Changes in the Balance of Phosphoinositide 3-Kinase/Protein Kinase B (Akt) and the Mitogen-activated Protein Kinases (ERK/p38MAPK) Determine a Phenotype of Visceral and Vascular Smooth Muscle Cells

Anti–PI3-K p85 subunit antiserum was purchased from Upstate Biotechnology. Polyclonal antibodies against PKB(Akt), ERK, JNK, p38MAPK, MEK1, and MKK6 were obtained from Santa Cruz Biotechnology. Monoclonal antibodies against c-Myc and Flag were purchased from Santa Cruz Biotechnology and Sigma Chemical Co., respectively.

Construction of the caldesmon promoter plasmid, GP3CAT, was described previously (Yano et al., 1995). The expression vector containing the constitutively active form of the c-Myc–tagged PI3-K p110α subunit (pCMV5p110αact) was kindly provided by Drs. H. Kurosu and T. Katada (Department of Physiological Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo). This cDNA was constructed by Hu et al. (1995) and was inserted downstream of the cytomegalovirus promoter, pCMV5. Expression vectors of constitutively active and dominant-negative forms of MEK1 and MKK6, and Flag-tagged ERK2 and Flag-tagged p38MAPK were kindly provided by Dr. K. Sugiyama (Boehringer Ingelheim), Drs. M. Hibi and T. Hirano (Osaka University, Medical School), and Dr. E. Nishida (Graduate School of Science, Kyoto University). Mutant cDNAs of MEK1 and MKK6 were constructed as described elsewhere (Mansour et al., 1994; Raingeaud et al., 1996), and were inserted downstream of the cytomegalovirus promoter of pCS2+ or the SRα promoter of pcDLSRα296. In this study, we used expression vectors constructed in pCS2+ for active and dominant-negative forms of MEK1 (pCS2+MEK1act and pCS2+MEK1DN, respectively) and MKK6 (pCS2+MKK6act and pCS2+MKK6DN, respectively). A PKB(Akt) cDNA was amplified by reverse transcriptase PCR using human placental mRNA as a template, and the accuracy of its sequence was checked. A PKB(Akt) cDNA thus obtained was inserted downstream of the cytomegalovirus promoter of pCS2+c-Myc–tagged (MT) (pCS2+MT-PKB(Akt)wt for expression of c-Myc–tagged wild-type PKB(Akt). The expression plasmid of c-Myc–tagged constitutively active form of PKB(Akt), pCS2+MT-PKB(Akt)act, was constructed as described previously (Dario et al. 1996).

Isolated gizzard SMCs were prepared from 15-d-old chick embryo gizzards as described elsewhere (Hayashi et al., 1998), and cultured on laminin-coated six-well plates with the indicated growth factors under kinase inhibited or stimulated conditions. Vascular SMCs were isolated from 5-wk-old rat aortae by enzyme-disperse methods as follows. Aortae were dissected under sterile conditions, minced well with scissors, and incubated at 37°C in 0.1% collagenase for 30 min, followed by incubation in the mixtures of 0.07% collagenase and 0.03% elastase for 90 min. Dispersed single cells were separated from undigested tissues by filtration, and were collected by centrifugation. The cells thus obtained were washed twice with growth factor–free basal medium (DME supplemented with 0.2% BSA), and were cultured in the medium containing IGF-I or PDGF-BB on laminin-coated culture plates. Treatment with specific inhibitors for ERK kinase (MEK1), PD98059 and/or for p38MAPK, SB203580, was performed as follows: gizzard or vascular SMCs were preincubated for 1 h in growth factor–free basal medium (DME supplemented with 0.2% BSA) containing the indicated amounts of inhibitors, and then stimulated with medium containing the indicated growth factors with or without inhibitors.

Ligand-induced contractility of cultured SMCs was monitored as follows. The SMCs were cultured under indicated conditions for 3 d, and then washed with PBS, followed by stimulation with basal culture medium containing 1 mM carbachol for 1 min. Contractility of cultured SMCs was observed with an Olympus microscope, and the same fields before and after carbachol treatment were photographed.

2 μg of total RNA from precultured or cultured SMCs under the indicated conditions were separated on 1.0% agarose-formaldehyde denaturing gels, and then transferred to nylon membranes. A caldesmon cDNA (GenBank M28417) fragment (nucleotides 286 to 810) and a calponin cDNA (GenBank M63559) fragment (nucleotides 1 to 867) were used as probes to monitor the expression of respective mRNAs. This caldesmon cDNA fragment, which contains parts of exons 2 and 3a is a common probe for the h- and l-caldesmons (Hayashi et al., 1991; Hayashi et al., 1992). In our previous studies using specific probes for h- or l-caldesmon, we demonstrated that the full lengths of h- and l-caldesmon mRNAs are 4.8 and 4.1 kb, respectively (Kashiwada et al., 1997; Obata et al., 1997). Probes were labeled with ³²P on the antisense strands and used for hybridization under the following conditions: 42°C for 16 h in 50% formamide, 6× SSC, 10× Denhardt's solution (1× Denhardt's solution is 0.02% polyvinylpyrrolidone and 0.02% BSA), 0.5% SDS, and 0.5 mg/ml denatured herring sperm DNA. The blots were washed in 0.1× SSC containing 0.1% SDS at 52°C, and visualized by autoradiography. To quantify the amount of RNA loaded, ribosomal RNAs were stained with 0.02% methylene blue.

Total protein of the cell lysates from SMC cultures was separated by SDS-PAGE and transferred to nitrocellulose membranes. Detection of target proteins on the membranes was performed using an ECL Western blotting detection kit (Amersham Pharmacia Biotech) with the indicated polyclonal antibodies.

Phospholipid mixtures (2 mg/ml) containing phosphatidylinositol (PI) and phosphatidylserine (PS) were dried under a stream of nitrogen, and sonicated in 10 mM Hepes (pH 7.4) in a bath sonicator at 0°C for 15 min. 10 μl of the resulting vesicles (PI/PS) were used as a substrate for PI3-K. The preparation of cell extracts and immunoprecipitation for PI3-K were performed at 4°C. The cultured cells were washed three times with ice-cold PBS, and then lysed in 550 μl of lysis buffer (20 mM Tris-HCl [pH 7.5], 137 mM NaCl, 1 mM MgCl₂, 1 mM CaCl₂, 1% NP-40, 50 mM NaF, 1 mM Na₃VO₄, 50 μg/ml PMSF, 1 μg/ml leupeptin, and 1 μg/ml pepstatin). After gentle shaking for 30 min, the cell extracts were obtained by centrifugation in a microfuge at 13,000 rpm for 5 min. The amount of PI3-K p85 subunit in the cell extracts was determined by Western blotting using antiserum against the PI3-K p85 subunit. The extracts containing equal amounts of PI3-K p85 subunit were precleaned with control rabbit IgG coupled protein A–Sepharose for 30 min. The PI3-K was immunoprecipitated with antiserum against the PI3-K p85 subunit followed by protein A–Sepharose. The immunoprecipitates were washed twice with lysis buffer, twice with 100 mM Tris-HCl (pH 7.5), 0.5 M LiCl, 1 mM DTT, and 0.2 mM Na₃VO₄, and three times with 10 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1 mM DTT, and 0.2 mM Na₃VO₄. All washes were performed at 4°C. The reaction mixtures (50 μl), containing the immunoprecipitates in 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM MgCl₂, 0.5 mM EGTA, 10 μM ATP, 5 μCi γ-[³²P]ATP, and 20 μg of PI/PS were incubated at 30°C for 10 min. The reactions were terminated and the lipids were extracted by the addition of CHCl₃/MeOH (1:2). The mixture was then vortexed and cleared by centrifugation. The extracted products were separated by thin-layer chromatography in a developing solution composed of CHCl₃/ MeOH/4 M NH₄OH (9:7:2). The production of phosphatidylinositol-3-phosphate was detected by autoradiography.

Cell lysis buffers were as follows: 50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 50 mM NaF, 1 mM Na₃VO₄, 10 mM β-glycerophosphate, 50 μg/ml PMSF, 1 μg/ml leupeptin, and 1 μg/ml pepstatin for the ERK and PKB(Akt) assays; 20 mM Tris-HCl (pH 7.5), 10% glycerol, 1 mM DTT, 120 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 50 mM NaF, 1 mM Na₃VO₄, 10 mM β-glycerophosphate, 50 μg/ml PMSF, 1 μg/ml leupeptin, and 1 μg/ml pepstatin for the JNK assays; and 20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM EDTA, 1% Triton X-100, 50 mM NaF, 2 mM Na₃VO₄, 10 mM β-glycerophosphate, 50 μg/ml PMSF, 1 μg/ml leupeptin, and 1 μg/ml pepstatin for the p38MAPK assays. The cell extracts were immunoprecipitated with specific antibodies against individual protein kinases, and the immunoprecipitates were washed thoroughly with their lysis buffers and then kinase assay buffers, and incubated with their respective substrates and 5 μCi γ-[³²P]ATP for 30 min at 30°C. The reaction products were analyzed by 15% SDS-PAGE. Reaction mixtures for the kinase assays were as follows: 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 1 μM protein kinase A inhibitor, 1 mM DTT, and 25 μg histone H2B for the PKB(Akt) assay; 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 1 μM protein kinase A inhibitor, 1 mM DTT, and 25 μg myelin basic protein (MBP) for the ERK assay; 50 mM Tris-HCl (pH 7.5), 10 mM MgCl₂, 1 μM protein kinase A inhibitor, 1 mM DTT, 1 mM Na₃VO₄, and 1 μg GST-Jun (1-79) for the JNK assay; and 20 mM Hepes (pH 7.4), 20 mM MgCl₂, 20 mM β-glycerophosphate, 1 μM protein kinase A inhibitor, 2 mM DTT, and 1 μg GST-ATF2 (1-96) for the p38MAPK assay.

The caldesmon promoter activity was analyzed using the chloramphenicol acetyltransferase (CAT) construct, GP3CAT, according to the method described previously (Yano et al., 1995; Obata et al., 1997). The SMCs prepared as described above were seeded onto laminin-coated six-well plates, and cultured in the indicated medium for 1 or 3 d. Transfection was carried out using Trans IT™-LT1, polyamine transfection reagents (Pan Vera Corporation). Complex mixtures composed of 10 μg of trans IT™- LT1 reagent and 2 μg of GP3CAT, 1 μg of control plasmid carrying the luciferase gene under the Rous sarcoma virus (RSV) promoter (RSV-luciferase), and 1 μg of either control expression plasmid (pCMV5, pCS2+, or pCS2+MT), expression plasmid carrying a c-Myc–tagged constitutively active form of PI3-K p110α subunit, a c-Myc–tagged wild-type or a constitutively active form of PKB(Akt) (pCMV5p110αact, pCS2+ MT-PKB(Akt)wt or pCS2+MT-PKB(Akt)act), or either or both of expression vectors carrying constitutively active and dominant-negative forms of MEK1 (pCS2+MEK1act and pCS2+MEK1DN) and MKK6 (pCS2+MKK6act and pCS2+MKK6DN), were added to the cells in Opti minimum Eagle's medium (GIBCO BRL). After a further 4-h incubation, the medium was replaced with DME supplemented with 0.2% BSA plus 2 ng/ml IGF-I or 20 ng/ml PDGF-BB, and the transfected cells were harvested 48 h later. Standardization of transfection efficiency was performed by measuring luciferase activity as described previously (Yano et al., 1995; Obata et al., 1997). The cell extracts containing equal amounts of luciferase activity were used for the CAT assay. The transfection experiments were repeated at least three times on duplicate cultures with two or three different plasmid preparations. The CAT activities were quantified by Scanning Imager (Molecular Dynamics).

The effects of forced expression of MEK1 and MKK6 in cultured SMCs were analyzed as follows. The indicated amounts of control expression plasmid and either or both of expression plasmids carrying active or dominant-negative MEK1 and MKK6 were transfected into cultured SMCs together with 1 μg of a reporter plasmid carrying the β-galactosidase gene downstream from the SV-40 early promoter. Total RNA was isolated from the transfected cells and the expression levels of caldesmon and calponin mRNAs were analyzed by Northern blotting as described above. Transfection efficiencies were determined by staining for β-galactosidase activity from the reporter plasmid using 5-bromo-4-chloro-3-indolyl-β-d-galactoside (X-gal) as a substrate.

Transfection was carried out as described above in Promoter Analysis and Transfection. In the cases of PI3-K and PKB(Akt) assays, SMCs were transfected with 3 μg of respective expression plasmids of constitutively active form of c-Myc–tagged PI3-K p110α subunit, pCMV5p110αact, or wild-type or constitutive active form of c-Myc–tagged PKB(Akt), pCS2+MT-PKB(Akt)wt or pCS2+MT-PKB(Akt)act. Two micrograms of expression plasmid of each Flag-tagged ERK2 or Flag-tagged p38MAPK was cotransfected with 2 μg of either expression plasmid of active or dominant-negative MEK1 or MKK6, or control plasmid. In both cases, SMCs were cultured under nonstimulated conditions after transfection. 2 d later, SMCs were stimulated under indicated conditions. The cell extracts containing the equal amounts of proteins were precleaned with control mouse IgG coupled protein G–Sepharose for 30 min and immunoprecipitated with monoclonal antibody against c-Myc or Flag followed by protein G–Sepharose. The kinase activities were determined as described above in PI3-K assay and other protein kinase assays.

Conditioned medium obtained from SMCs transfected with both expression plasmids carrying active MEK1 and MKK6 was filtered through a 0.22-μm membrane. The conditioned medium was heated to 100°C for 15 min or treated with trypsin (30 μg/ml) for 3 h at 30°C, followed by the addition of trypsin inhibitor at a 10-fold excess. The heat- or trypsin-treated conditioned medium was dialyzed against DME supplemented with 0.2% BSA at 4°C for 16 h, and adjusted to the concentration of IGF-I to 2 ng/ml. The SMCs were cultured in these medium for 3 d. Heparin–Sepharose affinity chromatography was carried out as follows. One ml of 50% slurry of heparin–Sepharose (Amersham Pharmacia Biotech) equilibrated with PBS was added to 20 ml of the conditioned medium and gently agitated for 5 h at 4°C. The mixture was poured over a 1-ml Prep Column (Bio-Rad Labs.), and the follow through fraction (non–heparin-binding) was collected. The column was rinsed with 10 vol of PBS and eluted stepwise with 1 ml of PBC containing NaCl (0.5, 1.0, and 1.5 M). Each fraction was collected and desalted by dialysis against DME supplemented with 0.2% BSA through a membrane of 3-kD cutoff (Spectrum). Aliquots of each fraction were diluted (1:4) with DME supplemented with 0.2% BSA and then added to SMC cultures. Treatment with specific inhibitor of EGF receptor kinase, AG1478, was performed as follows: SMCs cultured for 2 d under IGF-I–stimulated conditions were preincubated for 1 h in DME supplemented with 0.2% BSA containing 1 μM AG1478, and then stimulated with the conditioned medium containing the same concentration of this drug for 3 d. Total RNAs from cultured SMCs were extracted and the expression patterns of caldesmon and calponin mRNAs were analyzed by Northern blotting. Control medium was obtained from culture supernatant of SMCs transfected with expression plasmid alone.

We have recently established a novel culture system of gizzard SMCs in which they maintain a differentiated phenotype for a long culture period. Of growth factors and cytokines examined, IGF-I is the most potent for maintaining the differentiated SMC phenotype as defined by the expression of SMC-specific molecular markers, cell morphology, and function (Hayashi et al., 1998). On the other hand, PDGF-BB, bFGF, and EGF potently induce SMC de-differentiation (Fig. 1). In 3- and 5-d–cultured SMCs stimulated with PDGF-BB, bFGF, or EGF, h-caldesmon mRNA converts to l-caldesmon mRNA, and total h- and l-caldesmon mRNAs decrease to 20% of the levels seen in precultured SMCs. Calponin mRNA is also downregulated to a negligible level (Fig. 1 A). However, the levels of h-caldesmon and calponin mRNAs in cultured SMCs under IGF-I–stimulated conditions are identical with those seen in precultured cells (Fig. 1 A). Similar results are obtained using α- and β-tropomyosins (Kashiwada et al., 1997) and α1 integrin (Obata et al., 1997), which are also SMC-specific molecular markers (data not shown). With regard to cell morphology and function, cultured SMCs under IGF-I–stimulated conditions showed a spindle-like shape, formed a meshwork structure, and displayed carbachol-induced contraction. In contrast, SMCs stimulated with PDGF-BB, bFGF, or EGF showed a fibroblast-like shape and lost the contractility (Figs. 1 B and 4), indicating that PDGF-BB, bFGF, and EGF are potent factors for inducing SMC de-differentiation.

Our previous studies revealed that the PI3-K/PKB(Akt) pathway triggered by IGF-I plays a critical role in maintaining the differentiated SMC phenotype (Hayashi et al., 1998). To investigate the downstream signaling pathways involving in SMC de-differentiation triggered by PDGF-BB, bFGF, or EGF, we analyzed several kinases including ERK, JNK, p38MAPK, PI3-K, and PKB(Akt). PDGF-BB, bFGF, and EGF all activated ERK (Fig. 2 A) and p38MAPK (Fig. 2 C). Their maximum activations were found at 10 min after stimulation. bFGF and EGF also activated JNK, whereas PDGF-BB did not (Fig. 2 B). IGF-I had no effect on ERK (Fig. 2 A), JNK (Fig. 2 B), and p38MAPK (Fig. 2 C). These data suggest that growth factors inducing SMC de-differentiation coordinately activate the ERK and p38MAPK pathways.

As demonstrated previously (Hayashi et al., 1998), IGF-I potently activated PI3-K (Fig. 2 D) and PKB(Akt) (Fig. 2 E); the maximum activation of PI3-K was achieved at 10 min after IGF-I stimulation and this activation reduced thereafter, while the activation of PKB(Akt) by IGF-I (2 ng/ml) lasted for more than 180 min. The activation of PI3-K and PKB(Akt) by IGF-I was suppressed by specific PI3-K inhibitors, wortmannin or LY249002 (data not shown), indicating that the PKB(Akt) activation exclusively depends on the PI3-K activity. No significant activation of PI3-K and PKB(Akt) was observed in SMCs stimulated by either bFGF or EGF (Fig. 2, D and E). Among the three growth factors inducing SMC de-differentiation, PDGF-BB was the only one that could activate PI3-K and PKB(Akt) (Fig. 2, D and E). The PKB(Akt) activation by PDGF-BB was more potent than that by IGF-I at 15 min after growth factor stimulation, whereas this activation was transient, but retained at a substantial level for 180 min. By contrast, the PKB(Akt) activation by IGF-I was sustained at a high level for more than 180 min. These results suggest the possibility that in addition to PI3-K, PDGF-BB activates PKB(Akt) mediated through another unknown cascade.

It is curious that PDGF-BB, which is a potent factor inducing SMC de-differentiation (Fig. 1), triggered the dual signaling pathways mediated through both PI3-K/PKB(Akt) and two MAPKs, ERK and p38MAPK (Fig. 2). To simplify the PDGF-BB–triggered signaling pathways, we examined the effects of specific MAPK inhibitors, PD98059 for ERK kinase (MEK1) and SB203580 for p38MAPK, on the PDGF-BB–stimulated SMC phenotype. Either PD98059 or SB203580 specifically inhibited the PDGF-BB–induced activation of ERK or p38MAPK, respectively, to near basal levels (Fig. 3 A), but had no effect on PI3-K and PKB(Akt) (data not shown). Treatment with either PD98059 or SB203580 only slightly suppressed the PDGF-BB–induced isoform conversion of caldesmon mRNA and downregulation of caldesmon and calponin mRNAs (Fig. 3 B). However, simultaneous treatment with both drugs strongly suppressed the PDGF-BB–induced SMC de-differentiation as monitored by the expression of caldesmon and calponin mRNAs (Fig. 3 B). In addition to these molecular events, both drugs could also rescue the morphological alteration from a spindle-like shape to a fibroblast-like shape change and a loss of contractility induced by PDGF-BB (Fig. 4). As a control, treatment with individual drug resulted in less significant effect on cell morphology and function. Table I shows the effects of PD98059 and/or SB203580 on carbachol-stimulated contractility of SMCs under various culture conditions. Further, both drugs only slightly delayed the induction of SMC de-differentiation by bFGF or EGF, but did not prevent SMC de-differentiation (data not shown).

Promoter analyses of the caldesmon gene further support these findings (Fig. 5). We used the caldesmon promoter/CAT construct, GP3CAT, which produces the high promoter activity in differentiated SMCs (Yano et al., 1995). The promoter activity in SMCs stimulated by PDGF-BB reduced to 30% of that by IGF-I (Fig. 5 C). Even under PDGF-BB–stimulated conditions, inhibition of both the ERK and p38MAPK pathways by their specific inhibitors or the forced expression of active PI3-K (Fig. 5 A) or active PKB(Akt) (Fig. 5 B) could protect such reduction (Fig. 5 C). These results suggest that PDGF-BB displays the dual function in maintaining the differentiated SMC phenotype mediated through the PI3-K/PKB(Akt) pathway and inducing SMC de-differentiation by the ERK and p38MAK pathways. Thus, changes in the balance between the strengths of the PI3-K/PKB(Akt) pathway and the ERK and p38MAPK pathways would determine the SMC phenotype.

The MAPK signaling cascades are involved in a variety of cell functions (Force and Bonventre, 1998). Dual phosphorylation on Thr and Tyr within the Thr-Xaa-Tyr motif catalyzed by MAPK kinases is essential for MAPK activation (Davis, 1994). MEK1 and MKK6 are specific upstream kinases for ERK and p38MAPK, respectively (Cohen, 1997). To investigate the direct involvement of ERK and p38MAPK in SMC de-differentiation, we examined the effects of active or dominant-negative MEK1 and/or MKK6 on the caldesmon promoter activity in SMCs under IGF-I–stimulated conditions. We determined the respective MAPK kinase activity by in vitro kinase assay (Fig. 6 A). In this experiment, expression plasmids carrying active or dominant-negative MAPK kinases were cotransfected with expression plasmid carrying Flag-tagged ERK or Flag-tagged p38MAPK into cultured SMCs, and Flag-tagged proteins were immunoprecipitated from the cell lysates with anti-Flag monoclonal antibody. The ERK or p38MAPK activities were potently enhanced in SMCs cotransfected with active MEK1 or MKK6 under nonstimulated or PDGF-BB–stimulated conditions (Fig. 6 A, a and b). Even when SMCs were stimulated by PDGF-BB, their enhancements were strongly abolished by the forced expression of dominant-negative MEK1 or MKK6 (Fig. 6 A, a and b). The expressions of active and dominant-negative MEK1 or MKK6 proteins were confirmed by immunoblotting (data not shown). The caldesmon promoter activity in differentiated SMCs under IGF-I–stimulated conditions was analyzed by the forced expression of active or dominant-negative MAPK kinases (Fig. 6 B). The promoter activity was not affected by either or both dominant-negative MEK1 and/or MKK6. Transfection with either active MEK1 or MKK6 significantly reduced the caldesmon promoter activity, while cotransfection with both active kinases further suppressed the promoter activity. Since the SV-40 promoter was not affected by the forced expression of MEK1 and/or MKK6 (data not shown), suppression of the caldesmon promoter activity by active MEK1 and MKK6 was specific. These results indicate that the ERK and p38MAPK pathways are directly involved in the induction of SMC de-differentiation.

We further examined the effects of active and dominant-negative MEK1 or MKK6 on the endogenous expression of caldesmon and calponin mRNAs and on cell morphology. Transfection with either active kinase alone or cotransfection with dominant-negative kinases had less significant effects on caldesmon and calponin mRNAs in 2- or 4-d–cultured SMCs (Fig. 7 A, lanes 1, 2, 4–8, and 10– 12). Even under IGF-I–stimulated conditions, cotransfection of active MEK1 and MKK6 potently induced isoform conversion of caldesmon mRNA and downregulation of caldesmon and calponin mRNAs. These changes progressed during culture; the endogenous expressions of caldesmon and calponin mRNAs in 4-d–cultured SMCs after cotransfection were comparable to those seen in 4-d–cultured SMCs stimulated by PDGF-BB (Fig. 7 A, lanes 3, 9, and 13). These results indicate that after cotransfection with active MEK1 and MKK6, almost all of cultured SMCs alter their phenotype to de-differentiation. To compare transfection efficiency and cell morphology, cultured SMCs stimulated by IGF-I were transfected with control plasmid (Fig. 7 B, a) or both expression plasmids carrying active MEK1 and MKK6 (Fig. 7 B, b) together with β-galactosidase expression plasmid. As revealed by β-galactosidase staining, transfection efficiencies were ∼25%. The SMCs transfected with control vector remained to show a spindle-like shape (Fig. 7 B, a). In the case of SMCs cotransfected with active MEK1 and MKK6, all of the β-galactosidase-stained and -unstained cells converted from a spindle-like shape to a fibroblast-like shape (Fig. 7 B, b). The transfection efficiency of both active MEK1 and MKK6 was correlated with SMC-specific marker gene expression and cell morphology (Fig. 7, C and D). These data suggest that SMCs cotransfected with active MEK1 and MKK6 secrete some factor(s) which induces de-differentiation of surrounding normal SMCs.

To further characterize such a factor(s), conditioned medium (CM1) obtained from SMCs cotransfected with expression plasmids carrying active MEK1 and MKK6 was prepared. Fig. 8 A shows the expression of caldesmon and calponin mRNAs in 3-d–cultured SMCs stimulated with the conditioned medium. Even when SMCs were stimulated by IGF-I, the conditioned medium potently induced isoform conversion of caldesmon mRNA and downregulation of caldesmon and calponin mRNAs (Fig. 8 A, lane 1), whereas control medium (CM2) obtained from SMCs transfected with control plasmid alone did not (Fig. 8 A, lane 2). Further, the conditioned medium enhanced the ERK, JNK, and p38MAPK activities (Fig. 8 B). These results exclude the possibility of retransfection with residual expression vectors, because MEK1 and MKK6 proteins derived from expression plasmids were not detected in cultured SMCs by immunoblotting (data not shown). Thus, this study revealed that the forced activation of both ERK and p38MAPK in SMCs induces the secretion of some factor(s) which initiates de-differentiation of surrounding normal SMCs in a paracrine manner. Heat or trypsin treatment of the conditioned medium completely abolished the activity inducing SMC de-differentiation (Fig. 8, lanes 3 and 4), suggesting that the factor(s) is a protein in nature. The conditioned medium was further fractionated using a heparin–Sepharose column. The flow through (non–heparin-binding) fraction retained the potent de-differentiation activity, while the eluted fraction by 0.5, 1.0, and 1.5 M NaCl did not (Fig. 8 A, lanes 5–8). PDGFs and bFGF show heparin-binding abilities; the former was eluted with 0.5 M NaCl and the latter with 1.5 M NaCl (data not shown). By contrast, EGF is known as a non–heparin-binding growth factor, suggesting a candidate for the active protein factor. A specific inhibitor for EGF receptor kinase, AG1478, only slightly inhibited the conditioned medium-induced de-differentiation, but this effect was less significant (Fig. 8 A, lane 9). Therefore, the active protein factor(s), which induces SMC de-differentiation, in the conditioned medium is considered to be different from PDGFs, bFGF, and EGF.

To examine whether vascular SMCs could be regulated by the same signaling pathways as revealed in gizzard SMCs, we first applied our culture system of gizzard SMCs to vascular SMCs. We isolated rat vascular SMCs by the enzyme-disperse method, and cultured them on laminin-coated plates under IGF-I–stimulated conditions. Because of difficulty to obtain a lot of cell numbers from rat aortae, we observed cell morphology and ligand-induced contractility to determine the vascular SMC phenotype (Fig. 9). Vascular SMCs could also maintain a spindle shape for more than 2 weeks under IGF-I–stimulated conditions, and showed ligand-induced contractility (Fig. 9, a and f). A blockade of the PI3-K pathway by LY294002 resulted in a morphological change from a spindle shape to a fibroblast-like shape and a loss of contractility (Fig. 9, b and g). In-consistent with the case of gizzard SMCs, PDGF-BB rapidly induced de-differentiation of vascular SMCs as monitored by cell morphology and ligand-induced contractility (Fig. 9, c and h). Even under PDGF-BB–stimulated conditions, simultaneous treatment with PD98059 and SB203580 could retain a spindle shape and contractility (Fig. 9, d and i). As a control, treatment with each drug individually was less significant effect on the PDGF-BB– induced de-differentiation (data not shown). Furthermore, the conditioned medium obtained from gizzard SMCs transfected with both active MEK1 and MKK6 remarkably induced de-differentiation of vascular SMCs (Fig. 9, e and j). Therefore, these data also suggest that the PI3-K–mediated signaling pathway plays a vital role in maintaining a differentiated phenotype of vascular SMCs and the ERK and p38 MAPK pathways are coordinately involved in de-differentiation of vascular SMCs.

Under pathological conditions, phenotype of SMCs can change from a differentiated state to a de-differentiated state in vivo and in vitro. During de-differentiation, SMCs show dramatic and irreversible alterations in their cell shape, function, and expression of SMC-specific molecular markers. Long spindle-shaped cells change to fibroblast-like cells, accompanied by losses in a ligand-induced contractility and SMC-specific molecular marker expression. Since there has not been a primary culture system available for SMCs or SMC-derived cell lines in which they can maintain a fully differentiated phenotype, the intracellular signaling pathways regulating the SMC phenotype have not been well characterized. Recently, we established a novel culture system in which gizzard SMCs can maintain a differentiated phenotype for a long culture time (Hayashi et al., 1998). In this culture system, IGF-I is the most potent for maintaining the differentiated SMC phenotype, and the IGF-I–triggered signaling pathway, PI3-K/PKB- (Akt), plays a critical role in this maintenance. Here, we investigated the signaling pathways inducing SMC de-differentiation and compared them with the PI3-K/PKB- (Akt) pathway.

It has been reported that PDGF, EGF, bFGF, or angiotensin II enhance cell proliferation or hypertrophy through the activation of the ERK signaling cascade in passaged vascular SMCs (Force and Bonventre, 1998). MAPKs such as JNK and p38MAPK are also activated in response to various cellular stresses (Derijard et al., 1994; Kyrian and Avruch, 1996). Angiotensin II and phenylephrine, which induce acute hypertension, enhance the ERK and JNK activities in aortic, carotid and femoral arteries (Xu et al., 1996), and endothelin activates both of these kinases in proliferative airway SMCs (Shapiro et al., 1996). p38MAPK, ERK, and/or JNK are also activated by balloon injury of carotid arteries (Hu et al., 1997; Pyles et al., 1997; Larrivee et al., 1998). These findings suggest a close association of MAPK cascades with smooth muscle disorders. However, the direct involvement of these signaling cascades in regulating the SMC phenotype has been unknown. As a first step, we used a novel culture system of gizzard SMCs and demonstrated that activations of the PI3-K/PKB(Akt) pathway and the ERK and p38MAPK pathways are directly involved in maintaining SMC differentiation and inducing SMC de-differentiation, respectively. This conclusion is based on the following findings. First, the signaling pathways in regulating the phenotypic determination of gizzard SMCs were distinctly different; the PI3-K/PKB(Akt) pathway played a critical role in maintaining the differentiated SMC phenotype (Figs. 2 and 5) and the ERK and p38MAPK pathways triggered by PDGF-BB, bFGF, and EGF were closely associated with SMC de-differentiation (Figs. 1 and 2). Second, among the three growth factors inducing SMC de-differentiation, PDGF-BB only triggered both the PI3-K/PKB(Akt) pathway and the ERK and p38MAPK pathways. When both the MAPK pathways were blocked by their specific inhibitors, PD98059 and SB203580, or when SMCs were transfected with active PI3-K or PKB(Akt), PDGF-BB in turn initiated to maintain the differentiated SMC phenotype (Figs. 3–5 and Table I). Third, even when SMCs were cultured under IGF-I–stimulated conditions, the forced activation of both the MAPK pathways by the coexpression of active MEK1 and MKK6 potently induced SMC de-differentiation (Fig. 6). Fourth, SMCs cotransfected with active MEK1 and MKK6 secreted a nondialyzable and heat-labile protein factor(s), which induced de-differentiation of surrounding normal SMCs (Figs. 7 and 8). Finally, the same signaling pathways as described above were observed to be involved in regulating the vascular SMC phenotype (Fig. 9).

Since IGF-I enhances the proliferation and migration of cultured vascular SMCs (Clemmons 1985; Bornfeldt et al., 1990; Cercek et al., 1990; Delafontaine et al., 1991; Bornfeldt et al., 1994), it has been considered to be an important growth factor in the progression of atherosclerosis. These findings are, however, obtained using passaged SMCs. In this paper, we demonstrated using a novel SMC culture system that IGF-I solely triggers the PI3-K/PKB(Akt) pathway, but not the MAPK pathways (Fig. 2). Further, IGF-I did not affect the proliferation of SMCs (data not shown). These properties of IGF-I signaling made it possible to maintain the differentiated SMC phenotype for more than 2 wk in primary culture. Since rapamycin had no effect on the differentiated SMC phenotype under IGF-I–stimulated conditions (Hayashi et al., 1998), p70 ribosomal S6 kinase (p70^S6K) is unlikely to be a downstream target of PI3-K/PKB(Akt). Further study is required to identify the downstream targets of PI3-K/ PKB(Akt) in SMCs. It has been reported that IGF-I and its downstream signaling, PI3-K/PKB(Akt), play a role in protection against programmed cell death (Alessi and Cohen 1998). We observed neither a significant decrease in cell numbers nor DNA fragmentation in our SMC cultures even under nonstimulated conditions or in the presence of IGF-I neutralizing antibodies (data not shown). Therefore, cultured SMCs might secrete anti-apoptotic factor(s) in the absence of IGF-I, and the maintenance of the differentiated SMC phenotype by IGF-I would be distinct from an anti-apoptotic action. In passaged SMCs, IGF-I enhanced the JNK and p38MAPK activities (data not shown). These results suggest that the downstream signalings of IGF-I might be modulated during cell passage. Actually, differentiated SMCs rapidly change their phenotype under serum-stimulated conditions (Kashiwada et al., 1997; Hayashi et al., 1998). Since passaged SMCs do not represent a stable differentiated state, studies reported previously might not be able to reveal the IGF-I's function and signaling in differentiated SMCs. In this study, we used SMCs that showed well-characterized and stable phenotypes. We analyzed the relationship between the modulation of SMC phenotype and cell proliferation. Although serum-induced SMC de-differentiation was concomitant with cell proliferation, other growth factors that trigger SMC de-differentiation, such as PDGF-BB, bFGF, and EGF, did not significantly induce SMC proliferation (data not shown). This result also suggests that SMC de-differentiation is not essentially associated with cell proliferation.

It has been reported that de-differentiated SMCs produce and secrete PDGF which further promotes cell proliferation and migration in an autocrine/paracrine manner (Sjolund et al., 1988). PDGF is also known to promote the activation of ERK and p38MAPK in passaged SMCs (Bornfeldt et al., 1994; Pyne and Pyne, 1997). In our culture system, PDGF-BB triggered the dual signaling pathways mediated by PI3-K/PKB(Akt) and two MAPKs, ERK and p38MAPK. Under these culture conditions, PDGF-BB stimulation resulted in SMC de-differentiation. When the two MAPK pathways were blocked by their specific inhibitors, PD98059 and SB203580, PDGF-BB stimulation, in turn, initiated to maintain SMC differentiation (Figs. 3 and 4). bFGF and EGF are known to activate ERK and also to induce proliferation of SMCs (Berrou et al., 1996; Jones et al., 1997; Yu et al., 1997; Miyamoto et al., 1998). In our culture system, both growth factors activated ERK and p38MAPK and potently induced SMC de-differentiation (Figs. 1 and 2). However, PD98059 and SB203580 could not prevent such de-differentiation (data not shown). This is because the signaling pathway mediated by PI3-K/PKB(Akt) was not activated by bFGF or EGF (Fig. 2). In the present culture system, bFGF and EGF also activated JNK, but IGF-I and PDGF-BB did not (Fig. 2). It is, therefore, unlikely that JNK is involved in regulating the SMC phenotype. PDGF-BB reduced the caldesmon promoter activity and this reduction could be overcome by the forced expression of active PI3-K or PKB(Akt), or by treatment with both PD98059 and SB203580 (Figs. 3–5 and Table I). Further, the activation of ERK and p38MAPK by the forced expression of both active MEK1 and MKK6 could overcome the PI3-K/ PKB(Akt) pathway triggered by IGF-I, resulting in the induction of SMC de-differentiation (Fig. 6). These data support our hypothesis that the SMC phenotype would be determined by the balance between the strengths of the signaling pathways mediated by PI3-K/PKB(Akt) and by ERK and p38MAPK.

Curiously, even though the transfection efficiencies of both expression vectors carrying active MEK1 and MKK6 were only 25%, almost all of SMCs came to de-differentiate as monitored by cell morphology and endogenous expression of SMC-specific molecular markers (Fig. 7). These findings indicate that the production and secretion of some de-differentiation-inducing factor(s) occur in SMCs in which both the ERK and p38MAPK pathways are constitutively activated. We have not yet identified such a factor(s), but the conditioned medium from these cells activated three MAPK pathways (ERK, p38MAPK, and JNK). The activating factor(s) was a heat-labile, non– heparin-binding protein factor (Fig. 8). From these biochemical properties (Fig. 8), we excluded the possibility that PDGF-BB, bFGF, and EGF would be the main factor inducing SMC de-differentiation in the conditioned medium. Further study is required to identify such a factor(s). Anyway, this study provides a evidence that only a small portion of de-differentiated SMCs secretes a protein factor(s) for the surrounding normal SMCs to be de-differentiated. These findings could be helpful to understand the progression of smooth muscle disorders such as atherosclerosis.

We then applied a culture system of gizzard SMCs to that of vascular SMCs and investigated the signaling pathways in regulating the vascular SMC phenotype (Fig. 9). Like gizzard SMCs, IGF-I potently maintained a differentiated phenotype of vascular SMCs and a specific PI3-K inhibitors, LY24002, prevented this IGF-I's action. Treatment with two MAPK pathway inhibitors, PD98059 and SB203580, could rescue the PDGF-BB–induced de-differentiation of vascular SMCs. Further, the conditioned medium obtained from gizzard SMCs also induce de-differentiation of vascular SMCs. These results strongly suggest that a culture system of gizzard SMCs is applicable for that of vascular SMCs and that the distinct signaling pathways mediated by PI3-K/PKB(Akt) and two MAPKs are also involved in the phenotypic determination of vascular SMCs.

Our results presented here are summarized in Fig. 10. The signaling pathway mediated by PI3-K/PKB(Akt) is required to maintain a differentiated phenotype of visceral and vascular SMCs, while the activation of ERK and p38MAPK leads to SMC de-differentiation of both types of SMCs. Thus, the signaling pathways in regulating the phenotypic determination are considered to be essentially the same in visceral and vascular SMCs, and the SMC phenotype would be regulated by the balance between the strengths of these signaling pathways. Although visceral and vascular SMCs originate from different precursors, our present findings provide a further insight in the common molecular mechanism of phenotypic determination of two types of SMCs. This is because visceral and vascular SMCs have common characteristics with respect to cell structure, function, and expression of molecular markers as follows. The main function of both types of SMCs is contraction, which is Ca²⁺-dependent and controlled by a myosin-linked, actin-linked dual regulation (Sobue et al., 1988, 1991). Both types of SMCs are rich in myofibrils and are organized in a three-dimensional direction with two prominent electron dense structures such as the dense body in the cytoplasm and the dense membrane (dense plaque) in cell–cell contact. Further, contractile and cytoskeletal proteins are also specifically expressed in and serve as specific molecular markers for differentiated SMCs. The expression patterns of these molecular markers are identical in both visceral and vascular SMCs and their expression mechanisms, including transcription and splicing, are also regulated in common ways (Owens 1995; Sobue et al., 1998). Further studies are required to understand the detailed signaling pathways in regulating the vascular SMC phenotype and the functional linkage between such signaling pathways and SMC-specific gene regulation machineries, and to apply these findings to SMC disorders.

We thank Drs. H. Kurosu and T. Katada (Department of Physiological Chemistry, Graduate School of Pharmaceutical Sciences, University of Tokyo) for providing a constitutive active form of PI3-kinase p110α subunit cDNA, and also thank Dr. K. Sugiyama (Boehringer Ingelheim), Drs. M. Hibi and T. Hirano (Osaka University, Medical School), and Dr. E. Nishida (Graduate School of Science, Kyoto University) for kindly providing expression vectors carrying active and negative forms of MEK1 and MKK6, or Flag-tagged ERK2 and p38MAPK.

This work was supported by Grants-in-aid for COE Research (to K.S.) and in part by Grants-in-aid for Scientific Research from the Ministry of Education, Science, Sports and Culture of Japan.

bFGF

basic fibroblast growth factor

CM

conditioned medium

CAT

chloramphenicol acetyltransferase

ERK

extracellular signal-regulated kinase

IGF-I

insulin-like growth factor-I

JNK

c-Jun NH₂-terminal protein kinase

MAPKs

mitogen-activated protein kinases

MBP

myelin basic protein

MT

c-Myc-tag

p70^S6K

p70 ribosomal S6 kinase

PI

phosphatidylinositol

PI3-K

phosphatidylinositol 3-kinase

PKB

protein kinase B

PS

phosphatidylserine

RSV

Rous sarcoma virus

SMC

smooth muscle cell

X-gal

5-bromo-4-chloro-3-indolyl-β-d-galactoside