S-phase kinase-associated protein-2 (Skp2) promotes vascular smooth muscle cell proliferation and neointima formation in vivo.

OBJECTIVE: Vascular smooth muscle cell (VSMC) proliferation plays an important role in the development of postangioplasty or in-stent restenosis, venous graft failure, and atherosclerosis. Our previous work has demonstrated S-phase kinase-associated protein-2 (Skp2), an F-box subunit of SCF(Skp2) ubiquitin ligase, as an important mediator and common final pathway for growth factors, extracellular matrices, and cyclic-nucleotides to regulate VSMC proliferation in vitro. However, whether alteration of Skp2 function also regulates VSMC proliferation in vivo and neointimal thickening postvascular injury remains unclear. We investigated the effect of Skp2 on VSMC proliferation and neointimal formation in vivo. METHODS AND
RESULTS: Firstly, we demonstrated that Skp2-null mice developed significantly smaller neointimal areas than wild-type mice after carotid ligation. Secondly, to further identify a local rather than a systemic effect of Skp2 alteration, we demonstrated that adenovirus-mediated expression of dominant-negative Skp2 in the balloon-injured rat carotid artery significantly increased medial p27(Kip1) levels, inhibited VSMC proliferation, and the subsequent neointimal thickening. Lastly, to determine if Skp2 alone is sufficient to drive VSMC proliferation and lesion development in vivo, we demonstrated that adenovirus-delivery of wild-type Skp2 to the minimally-injured rat carotids is sufficient to downregulate p27(Kip1) protein levels, enhanced medial VSMC proliferation, and the neointimal thickening.
CONCLUSION: This data provides, we believe for the first time, a more comprehensive understanding of Skp2 in the regulation of VSMC proliferation and neointimal formation and suggests that Skp2 is a promising target in the treatment of vasculoproliferative diseases.

Vascular smooth muscle cell (VSMC) proliferation is rare in normal adult arteries but increases during atherosclerosis. The resulting fibrous cap is vital in protecting the plaque from rupture and subsequent myocardial infarction. On the other hand, excessive VSMC proliferation underlies the development of postangioplasty or post-stenting restenosis, venous graft failure, and transplant arteriosclerosis. Tight regulation of VSMC proliferation is therefore essential. Although a myriad of factors are able to regulate VSMC proliferation, cell-cycle activation is the final common pathway, and targeting cell-cycle regulation has emerged as a promising strategy in the treatment of vasculoproliferative diseases.

Progression through the G1 phase of the cell-cycle is regulated by phosphorylation and inactivation of retinoblastoma (Rb) proteins. This is promoted by the cyclins (cyclin A, D, and E), which associate with and activate the cyclin-dependent kinases (CDK2, CDK4, and CDK6). These CDK:cyclin complexes are subject to negative regulation by the cyclin-dependent kinase inhibitors (CDKIs), such as the Cip/Kip family of CDKIs (p21Cip1, p27Kip1, and p57Kip2). Downregulation of CDKIs during late G1 after addition of growth factors in vitro or balloon injury in vivo is a critical step in allowing activation of cyclin:CDK complexes and subsequent G1/S phase transition. Skp2 is an F-box component of SCFSkp2 ubiquitin ligase, implicated in the polyubiquitination and proteasomal degradation of many cell-cycle regulators, including p27Kip1. Although many cell-cycle proteins are potential substrates of Skp2, p27Kip1 is considered most important, because it accumulates in Skp2-null cells and the phenotypic and histologic abnormalities in Skp2-null mice are almost completely reversed in Skp2/p27Kip1 doubly-null mice.

Increased Skp2 expression along with the related p27Kip1 deregulation has been deemed a marker of poor prognosis in cancer disease. However, the role of Skp2 in vascular disease is less well established. Our previous work demonstrated that isolated VSMC in culture express high levels of Skp2. Silencing of Skp2 in these cells with siRNA or adenovirus-mediated over-expression of dominant-negative Skp2 (DN-Skp2), increases p27Kip1 levels, and inhibits VSMC proliferation, consistent with an essential role for Skp2. Moreover, both positive (eg, growth factors and matrix contact) and negative (eg, cyclic-nucleotides) factors for VSMC growth also regulate Skp2 expression in a way reflecting their effects on VSMC proliferation in vitro. Despite this, evidence that Skp2 regulates VSMC proliferation in vivo remains indirect. For example, we showed that elevation of 3′-5′-cyclic adenosine monophosphate (cAMP) in rat carotid arteries after balloon injury suppresses Skp2 expression, increases p27Kip1 levels, and inhibits both VSMC proliferation and neointima formation. Over-expression of a dominant-negative form of the Rac1 GTPase in the same model showed the same association between decreased Skp2 and increased p27Kip1 levels with suppression of VSMC proliferation. In the present study, we sought for the first time to provide direct evidence that Skp2 expression regulates VSMC proliferation in vivo sufficiently to affect the final neointimal thickening.

Materials

All reagents, except for those stated otherwise, were purchased from Sigma-Aldrich Co (St. Louis, Mo). Antibodies for Skp2 were obtained from US Biological Inc (Swampscott, Mass) and Invitrogen (Carlsbad, Calif). Mouse anti-p27Kip1 and anti-p21Cip1 antibodies were from BD Transduction Laboratories (Oxford, United Kingdom). Rabbit anti-phospho-Rb and mouse anti-bromodeoxyuridine (BrdU) antibodies were bought from Cell Signaling Technology (Danvers, Mass) and MP Biomedicals (Solon, Ohio), respectively. All secondary antibodies were purchased from Dako (Burlingame, Calif) and NovaRed from Vector Laboratories (Burlingame, Calif).

Methods

Animals

Homozygous Skp2−/− and wild type control mice were obtained by crossing heterozygous Skp2+/− mice. Genotypes of mice were determined by polymerase change reaction (PCR) of tail tip DNA. Male Sprague-Dawley (SD) rats were obtained from Charles River (Margate, United Kingdom) or BioLASCO Taiwan Co Ltd (Taiwan). The housing and care of the animals and all procedures used adhered to the guidelines and regulations of the United Kingdom Animal (Scientific Procedures) Act 1986, United Kingdom, and Mackay Memorial Hospital, Taiwan.

Ligation of mouse common carotid artery

Male Skp2−/− and Skp2+/+ mice (20-30 grams) were anesthetized by inhalation of isoflurane and oxygen (0.2 L/minute). The left common carotid artery was ligated near the carotid bifurcation as described. All animals were killed 28 days later and perfusion fixed with 4% formaldehyde/phosphate buffered saline (PBS) for 7 minutes before excision of ligated carotid arteries. Vessels were cut into longitudinal sections to increase the accuracy of neointimal area calculation. Sections were stained with Elastic Van Gieson (EVG) and hematoxylin and eosin (H&E) stain and neointimal areas, defined as the area of tissue within the internal elastic lamina, quantified using Image Pro analysis software (MediaCybernetics, Bethesda, Md). Vessel diameter (distance between the external elastic lamellae) and medial thickness (distance between the internal and external elastic lamellae) were measured 300-400 μm distal to the ligation point.

Balloon injury of rat common carotid artery

Balloon injury of left common carotid artery was performed as described by Clowes et al. Briefly, male SD rats (∼400 grams) were anesthetized with an intraperitoneal injection of a mixture of ketamine hydrochloride (100 mg/kg) and xylazine (10 mg/kg). The bifurcation of the left common carotid artery was surgically exposed, and a 2F arterial embolectomy balloon catheter (Edwards Lifesciences Co, Irvine, Calif) was introduced through an arteriotomy on left external carotid artery. The balloon was inflated with saline and pulled back with rotation through common carotid artery three times. The injured common carotid arteries were then subjected to adenovirus-mediated gene delivery. For cell proliferation assay, animals were injected subcutaneously with BrdU solution (25 mg/kg) at 17, 9, and 1 hour before euthanasia.

Filament de-endothelialization of rat common carotid artery

The procedures of filament de-endothelialization of rat common carotid artery were modified from the methods developed by MA Reidy's group. Male SD rats were anesthetized and the left common carotid arteries were approached as described in balloon injury. The endothelium was then denuded from a carotid artery by a 5-0 nylon loop. Briefly, the nylon loop was introduced into the left external carotid artery through a Trocar. The loop was then slowly pulled back along the common carotid artery with constant rotation for ∼30 seconds. The Trocar was retrieved and the artery underwent subsequent adenoviral infection for specific gene expression. Animals were killed and cell proliferation was evaluated as described in the previous section.

Adenovirus-mediated gene delivery and expression

Recombinant replication-defective adenoviruses encoding wild-type Skp2 (Ad:WT-Skp2), dominant-negative Skp2 (Ad:DN-Skp2), and β-galactosidase (Ad:βGal) have been described previously. To infect rat common carotid artery with adenoviruses for gene expression, both proximal end of common carotid artery and internal carotid artery were temporarily clamped and an aliquot (100 μl) of adenovirus (1 × 109 pfu/mL)/PBS solution was infused into the filament- or balloon-injured common carotid artery through the external carotid artery. Heparin (500 IU/kg) was given intraperitoneally to prevent arterial thrombosis. The adenoviral solution was retained in the vessel for 30 minutes and then withdrawn. After ligation of external carotid artery, the clamps were removed and the blood flow was restored. The infection rate using this protocol was ∼40% of medial smooth muscle cells according to the extent of β-galactosidase expression (data not shown).

Immunohistochemical analysis

Rat carotid arteries were perfused fixed in formalin and cut into ∼0.3-cm-long pieces (three pieces per artery) before being embedded in paraffin wax. Each paraffin block containing three pieces of the same carotid artery was cut into 4-μm-thick sections so that each section had three arterial rings representing arterial lesion 0.3 cm apart in the actual artery. For analysis of cell proliferation (BrdU) and cell-cycle regulators (hyperphosphorylated Rb, p27Kip1, p21Cip1, and Skp2), all cells in each arterial ring (∼300-450 cells per ring) were counted and three arterial rings were analyzed for each carotid artery. The number of positive cells was expressed as a percentage of total cell number calculated by counting hematoxylin counterstained nuclei.

Cell apoptosis was assessed by terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) method using DNA Fragmentation Kit (Calbiochem) according to the manufacturer's instruction. Briefly, arterial sections were incubated with terminal deoxynucleotidyl transferase (TdT) end-labeling cocktail containing TdT, biotin-dUTP, and TdT buffer. The DNase I (0.375 Kunitz unit/μl) treated arterial rings were served as positive control. The arterial rings were then incubated with fluorescein isothiocyanate (FITC) conjugated avidin. Total cell nuclei visualized by Hoescht 33258 and nuclei containing cleaved DNA (FITC-labeled) cells were counted.

For the evaluation of neointimal thickening, arterial sections were subjected to H&E staining, followed by neointimal (the tissue within the ring of internal elastic lamina) and medial areas measured using Image Pro software system (MediaCybernetics). The extents of neointimal thickening were presented as intima/media (I/M) ratios.

Statistical analysis

All data were expressed as mean ± SEM. Unpaired two-tailed t test was adopted to test the significance, if appropriate. Where indicated, data was analyzed using nonparametric Mann-Whitney test. A P value less than .05 was considered significant.

Results

To demonstrate the overall effect of Skp2 gene deletion on the neointimal thickening, Skp2−/− and Skp2+/+ mice were subjected to common carotid arterial ligation, which induces arterial remodeling, VSMC proliferation, and neointimal thickening. Compared with wild-type Skp2 mice (n = 11), Skp2−/− mice (n = 8) developed significantly smaller neointimal areas (130,371 ± 29,054 vs 52,903 ± 35,291 μm2; Mann-Whitney test, P = .026) 28 days after common carotid ligation (Fig 1). However, vessel diameter (309.5 ± 19.5 μm vs 384.8 ± 40.6 μm, P = .086 for wild-type and Skp2-null, respectively) and medial wall thickness (40.6 ± 2.1 μm vs 41.0 ± 2.2 μm, P = .9) was not significantly different between wild-type and Skp2-null mice.

Fig 1

Neointimal thickening 28 days after common carotid ligation in Skp2+/+ and Skp2−/− mice. Carotid ligation was performed on wild-type or homozygous Skp2-null mice. A, Representative longitudinal sections of from both groups 28 days post-ligation. Arrows locate ligation position. B, Quantification of neointimal areas. The comparison was made using Mann-Whitney test.

Further mechanistic studies in the mouse model were impossible because of the small number of longitudinal sections that could be cut from each lesion and the difficulties of breeding additional Skp2-null mice due to low fertility and fecundity. To more specifically elucidate the local effect of Skp2 on VSMC proliferation and neointimal thickening, and to probe more deeply into the mechanisms, we used adenovirus-mediated delivery of either wild-type Skp2 or DN-Skp2 in balloon- or filament-injured rat carotid arteries, respectively. DN-Skp2 is an F-box-deleted Skp2 mutant, which preserves substrate-binding ability but loses its ability to form SCFSkp2 complex, and, therefore, inhibits Skp2 function by competing with endogenous Skp2. To test whether local inhibition of Skp2 function also reduces VSMC proliferation in vivo, DN-Skp2 was expressed in balloon-injured rat carotid arteries using an adenoviral vector. The DN-Skp2 expression significantly enhanced medial VSMC p27Kip1 expression (from 15.7 ± 1.3% to 24.1 ± 2.5% of medial VSMC, P = .012, n = 6 and 5, respectively) but not p21Cip1 expression (from 20.5 ± 3.2% to 30.2 ± 3.9% of medial VSMC, P = .08). DN-Skp2 also inhibited VSMC proliferation measured as BrdU index (from 37.3 ± 2.5% to 17.0 ± 3.2%, P < .01, n = 6 and 5, respectively) and hyper-phosphorylation of Rb protein (from 43.5 ± 4.5% to 25.9 ± 2.8% of medial VSMC, P = .012), a marker of G1-S phase transition, 4 days after injury, compared with vessels infected with a β-galactosidase expressing control adenovirus (Fig 2). In addition, subsequent neointimal thickening 14 days postinjury was also significantly reduced by DN-Skp2 (I/M ratios, from 1.58 ± 0.20 to 0.49 ± 0.21, P < .01, n = 5) (Fig 2). Importantly, frequency of apoptosis, detected by TUNEL staining (Fig E1, online only), was not increased by expression of dominant-negative Skp2 (less than 0.15% in both Ad: βGal and Ad:DN-Skp2-infected vessels). This data demonstrates that Skp2 activity is required for maximal VSMC proliferation and neointimal lesion development in vivo.

Fig 2

DN-Skp2 up regulates p27Kip1 levels, inhibits VSMC proliferation and neointima formation in balloon-injured rat common carotid arteries. Balloon-injured rat common carotid arteries were infected with either Ad:βGal (control) or Ad:DN-Skp2 by the luminal retention of 100 μl adenoviral solution (1 × 109 pfu/mL) for 30 minutes. The animals were injected with BrdU (25 mg/kg) at 17, 9, and 1 hour before euthanasia on the day 4 post balloon-injury. A, Representative pictures of arterial sections undergoing staining for p27Kip1, BrdU, phospho-Rb, p21Cip1, 4 days post injury and hematoxylin and eosin staining 14 days post injury, as indicated. Scale bars = 50 μm. Arrows indicate labeled cells. B, Quantified results of the immunohistochemistry and I/M ratios. Data are expressed as % of labeled cells. ** indicates P < .01, * indicates P < .02.

We next sought to determine if increased Skp2 activity alone, in the absence of a large medial injury inflicted by balloon angioplasty, can induce VSMC proliferation and neointimal thickening in vivo. To this end, we used adenoviral-delivery of wild-type Skp2 to carotids subjected to filament de-endothelialization. Unlike balloon injury, filament de-endothelialization results in minor medial injury, minimal VSMC proliferation, and thus, a limited thickness of neointima. Compared with filament-injured carotids infected with control adenovirus (Ad:βGal), those infected with wild-type Skp2 adenovirus (Ad:WT-Skp2) had substantially increased Skp2 expression (5 ± 1% compared to 43 ± 5% of medial VSMC, P < .01), accompanied by reduced expression of p27Kip1 (from 18.4 ± 5.4% to 0.6 ± 0.3%, P = .016, n = 4 and 6, respectively) but not p21Cip1 (from 25.8 ± 4.8 to 18.0 ± 2.4%, P = .26). Wild-type Skp2 also enhanced VSMC proliferation measured by BrdU incorporation (6 ± 2% to 32 ± 5%, P < .01, n = 4 and 6, respectively) and hyper-phosphorylation of Rb (from 9.5 ± 2.1% to 28.3 ± 5.5%, P = .015) 4 days after injury (Fig 3). Eight days after filament injury, Ad:WT-Skp2-infected carotids gave rise to significantly higher I/M ratios than those infected with Ad:βGal (0.11 ± 0.03 compared to 0.42 ± 0.09, P = .024, n = 4 and 5, respectively) (Fig 4).

Fig 3

Skp2 over expression down regulates p27Kip1 levels and increases VSMC proliferation in minimally-injured rat common carotid arteries. After filament de-endothelialization, carotid arteries were infected with Ad:βGal or Ad:Skp2. A, Representative pictures of both groups. Carotid arterial sections were subjected to immunohistochemistry for the staining of Skp2, p27Kip1, phospho-Rb, p21Cip1 and BrdU 4 days post infection. Of note, Skp2- and BrdU-labeled cells were detected in Ad:βGal infected vessels mostly over the superficial layer of media (arrows), whereas those infected with Ad:Skp2 tended to have Skp2 and BrdU-positive cells trans-medially. Scale bars = 50 μm. Arrows indicate labeled cells. B, Quantification of the immunohistochemistry. Data are expressed as the percentage of labeled cells. ** indicates P < .01, * indicates P < .02.

Fig 4

Wild-type Skp2 enhances neointimal thickening in minimally-injured rat carotid arteries. Carotid arteries were treated as in Fig 4. Eight days after filament de-endothelialization, the carotids from both treatment groups were subjected to hematoxylin and eosin staining as described in Fig 2. A, Representative pictures of both groups. Scale bar = 100 μm. B, Quantified results of I/M ratios.

Discussion

Although the role of Skp2 in the oncogenesis is well recognized, its role in the regulation of VSMC proliferation is less established. The current studies directly demonstrate for the first time that Skp2 function is involved in injury-induced VSMC proliferation and neointima development in vivo. Furthermore, we demonstrate that Skp2 over-expression is sufficient to drive both proliferation and lesion development after endothelial denudation in the absence of major medial injury. These experiments implicate Skp2 as a novel target for the treatment of vasculoproliferative diseases, such as vascular restenosis after percutaneous vascular intervention.

Numerous factors, such as growth factors, inflammatory mediators, extracellular matrix attachment, cell-cell contact, and cyclic nucleotides are involved in the regulation of VSMC proliferation. Nevertheless, the cell-cycle is the final machinery to integrate and orchestrate intricate upstream signals to delicately tune VSMC proliferation. Recent encouraging clinical reports show that stents coated with rapamycin or paclitaxel, which target G1-S or G2-M transition, respectively, prevent the occurrence of restenosis after coronary intervention. This strongly supports the notion that the cell-cycle, the common final pathway for VSMC proliferation, is an ideal target for the treatment of vascular restenosis and, by implication, other vasculoproliferative diseases.

Our previous studies established Skp2 as an important mediator in the final common pathway for VSMC proliferation in vitro. For example, we showed that positive mediators for VSMC proliferation (growth factors and matrix contact) increase Skp2 levels, whereas negative regulators (cyclic nucleotides) decrease them. These effects are mediated through regulation of focal adhesion kinase, phosphoinositide 3-kinase, and Rac1 signaling pathways. DN-Skp2 blocks the stimulation of VSMC proliferation after growth factor addition, whereas overexpression of native Skp2 reverses the inhibitory effects of cyclic-nucleotides.

In vivo we also demonstrated that Skp2 expression levels are strikingly enhanced after balloon injury with a time course that perfectly mirrors that of VSMC proliferation. Moreover, adventitial application of forskolin (an activator of endogenous cyclic AMP levels) inhibited Skp2 expression, VSMC proliferation, and neointima formation. This indirect evidence supports the role of Skp2 in VSMC proliferation in vivo. Justified by this correlative evidence, we directly tested for the first time whether Skp2 is necessary and sufficient for VSMC proliferation and neointima formation in vivo after vascular injury in mice and rats. Initially, we demonstrated that neointima formation after carotid ligation was significantly attenuated in Skp2−/− mice, compared with that in Skp2+/+ mice. Although Skp2−/− mice tend to be smaller than their wild-type counterparts we found that carotid diameter and medial thickness was not different and no gross vessel abnormalities were observed at the time of surgery. Impaired lesion formation in these mice is consistent with previous reports of reduced rates of cellular proliferation. It is important to note that impaired lesion formation in Skp2−/− mice could be explained in part by actions of Skp2 in other cells, such as bone marrow progenitor cells, inflammatory cells, fibroblasts, and endothelial cells, all of which may interact with VSMC and contribute significantly to the final neointimal thickening. To more specifically pinpoint the effects of Skp2 inhibition on local VSMC and neointimal thickening, DN-Skp2 was expressed in balloon-injured rat carotid arteries via adenovirus-mediated gene transfer. The expression of DN-Skp2 resulted in elevated levels of the CDKI p27Kip1, which is normally rapidly downregulated soon after vascular injury. These observations are consistent with the hypothesis that injury-induced expression of Skp2 is responsible for the downregulation of p27Kip1 after vascular injury. Downregulation of p27Kip1 is thought to be an important step in removing the brake on proliferation, because forced expression of p27Kip1 inhibits S-phase entry and neointima formation after balloon injury and p27Kip1-null mice display a hyper-proliferative phenotype and exaggerated responses to vascular injury. We showed here that expression of DN-Skp2 prevented downregulation of p27Kip1 and resulted in significantly reduced levels of hyperphosphorylated Rb protein, a marker of G1-S transition, reduced rates of VSMC proliferation and smaller neointimal lesions. Expression of DN-Skp2 did not, however, significantly effect the expression of p21Cip1, even though our previous data demonstrated that Skp2 regulates p21Cip1 levels in VSMC in vitro. However, p21Cip1 regulation seems to be more complicated than that of p27Kip1 in these cells. The p21Cip1 levels actually increase in response to mitogen stimulation, coordinately with increases in Skp2, which functions to maintain p21Cip1 within growth permissive levels. This complex pattern of regulation may account for the lack of a significant change in p21Cip1 in this study. Importantly, apoptosis rates were not increased in vessels expressing DN-Skp2, indicating that reduced lesion formation was due to inhibition of proliferation rather than increased cell death. Likewise, expression of DN-Skp2 in VSMC in vitro does not result in increased rate of apoptosis (unpublished data). Taken together with our observations from Skp2−/− mice, our data demonstrates that injury-induced expression of Skp2 is a major mechanism regulating p27Kip1 levels and is essential for VSMC proliferation and the final neointimal thickening.

We previously demonstrated that exogenous expression of Skp2 is able to force VSMC to enter S-phase in the uninjured rat aorta in organ cultures. In the present study, we asked whether increased Skp2 expression would be sufficient to induce VSMC proliferation and neointimal lesion development, in the absence of major medial injury in vivo. Consistent with previous reports, we showed that filament de-endothelialization of rat carotid artery caused minimal levels of VSMC proliferation and neointimal thickening, and only a small increase in Skp2 expression. However, forced expression of wild-type Skp2 resulted in a remarkable reduction of p27Kip1 expression and a transmedial increase of BrdU-labeled cells. This demonstrates that Skp2 alone is sufficient to promote p27Kip1 downregulation and VSMC proliferation in the absence of major medial injury.

Taken together, our data provides the first direct and robust evidence that Skp2 plays an important role in promoting VSMC proliferation and neointimal thickening in vivo in response to vascular injury. These results suggest that Skp2 may represent an attractive target for future gene or pharmacologic-based therapies to combat vasculoproliferative diseases. Targeting Skp2 in this way may be more effective than approaches designed to block the activity of single downstream cell-cycle regulators, such as p27Kip1, because Skp2 regulates the levels and activity of multiple cell-cycle proteins. However, more work is required to elucidate the role and regulation of Skp2 in other vascular cells, particularly endothelial cells, before this may become a reality.

Author contributions

Conception and design: MB, AN

Analysis and interpretation: YW, MB

Data collection: YW, GS, KS, KIN, KN, HY, MB

Writing the article: YW, AN, MB

Critical revision of the article: YW, MB

Final approval of the article: MB

Statistical analysis: YW, MB

Obtained funding: MB, AN

Overall responsibility: MB