Cell Physiology

The mTOR/p70 S6K1 pathway regulates vascular smooth muscle cell differentiation

Kathleen A. Martin, Eva M. Rzucidlo, Bethany L. Merenick, Diane C. Fingar, David J. Brown, Robert J. Wagner, Richard J. Powell


Vascular smooth muscle cells (VSMC) in mature, normal blood vessels exhibit a differentiated, quiescent, contractile morphology, but injury induces a phenotypic modulation toward a proliferative, dedifferentiated, migratory phenotype with upregulated extracellular matrix protein synthesis (synthetic phenotype), which contributes to intimal hyperplasia. The mTOR (the mammalian target of rapamycin) pathway inhibitor rapamycin inhibits intimal hyperplasia in animal models and in human clinical trials. We report that rapamycin treatment induces differentiation in cultured synthetic phenotype VSMC from multiple species. VSMC treated with rapamycin assumed a contractile morphology, quantitatively reflected by a 67% decrease in cell area. Total protein and collagen synthesis were also inhibited by rapamycin. Rapamycin induced expression of the VSMC differentiation marker contractile proteins smooth muscle (SM) α-actin, calponin, and SM myosin heavy chain (SM-MHC), as observed by immunoblotting and immunohistochemistry. Notably, we detected a striking rapamycin induction of calponin and SM-MHC mRNA, suggesting a role for mTOR in transcriptional control of VSMC gene expression. Rapamycin also induced expression of the cyclin-dependent kinase inhibitors p21cip and p27kip, consistent with cell cycle withdrawal. Rapamycin inhibits mTOR, a signaling protein that regulates protein synthesis effectors, including p70 S6K1. Overexpression of p70 S6K1 inhibited rapamycin-induced contractile protein and p21cip expression, suggesting that this kinase opposes VSMC differentiation. In conclusion, we report that regulation of VSMC differentiation is a novel function of the rapamycin-sensitive mTOR signaling pathway.

  • rapamycin
  • contractile proteins
  • phenotypic modulation
  • signal transduction
  • intimal hyperplasia

intimal hyperplasia is an important clinical problem associated with most vascular interventions. This can lead to restenosis in up to 40% of coronary balloon angioplasties or arterial bypass grafts (25). These secondary lesions are often of greater severity than the original lesion and require repeated intervention. Intimal hyperplasia is precipitated by vessel injury and may be the result of an uncontrolled healing response. Vascular smooth muscle cells (VSMC) in normal, mature vessels are quiescent, contractile, and differentiated but assume a proliferative, migratory, and protein synthetic phenotype in response to injury (2). VSMC proliferation, migration to the intima, and synthesis of extracellular matrix (ECM) after injury is an important component of intimal hyperplasia. Thus understanding the processes that control VSMC differentiation may contribute to a better understanding of how pharmacological interventions may limit restenosis.

VSMC do not terminally differentiate but retain plasticity to dedifferentiate toward a proliferative, synthetic phenotype in response to injury. As the primary function of mature VSMC is contraction, the complement of contractile, structural, and regulatory proteins expressed in fully differentiated VSMC provides markers of differentiation status (37). These include smooth muscle (SM) myosin heavy chain (SM-MHC), calponin, and SM α-actin. The SM-MHC SM2 isoform is considered the most stringent marker of the differentiated state (19). Expression of these proteins decreases as VSMC dedifferentiate and myofilaments break down. Differentiation is often accompanied by expression of the cyclin/CDK inhibitors p21cip (waf-1) and p27kip (44).

Many factors may contribute to VSMC dedifferentiation after injury, including denudation of the endothelium, altered contacts with ECM and integrins, and exposure to growth factors. Culturing VSMC in vitro mimics this progression, as primary cultures rapidly lose differentiation markers and exhibit a synthetic phenotype (10). Such cell culture models have provided insights into factors that influence VSMC phenotype. Exposure to growth factors including PDGF, EGF, and TGFβ, or unsaturated lysophosphatidic acid promotes dedifferentiation (10, 12, 37), but G protein-coupled receptor agonists in serum, including thrombin, can induce VSMC differentiation via Gβγ subunits (32). TGFβ can also promote SMC differentiation in pluripotent stem cells (14, 35). ECM components influence VSMC phenotype, as culture on laminin, especially in combination with insulin-like growth factor I (IGF-I) (10, 11), promotes a differentiated phenotype. Neighboring cells also provide important inputs, as bilayer coculture of VSMC opposite endothelial cells induces VSMC differentiation (5). Although little is known about the intermediate signaling mechanisms underlying these effects, IGF-I-mediated SMC differentiation is inhibited by wortmannin, suggesting the involvement of phosphatidylinositol 3-kinase (PI 3-K) (10, 11).

Here, we investigate the effects of the macrolide antibiotic rapamycin on VSMC differentiation. Rapamycin has been previously shown to inhibit VSMC proliferation, protein synthesis, and migration in vitro, and intimal hyperplasia in animal models (1, 3, 7, 24, 30, 33, 40) and in human clinical trials (26, 38). Rapamycin binds FKBP12, and this complex inhibits mTOR, the mammalian target of rapamycin (also known as FRAP or RAFT) (21). mTOR is a ubiquitously expressed protein kinase that regulates translation initiation of specific growth-related mRNA subsets through the effectors p70 S6 kinase 1 (S6K1) and 4E-BP1/eIF4E. Other mTOR-regulated effectors included S6K2, PKCδ, and eIF2 kinase. The mTOR signaling pathway regulates protein synthesis in response to amino acids, and these effectors integrate nutrient sufficiency signals from mTOR with growth factor signals via PI 3-K to coordinately regulate protein synthesis, cell cycle progression, and proliferation (21). No studies of rapamycin in VSMC have addressed differentiation, which is characterized by major changes in gene expression and morphology. We report that the mTOR inhibitor rapamycin induces biochemical differentiation of primary VSMC in culture and that one mTOR effector, S6K1, may play an important role in this process.


Cell culture. Bovine thoracic aortic endothelial cells (EC) and bovine aortic VSMC were isolated and cultured in Dulbecco's modified Eagle's medium (DMEM) with 10% calf serum, l-glutamine, and penicillin-streptomycin. Rat and porcine VSMC were isolated from thoracic aorta and cultured as above but with 10% fetal bovine serum (36). Human VSMC were isolated from arterial or venous samples from vascular surgery patients or organ donors and cultured in M199 medium with 10% fetal bovine serum, l-glutamine, penicillin-streptomycin, insulin, hEGF, and hFGF (Clonetics). Human aortic VSMC were also purchased from Clonetics. EC were harvested by the scrape technique and VSMC by the explant method (5). VSMC at passage 2-6 were used for experiments. For experiments, cells were cultured in 2.5% calf serum for the duration of drug treatment. Untreated VSMC proliferate and do not spontaneously differentiate under these conditions. Rat VSMC were used to assess SM2-MHC expression due to increased immunoreactivity with the antibody relative to bovine or human VSMC.

VSMC were treated with vehicle (ethanol) or drug as indicated in the figure legends. For vehicle controls, cells were incubated with ethanol for the maximum duration of the experimental treatment. Rapamycin was purchased from Sigma. LY-294002, SB-203580, and U-0126 were purchased from Calbiochem.

Coculture model. Bovine EC (∼250,000) were plated on the outer side of Cyclopore membrane tissue culture inserts with 0.4-μm pores in six-well plates (Becton-Dickinson) and grown to confluence for 5 days. Bovine VSMC were then plated (1 × 105 cells) on the inner side of the membrane and cultured for 72 h. For VSMC cultured alone controls, VSMC were plated on the inner membranes of inserts lacking EC (5).

Cell area measurement. VSMC grown on Cyclopore membranes were stained with Mayer's hematoxylin and mounted on slides. Five separate high power fields were scanned in the same region of each slide in a blinded manner, and cell area was determined by planimetry, outlining cell dimensions, and computing two-dimensional area using NIH Image software. A student's t-test was used to determine statistical significance.

Immunohistochemistry. VSMC were cultured on glass coverslips. Cells were washed with phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde in PBS. Cells were permeabilized in 0.1% Triton in PBS. For horseradish peroxidase (HRP) staining, cells were treated with 3.0% hydrogen peroxide to quench endogenous peroxidase. Cells were blocked with 1.0% goat serum in PBS and then incubated with primary antibodies (as described in Western blotting) in PBS for 1 h. After washing, anti-mouse biotin-conjugated secondary antibody (for HRP stain) or anti-mouse TRITC conjugate for immunofluorescence, was incubated for 1 h and washed. Coverslips were then processed for immunofluorescent microscopy or treated with a peroxidase conjugate (Sigma) before mounting.

Protein and collagen synthesis. VSMC were treated with 4 μCi/ml [3H]proline (for collagen assay) or 2 μCi/ml [3H]leucine (for total protein) for 24 h, and then medium was replaced with fresh 3H-free medium. After 48 h, cells were subjected to trichloroacetic acid precipitation and scintillation counting. Collagens account for 80% of [3H]Pro incorporation into soluble ECM protein (31), whereas [3H]Leu counts reflect total protein synthesis. DNA content was measured by fluorescence spectroscopy of cell lysates using Hoescht reagent. Data are expressed as cpm (protein) per ng DNA.

Semiquantitative RT-PCR. Total RNA was isolated using the Qiagen RNeasy kit with DNase I and quantitated in duplicate spectro-photometrically. RNA (500 ng) was reverse transcribed using Superscript II (Life Technologies) or MMLV reverse transcriptase (Invitrogen) and oligo dT primer. Titrations using dilutions of the reverse transcribed cDNA were performed to determine the linear range of the PCR using the following primers to specifically amplify the human basic calponin gene transcript: sense 5′ TAACCGAGGTCCTGC-CTACG, antisense 5′ TGTGGGTGGGCTCACTCAGC. The 5′ primer sequence is common among calponin isoforms (nt 187-206 in NCBI nucleotide sequence #NM-001299), but the 3′ primer is specific to the unique COOH terminus of basic calponin (nt 1,009-1,028). The COOH-terminal sequences of neutral or acidic calponin are greatly divergent in this region. PCR was performed using 20 pmol each primer, 0.04 mM dNTPs, Red taq (Sigma), and TaqStart antibody (Clontech) for 32 cycles: 94°C for 30 s, 60°C for 45 s, and 72°C for 1 min. The following primers were used to specifically amplify SM-MHC: sense 5′ CGCTGAATGACAACGTGACTTCC, antisense 5′ CCAGTTCCGCAGCTTGAGGTA. The linear range of the PCR assay was again determined by template titration. PCR was performed as above for 30 cycles: 95°C for 30 s, 62°C for 30 s, and 68°C for 1 min. For each experiment, control reactions with primers to the pyruvate dehydrogenase (PDH) gene were performed for quality control. The linear range of the PDH PCR was similarly verified using template titrations. PCR products were resolved on 2% agarose gels with ethidium bromide and quantitated using a Typhoon Scanner (Molecular Dynamics).

Western blotting. Lysates were harvested for Western analysis as in Ref. 23. Bradford assay was used to determine total protein concentration in lysates. Equal amounts of protein per lane were subjected to SDS polyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membrane, and immunoblotted using antibodies against SM-MHC SM2 (Seikagaku America), SM α-actin, calponin (Sigma), p27kip1, or p21cip (Santa Cruz). The calponin antibody does not cross-react with skeletal, cardiac, or nonmuscle tissue calponin and exhibits smooth muscle specificity when used in immunohistochemistry (Sigma catalog no. C2687).

Adenoviral overexpression of S6K1. We have generated recombinant adenoviruses encoding HA-tagged wild-type or rapamycin-resistant mutant (ED3E) (41) p70 S6K1. These viruses coexpress green fluorescent protein (GFP) to easily assess infection efficiency using the AdEasy system (see www.coloncancer.org/adeasy for methods) (13). Virus encoding GFP alone serves as a control.

VSMC were infected with adenoviral supernatant (at 1:20-1:100 dilutions) in DMEM with 10% calf serum overnight, washed, and treated with 20 nmol/l rapamycin or vehicle for 48 h in 2.5% calf serum medium. Optimal titers have been predetermined for each virus to yield infection of ∼75% of VSMC. Cells were assayed for infection efficiency (GFP fluorescence), S6K1 expression [Western blotting with anti-HA and anti-S6K1 antibodies (23)], and S6K1 activity.

Immune complex kinase assay. Infected VSMC were lysed (as for Western blotting) and S6K1 immunoprecipitated with an antibody to the HA-epitope tag and protein A-Sepharose. Kinase activity toward recombinant GST-S6 peptide in washed immunoprecipitates was assayed as in Martin et al. (23).


Rapamycin induces contractile morphology. We assessed the effects of rapamycin on VSMC differentiation using primary aortic smooth muscle cell cultures from multiple species. VSMC in primary culture rapidly dedifferentiate in early passages from their in vivo contractile phenotype to a synthetic phenotype, characterized by a flattened morphology, increased protein synthesis, and loss of contractile protein expression (10). However, VSMC treated with 20 nmol/l rapamycin (a maximal concentration for inhibition of S6K1 and 4E-BP1) for 48 h underwent a phenotypic modulation with morphology characteristic of the differentiated, contractile phenotype (Fig. 1A). As has been previously reported (24), rapamycin inhibited proliferation was determined by cell number (data not shown). We have previously demonstrated that VSMC phenotype is influenced by the presence of EC (5). Bilayer coculture of VSMC and EC on opposite sides of a porous membrane inhibited VSMC extracellular matrix synthesis and induced a contractile morphology (5). The effect of rapamycin on VSMC cultured alone was similar to the effects of EC coculture (Fig. 1B). This dramatic change in VSMC morphology is quantitatively reflected in measurements of cell area (two dimensional). VSMC treated with rapamycin for 48 h were 67% smaller in area than VSMC treated with ethanol vehicle (Fig. 1B).

Fig. 1.

Effect of rapamycin on vascular smooth muscle cell (VSMC) morphology and cell area. A: bovine VSMC were cultured with ethanol vehicle (left) or 20 nmol/l rapamycin (right) for 48 h and subjected to immunohistochemical staining with an anti-calponin primary antibody and horseradish peroxidase-conjugated secondary antibody. B: bovine VSMC were plated on membranes alone or in coculture (CC) opposite confluent endothelial cells (EC). Rapamycin (20 nmol/l) or ethanol vehicle was added for 48 h. VSMC alone, VSMC cultured alone, treated with vehicle. VMSC Rap, VSMC cultured alone, treated with rapamycin. VSMC CC, VSMC cocultured with EC, treated with vehicle. VSMC cell area was determined in μm2 as described in materials and methods. n = 400 cells per condition; error bars = SE.

Rapamycin inhibits VSMC protein and extracellular matrix synthesis. Rapamycin is a known inhibitor of specific translation initiation and thus inhibits protein synthesis in many cell types (8, 21). Enhanced synthesis of collagen and other extracellular matrix proteins is a characteristic of the synthetic, dedifferentiated VSMC phenotype and is a major component of intimal hyperplasia (5). We labeled VSMC with 3H-amino acids to assess total protein and collagen synthesis after rapamycin treatment. As expected, rapamycin inhibited both total protein and collagen synthesis after 24-48 h (Fig. 2A). Notably, rapamycin treatment inhibited total protein synthesis by only 30% after 48 h, whereas cycloheximide treatment inhibited total protein synthesis by 93% at this time point (Fig. 2B). This is consistent with rapamycin inhibition of translation of a subset of highly structured mRNAs, whereas cycloheximide inhibits global protein synthesis.

Fig. 2.

Rapamycin inhibits total protein and collagen synthesis. A: human aortic VSMC were cultured alone in the presence or absence of 20 nmol/l rapamycin for 24 or 48 h with either [3H]leucine (total protein) or [3H]proline (collagen) as tracers. TCA precipitates were counted, and DNA content was determined (Hoescht). Data were expressed as cpm protein/ng DNA and then normalized to percent maximal based on values for vehicle-treated samples. Average and SE from duplicate samples is presented, representative of 3 experiments. One-way analysis of variance P value = 0.0037 for total protein synthesis, P = 0.019 for collagen synthesis. P values for Newman-Keuls multiple comparison posthoc test are indicated above the bars: *P < 0.05; **P < 0.01 vs. vehicle. B: bovine aortic VSMC were treated with 1 μg/ml cycloheximide for 24 or 48 h and [3H]leucine tracer. Cells were processed as above and total protein synthesis was normalized to percent maximal. Average and SE from duplicate samples are presented, representative of 2 experiments. One-way analysis of variance P value = 0.0038. P values for Newman-Keuls multiple comparison post hoc test are indicated above the bars: **P < 0.01 vs. vehicle.

Rapamycin induces protein markers of VSMC differentiation. Given the striking effects on VSMC morphology, we assessed modulation of VSMC-specific contractile protein gene expression, which serves as a biochemical marker of differentiation status. Rapamycin treatment induced a rapid upregulation of the mRNA for basic calponin in human VSMC as determined by semiquantitative RT-PCR (Fig. 3B). This isoform is unique to vascular smooth muscle and is down-regulated upon dedifferentiation (37). A sharp 20-fold induction of the calponin message was observed after 2 h of rapamycin treatment. A maximal 35-fold induction was reached after 6 h of rapamycin treatment. The elevated mRNA levels persisted through 24 h. We similarly evaluated mRNA levels for SM-MHC, the most stringent contractile protein marker of VSMC differentiation (19). Rapamycin also induced SM-MHC mRNA levels, with a maximal sixfold induction after 1 h of treatment (Fig. 3C). These data suggest that rapamycin may promote a differentiated phenotype by inducing a new program of gene expression at the level of transcription, a new role for the mTOR pathway in mammalian cells.

Fig. 3.

Rapamycin induces calponin mRNA in VSMC. A: total RNA was prepared from human VSMC treated with 20 nmol/l rapamycin for 24 h, and 500 ng were subjected to reverse transcription as described in materials and methods. Serial dilutions (expressed in terms of RNA input) of the reverse transcription reaction were used as a PCR template to define the linear range of the assay using primers to the basic calponin gene (top), SM-MHC gene (middle), or to the housekeeping gene pyruvate dehydrogenase (PDH) (bottom). B: total RNA was prepared from human VSMC treated with vehicle (24 h) or with 20 nmol/l rapamycin for the times indicated and subjected to reverse transcription and semiquantitative RT-PCR as described using primers to the basic calponin gene. The same samples were subjected to RT-PCR using primers to PDH as a control for RNA quantitation and reverse transcription. A representative experiment is shown at top. Experiments were repeated as above 3 times and mean fold induction (calponin PCR product divided by PDH PCR product, expressed relative to vehicle treatment) plus standard error of the mean is presented as a bar graph (bottom). The number of replicates (n) at each time point is indicated above the bars. One-way analysis of variance P value = 0.0091. P values for Newman-Keuls multiple comparison post hoc test are indicated above the bars: *P < 0.05; **P < 0.01 vs. vehicle. C: human VSMC were treated with 20 nmol/l rapamycin for the times indicated and subjected to RT-PCR as described above using primers to the smooth muscle-myosin heavy chain (SM-MHC) (top) or PDH (bottom) genes. The bar graph shows densitometric quantitation of the above gels, with SM-MHC normalized to PDH and expressed as fold induction relative to time zero. A representative experiment is shown.

Inhibition of mTOR with rapamycin induced expression of contractile protein markers of the differentiated phenotype, including SM α-actin, calponin, and the SM-MHC SM2 isoform. Notably, the kinetics of induction of these proteins closely followed those observed for induction of calponin message (Fig. 4A). With all contractile proteins assayed, protein levels continued to accumulate with increasing time of rapamycin treatment.

Fig. 4.

Rapamycin induces VSMC contractile protein expression. A: human iliac or bovine aortic VSMC were cultured alone in the presence of vehicle (24 h) or 20 nmol/l rapamycin for the times indicated. Cell lysates normalized for total protein content were subjected to Western blotting using antibodies against SM2-MHC, SM α-actin, or calponin, as indicated. A Western blot for total ERK1/2 is included as a loading control (human). B: rat VSMC were cultured alone in the presence of 20 nmol/l rapamycin or vehicle for 24 or 48 h. Lysates were subjected to Western blotting as above using an anti-SM-MHC SM2 antibody. Data are representative of 3 experiments. C: bovine VSMC were cultured as above in the presence of vehicle or the indicated drug for 48 h: Rap, 20 nmol/l rapamycin; LY, 10 μmol/l LY-294002; SB, 10 μmol/l SB-203580; U0, 5 μmol/l U-0126. Lysates were analyzed by Western blotting with anti-calponin antibody as above. Data are representative of at least 3 experiments. D: rat (top and bottom) and bovine (middle) VSMC were cultured alone in the presence of 20 nmol/l rapamycin or vehicle for 48 h and then subjected to immunohistochemical analysis using antibodies against SM α-actin, calponin, or SM2-MHC and TRITC-conjugated secondary antibodies (red). Photographs from standard (middle) or confocal (top, bottom) immunofluorescence microscopy are shown. Data are representative of at least 3 experiments.

Confirming the effects of rapamycin on contractile protein gene expression observed at early time points, 48 h of rapamycin treatment induced expression of SM α-actin, calponin, and SM2-MHC in VSMC in Western blot analyses (Figs. 4, B and C, and 6, D and E). We have also observed this new contractile protein expression by immunofluorescent staining (Figs. 4D and 6, B and C). Enhanced expression of these proteins is consistent with the differentiated morphology observed after 48 h of rapamycin treatment in Fig. 1. Notably, induction of contractile proteins was specific to inhibition of the mTOR pathway, as treatment with pharmacological inhibitors of other signaling pathways, including the PI 3-K, p38, and ERK1/2 pathways, did not induce contractile protein expression (Fig. 4C). Furthermore, rapamycin fold induction of contractile protein expression relative to vehicle was highly reproducible in VSMC from all species tested (Table 1).

Fig. 6.

S6K1 overexpression inhibits rapamycin-induced VSMC differentiation. A, top: bovine VSMC were infected with adenovirus co-overexpressing HA-S6K1 and GFP, or GFP alone (control) and treated with 20 nmol/l rapamycin or vehicle for 48 h as described in materials and methods. Western blotting with anti-HA antibody (top) and activity measured in immune complex kinase assay of HA-S6K1 toward recombinant GST-S6 substrate (bottom) is shown. Bottom: rat VSMC were infected in duplicate with adenovirus co-overexpressing wild-type HA-S6K1 and GFP (WT), or the rapamycin-resistant HA-S6K1 ED3E mutant and GFP (ED3E) and treated with 20 nmol/l rapamycin or vehicle for 48 h and subjected to immune complex kinase assay as above. Kinase activity (32P incorporated into the GST-S6 substrate measured by phosphorimaging) is plotted with vehicle treatment set equal to 100%. Error bars represent SE. Expression of each construct was verified by anti-HA Western blot (not shown). B: confocal immunofluorescent micrographs of HA-S6K1 wild-type-infected rat VSMC treated with rapamycin for 48 h (both panels). GFP (green) fluorescence indicates HA-S6K1-overexpressing cells. Red punctate staining marks cells expressing SM2-MHC (anti-SM2 primary and TRITC-conjugated secondary antibodies). Note mutually exclusive green/red staining pattern. C: rat VSMC infected and processed as above, but with primary antibody toward SM α-actin. Note that in control infected cells (expressing GFP only, far right) and in HA-S6K1 wild-type infected cells treated with vehicle only (far left), green and red stains overlap, indicating that GFP expression does not preclude SM α-actin staining. There is no overlap in rapamycin-treated HA-S6K1-infected cells (middle) expressing high (induced) levels of SM α-actin. D: bovine VSMC were infected as above with HA-S6K1 wild-type or GFP control (ctrl) adenovirus and treated with ethanol or 20 nmol/l rapamycin for 48 h. Lysates normalized for total protein content were analyzed by Western blotting using anti-SM α-actin, anti-calponin, anti-HA (top), or anti-p21cip (bottom) antibodies. E: rat VSMC were infected as above with HA-S6K1-ED3E (rapamycin resistant) or GFP control adenovirus and treated with vehicle or 20 nmol/l rapamycin for 48 h. Left: lysates normalized for total protein content were analyzed by Western blotting using anti-SM2-MHC (top), anti-S6K1 (middle), or ERK1/2 as a loading control (bottom). Note that endogenous S6K1 is not detected in the control infected lanes at this film exposure relative to the overexpressed ED3E. Right: for statistical analysis, 2 additional experiments were performed as above, in duplicate. Western blots for SM2-MHC were digitized and expression levels quantitated using Scion Image 1.62 software. Gray bars indicate vehicle treatment, and black bars indicate rapamycin treatment. Data (means, n = 4) are expressed relative to vehicle treatment for each adenovirus condition. Error bars = SE of the mean. P values were determined by a paired t-test. *P = 0.05; **P = 0.01.

View this table:
Table 1.

Statistical analysis of contractile protein induction by rapamycin

Rapamycin induces expression of antiproliferative cyclin-dependent kinase inhibitors. Because proliferation and differentiation are not necessarily mutually exclusive processes in VSMC, we assessed expression levels of cyclin-dependent kinase inhibitory proteins after rapamycin treatment. Rapamycin treatment induced expression of the p27kip and p21cip cyclin-dependent kinase inhibitors in VSMC (Fig. 5), consistent with rapamycin-inhibited proliferation. Notably, the kinetics of rapamycin-induced p27kip and p21cip expression closely parallel the kinetics of contractile protein induction, suggesting a coordinated regulation of proliferation and differentiation. Rapamycin induction of these cyclin-dependent kinase inhibitors was also highly reproducible in VSMC from all species tested (Table 1).

Fig. 5.

Rapamycin induces expression of cyclin dependent kinase inhibitor proteins. Human iliac (top) or bovine or rat aortic (bottom) VSMC were cultured in the presence of vehicle or 20 nmol/l rapamycin over the indicated time courses and subjected to western blotting as in Fig. 4 using anti-p27kip or -p21cip antibodies.

The mTOR effector S6K1 opposes rapamycin-induced VSMC differentiation. Our data suggest that activation of the mTOR pathway opposes VSMC differentiation. Rapamycin inhibits the activity of mTOR, which in turn regulates multiple downstream effector proteins. To determine whether S6K1, one of the best-characterized effectors of mTOR, is a critical regulator of VSMC differentiation, we employed an adenovirus strategy to overexpress HA-tagged S6K1 constructs. The virus encodes green fluorescent protein (GFP) driven by the same promoter, such that S6K1-overexpressing cells are positive for GFP under fluorescent microscopy. A negative control virus expresses only GFP. HA-tagged S6K1 was detected by immunoblotting in infected bovine cells, and immune complex kinase assay revealed that a modest S6K1 kinase activity persisted in rapamycin-treated cells overexpressing wild-type S6K1 (Fig. 6A, top). In rat VSMC adenovirally overexpressing wild-type S6K1, 25% of the initial activity persisted in the presence of rapamycin (Fig. 6A, bottom). After 48 h of rapamycin treatment, individual rat cells overexpressing wild-type S6K1 (GFP-positive) were negative for contractile protein expression, whereas uninfected cells in the same field demonstrated strong positive staining for SM2-MHC and SM α-actin (Fig. 6, B and C). As a control, GFP expression did not preclude contractile protein staining in rapamycin-treated cells infected with control virus (Fig. 6C), verifying that overexpression of GFP itself is not incompatible with contractile protein expression.

Western blot analysis of similar experiments confirmed that rapamycin induced expression of the contractile proteins SM α-actin and calponin and the G1 cyclin inhibitor p21cip in cells infected with control virus but not in cells infected with the HA-S6K1 adenovirus (Fig. 6D). To further enhance S6K1 activity in the presence of rapamycin, we employed an adenovirus encoding an HA-tagged rapamycin-resistant mutant of S6K1 (ED3E) (41). This mutant S6K1 retained 75% of its initial activity in the presence of rapamycin when expressed in rat VSMC (Fig. 6A, bottom). Rapamycin similarly induced expression of SM2-MHC in control infected cells, but not in S6K1-ED3E-infected cells, as determined by Western blotting (Fig. 6E). The difference in SM2-MHC induction between control and S6K1-ED3E-infected cells was statistically significant (P = 0.01). Immunoblots confirmed overexpression of S6K1 (note that mobility shift of ED3E in Fig. 6E is less sensitive to rapamycin than wild-type S6K1 in Fig. 6D), and ERK1/2 blots verified even lane loading (Fig. 6E). These data indicate that overexpression of the mTOR effector S6K1 inhibits rapamycin-induced VSMC differentiation.


We demonstrate for the first time that inhibition of the mTOR pathway induces VSMC differentiation as determined by induction of contractile protein expression and morphology. Importantly, our data provide evidence suggesting that this involves transcriptional regulation, a largely unexplored function for the mTOR signaling pathway, which is best known for translational regulation. We have also employed a unique approach allowing assignment of a role to one particular mTOR effector, p70 S6K1, in VSMC phenotypic modulation.

The mTOR inhibitor rapamycin has previously been shown to inhibit VSMC proliferation. However, proliferation and differentiation are not mutually exclusive processes in VSMC (5). VSMC differentiation is characterized by induction of a novel program of gene expression resulting in new protein expression and morphology (37). Therefore, the finding that rapamycin promotes a differentiated phenotype is compelling. Rapamycin induced expression of the differentiation marker contractile proteins, including α-actin, calponin, and myosin heavy chain. Notably, rapamycin also induced expression of the cyclin-dependent kinase inhibitors p27kip and p21cip with similar kinetics. Thus mTOR inhibition promotes coordinated regulation of inhibitors of cell cycle progression and contractile proteins to induce the differentiated VSMC phenotype.

The most intriguing aspect of this work was the finding that inhibition of mTOR by rapamycin induces mRNAs encoding contractile proteins. Because calponin mRNA was undetectable in the vehicle treated dedifferentiated VSMC and was rapidly induced 20-fold after 2 h of rapamycin treatment, this induction is likely due to new transcription. We are investigating the possibility that rapamycin may additionally regulate message stability. Similar induction of SM-MHC mRNA suggests that mTOR-dependent transcriptional regulation may be a common mechanism among VSMC differentiation marker proteins. Notably, the kinetics of contractile protein expression closely follow the induction of these messages.

Although the TOR pathway is known to regulate nutrient-sensitive transcription factors in yeast (15), very little is known regarding transcriptional control by mTOR in mammalian cells. However, several studies have reported that rapamycin induces hematopoietic differentiation (11, 16, 43) but inhibits adipocyte differentiation (6) and chondrogenesis (27). Differential effects of rapamycin on differentiation in other cell lines suggests that mTOR may indeed be an important regulator of cell type-specific transcriptional processes. We are currently evaluating mTOR-regulated VSMC-specific transcription factor candidates (manuscript in preparation).

To address potential mechanisms of rapamycin-induced biochemical differentiation, we evaluated the role of S6K1, a key mTOR target, in VSMC differentiation. S6K1 promotes cell growth and proliferation, in addition to protein synthesis via 5′TOP mRNA regulation (21). As these processes are characteristic of dedifferentiated VSMC, S6K1 activation may antagonize differentiation. However, because there are multiple effector proteins, including 4E-BP1 (8) and S6K2 (23), down-stream of mTOR subject to rapamycin inhibition, one cannot immediately attribute any and all effects of rapamycin to S6K1. We overexpressed S6K1 to determine whether the effects of rapamycin on VSMC differentiation are mediated by this kinase. Although wild-type S6K1 retains rapamycin-sensitivity, high-level overexpression confers a degree of rapamycin-resistant activity (Fig. 6A), perhaps by exceeding limiting quantities of an mTOR-regulated inhibitor (29). The persistence of S6K1 activity in the presence of rapamycin, when other mTOR target proteins are inhibited, or overexpression of a rapamycin-resistant S6K1 mutant, allows us to assign a role specifically to S6K1. In these studies, overexpression of S6K1 was mutually exclusive with rapamycin-induced contractile protein and p21cip expression. These data suggest that mTOR normally opposes VSMC differentiation through activation of S6K1 and that rapamycin inhibition of this kinase promotes differentiation.

The effect of S6K1 overexpression on basal levels of differentiation markers varied somewhat in that basal expression of SM2-MHC and calponin, but not SM α-actin and p21cip, was inhibited by S6K1. We attribute this variation to two factors. It is likely that the enhanced kinase activity of the ED3E mutant as opposed to wild-type S6K1 contributed to inhibition of basal SM2-MHC expression (Fig. 6E, left). It is also likely that the effect correlates with the temporal expression of contractile proteins during the differentiation process. SM α-actin is the earliest marker of the VSMC lineage, and its expression is reduced, but not completely inhibited, even in the synthetic phenotype (Figs. 4D and 6C). Thus it is perhaps not surprising that (wild-type) S6K1 failed to inhibit basal expression of this protein but did inhibit that of calponin, an intermediate stage marker, and of SM2-MHC, the most stringent late-stage differentiation marker.

Other than the ribosomal S6 protein, few substrates for S6K1 have been identified. Notably, the testis-specific transcription factor CREMτ has been identified as an S6K1 substrate with rapamycin-sensitive transactivation function in vivo (4). An attractive hypothesis that we aim to pursue is that S6K1 may oppose VSMC differentiation by negatively regulating cell type-specific transcription factors through its kinase activity. Alternately, S6K1-mediated modulation of proliferation and protein synthesis may contribute to the differentiation program and indirectly regulate transcription factors. An additional possibility is that altered translation of 5′TOP-containing mRNAs may contribute to the differentiated phenotype, as rapamycin-inhibited induction of G1 cyclins and Cdks in coronary VSMC was found to occur through a posttranscriptional mechanism (1). Interestingly, a 5′ UTR stem-loop structure has recently been implicated in regulation of collagen I translation (39), suggesting that mTOR-dependent translational regulation may contribute to rapamycin inhibition of collagen synthesis. However, the contractile proteins that undergo dramatic upregulation in differentiated VSMC are not known to be 5′TOP-containing messages.

S6 kinases are found in several cellular locations. p70 S6K1 is found in both the nucleus and the cytoplasm, whereas the p85 S6K1 isoform is nuclear (17). Both isoforms of the related kinase S6K2 are nuclear (9, 18). The roles of the nuclear S6 kinases are largely unknown, but we speculate that S6Ks could potentially regulate differentiation through as yet undetermined nuclear targets.

Our data demonstrate that S6K1 opposes VSMC differentiation, and previous studies have reported rapid activation of S6K1 after vascular injury (1). However, it is likely that inhibition of S6K1 alone is not sufficient for VSMC differentiation: we found that PI 3-K inhibition with LY-294002, which, in turn, leads to S6K1 inhibition (21), did not promote VSMC differentiation (Fig. 4C). We postulate that whereas S6K1 may oppose differentiation, promotion of differentiation also requires PI 3-K-dependent prodifferentiation signals. This is supported by work from Sobue and coworkers, who have demonstrated that IGF-I-induced VSMC differentiation requires PI 3-K (11) and that Akt is an important mediator of visceral SMC differentiation (28). It is likely that the full complement of signaling effectors activated downstream of IGF-I and PI 3-K results in a prodifferentiative signal, despite activation of S6K1.

Using inhibitors of multiple signaling pathways, we found that only inhibition of the mTOR pathway induced VSMC differentiation. Previous studies suggest that the ERK MAPK pathway is proproliferative in VSMC (11). Although inhibition of this pathway could conceivably contribute to differentiation, we did not observe such an effect. It may be that ERK signals are required for proper function of the CArG DNA regulatory elements common to most muscle-specific promoters, including those of the VSMC contractile protein genes (20, 34, 42).

In response to vessel injury, VSMC dedifferentiate to a “synthetic” phenotype, which is named for the copious extracellular matrix synthesis and secretion characteristic of these cells. Because the major known function of the mTOR pathway is regulation of protein synthesis, rapamycin might be expected to inhibit the synthetic phenotype primarily at the level of inhibition of translation initiation. We found that rapamycin indeed inhibited total protein synthesis, including the major secreted extracellular matrix component collagen. However, despite attenuated protein synthesis, rapamycin promoted a new program of gene expression characteristic of the differentiated phenotype. Because rapamycin initially inhibits only translation of a subset of 5′ structured mRNAs, sufficient translational activity persists to synthesize the new proteins.

The plasticity of VSMC phenotype is important in clinical conditions such as intimal hyperplasia, in which dedifferentiated VSMC migrate, proliferate, and secrete extracellular matrix protein in response to vessel injury. Rapamycin was first investigated as a potential antirestenotic drug due to its known antiproliferative and protein synthesis effects. However, our study suggests that rapamycin may be a particularly effective therapeutic because, in addition to inhibiting these components of intimal hyperplasia, it also promotes a differentiated, contractile phenotype. This is in contrast to other drugs that act as cytotoxic agents, which inhibit proliferation but result in the death of the cells, as opposed to maintenance of functional contractile quiescent VSMC at the site of injury. Recent clinical trials of rapamycin-coated stents have demonstrated promise for prevention of restenosis (26, 38). Our finding that mTOR inhibition promotes VSMC differentiation furthers our understanding of the basic signaling mechanisms regulating VSMC phenotype and suggests new avenues for discovery of cardiovascular therapeutics.


We thank Dr. Kenneth Walsh for generous gifts of anti-SM2-MHC antibody and rat ASMC and for critical review of this manuscript. We thank Drs. John Hwa and Stefanie Schalm for helpful comments and Dr. John Blenis and Kristen L. Keon for assistance with construction of recombinant adenoviruses. We thank Dr. Raymond Perez for assistance and reagents in developing RT-PCR assays.


Supported by a grant from the National Heart, Lung, and Blood Institute (R-01-HL-59590), a Wylie Scholar in Academic Surgery Award from the Pacific Vascular Research Foundation to R. J. Powell, and a Hitchcock Foundation Award and an American Heart Association Scientist Development Grant to K. A. Martin.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • * K. A. Martin and E. M. Rzucidlo contributed equally to this work.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 23.
  23. 24.
  24. 25.
  25. 26.
  26. 27.
  27. 28.
  28. 29.
  29. 30.
  30. 31.
  31. 32.
  32. 33.
  33. 34.
  34. 35.
  35. 36.
  36. 37.
  37. 38.
  38. 39.
  39. 40.
  40. 41.
  41. 42.
  42. 43.
  43. 44.
View Abstract