Platelet-derived Growth Factor-BB-mediated Activation of Akt Suppresses Smooth Muscle-specific Gene Expression through Inhibition of Mitogen-activated Protein Kinase and Redistribution of Serum Response Factor*

Platelet-derived growth factor (PDGF) inhibits expression of smooth muscle (SM) genes in vascular smooth muscle cells and blocks induction by arginine vasopressin (AVP). We have previously demonstrated that suppression of SM-α-actin by PDGF-BB is mediated in part through a Ras-dependent pathway. This study examined the role of phosphatidylinositol 3-kinase (PI3K)y and its downstream effector, Akt, in regulating SM gene expression. PDGF caused a rapid sustained activation of Akt, whereas AVP caused only a small transient increase. PDGF selectively caused a sustained stimulation of p85/p110α PI3K. In contrast, p85/110β PI3K activity was not altered by either PDGF or AVP, whereas both agents caused a delayed activation of Class IB p101/110γ PI3K. Expression of a gain-of-function PI3K or myristoylated Akt (myr-Akt) mimicked the inhibitory effect of PDGF on SM-α-actin and SM22α expression. Pretreatment with LY 294002 reversed the inhibitory effect of PDGF. Expression of myr-Akt selectively inhibited AVP-induced activation of c-Jun N-terminal kinase and p38 mitogen-activated protein kinases, which we have shown are critical for induction of these genes. Nuclear extracts from PDGF-stimulated or myr-Akt expressing cells showed reduced serum response factor binding to SM-specific CArG elements. This was associated with appearance of serum response factor in the cytoplasm. These data indicate that activation of p85/p110α/Akt mediates suppression of SM gene expression by PDGF.

Phenotypic modulation of vascular smooth muscle cells (VSMC) 1 is critical during development and in the onset of diseases such as atherosclerosis and hypertension. Unlike myo-cardial and skeletal muscle cells, VSMC are highly plastic and retain the ability to modulate their phenotype from a contractile to a more proliferative synthetic state (1,2). Retention of this plasticity is important in that it permits processes such as vessel repair after injury, but it may also be detrimental in that it increases susceptibility to atherogenic risk factors. Understanding the molecular control of phenotypic modulation of VSMC will be dependent on the identification of signaling mechanisms that control the transcription of genes encoding proteins necessary for the differentiated function of VSMC. Several smooth muscle (SM)-specific contractile or contractileassociated proteins, including SM-␣-actin, myosin heavy chain, h-caldesmon, and SM-22␣, are useful markers for studying VSMC differentiation (3). Of these, SM-␣-actin has been the most extensively studied, because it is the most abundant VSMC protein, and the first known marker to appear during vessel development (4). Regulation of SM-␣-actin expression involves a complex interaction of multiple positive and negative cis-elements that act in a cell type-specific fashion (5). SM22␣, another early marker, is a calponin-related protein that may be involved in contraction of VSMC because of its interactions with tropomyosin and F-actin (6).
Our laboratory has examined the regulation of SM-specific gene expression in cultured adult VSMC. These cells have been widely studied and are likely to reflect most of the critical features of VSMC in vivo (7). The proximal region of the rat SM-␣-actin promoter contains two CArG elements, which are critical for regulation of promoter activity by vasoconstrictor hormones and growth factors (8,9). CArG elements are the binding sites for the MADS (MCM1 agamous deficiens SRF) box transcription factor, serum response factor (SRF) (10). Truncation mutations (8) or point mutations (11) in individual CArG elements of the SMA promoter abolish both basal and AVP-stimulated activation of the promoter. The SM22␣ gene, like many SM genes, also contains multiple CArG boxes in the proximal region of its promoter (12).
In VSMC, platelet-derived growth factor (PDGF) mediates a variety of biological effects through activation of intracellular signal transduction pathways including the MAP kinase cascade, phosphatidylinositol turnover, and calcium mobilization. These effects contribute to smooth muscle cell proliferation and directed migration (13). Earlier work from our laboratory (14) and others (15) reported that PDGF suppressed expression of SM-␣-actin in rat aortic VSMC. This effect is mediated at least in part through suppression of SM-␣-actin promoter activity and involves PDGF-induced activation of the low molecular weight G protein Ras (14). Activation of Ras engages multiple downstream effectors, including the Raf/MEK/ERK pathway. However, suppression of SM-␣-actin appeared to be mediated through a Ras-dependent/Raf-independent pathway involving effector pathways that are not well characterized.
Activated PDGF receptors and other receptor tyrosine kinases have been shown to bind and activate Class IA phosphatidylinositol 3-kinase (PI3K) (16), which in turn results in the local accumulation of polyphosphoinositide phosphatidylinositol 3,4,5-trisphosphate at the plasma membrane (17). Newly synthesized phosphatidylinositol 3,4,5-trisphosphate recruits pleckstrin homology domain-containing signaling molecules such as phosphoinositide-dependent kinase-1, Akt/PKB, and possibly protein kinase C, to the plasma membrane, where the combination of lipid binding and phosphorylation by phosphoinositide-dependent kinase-1 serves to activate these enzymes (18,19). Activated Akt can phosphorylate and inhibit members of the forkhead family of transcription factors (20,21). G protein-coupled receptors have been shown to activate a distinct class of PI3K (Class IB, p101/p110␥) through ␤/␥ subunits of heterotrimeric G proteins (22,23).
The goal of this study was to examine the role of the PI3K/ Akt pathway in PDGF-induced down-regulation of SM-specific gene expression in VSMC. We investigated the contribution of this pathway to the regulation of SM-␣-actin and SM22␣ expression and showed for the first time that PDGF-BB-induced suppression of SM-specific genes is mediated through activation of a PI3K-dependent Akt signaling pathway. We also report a novel PI3K-dependent Akt signaling pathway activated by PDGF, leading to changes in the subcellular localization of SRF.

EXPERIMENTAL PROCEDURES
Materials-Trypsin/EDTA and Eagle's minimal essential medium (MEM) were from Invitrogen. Antibodies to Akt, JNK, p38, and ERK1/2 and to their phosphorylated forms were purchased from Cell Signaling Technologies (Beverly, MA). Monoclonal antibody against SM-␣-actin was obtained from Sigma. Monoclonal antibody against SM22␣ was a gift of Dr. Angela Chiavegato (University of Padua, Padua, Italy). Horseradish peroxidase-labeled IgG was from Amersham Biosciences. Expression plasmid encoding myristoylated Akt (myr-Akt) was a generous gift of Dr. Lynn Heasley (University of Colorado Health Sciences Center, Denver, CO). Expression plasmid for myr-p110␣ PI3K and antibodies against p110␣, p110␤, and p101 were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY).

FIG. 1. Sustained activation of Akt by PDGF in VSMC.
A, VSMC were serum-restricted overnight and stimulated with AVP (10 Ϫ6 M) or PDGF (20 ng/ml) for the indicated lengths of time. The cell extracts were prepared and immunoblotted with antibodies specific for Akt phosphorylated at Ser 473 (phospho-Akt) or with antibodies that react with Akt regardless of phosphorylation status (Akt). A representative of four experiments is shown. B, Akt kinase assay was performed by immunoprecipitation of Akt from VSMC stimulated with AVP or PDGF for the indicated lengths of time. A fragment of GSK3 containing Akt phosphorylation sites was used as a substrate in a kinase reaction and immunoblotted with a phospho-specific GSK3 antibody. A representative of two independent experiments is shown. C, VSMC preincubated for 30 min with or without 10 M LY 294002 and then stimulated with PDGF for the indicated times. The extracts were prepared and immunoblotted with phospho-Akt antibodies. A representative of three independent experiments is shown. D, two clones (Clones 11 and 16) of VSMC stably expressing constitutively active H-Ras and cells stably transfected with empty vector (Neo Clone 2) were stimulated with AVP, PDGF, or vehicle for 5 min. The extracts were prepared and immunoblotted for phospho-Akt. MEM with 10% fetal calf serum and then placed in Eagle's MEM with 0.2% fetal calf serum with or without agonists for 72 h. The transiently transfected VSMC were then washed twice with ice-cold PBS and harvested in luciferase reporter lysis buffer (Promega, Madison, WI). The Cell lysates were centrifuged, and the supernatants assayed for luciferase and ␤-galactosidase activities as previously described (11).
Stable Transfection of VSMC-A cDNA encoding Akt containing the Src myristoylation sequence at the N terminus and the hemagglutinin tag at the C terminus (25) was ligated between the HindIII and HpaI sites of the retroviral expression vector pLNCX. The resulting vector, LNCX-myr-Akt, was packaged into replication-defective retrovirus using 293T cells and the retrovirus-component expression plasmids SV-⌿ Ϫ -A-murine leukemia virus and SV-⌿ Ϫ -env Ϫ -murine leukemia virus by the procedure described previously (26). Secreted retrovirus was supplemented with polybrene (8 g/ml), filtered (0.45 m), and incubated with VSMC for 48 h. The cells expressing myristoylated Akt were selected by culturing them in medium containing G418 (500 g/ml). Individual clones were screened by immunoblotting with anti-hemagglutinin, anti-Akt, and anti-phospho-Akt antibodies.
Akt Kinase Assay-VSMC were plated at 1.5 million cells/60-mm dish. The cells were made quiescent by incubation in Eagle's MEM containing 0.2% fetal bovine serum for 24 h and then stimulated with AVP or PDGF-BB for the indicated times. The cells were harvested in Akt kinase lysis buffer (25 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 ␤-glycerolphosphate, 1 mM Na 3 VO 4 , and protease inhibitor mixture). Akt kinase activity was assayed with a kit from Cell Signaling Technologies (Beverly, MA), using GSK3 as a substrate as per the manufacturer's instructions.
PI3K Assay-VSMC were treated with AVP or PDGF for the indicated times and harvested in Akt kinase lysis buffer. Extracts were microcentrifuged for 10 min, and 0.5 mg of protein from each extract was subjected to immunoprecipitation for 16 -18 h at 4°C with antibodies against specific isoforms of PI3K (p110␣, p110␤, and p101). Protein A-Sepharose (Sigma) beads were added for 1-2 h at 4°C. The beads were rinsed three times with lysis buffer and once with wash buffer (0.1 M NaCl, 1 mM EDTA, 20 mM Tris-HCl, pH 7.5). Phosphatidylinositol micelles were prepared in 10 mM HEPES, pH 7.4, as 20 mg/ml stock by sonication in a bath sonicator for 15 min at 4°C. Washed Sepharose beads containing immunoprecipitated PI3K isoforms were incubated with phosphatidylinositol micelles (20 g/reaction) for 10 min on ice, followed by the addition of kinase reaction mix (final concentration, 20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 , 0.5 mM EGTA, 20 M ATP, 5 Ci of [␥-32 P]ATP). Kinase reactions were run for 20 min at 25°C and stopped by the addition of 100 l of 1 N HCl. Phosphorylated phosphatidylinositol was extracted once with 160 l of chloroform:methanol (1:1) and once with 80 l of MeOH with 100 mM HCl, 2 mM EDTA (1:1). Ten l of the organic phase was spotted on TLC plates and separated for 2 h in chloroform:methanol:ammonium hydroxide (29.5%):water (42:38:4.7:7.3). TLC plates were dried and autoradiographed. The results were quantitated by densitometry.
Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared from cells either stimulated with PDGF for 96 h or cells stably expressing gain-of-function Akt using a variation of the procedure of Dignam et al. (27) as described previously (11). A double-stranded 94-bp DNA piece (Ϫ184 to Ϫ92), which included the near CArG and GC boxes of the SM22␣ promoter was made by PCR amplification. Primers used for PCR amplifications included a 3Ј MluI site for the purpose of fill-in labeling with [ 32 P]dCTP using Klenow DNA polymerase. Unincorporated [ 32 P]dCTP was removed by polyacrylamide gel electrophoresis. Binding was performed for 30 min at 4°C with 2 g of nuclear binding proteins and 1 ng of labeled DNA probe in 20 l of total volume containing 0.125 mg of poly(dI-dC), 12 mM HEPES-KOH, pH 7.9, 150 mM KCl, 1.0 mM EDTA, 0.3 mM dithiothreitol, 12% glycerol, and protease inhibitors. For supershift experiments, 2 g of SRF antibody was added at 4°C 30 min prior to the addition of radiolabeled probe. Protein-DNA complexes were resolved on a 24 cm, 5% acrylamide gel (29:1 acrylamide:bisacrylamide; Fisher) at 25 mAmp/gel in 1ϫ TGE (25 mM Tris, 1.0 mM EDTA, 190 mM glycine, pH 8.0) for 3 h. The gels were dried and autoradiographed.
Immunofluorescence-VSMC were cultured on 18-mm glass coverslips, before being stimulated either with AVP for 72 h or with PDGF for 72 or 96 h. Stimulated cells or cells stably expressing gain-of-function Akt (myr-Akt2 and myr-Akt9), along with their control cells, were then fixed in 3.7% formaldehyde in PBS for 10 min at room temperature. The cells were rinsed and permeabilized with 0.1% Triton in PBS for 5 min and rinsed three times with PBS. Prior to labeling with SRF antibody (Santa Cruz) for 1 h at room temperature in 1% bovine serum albumin in PBS, the cells were blocked with 5% bovine serum albumin for 30 min. SRF labeling was detected using an anti-rabbit antibody coupled to Texas Red (Molecular Probes). After the secondary antibody labeling, the cells were rinsed extensively and incubated with ALEXA Fluor 488 Phalloidine (Molecular Probes) in 1% bovine serum albumin for 20 min for detection of F-actin. The coverslips were mounted with VectaShield medium containing 4Ј-6-diamino-2-phenylindole (DAPI) (Vector laboratories), and the cells were visualized using a Zeiss Axiovert inverted fluorescence microscope. The cells stably expressing myr-Akt (Clones 2 and 9) or cells transfected with empty vector (Clone 10) were serum-restricted overnight, and stimulated with AVP (10 Ϫ6 M) or PDGF (20 ng/ml) for 5 min. The cell extracts were prepared and immunoblotted with antibodies specific for Akt phosphorylated at Ser 473 . B, cells stably expressing myr-Akt (Clones 2 and 9) or cells transfected with control vector (Clone 10) were serum-restricted overnight and exposed for 96 h to AVP (10 Ϫ6 M) or PDGF (20 ng/ml). SM-␣-actin and SM22 protein expression was determined by immunoblotting. ing agonist stimulation, using a phospho-specific Akt antibody, which detects phosphorylation at serine 473. PDGF rapidly increased Akt phosphorylation by 5 min. Maximal stimulation (ϳ6-fold) was detected at 5-30 min, and stimulation was sustained up to 4 h with no significant decrease (Fig. 1A). In contrast, exposure of VSMC to AVP resulted in a very modest increase in phospho-Akt (Ͻ2-fold), which was more transient in nature (Fig. 1A), and returned to basal levels by 15 min. Activation of Akt was confirmed by directly measuring kinase activity after immunoprecipitation with anti-Akt antibodies, using GSK3 as a substrate. PDGF gave a sustained increase in Akt activity ϳ10-fold, whereas a more transient and modest increase was seen with AVP stimulation (Fig. 1B). Treatment of cells with LY 294002 (10 M), a PI3K inhibitor, blocked PDGF-induced Akt phosphorylation by greater than 80% (Fig.  1C), indicating that activation of Akt is mediated through a PI3K-dependent pathway as shown by other workers (28). To test the role of Ras as an upstream activator of Akt, we determined Akt activation in VSMC stably expressing H-Ras (14). In two independent clones, basal phospho-Akt levels were not significantly elevated compared with control cells (Fig. 1D), and PDGF but not AVP was still able to activate Akt. These data indicate that PDGF-induced stimulation of Akt is likely mediated directly through the PDGF receptor and does not require Ras.

PDGF
Class IA PI3K, which is regulated by receptor tyrosine kinases, consists of a heterodimer of a regulatory 85-kDa and a catalytic 110-kDa subunit (17). To examine specific activation of this form, extracts from control or stimulated VSMC were immunoprecipitated with an antibody against the p110␣ catalytic subunit of PI3K and phosphatidylinositol kinase activity measured in the immunoprecipitates (see "Materials and Methods"). PDGF-induced formation of phosphatidylinositol 3,4,5-trisphosphate by p110␣ was apparent by 5 min of stimulation and the response was still sustained after 1 h of stimulation ( Fig. 2A). AVP did not stimulate phosphatidylinositol 3,4,5-trisphosphate formation by p110␣ at any time points that were tested. Immunoprecipitation with an antibody against p110␤ showed constitutive PI3K activity, with no significant changes in response to either PDGF or AVP stimulation (Fig.  2B). Class IB PI3K consists of a dimer of a 101-kDa regulatory subunit and a p110␥ catalytic subunit (p101/p110␥). To assess regulation of this form, the extracts were immunoprecipitated with an antibody against p101. Stimulation of cells by either PDGF or AVP caused a delayed increase in activity, which was only observed after 30 min of stimulation (Fig. 2C). Thus, although multiple forms of PI3K may be active in VSMC, the selectivity of p85/p110␣ for PDGF and the kinetics of activation suggest that p85/p110␣ mediates the activation of Akt by PDGF in VSMC.
Activation of p85/p110␣ PI3K Mediates Suppression of SM-␣-Actin Expression-To establish the role of the PI3K/Akt pathway on SM-specific gene expression, we examined the ability of the PI3K inhibitor to modulate the effects of PDGF on SM-␣actin promoter activity. Consistent with our earlier studies, PDGF suppressed SM-␣-actin promoter activity, and this was blocked by exposure of the cells to 10 M LY 294002 (Fig. 3A). Changes in protein expression were determined by immunoblotting with a specific SM-␣-actin antibody. PDGF decreased expression of SM-␣-actin, and exposure to LY 294002 almost completely reversed this inhibition (Fig. 3B). Conversely, VSMC were transiently co-transfected with a gain-of-function, myristoylated form of p110␣ (myr-p110␣), along with the SM-␣-actin promoter. Expression of myr-p110␣ reduced both basal and AVP-stimulated increases in promoter activity (Fig. 3C), implicating p85/p110␣ PI3K in the suppressive pathway regulated by PDGF. The addition of PDGF caused a further suppression of promoter activity, suggesting the involvement of additional pathways.
Akt Mediates Suppression of SM Gene Expression-To assess the role of Akt as a downstream effector of PDGF-induced PI3K activity in the regulation of SM-specific gene expression, VSMC were transiently transfected with a plasmid encoding an activated form of Akt in which the c-Src myristoylation signal is fused to the N terminus of the c-Akt sequence (myr-Akt) (25), leading to enhanced association of the protein kinase with the plasma membrane and constitutive activation. Co-transfection of myr-Akt with the SM-␣-actin promoter had minimal effects on basal promoter activities but markedly inhibited the induction by AVP (Fig. 4A). We examined the effects of Akt on another SM marker, SM22␣, using a 505-bp region of the SM22␣ promoter (29). AVP increased SM22␣ promoter activity 4 -6-fold, and this was suppressed by PDGF. Expression of myr-Akt inhibited AVP induction of SM22␣ promoter activity to a similar extent as with the SM-␣-actin promoter (Fig. 4B).
The effects of Akt on SM protein expression were confirmed by stable expression of myr-Akt using retroviral-mediated gene transfer (see "Materials and Methods"). G418-resistant clones were screened for expression of Akt. Of nine clones screened, one clone showing the highest level of expression (Clone 9) and another showing more modest expression (Clone 2) were selected for further characterization. All transfected cells had normal morphology by phase contrast (data not shown). myr-Akt9 showed constitutively high levels of phospho-Akt compared with cells stably transfected with empty vector (Neo Control 10), and this was not significantly increased by PDGF stimulation (Fig. 5A). Clone 2 also showed constitutive increases in basal phospho-Akt, but PDGF stimulation resulted in a further increase. Levels of SM-␣-actin and SM22␣ protein were determined by immunoblotting following exposure to either AVP or PDGF. In Neo controls, AVP increased expression of both SM-␣-actin and SM22␣, whereas PDGF suppressed expression (Fig. 5B), similar to what has been observed in wild-type cells. In both clones expressing myr-Akt, basal expression of both proteins was decreased. AVP induction was blunted in Clone 2, which had moderate levels of overexpression. In Clone 9, which had the highest expression levels, AVP failed to increase SM22␣ expression and in some experiments inhibited SM-␣-actin expression. PDGF inhibited expression in both clones. Consistent with the results obtained with transient co-transfections with myr-Akt, the AVP-mediated increase in SM-␣-actin and SM22␣ promoter activity was also blocked in cells stably expressing myr-Akt (Fig. 6).
Akt Activation Inhibits AVP-mediated Stimulation of JNK and p38 MAP Kinases-We sought to identify the downstream effectors of Akt activation, which mediate suppression of SMspecific gene expression. Previous work from our laboratory has shown that both JNK and p38 MAP kinases are necessary for induction of SM-␣-actin expression by AVP, whereas activation of ERK does not play a major role in this regulation (11). In cells stably transfected with empty vector, AVP rapidly stimulated ERK, JNK, and p38 activity (Fig. 7A), similar to what we have previously reported (11). However, in both clones stably expressing myr-Akt, AVP-induced activation of JNK was completely blocked, and activation of p38 MAP kinases was significantly reduced (Fig. 7A). This was not due to a general impairment of AVP signaling, because no significant decrease in ERK activation by AVP was observed. To determine whether the suppressive effects of PDGF could be mediated through a similar mechanism, wild-type VSMC were pretreated with PDGF for 7 h and then acutely stimulated with AVP. Similar to what was observed in cells expressing myr-Akt, pretreatment with PDGF inhibited AVP-mediated activation of JNK by ϳ50% and p38 activation by 70% but had no significant effect on ERK activation (Fig. 7B).
PDGF Stimulation and Akt Activation Localizes SRF to Cytosol and Inhibits Its DNA Binding-To determine possible effects of Akt activation on DNA-protein interactions, nuclear extracts were prepared from control cells, cells were stimulated with PDGF for 96 h or cells stably expressing myr-Akt, and DNA binding of SRF was determined by electrophoretic mobility shift assay (see "Materials and Methods"). As seen in Fig.  8A, nuclear extracts from PDGF-stimulated cells and both of the myr-Akt clones (myr-Akt2 and myr-Akt9) showed reduced binding of a band that was supershifted by an SRF antibody. Immunoblotting of SRF expression in whole cell and nuclear extracts showed that although overall cellular expression of SRF was unchanged (Fig. 8B, Whole cell), expression levels of SRF in nuclear extracts were reduced in both PDGF-stimulated cells and Akt clones (Fig. 8B, Nuclear). We have detected decreased expression of SRF in nuclear extracts as early as after 72 h of PDGF treatment (data not shown), consistent with the time course of suppression of SM genes. These results suggested that the subcellular localization of SRF may differ in cells treated with PDGF versus control. Consistent with this idea, by immunofluorescence SRF staining was exclusively nu- FIG. 7. Myristoylated Akt and PDGF inhibit activation of JNK and p38 MAP kinases by AVP. A, VSMC were stably transfected with plasmid encoding myristoylated Akt (myr-Akt) by retrovirus-mediated gene transfer as described under "Materials and Methods." The cells stably expressing myr-Akt (Clones 2 and 9) or cells transfected with empty vector (Clone 10) were serum-restricted overnight and exposed to AVP (10 Ϫ6 M) for 5 min. The proteins were extracted and resolved on SDS-PAGE as described under "Materials and Methods" and immunoblotted for phosphorylated forms and protein forms of ERK1/2, JNK, and p38 MAP kinases. B, VSMC were serum starved for 24 h and then pretreated with or without PDGF for 7 h prior to stimulation with AVP for the indicated lengths of time. The proteins were extracted and resolved on SDS-PAGE as described under "Materials and Methods" and immunoblotted for phosphorylated forms of ERK1/2, JNK, and p38 MAP kinases. The intensity of the phosphorylated bands for each MAP kinase was compared with the same time point of AVP stimulation in cells without the prior PDGF stimulation. A representative experiment is shown. clear in control VSMC. However, VSMC stimulated with PDGF for 24, 48 (not shown), 72, and 96 h (Fig. 8C) and myr-Akt clones (Fig. 8D) showed the accumulation of SRF in the cytosol, detected as punctate staining mostly around the nucleus but also spread toward the periphery of the cells. Finally, the cytosolic localization of SRF was specific for PDGF, because stimulation of cells with AVP resulted in exclusively nuclear staining of SRF, similar to control cells (Fig. 8E). DISCUSSION VSMC differentiation is important during vascular development, and alterations in VSMC phenotype play a major role in the progression of cardiovascular diseases including atherosclerosis, hypertension, and restenosis (2,30,31), all of which are diseases of mature vessels. It is well established that VSMC do not terminally differentiate and that VSMC phenotype in mature vessels is regulated by a complex array of local FIG. 8. Myristoylated Akt and PDGF inhibit SRF-DNA interaction by localizing SRF out of nucleus. A, labeled 94-bp SM22␣ promoter probe (Ϫ184 to Ϫ92) was incubated with 2 g of nuclear binding proteins prepared from untreated VSMC or cells stimulated with PDGF-BB for 96 h or VSMC stably expressing myr-Akt (Clones 2 and 9) or their control cells (Neo Clone 10) and analyzed by electrophoretic mobility shift assay. For supershift assay, 2 g of antibody against SRF were incubated with nuclear extracts for 30 min at 4°C prior to the addition of probe. In the presence of SRF antibody, specific band a was shifted to give band b. These data are representative of three independent experiments. B, either untreated VSMC or cells stimulated with PDGF-BB for 96 h or VSMC stably expressing myr-Akt (Clones 2 and 9) or their control cells (Neo Clone 10) were harvested in RIPA buffer, or nuclear extracts from similarly treated cells were subjected to immunoblotting for SRF. C, VSMC were cultured on 18-mm glass coverslips before being stimulated with PDGF-BB for 72 and 96 h. Stimulated cells along with their control cells, were then fixed and permeabilized and processed for immunofluorescence as described under "Materials and Methods." SRF staining is shown as red, and F-actin staining is green. Nuclei of the cells were labeled with DAPI. The cells were visualized using a Zeiss Axiovert inverted fluorescence microscope. D, VSMC stably expressing myr-Akt (Clones 2 and 9) along with their control cell (Neo Clone 10) were prepared and processed as in C. E, VSMC were prepared and processed as in C except they were stimulated with AVP for 72 h. environmental cues including humoral factors, cell-cell and cell-matrix interactions, mechanical stresses, and inflammatory stimuli (3,32,33). Although much has been learned regarding critical regulatory promoter elements and transcription factors that control SM-specific gene expression (34), the signal transduction pathways that act on these factors are less well understood. PDGF induces a variety of cellular responses in VSMC including cell proliferation, migration, dedifferentiation, and extracellular matrix deposition. Like many receptor tyrosine kinases, the activated PDGF receptor transmits signals by recruiting a variety of SH2 and phosphotyrosine binding domain-containing proteins to multiple phosphorylated tyrosine residues, including PI3K, phospholipase C␥, Ras GTPactivating protein (RasGAP), the SHP2 phosphatase, and the Src family kinases (35,36). A long standing question regarding these receptor tyrosine kinase signals is whether each pathway transmits a unique signal or whether integration of multiple signals leads to unique biological outcomes. We have previously presented data suggesting that a Ras-dependent/Raf-independent pathway involving the induction of cytosolic phospholipase A 2 participates in PDGF-induced inhibition of SM-␣-actin expression (14). In the present study, we have identified a role for PDGF-dependent activation of Akt in this process.
PDGF induced a more robust and sustained activation of Akt compared with that seen with AVP. Although multiple forms of PI3K are expressed in VSMC, based on selectivity for PDGF and kinetics of activation, we propose that sustained Akt activation is mediated by selective activation of the p85/p110␣ isoform of PI3K. Our data support a role for this pathway in suppression of SM gene expression. Inhibition of PI3K with LY 294002, which blocked Akt activation, was able to reverse the suppressive effects of PDGF on SM gene expression, whereas transient expression of a myristoylated, gain-of function form of p110␣ PI3K mimicked the suppressive effect of PDGF. Expression of a myristoylated gain-of-function form of Akt decreased protein expression and promoter activity of both SM-␣-actin and SM22␣ and blocked the AVP-mediated increase in expression. Other isoforms of PI3K, specifically p85/p110␤ or Class I B (p101/p110␥), do not appear to contribute to this sustained activation of Akt in VSMC. In fact, in hypertensive rats, increased activation of p110␤ and p110␦ was shown to couple not to Akt activation but rather to regulation of intracellular calcium handling (37). In addition, PDGF-or vasoconstrictorinduced activation of PI3K␤ and ␥ but not PI3K␣ were shown to couple to L-type calcium channels in VSMC (38). In our cells, Akt phosphorylation was not increased in VSMC stably expressing constitutively active Ras, indicating that Akt activation is not mediated by Ras activation but proceeds directly through receptor-mediated activation of PI3K. Thus, we propose that PDGF binding to its receptor engages multiple downstream effectors, specifically Ras and p85/p110␣/Akt, which cooperate to suppress expression of multiple SM genes (Fig. 9).
Although activation of Akt by PDGF is likely to have multiple effects in VSMC, we have defined two events that are likely to be important in the regulation of SM gene expression. First, activation of Akt leads to selective inhibition of AVP-induced stimulation of two MAP kinase pathways, JNK and p38, which we have shown to be required for induction of SM genes (Fig. 9). Because AVP-stimulated ERK activation is not affected, inhibition of JNK and p38 MAP kinases must be at post-receptor level. Previously, activation of Akt has been shown to inhibit JNK activation in other cell types (39 -41), and this occurs at the level of the upstream kinase, MKK4. However, this effect appears to be cell specific, because other workers have found positive effects of Akt on JNK activation (42). Additional studies will be required to define whether the effects of Akt are specific for individual JNK isoforms. We are currently attempting to define the isoforms activated by AVP in VSMC. We propose that the ability of PDGF to block AVP induction of SM genes is mainly mediated through this pathway. Studies by Reusch et al. (43) have examined the role of Akt in controlling differentiation in neonatal VSMC. In those cells, stimulation by serum or thrombin leads to a biphasic activation of ERKs (44). Activation of Akt inhibited sustained Raf activation, resulting in abolishment of the second, delayed phase of ERK activation critical for induction of SM myosin heavy chain promoter (43). Differences between these studies and our results may reflect inherent differences in the factors and pathways controlling VSMC differentiation between adult and neonatal cells. In adult VSMC, we have not observed a biphasic ERK activation, and chronic inhibition of ERKs does not affect regulation of SM-specific genes (11).
Second, we have shown that PDGF, again through an Aktdependent pathway decreases nuclear levels of SRF by accumulation of this transcription factor in a cytosolic compartment. The decreased levels of nuclear SRF would be expected to result in decreased binding to specific CArG boxes in the promoters of SM genes and suppression of expression of these genes. We propose that this mechanism accounts at least in part for the ability of PDGF to decrease basal expression of SM genes and accentuates the inhibition on AVP-induced activation of SM-specific genes. The time course of redistribution of SRF is much slower than that reported for Forkhead family members that are phosphorylated by Akt. This suggests that additional intermediate signaling events are involved. The FIG. 9. Signaling pathways controlling smooth muscle gene expression. Activation of PI3K pathway (p85/p110␣) by PDGF leads to activation of Akt. Activated Akt has inhibitory effects on the phosphorylation and activation of JNK and p38 MAP kinases, which are involved in the AVP-induced up-regulation of SM-specific gene expression, such as SM-␣-actin. Akt may also have direct effects on the promoter regulation through phosphorylation of transcription factors. Stimulation by AVP or PDGF activates p101/p110␥ isoform of PI3K, which does not lead to activation of Akt. Activation of PDGF receptors also leads to activation of the ras/cPLA2/eicosanoid pathway, inhibiting SM-specific gene expression. identity of these pathways is not known and is currently under investigation in numerous laboratories.
Changes in actin dynamics regulate the transcriptional activity of SRF-dependent genes in several muscle types (45,46) and nonmuscle cells (47). These effects are thought to be regulated by the Rho family of GTPases (45)(46)(47), which increases F-actin formation, eventually reducing the G-actin pool in the cell (48). Redistribution of SRF from the nucleus to the cytosol through changes in actin dynamics and by reduced Rho activity suggests that reduced SRF-dependent transcriptional activation may be correlated with its localization out of the nucleus (46,49). Based on these findings, we can propose that the activation of Akt by PDGF might be involved in the regulation of cytoskeletal rearrangement in our adult VSMC, an effect opposing Rho activation, leading to SRF redistribution to the cytosol. Alternatively, the effects of the PDGF/PI3K/Akt pathway may be direct ones where post-translational modifications of SRF either by Akt or by an Akt effector might initiate the translocation of SRF out of the nucleus. Although SRF has no reported nuclear export sequences, it was shown to have several nuclear localization sequences surrounded by putative phosphorylation sites (50, 51), one of which was shown to be phosphorylated by MAP kinase-associated kinase 2, a downstream effector of p38 MAP kinases (52). Thus, phosphorylation of SRF at Ser 103 , which we have shown to be increased in AVP-stimulated VSMC (11), may be required for its nuclear localization. Therefore, the inhibition of p38 MAP kinases by the PDGF/Akt pathway may cause redistribution of SRF out of the nucleus.
In summary, earlier studies have indicated that activation of the PI3K/Akt pathway in adult VSMC is involved in controlling growth and cell survival (53). Our data provide novel findings that this pathway also controls the phenotype of these cells through direct effects on SM-specific transcription.