and Lipofectin. The cells were washed with Opti-MEM I and then incubated with 2.5 ml of the above mixture at 37 °C, 5% CO₂. After 4.5 h, Opti-MEM I was removed, and cells were incubated in SmGM medium for 24-48 h, as indicated. For co-transduction with adenovirus and siRNA, the cells were plated and transduced with the viruses expressing either control β-galactosidase (Adv) or TM-FKHRL1 on day 1 and then transfected with either NS or p27^kip1 siRNA on the next day.

Gene Transfer and in Vivo Balloon Angioplasty—The rat carotid balloon angioplasty model was performed according to the method described by Clowes et al. (57, 58). Briefly, 4-month-old Sprague-Dawley rats (350-400 g) were heparinized systemically (100 units/kg). 1-1.5 cm of the left carotid arteries of adult males were injured with a 2F embolectomy catheter (Baxter Edwards Healthcare, Irvine, CA), which was inflated in the carotid and passed three times to achieve significant injury. The vessel was filled with 50 μl of normal saline (n = 8) or with 10⁸ to 5 × 10⁸ plaque-forming units of rAd.TM-FKHRL1 (n = 14) or control rAd.β-gal (n = 6) for 20 min. The adenovirus solution was flushed out via the opened external carotid artery. The external carotid was then ligated, and blood flow was restored. All rats were housed at the BIDMC animal facility and were treated according to published National Institutes of Health guidelines.

Tissue Collection, Morphology, and Immunocytochemistry—Rats subjected to balloon injury were monitored for 14 days and subsequently sacrificed. Injured carotids were retrieved, fixed in 10% formalin, embedded in paraffin, sectioned, and stained with hematoxylin/eosin for histomorphometric analysis. The ratio of intima versus media (I/M) in each carotid was measured by morphometric analysis using the NIH Image version 1.62 public domain software. I/M ratios from 10 serial sections 50 μm apart were averaged for each animal. To detect PDGF-B in the tissue sections, immunochemical staining was performed using biotinylated anti-IgG as a secondary antibody. Apoptotic cells were identified using the Vectastain Elite ABC kit. Immunofluorescent staining of 5-μm frozen sections was employed to detect Ki-67, smooth muscle actin, HA, FKHRL1, and p27^kip1. The snap frozen tissue sections were fixed with 2% paraformaldehyde followed by the heat treatment in 10 mm citrate buffer.

Statistical Analysis—Data are given as mean ± S.E. of at least three independent experiments. Statistical analysis was performed by analysis of variance as appropriate.

RESULTS

PDGF-BB, TNF-α, and IGF-1 Induce Phosphorylation of FoxO Proteins in CASMC—To determine whether PDGF-BB, TNF-α, and IGF-1 modulate the phosphorylation of forkhead proteins in VSMC, we carried out Western blot analyses of CASMC treated in the absence or presence of the cytokine/growth factor. As shown in Fig. 1, incubation of CASMC with PDGF-BB (10 ng/ml for 15 min), TNF-α (10 ng/ml for 30 min), or IGF-1 (100 ng/ml for 15 min) each resulted in phosphorylation of Akt at serine 473, FKHR at serine 256, and AFX at serine 193. PDGF-BB-mediated induction of FKHR phosphorylation was blocked by pretreatment with the PI3K inhibitors, LY294002 and wortmannin, but not by the MEK1/2 inhibitor, PD980059 (Fig. 1A shows LY294002 and PD980059). In contrast, PDGF-BB-induced phosphorylation of AFX was only partially inhibited by LY294002. The effect of TNF-α on the phosphorylation of FKHR and AFX was partially attenuated by LY294002 and PD980059 (Fig. 1B). Finally, IGF-1-mediated phosphorylation of FKHR and AFX was blocked by LY294002 (Fig. 1C). In time course experiments, the effect of TNF-α on phosphorylation of FKHR and AFX was delayed (maximum at 30 min) compared with PDGF-BB and IGF-1 (maximum at 15 min) (data not shown). Taken together, these data indicate that PDGF-BB, TNF-α, and IGF-1 each induces phosphorylation of forkhead in CASMC through MEK1/2 and/or PI3K-dependent signaling pathways.

PDGF-BB, TNF-α, and IGF-1 Induce Nuclear Export of FKHR and FKHRL1 in CASMC—To determine the effect of PDGF-BB, TNF-α, or IGF-1 on nuclear localization of FoxO, CASMC were cultured in 4-well chamber slides, serum-starved for 24 h, and then treated with or without 10 ng/ml PDGF-BB, 10 ng/ml TNF-α, or 100 ng/ml IGF-1 for 30 min. Subcellular distribution of FKHR and FKHRL1 was assayed by immunofluorescence. Under serum starvation conditions, FKHR and FKHRL1 were predominantly localized in the nucleus. Treatment with PDGF-BB, TNF-α, or IGF-1 resulted in nuclear export of FKHR and FKHRL1 (Figs. 2, A and B). Preincubation of CASMC with 50 μm LY294002 significantly inhibited PDGF-BB-, TNF-α-, and IGF-1-mediated nuclear exclusion of FKHR, whereas the inhibitory effect of 50 μm PD98059 was limited to TNF-α-mediated translocation of FKHR (Fig. 2C). These results are consistent with the phosphorylation data (see Fig. 1) and collectively suggest that PDGF-BB-, TNF-α-, and IGF-1-mediated phosphorylation of FoxO results in cytoplasmic translocation/nuclear exclusion of the transcription factors.

PDGF-BB Suppresses p27^kip1 Gene Expression in CASMC by a Forkhead-dependent Pathway—Previous studies in non-VSMC cells have demonstrated a link between forkhead activity and expression of the cell cycle inhibitor/apoptotic gene, p27^kip1 (53, 59-64). Since PDGF-BB, TNF-α, and IGF-1 each phosphorylated FKHR and AFX and resulted in nuclear exclusion of FKHR and FKHRL1 (see Figs. 1 and 2), we examined the effects of these cytokines/growth factors on forkhead-mediated target gene expression in CASMC. Fig. 3A shows that PDGF-BB significantly inhibited endogenous p27^kip1 mRNA expression. TNF-α resulted in a slight increase in p27^kip1 transcript levels, whereas IGF-1 had little or no effect. As expected, adenovirus-mediated overexpression of TM-FKHRL1 increased basal levels of p27^kip1 mRNA in CASMC (Fig. 3B). The constitutively active TM-FKHRL1, but not WT-FKHRL1, blocked the inhibitory effect of PDGF-BB on p27^kip1 expression. Similar results were observed at the level of p27^kip1 protein (Figs. 3, C and D). To confirm transduction efficiency, TM-FKHRL1-transduced CASMC were assayed for HA expression by immunofluorescence and Western blots (data not shown). As shown in Fig. 3E, the majority (>95%) of cells were transduced. Together, these results suggest that PDGF-BB inhibits p27^kip1 expression via a forkhead-dependent signaling pathway.

Forkhead Protein Regulates Cell Apoptosis and Cell Survival in CASMC—To determine whether FoxO regulates apoptosis of VSMC, CASMC were transduced with adenovirus expressing β-galactosidase (Adv) or WT- or TM-FKHRL1 (each at a multiplicity of infection of 20) and assayed by FACS analysis for annexin-V and propidium iodide as markers of early apoptosis (annexin-V-positive cells and propidium iodide-negative cells) and late apoptosis (both annexin-V- and propidium iodide-positive cells). Overexpression of WT- or TM-FKHRL1 increased apoptosis by more than 14 or 22%, respectively, compared with β-galactosidase control (Fig. 4). Incubation of CASMC with PDGF-BB or 10% fetal bovine serum (FBS) reduced the number of apoptotic cells by more than 50% in cells expressing either β-galactosidase (Adv) or WT-FKHRL1 but not constitutively active TM-FKHRL1. IGF-1 resulted in a small but significant reduction of apoptosis in WT-FKHRL1-transduced cells, whereas TNF-α had no effect. Taken together, these results suggest that forkhead induces apoptosis in CASMC and that PDGF-BB may enhance cell survival through the phosphorylation and inactivation of forkhead.

TM-FKHRL1 Blocks PDGF-BB-mediated Stimulation of CASMC Proliferation and Cell Cycle Progression—PDGF-BB and 10% serum induced cell proliferation by 3- and 9-fold, respectively, as determined by [³H]thymidine incorporation (Fig. 5A). IGF-1 had a small but significant effect on DNA synthesis. TNF-α failed to increase thymidine uptake. PDGF-BB-mediated induction of cell proliferation was completely inhibited by overexpression of TM-FKHRL1 but not by WT-FKHRL1 (Fig. 5B). The proliferative effect of 10% FBS was only partially inhibited (by 50%) by overexpression of TM-FKHRL1 (Fig. 5B). Similar results were obtained using the XTT proliferation assay (Fig. 5, C and D) and cell counting by the trypan blue exclusion method (data not shown). In FACS analyses, PDGF-BB was shown to promote cell cycle progression into S phase, with a corresponding decrease in the number of cells in G₀/G₁ phase and an increase in the cells traversing G₂/M phase (Fig. 6A). These effects were completely abrogated by TM-FKHRL1 but not WT-FKHRL1 (Fig. 6, B-D). Similar results were observed with FBS-treated cells. Neither TNF-α nor IGF-1 had a significant effect on VSMC cell cycle (Fig. 6A). Taken together, these data indicate that forkhead protein plays a role in PDGF-BB and serum-mediated progression of the cell cycle.

Antiproliferative Effect of TM-FKHRL1 Is Partially Reversed by siRNA against p27^kip1—Given that forkhead is coupled to the expression of p27^kip1 in CASMC, we hypothesized that the inhibitory effects of TM-FKHRL1 on cell proliferation may be mediated, at least in part, by this cell cycle inhibitor. To test this hypothesis, CASMC were transduced with TM-FKHRL1 or control Adv, transfected with siRNA against p27^kip1 or nonsilencing control (NS), treated in the absence or presence of PDGF-BB, and then assayed for cell proliferation. Efficacy of siRNA against p27^kip1 in CASMC was confirmed by Western blot analyses (Fig. 7A). As shown in Fig. 7, B and C, TM-FKHRL1-mediated inhibition of CASMC proliferation by PDGF-BB was indeed partially reversed by siRNA-mediated down-regulation of p27^kip1. Taken together with our earlier findings, these data suggest that PDGF-BB-mediated proliferation of VSMC is dependent on nuclear exclusion of forkhead and secondary reduction of p27^kip1 levels.

Adenovirus-mediated Gene Transfer of FoxO Protein Inhibits Neointimal Hyperplasia in Rat Carotid Artery Injury Remodeling—To confirm the above results in vivo, we studied a rat carotid balloon angioplasty model for neointimal hyperplasia. Three days after mechanical injury to the intimal layer of the rat carotid artery, endogenous PDGF-BB was found to be highly expressed at the site of injury (Fig. 8). In the next set of experiments, we delivered adenovirus expressing β-galactosidase (rAd.β-gal) or TM-FKHRL1 (rAd.FKHRL1-TM) to balloon-injured carotid arteries of the rat. At day 14 following injury, significant neointimal hyperplasia was observed in the absence of adenovirus transduction (n = 8) or in the presence of rAd.β-gal (n = 6) with intima over media ratios (I/M) reaching 0.87 ± 0.33 and 0.92 ± 0.36, respectively (Fig. 9). Overexpression of rAd.FKHRL1-TM (n = 14) resulted in a dose-dependent inhibition of neointimal hyperplasia with an I/M ratio of 0.35 ± 0.25 at the higher concentration (p < 0.01) (Fig. 9). In immunofluorescent studies of injured vessels at days 1 and 3, HA co-localized with smooth muscle cell antigen (SMA), confirming expression of exogenous HA-tagged TM-FKHRL1 in VSMC (Fig. 10, A-D). Consistent with these results, TM-FKHRL1-transduced vessels demonstrated a marked increase in FKHRL1 immunostaining, compared with β-galactosidase control (Fig. 10, E-H).

To gain insight into the mechanism of inhibition of neointimal hyperplasia by TM-FKHRL1, VSMC were assayed for proliferation and apoptosis in vivo, using Ki-67 and the Vectastain Elite ABC kit, respectively. At day 3 following injury, Ki-67 expression was significantly reduced in HA-TM-FKHRL1-transduced arteries compared with the control (β-galactosidase-expressing) virus-transduced vessels (from 35 to 12% of DAPI-positive VSMC) (Fig. 10, I-L). In contrast, there was no increase in VSMC apoptosis in TM-FKHRL1-transduced vessels compared with the control (Fig. 10, M-P). Similar results were obtained using a terminal dUTP nick-end labeling assay (data not shown). Taken together, these data suggest that FKHRL1 attenuation of neointimal hyperplasia is mediated by inhibition of cell proliferation.

As an extension of our in vitro studies, we assayed for expression of p27^kip1 in the injured arteries. As shown in Fig. 10, Q-T, a significant increase in the expression of p27^kip1 in the smooth muscle cells of the TM-FKHRL1-transduced carotid artery was observed compared with the rAd.β-gal control at day 3 following vessel injury. These findings suggest that p27^kip1 may play a role in mediating the in vivo antiproliferative effect of constitutively active forkhead.

DISCUSSION

Previous studies in other cell types have demonstrated a role for the forkhead family of transcription factors in mediating the effects of growth factors or serum on cell proliferation, cell cycle progression, and/or cell survival (49, 50, 53). When cells are exposed to growth factors or serum, the PI3K/Akt pathway is activated, resulting in the phosphorylation (at Thr-32, Ser-253, and Ser-315 in FKHRL1; at Thr-24, Ser-256, and Ser-319 in FKHR; and at Thr-28, Ser-193, and Ser-258 in AFX) and nuclear exclusion of forkhead transcription factors, down-regulation of target gene expression, and enhanced cell survival and proliferation. Putative forkhead-responsive genes include IG-FBP-1, glucose-6-phosphatase, FasL, TRAIL, GADD45A, BCL-6, BIM, and p27^kip1 (51, 59, 65-70). In contrast, the withdrawal of serum or growth factors leads to nuclear translocation of forkhead proteins and transcriptional activation of target genes, with subsequent cell cycle arrest and/or apoptosis.

The present study is the first to show that VSMC agonists, including PDGF-BB, TNF-α, and IGF-1, promote phosphorylation of FoxO proteins in VSMC, resulting in nuclear exclusion of the transcription factor. Moreover, the data suggest that forkhead activity is a critical determinant of VSMC proliferation both in vitro and in vivo.

PDGF-BB-mediated phosphorylation of forkhead proteins occurred primarily through PI3K, whereas the TNF-α and IGF-1 effect was dependent on both MEK1/2 and PI3K signaling. These findings are novel in that they provide the preliminary evidence for a PI3K-independent mechanism of forkhead regulation in VSMC. Whereas previous studies have implicated non-AKT kinases in forkhead signaling (61, 71-73), they have consistently demonstrated an absolute requirement for PI3K.

The forkhead-responsive gene, p27^kip1, encodes a cyclin-inhibitory protein that plays an important role in regulating proliferation of many different cell types. Indeed, in a previous study, overexpression of p27^kip1 was reported to inhibit VSMC growth (74). Here, we show that whereas all three agonists (PDGF-BB, TNF-α, and IGF-1) induced phosphorylation and nuclear exclusion in VSMC, only PDGF-BB had a significant inhibitory effect on p27^kip1. The reason for this discrepancy is not clear. Perhaps, PDGF-BB results in more intense and/or prolonged inactivation of FoxO in VSMC. Alternatively, TNF-α and IGF-1 may activate parallel signaling pathways that promote expression of p27^kip1. Importantly, siRNA-mediated down-regulation of p27^kip1 partially reversed the ability of TM-FKHRL1 to block PDGF-BB mitogenicity. Together, these data suggest that PDGF-BB-mediated nuclear exclusion of forkhead and subsequent reduction of p27^kip1 expression is responsible, at least in part, for its growth-promoting activities.

In addition to p27^kip1, another cell cycle inhibitor/anti-proliferative gene, BTG1, has been shown to be a target of the forkhead transcription factors in erythroid progenitors (75) and in human endothelial cells.² We have found that PDGF-BB decreases expression of BTG1, like p27^kip1, in CASMC in a forkhead-dependent manner, whereas IGF-1 and TNF-α had no such effect (data not shown). A role for BTG1 may explain why siRNA-mediated down-regulation of p27^kip1 only partially reversed the effect of TM-FKHLR1 on PDGF-BB mitogenicity.

The importance of p27^kip1 in mediating the inhibitory effects of FKHRL1 is further supported by our in vivo findings. In the carotid injury model, VSMC overexpressing FKHRL1 demonstrated significantly higher levels of p27^kip1 expression and lower levels of proliferation, compared with β-galactosidase (Adv) control. Consistent with these results, Braun-Dallaeus et al. (76) demonstrated that p27^kip1 is expressed at high levels in uninjured arteries and is rapidly down-regulated after balloon injury. Overexpression of p27^kip1 in cultured VSMC resulted in G₁ arrest and complete inhibition of proliferative response to serum, whereas adenoviral transfer of p27^kip1 in a rabbit model of balloon injury resulted in significant inhibition of intimal cell proliferation (77). However, our in vivo data do not exclude a role for other forkhead target genes (e.g. BTG1) in mediating the inhibitory effects of TM-FKHRL1 on neointimal hyperplasia.

Following the original submission of this manuscript, a study was published describing that constitutively active TM-FKHRL1 induces p27^kip1 protein in rat aortic vascular smooth muscle cells and that overexpression of TM-FKHRL1 leads to cell cycle arrest and apoptosis (78). Balloon injury resulted in rapid induction of phospho-FKHRL1 and a delayed reduction in p27^kip1 in the blood vessel wall. Finally, overexpression of TM-FKHRL1 resulted in a significant increase in p27^kip1 expression, lower proliferative index, increased apoptosis, and reduction in neointimal area (78).

The findings of our present study add to the latter report in several important ways. First, we have shown that human coronary artery vascular smooth muscle cells express not only FKHRL1, but also FKHR and AFX. Second, we have demonstrated that growth factors induce phosphorylation and nuclear exclusion of endogenous forkhead proteins in VSMC. Third, we have delineated the signaling pathways involved in agonist-mediated phosphorylation of endogenous forkhead and consequent down-regulation of the forkhead target gene, p27^kip1, at the transcript and protein levels. Fourth, we have employed an siRNA approach to provide direct evidence for the role of p27^kip1 in the PDGF-BB-forkhead-cell proliferation pathway. Finally, our in vivo results contrast with those of the previous study in that they fail to demonstrate TM-FKHRL1-mediated changes in VSMC apoptosis, using two different assays. It is conceivable that TM-FKHRL1 induces cell cycle arrest in vivo, without promoting apoptosis. Alternatively, apoptotic cells may be promptly removed from the site of injury.

There is increasing evidence that bone marrow-derived progenitors contribute to neointimal formation in response to vascular injury (79-83). An interesting question is whether progenitor-derived VSMC are more or less prone to the inhibitory effects of forkhead overexpression, compared with locally derived VSMC. Moreover, it is interesting to speculate whether forkhead-mediated changes in gene transcription in the native vessel wall results in reduced homing of circulating progenitor cells to the injury site.

In conclusion, we have shown that growth factor and cytokine signaling is linked to forkhead transcription factor(s) in VSMC. Our data suggest that forkhead signaling pathway contributes to the regulation of VSMC proliferation, cell cycle progression, and neointimal hyperplasia. Thus, the forkhead pathway may represent a novel therapeutic target in vasculopathic diseases.