Focal Adhesion Kinase (FAK)-dependent Regulation of S-phase Kinase-associated Protein-2 (Skp-2) Stability

Smooth muscle cell (SMC) proliferation is suppressed in intact blood vessels but stimulated in atherosclerosis, restenosis after angioplasty, and vein graft disease. The cyclin-dependent kinase inhibitors, including p27Kip1, play important roles in maintaining SMC quiescence. Levels of p27Kip1 are dependent on attachment to and the composition of the extracellular matrix (ECM). Here we sought to elucidate mechanisms underlying the ECM-dependent regulation of p27Kip1 and hence, SMC proliferation. Serum stimulation decreased p27Kip1 levels in isolated SMC but not in rat aorta. The effect was post-translational and mediated by proteasomal degradation. We studied the S-phase-associated kinase protein-2 (Skp-2), an F-box protein involved in ubiquitination and proteasome-mediated degradation. Skp-2 protein is strongly induced by serum from undetectable levels in isolated SMCs but remains undetectable in aorta; Skp-2 mRNA is also lower in aorta. Overexpression of wild-type Skp-2 in SMCs decreased p27Kip1 levels, whereas dominant negative F-box deleted mutant (ΔF-Skp-2) Skp-2 increased p27Kip1 levels. Furthermore, hyperphosphorylation of retinoblastoma protein and SMC proliferation were also reciprocally affected by wild-type and dominant negative Skp-2. Skp-2 expression was absolutely dependent on cell attachment to the ECM and was inhibited by laminin and type-1 fibrillar collagen but increased by fibronectin. Expression of Skp-2 protein, but not mRNA, was associated with focal adhesion kinase (FAK) activity and inhibited by overexpression of FAK-related non-kinase and a dominant negative FAKY397F mutant. Furthermore, the inhibition of Skp-2 expression by dominant negative FAK was reversed by the proteasome inhibitor MG-132. Taken together, these data demonstrate that the vascular ECM controls SMC proliferation via FAK-dependent regulation of Skp-2 protein stability.

Smooth muscle cell (SMC) proliferation is suppressed in intact blood vessels but stimulated in atherosclerosis, restenosis after angioplasty, and vein graft disease. The cyclin-dependent kinase inhibitors, including p27 Kip1 , play important roles in maintaining SMC quiescence. Levels of p27 Kip1 are dependent on attachment to and the composition of the extracellular matrix (ECM). Here we sought to elucidate mechanisms underlying the ECMdependent regulation of p27 Kip1 and hence, SMC proliferation. Serum stimulation decreased p27 Kip1 levels in isolated SMC but not in rat aorta. The effect was posttranslational and mediated by proteasomal degradation. We studied the S-phase-associated kinase protein-2 (Skp-2), an F-box protein involved in ubiquitination and proteasome-mediated degradation. Skp-2 protein is strongly induced by serum from undetectable levels in isolated SMCs but remains undetectable in aorta; Skp-2 mRNA is also lower in aorta. Overexpression of wild-type Skp-2 in SMCs decreased p27 Kip1 levels, whereas dominant negative F-box deleted mutant (⌬F-Skp-2) Skp-2 increased p27 Kip1 levels. Furthermore, hyperphosphorylation of retinoblastoma protein and SMC proliferation were also reciprocally affected by wild-type and dominant negative Skp-2. Skp-2 expression was absolutely dependent on cell attachment to the ECM and was inhibited by laminin and type-1 fibrillar collagen but increased by fibronectin. Expression of Skp-2 protein, but not mRNA, was associated with focal adhesion kinase (FAK) activity and inhibited by overexpression of FAK-related non-kinase and a dominant negative FAK Y397F mutant. Furthermore, the inhibition of Skp-2 expression by dominant negative FAK was reversed by the proteasome inhibitor MG-132. Taken together, these data demonstrate that the vascular ECM controls SMC proliferation via FAK-dependent regulation of Skp-2 protein stability. 1 is an important event during normal physiological development and the pathological responses to vessel injury, including angioplasty restenosis, late vein graft failure, and atherosclerosis. A number of peptide growth factors released from platelets and SMCs in response to vascular injury are known to play an important role in stimulating this proliferative response (1). However, it has become increasingly apparent in recent years that the presence of growth factors alone is insufficient to stimulate SMC proliferation (2)(3)(4)(5). For example, healthy vessel SMCs exhibit extremely low rates of proliferation, even though endogenous growth factors are present (6). Furthermore, exposure of arterial segments to exogenous growth factors in vitro (7) or in vivo does not lead to high rates of SMC proliferation (8 -10). These observations imply that growth factors work in concert with additional factors to regulate SMC proliferation. The fact that SMCs are capable of rapidly proliferating in vitro when enzymatically digested free of their native extracellular matrix suggests that the nature of the local environment, and in particular the composition of the extracellular matrix, is an important factor determining SMC response to mitogens. Several vascular ECM components, including fibrillar type-1 collagen, fibronectin, and laminin, have been demonstrated to regulate SMC proliferation (5,(11)(12)(13)(14). However, the mechanisms underlying this regulation are unclear.

Proliferation of vascular smooth muscle cells (SMCs)
The cyclin-dependent kinases (CDK2, CDK4) and their regulatory partners, the cyclins (cyclin E, cyclin D) control progression through the cell cycle (15). The activity of the CDKs is further regulated by the CIP/Kip family and the INK family of cyclin-dependent kinase inhibitors (CDKIs) (16). A large body of research has documented the role played by the CDKIs, in particular p27 Kip1 , which can inhibit the activity of the CDKcyclin E and CDK4-cyclin D complexes (17)(18)(19)(20). Typically, in response to mitogenic stimulation, the levels of p27 Kip1 are reduced during G 1 , relieving inhibition of CDK activity and allowing retinoblastoma protein (Rb) hyperphosphorylation. The precise mechanisms that control CDKI levels during the SMC cell cycle remain unclear.
Several lines of evidence support the hypothesis that the ECM regulates p27 Kip1 levels in SMC and other cells (21,22). First, fibroblasts forced into suspension fail to down-regulate p27 Kip1 and arrest in G 1 (2,24). Second, p27 Kip1 levels are dependent on the type of ECM substratum (25)(26)(27). Last, SMCs in intact rat aorta and hence interacting with their native basement membrane ECM fail to degrade p27 Kip1 in response to mitogen stimulation, whereas SMC digested free of their basement membranes and cultured are able to rapidly degrade p27 Kip1 (28). Taken together, these observations suggest that the ECM controls SMC proliferation, at least in part by regulating the levels of p27 Kip1 . However, the mechanism underlying this regulation is unclear. This prompted us to investigate the mechanisms controlling the ECM-dependent degradation of p27 Kip1 in vascular SMCs and in particular the role played * This work was supported by the British Heart Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
by Skp-2. Here we have tested the hypothesis that the vascular ECM regulates the expression of Skp-2 and hence p27 Kip1 degradation.

Materials
Male Wistar rats were obtained from Charles River culture media, and additives were obtained from Invitrogen. Monoclonal antibody to Skp-2 was obtained from Zymed Laboratories Inc. Inc. (Cambridge, UK). Monoclonal antibody to p27 Kip1 was obtained from Transduction Laboratories (San Diego, CA). pCDNA3.1 plasmids containing wildtype and dominant negative F-box-deleted Skp-2 were generously provided by Dr. M. Pagano (New York, NY). Recombinant adenovirus expressing FAK-related-non-kinase (FRNK) was provided by Prof. A. M Samarel (Loyola University, Chicago, IL) and cDNA encoding FAK Y397F was kindly provided by Prof. J. T. Parsons (University of Virginia).

SMC Culture and [ 3 H]Thymidine Incorporation-Male
Wistar rats (300 -400 g) were anesthetized with sodium pentabarbitone followed by retrograde perfusion with phosphate-buffered saline via the abdominal aorta. The thoracic aorta was excised, cut into 4-mm sections, and cultured as described previously (28) in DMEM (100 units/ml streptomycin and 100 mg/ml penicillin) containing 10% FCS. Isolated SMCs were prepared using a modification of the explant technique described previously (29). Isolated SMCs were cultured in DMEM containing 2 mM glutamine, 100 units/ml streptomycin, 100 mg/ml penicillin, and 10% fetal bovine serum. Cells were passaged by trypsin/EDTA treatment and used between passages 2-6. Cells were rendered quiescent by serum deprivation for 72 h. Where indicated, tissue culture plates were coated with 20 g/ml of Engelbreth-Holm-Swarm-derived laminin (Invitrogen) or 20 g/ml fibronectin (Sigma) for 2 h at room temperature. Plates were blocked with 0.1% bovine serum albumin/phosphate-buffered saline for 30 min and washed three times in phosphate-buffered saline before cell seeding. Cells cultured on a laminin matrix were also cultured in the presence of 10 M cyclic RGDFV peptide to block interaction with endogenously produced fibronectin as previously described (30). Type-1 collagen gels were formed by neutralization of acidic mo-nomeric collagen (Vitrogen) with 0.1 M NaOH. To quantify proliferation, cells were cultured in the presence of 0.5 Ci/ml [ 3 H]thymidine for 18 h. Cells were then washed in phosphate-buffered saline and incubated with 10% trichloroacetic acid at 4°C for 30 min. Trichloroacetic acid precipitates were collected and analyzed for DNA [ 3 H]thymidine incorporation and total DNA content as described previously (29). For suspension cultures, SMCs were cultured over 10% agarose at a density of 1 ϫ 10 6 cells/ml in DMEM (100 units/ml streptomycin and 100 mg/ml penicillin) containing 10% FCS.
Adenovirus Infection of Rat Aorta-Sections of rat aorta (4 mm) were infected with 1 ϫ 10 10 plaque-forming units of recombinant adenovirus in DMEM containing 10% FCS for 4 h. Aortic sections were then cultured in DMEM/10% FCS containing 10 M BrdUrd for 72 h. Immunohistochemical staining for Skp-2 expression and BrdUrd incorporation was performed on transverse sections (3 M) of formalin-fixed paraffin-embedded sections using specific antibodies for Skp-2 and BrdUrd. Specific staining was detected with ExtrAvidin-horseradish peroxidase conjugate and diaminobenzidine. Nuclear counterstaining was performed with hematoxylin.
Western Blotting-Isolated rat SMCs were plated at a density of 4 ϫ 10 4 /cm 2 . Where indicated, cells were synchronized in G 0 by serum deprivation for 72 h. Total cell lysates were prepared at the indicated times using SDS-lysis buffer (50 mM Tris-HCl, pH 6.8, 10% glycerol, 1% SDS). Lysates were analyzed for protein content (Micro BCA assay kit, Pierce), and equal amounts of reduced protein (50 -100 g) were separated by polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membrane (Bio-Rad). Membranes were blocked with TBS-T (20 mM Tris-HCl, pH 7.6, 137 mM NaCl, 25 mM KCl, 0.25% Tween 20) containing 6% milk powder and incubated in primary antibody for 3 h at room temperature. Specific proteins were detected using horseradish peroxidase-conjugated secondary antibodies (Dako, Ely, UK). Peroxidase activity was detected using enhanced chemoluminescence (Amersham Biosciences).
Recombinant Adenoviruses-Control adenovirus (Ad:control) was a gift from Dr. G. W. G. Wilkinson (University of Wales College of Medicine, Heath Park, Cardiff, UK). Coding sequences isolated from donated plasmids excised and ligated into the adenovirus shuttle vector pDC515 (Microbix Biosystems Inc.). Replication-deficient adenoviruses were generated by site-specific recombination of the co-transfected FIG. 1. p27 Kip1 expression in intact rat aorta and isolated rat aortic SMCs. Freshly isolated rat thoracic aortas and isolated SMCs were stimulated with serum mitogens (10% FCS) for the indicated times. A, total cell lysates were analyzed by Western blotting for p27 Kip1 , phosphorylated and total retinoblastoma and GAPDH. The histogram shows the mean Ϯ S.E. of three experiments. B, total RNA from unstimulated aorta and isolated quiescent SMCs was analyzed by RT-PCR for p27 Kip1 mRNA, p27 Kip1 hnRNA, and GAPDH mRNA expression. Bars with the same symbols are significantly different from each other, p Ͻ0.05. The histogram shows the mean Ϯ S.E. of three experiments. C, cultured SMCs rendered quiescent for 72 h of serum deprivation were stimulated for 24 h with 10% FCS either in the presence or absence of 10 M MG-132. Lysates were analyzed for p27 protein by Western blotting.
shuttle and genomic plasmids in HEK293 cells. Viral stocks were plaque-purified, amplified, CsCl-banded, and titrated as previously described (31). Rat SMCs were infected with adenovirus at 200 plaqueforming units/cell for 3 h.

Different Regulation of p27 Kip1 in Intact Aorta and Isolated
SMCs-The effect of mitogen stimulation (10% FCS) on p27 Kip1 levels was analyzed in intact rat thoracic aortic segments in which SMC interact with their native basement membranes and vascular ECM and in cultured rat aortic SMCs that have been isolated from the native vascular ECM. In isolated SMCs, mitogen stimulation resulted in an early but partial reduction in p27 Kip1 levels after 4 h (p ϭ 0.0017, n ϭ 6) (Fig. 1A) and a more dramatic reduction (to 17.1% Ϯ 4.7% of control, p ϭ 0.0001, n ϭ 6) after 24 h (Fig. 1A). However, no reduction in p27 Kip1 levels was observed in intact aortic segments in response to mitogen stimulation (Fig. 1A), in agreement with our previous observations (28). The degradation of p27 Kip1 in isolated SMCs after 24 h of stimulation was associated with hyperphosphorylation of Rb protein (Fig. 1A), a critical step in the transition through the G 1 restriction. No Rb phosphorylation was detected in intact aorta (Fig. 1A). Although p27 Kip1 levels remained higher in intact aorta compared with isolated SMCs, the expression of p27 Kip1 mRNA was found to be equal in both unstimulated intact aorta and isolated cells (Fig. 1B). In addition, there was no difference in the rate of p27 Kip1 transcription, measured by RT-PCR analysis of prespliced p27 Kip1 hnRNA (Fig. 1B). Taken together, these data imply that the lower p27 Kip1 protein levels in isolated SMCs late in G 1 are mediated by post-transcriptional mechanisms. Previous work in other cells suggests that the late G 1 decline in p27 Kip1 may be mediated by enhanced 26 S proteasome-mediated degradation (32,33). Indeed, treatment of isolated SMCs with the 26 S proteasome inhibitor, MG-132, prevented the mitogeninduced reduction in p27 Kip1 (Fig. 1C).
Relationship between p27 Kip1 Expression and Skp-2 Expression-The above data demonstrating proteasome-mediated degradation of p27 Kip1 prompted us to investigate the role of Skp-2, a component of the SCF Skp-2 ubiquitin-protein isopeptide ligase that might mediate p27 Kip1 degradation. Skp-2 protein was undetectable in un-stimulated isolated SMCs but was strongly induced after 16 -24 h of stimulation with serum mitogens (Fig. 2, A and B). Interestingly, the appearance of Skp-2 protein after 16 -24 h of stimulation coincided with the late decrease in p27 Kip1 levels ( Fig. 2A) and the time when we previously demonstrated the cells are entering S-phase (28). Immunofluorescent staining revealed a predominantly nuclear localization of Skp-2 in serum-stimulated isolated SMC (Supplemental Fig. S1). Importantly, Skp-2 protein expression was undetected in either unstimulated or serum-stimulated intact aorta (Fig. 2B). Skp-2 mRNA expression was also found to be significantly lower in intact aorta (17.4 Ϯ 4%, p ϭ 0.0024, n ϭ 3) compared with isolated SMCs (Fig. 2C). This implies, first, that the altered Skp-2 protein levels between isolated SMC and aortas result from differences in steady-state mRNA levels. Secondly, the failure to up-regulate Skp-2 in intact aorta could underlie the failure to down-regulate p27 Kip1 protein levels.
Modulation of Skp-2 Activity by Gene Transfer Affects p27 Kip1 Degradation and Regulates SMC Proliferation-To investigate a causal role for Skp-2 in p27 Kip1 degradation and SMC proliferation, we constructed adenovirus vectors capable of expressing wild-type Skp-2 (Ad:WT-Skp-2) and dominant negative F-box-deleted mutant (Ad: ⌬F-Skp-2). Previous work showed that F-box-deleted Skp-2 retains the ability to interact with p27 Kip1 but does not recruit it to the SKP1-CUL1 complex, thus preventing ubiquitination of the substrate protein (34,35). Infection with Ad:WT-Skp-2 or Ad: ⌬F-Skp-2 resulted in elevated expression of 45-kDa wild-type Skp-2 or 35-kDa Fbox-deleted Skp-2, respectively (Fig. 3A, a and b). These blots were underexposed so as to clearly visualize the exogenous proteins; hence, endogenous levels of Skp-2 were undetectable. Immunofluorescent staining revealed that exogenously expressed Skp-2 was localized both in the nucleus and the cytoplasm (Supplemental Fig. S1). Infection with Ad:WT-Skp-2 resulted in a significant decrease (to 30.8 Ϯ 10% of control; p ϭ 0.0204, n ϭ 3) in p27 Kip1 levels compared with Ad:controlinfected cells, indicating that exogenous Skp-2 promotes p27 Kip1 degradation in vascular SMC (Fig. 3A, c). These Western blots are overexposed to detect low levels of p27 Kip1 in control cells. Importantly, infection with Ad:⌬F-Skp-2 resulted in a significant increase (to 156 Ϯ 13.5% of control; p ϭ 0.0253, n ϭ 3) in p27 Kip1 levels (Fig. 3A, d), further supporting a role for endogenous Skp-2 in p27 Kip1 degradation.
To determine whether the lack of Skp-2 expression in rat aorta is an important factor in limiting SMC proliferation, segments were infected in in vitro organ culture with 1 ϫ 10 10 plaque-forming units of either Ad:WT-Skp-2 or Ad:control. Three days postinfection Ad:WT-Skp-2, but not Ad:control, strongly up-regulated Skp-2 by Western blotting (Fig. 3B, inset); immunohistochemistry showed detectable Skp-2 in medial SMCs compared with none after control infection (Supplemental Fig. S2). Infection with Ad:WT-Skp-2 significantly increased SMC proliferation (9.64 Ϯ 1.96% BrdUrd-positive cells for Ad:WT-Skp-2 compared with 1.24 Ϯ 0.46% for Ad:control, n ϭ 6, p ϭ Ͻ0.0162) (Fig. 3B). Hence, adenovirus-mediated expression of wild-type Skp-2 rescued the ability of SMCs in intact aorta to proliferate in response to serum.
Skp-2 Expression Depends on the ECM-The difference in Skp-2 expression between SMCs in intact aorta and isolated SMC led us to investigate whether Skp-2 expression depends on the nature of the ECM. First, to determine whether the expression of Skp-2 depends on attachment to the ECM, isolated SMCs were synchronized in G 0 /G 1 by serum deprivation for 72 h, trypsinized, and replated either in suspension (over 10% agarose at high density to maintain cell:cell contacts) or adherent (on plastic) in the presence of 10% FCS. After 18 h of stimulation with serum mitogens, expression of Skp-2 protein was induced in the adherent cells but not in cells cultured in suspension (Fig. 4A). Expression of Skp-2 protein in suspension cultures was reduced by 98.1 Ϯ 1.91% (p ϭ 0.0004, n ϭ 3) compared with adherent cultures, whereas cell viability in suspension cultures, measured by trypan blue exclusion, was only reduced by 14 Ϯ 3.7%. This relatively small reduction in viability is therefore unlikely to account for the large reduction in Skp-2 expression observed in suspension cultures, indicating a requirement for adhesion-dependent signaling for Skp-2 expression.
To further test whether the nature of the ECM plays an important role in regulating Skp-2 expression, we cultured isolated SMCs on a fibrillar collagen matrix, which has previously been demonstrated to inhibit SMC proliferation by increasing p27 Kip1 levels (22). SMCs synchronized in G 0 /G 1 were plated either onto plastic or type-1 fibrillar collagen-coated plates in the presence of 10% FCS. Expression of Skp-2 protein 24 h later was found to be significantly lower (43.9 Ϯ 4.3%, p ϭ 0.0002, n ϭ 5) in cells cultured on a fibrillar type-1 collagen matrix compared with cells cultured on plastic, implying that the composition of the ECM is important in regulating Skp-2 expression (Fig. 4B). Similarly, Skp-2 steady-state mRNA expression was significantly lower (57.9 Ϯ 14.5%, p ϭ 0.0335, n ϭ 6) when cells were cultured on a collagen matrix (Fig. 4B). To further investigate the importance of ECM composition we compared the relative effects of fibronectin and laminin on Skp-2 expression. Laminin, a component of SMC basement membranes, has been implicated in promoting SMC differentiation into a non-proliferative phenotype, whereas fibronectin promotes SMC differentiation and proliferation (30). SMCs plated on laminin in the presence of 10% FCS expressed significantly less Skp-2 protein (45.3 Ϯ 15.2% compared with fibronectin, p ϭ 0.0361, n ϭ 4) and steady-state mRNA (17.4 Ϯ 4.9% compared with fibronectin, p ϭ 0.0001, n ϭ 6) than cells cultured on fibronectin, again demonstrating that ECM composition regulates Skp-2 expression (Fig. 4C).
Skp-2 Expression Is Dependent on Focal Adhesion Kinase-Focal adhesion kinase (FAK) is an important mediator of ECMdependent signals in numerous adhesion-dependent cell types, including vascular SMCs. To investigate the role played by FAK in the ECM-dependent regulation of Skp-2, we measured FAK activity (measured by phosphorylation of Tyr-397) in intact rat aorta and isolated rat SMC in culture. Phosphorylation of FAK Tyr-397 was found to be significantly lower in intact aorta compared with SMC cultured on plastic (Fig. 5, A and B).
However, FAK Tyr-397 phosphorylation was not affected by serum stimulation in aorta or culture cells (Fig. 5, A and B), implying ECM-dependent but not mitogen-dependent regulation. Total levels of FAK protein were equal in both aorta and cells (Fig. 5B).
We employed two approaches to test whether Skp-2 expression depends on FAK signaling. First, we used an adenovirus vector (Ad:FRNK) to overexpress FRNK, an endogenous protein homologous to the C-terminal domain of FAK that acts as an inhibitor of FAK activity (36). Infection with Ad:FRNK elevated levels of FRNK and inhibited global phosphorylation of FAK without affecting total FAK protein expression (Fig.  6A). Infection with Ad:FRNK but not Ad:control also resulted in a significant inhibition (to 53.1 Ϯ 10.7%, p Ͻ0.05, n ϭ 4) of Skp-2 expression (Fig. 6A). Second, we used a FAK mutant (Ad:FAK Y397F ) in which the tyrosine 397 autophosphorylation site that triggers FAK activation is mutated to a phenylalanine. Infection of cultured SMC with Ad:FAK Y397F , but not Ad:control, elevated FAK Y397F expression levels and inhibited wild-type FAK phosphorylation at Tyr-397 (Fig. 6B). Infection with Ad:FAK Y397F , but not Ad:control, also resulted in a significant inhibition (to 15.1 Ϯ 2.1%, p Ͻ0.05, n ϭ 3) of Skp-2 protein expression (compare with Ad:control-infected cells) (Fig. 6B). Neither Ad:FRNK nor Ad:FAK 397 infection affected Skp-2 mRNA expression (Fig. 6C). Interestingly, treatment with the proteasome inhibitor MG-132 for 6 h blocked the down-regulation of Skp-2 induced by Ad:FAK 397 (Fig. 6D).

DISCUSSION
In this study, we sought further information regarding the mechanisms by which the ECM controls SMC proliferation in FIG. 4. Skp-2 expression in SMCs is adhesion-dependent. Rat SMCs were rendered quiescent by serum deprivation for 48 h. Cells were then seeded in the presence of serum mitogens (10% FCS) onto tissue culture plastic (Adherent) or over 10% agarose (Suspension) at high density (A). Quiescent SMCs were seeded onto tissue-culture plastic or tissue culture plastic coated with a type-1 fibrillar collagen matrix (B) or tissue culture plastic coated with either 20 g/ml laminin or 20 g/ml fibronectin in the presence of serum mitogens (C). Cell lysates and total RNA were prepared 24 h later and analyzed for Skp-2 expression by Western blotting and RT-PCR. The histogram shows the mean Ϯ S.E. of at least four experiments.

FIG. 5. FAK phosphorylation is associated with Skp-2 expression.
Freshly isolated rat thoracic aortas and isolated rat SMCs were stimulated with serum mitogens (10% FCS) for the indicated times. Total cell lysates were analyzed for phospho-FAK Y397 and total FAK protein by Western blotting (A) and densitometry (B). response to mitogens and, hence, the basis of SMC quiescence in intact aortas. Our data confirmed that SMC in intact rat aortas fails to down-regulate p27 Kip1 or proliferate in response to serum. More importantly, our data show for the first time that late G 1 levels of p27 Kip1 are down-regulated in isolated SMC, but not aortas, by proteasomal degradation mediated by the Skp-2 ubiquitin-protein ligase subunit. First, we show that Skp-2 protein is undetectable after serum stimulation of aortas but up-regulated in isolated cultured SMC. Second, adenovirus-mediated expression of wild-type Skp-2 or a dominant negative form decrease or increase, respectively, p27 Kip1 levels. Third, wild-type Skp-2 promotes and dominant negative Skp-2 inhibits SMC proliferation. Finally, we present new data that Skp-2 expression is adhesion-dependent in SMC and negatively regulated by the ECM components collagen and laminin, which do not support proliferation, but up-regulated by fibronectin, which does. This ECM-dependent regulation of Skp-2 is mediated, at least in part, by signaling through focal adhesion kinase. Our results imply that Skp-2 plays a significant role in the regulation of SMC proliferation by the ECM and that this, at least in part, explains why SMC in intact aortas are maintained quiescent.
It has been appreciated for many years that the rates of SMC proliferation in intact arteries are very low. Although this may partly reflect low free concentrations of mitogens, it is clear that uninjured SMCs, unlike fibroblasts, remain quiescent even when exposed to mitogens in organ culture or in vivo. Nevertheless, the potential of SMC to proliferate rapidly can be unlocked by digesting arterial tissue with collagenase, which frees SMC from contacts with their normal basement membranes and allows new contacts with existing and nascent interstitial matrix compo-nents. Fibroblasts by contrast do not express basement membranes. It has been postulated therefore that the ECM is an important regulator of SMC proliferation. Moreover, components of the native ECM (e.g. polymerized collagen, laminin) have negative growth regulatory properties, whereas those up-regulated in remodeling ECM (e.g. monomeric collagen, fibronectin) have positive effects (22,30,37). The data are consistent with the concept that vascular ECM in intact aortas either inhibits or is not permissive to proliferation.
Koyama et al. (22) observed that the suppression of SMC proliferation by a type-1 fibrillar collagen is mediated, at least in part, by an increase in the level of the cyclin-dependent kinase inhibitor p27 Kip1 . We also found previously that low rates of SMC proliferation in intact rat aorta are associated with constitutively elevated levels of p27 Kip1 (28). In contrast, isolated cultured SMC acquired the ability to down-regulate p27 Kip1 and proliferate in response to mitogen stimulation. The importance of p27 Kip1 as a regulator of SMC proliferation in vivo is clear from studies of balloon injury to porcine femoral arteries, which induced SMC proliferation, coincident with an early reduction of p27/Kip levels (20). The later injury-induced overexpression of p27 Kip1 occurred coordinately with a reduction in proliferation and a return to quiescence (20). Furthermore, gene transfer of p27 Kip1 to porcine femoral arteries significantly reduced SMC proliferation and neo-intima size after balloon injury (19). Taken together, these data show that p27 Kip1 is an important regulator of SMC proliferation and suggest that the vascular ECM regulates SMC proliferation at least in part by controlling the levels of p27 Kip1 . However, it is not clear how signals from the ECM control the levels of p27 Kip1 in SMCs. Recent research on other cell types (32,35) led us to consider a role for the F-box protein Skp-2 in the regulated ubiquitination of p27 Kip1 and cellular proliferation. Skp-2, originally identified as a protein interacting with cyclin A and CDK2 is expressed at high levels in numerous tumors and is often associated with advanced clinical stages (38 -41). Levels of Skp-2 in these tumors are often inversely related to the levels of p27 Kip1 (38 -41). Our results show a similar relationship in SMCs where expression of Skp-2 during the G 1 phase of the cell cycle is inversely related to the level of p27 Kip1 expression. Moreover up-regulation of Skp-2 parallels the ability of SMC in aortas or in isolated cell culture to degrade p27 Kip1 in response to mitogen stimulation. Importantly, exogenous expression of Skp-2 in aortas rescues the ability of SMCs to proliferate in response to growth factors, implying that low levels of endogenous Skp-2 expression in intact arteries are an important factor limiting proliferation and maintaining SMC quiescence.
SMC in intact aorta fail to up-regulate Skp-2 in response to mitogens despite, as we have previously demonstrated, equal mitogen-activated protein kinase activation and cyclin D and cyclin E up-regulation compared with isolated SMC (28). One explanation could be that signals from the vascular ECM regulate Skp-2 expression coordinately with signals from growth factors. The requirement for ECM signals is clearly demonstrated by the complete loss of Skp-2 expression in SMC forced into suspension. Similar adhesion-dependent regulation of Skp-2 was recently demonstrated in lung fibroblasts (2), implying that at least this aspect of matrix regulation is common to both cell types. We went on to show up-regulation of Skp-2 expression by fibronectin, a component of the vascular interstitial matrix that is up-regulated during matrix remodeling after injury (14), but not fibrillar type-1 collagen. SMC cultured on a laminin matrix, albeit a mixture of laminin-1 isoforms that may not completely mirror those present in SMC basement membrane, also show reduced Skp-2 expression. These data imply that regulation of Skp-2 expression by the ECM is likely to be an important mechanism maintaining SMC quiescence in healthy vessels and initiating SMC proliferation during vascular diseases, where extensive remodeling of the ECM is known to occur.
Focal adhesion kinase plays an important role in regulating cellular proliferation by integrating signals from growth factor receptors and the ECM in numerous adhesion-dependent cell types. Concurrent with this, phosphorylation of FAK at Tyr-397, an auto-phosphorylation event responsible for the activation of FAK, is significantly lower in rat aorta compared with isolated SMCs in culture and correlates with the ability of SMCs to express Skp-2 and proliferate. Phosphorylation at this residue creates binding sites for other SH2 domain-containing kinases that trigger further phosphorylation events within the FAK C-terminal domain. The potent inhibition of Skp-2 protein, but not mRNA expression in SMCs overexpressing FAK mutated at Tyr-397 or FRNK, clearly demonstrates the importance of FAK activity for Skp-2 protein expression and implies regulation at a post-transcriptional level. The ability of the proteasome inhibitor MG-132 to block the down-regulation of Skp-2 protein induced by FAK 397 demonstrates that FAK signals regulate Skp-2 at the level of Skp-2 protein stability. Taken together, our data demonstrate that FAK activity is essential for the stability of Skp-2 protein and, hence, progression through G 1 -S phases of the cell cycle.
A number of studies have demonstrated that general proteasome inhibitors have potent antiproliferative effects, largely in the cancer field (42). The proteasome inhibitor MG-132 has also been shown to effectively inhibit restenosis in the rat carotid balloon injury model (23). As these general proteasome inhib-itors block degradation of all ubiquitinated proteins, they often have multiple effects on the cell, including induction of apoptosis and inhibition of inflammatory pathways, such as NFB activation. Our data demonstrating the important role of Skp-2 in controlling SMC proliferation suggest that agents targeting Skp-2 function could prove more selective antiproliferative therapies for the treatment of restenosis and late vein graft failure.