Regulation of Vascular Smooth Muscle Proliferation by Heparin

Uncontrolled proliferation of vascular smooth muscle cells (VSMCs) contribute to intimal hyperplasia during atherosclerosis and restenosis. Heparin is an antiproliferative agent for VSMCs and has been shown to block VSMC proliferation both in tissue culture systems and in animals. Despite the well documented antiproliferative actions of heparin, its cellular targets largely remain unknown. In an effort to characterize the mechanism of the antiproliferative property of heparin, we have analyzed the effect of heparin on cell cycle in VSMC. Our results indicate that the heparin-induced block in G1 to S phase transition is imposed by p27kip1-mediated inhibition of cyclin-dependent kinase 2 activity. Further analysis of p27kip1 mRNA levels showed that the increase in p27kip1 protein levels in heparin-treated VSMC occurs at posttranscriptional levels. We present evidence that heparin causes stabilization of p27kip1 protein during G1 phase and thereby prevents activation of cyclin-dependent kinase 2.

Uncontrolled proliferation of vascular smooth muscle cells (VSMCs) contribute to intimal hyperplasia during atherosclerosis and restenosis. Heparin is an antiproliferative agent for VSMCs and has been shown to block VSMC proliferation both in tissue culture systems and in animals. Despite the well documented antiproliferative actions of heparin, its cellular targets largely remain unknown. In an effort to characterize the mechanism of the antiproliferative property of heparin, we have analyzed the effect of heparin on cell cycle in VSMC. Our results indicate that the heparin-induced block in G 1 to S phase transition is imposed by p27 kip1mediated inhibition of cyclin-dependent kinase 2 activity. Further analysis of p27 kip1 mRNA levels showed that the increase in p27 kip1 protein levels in heparin-treated VSMC occurs at posttranscriptional levels. We present evidence that heparin causes stabilization of p27 kip1 protein during G 1 phase and thereby prevents activation of cyclin-dependent kinase 2.
The proliferation of vascular smooth muscle cells (VSMCs) 1 is a key event in the development of atherosclerotic lesions and postangioplasty restenosis (1). In a normal artery, the VSMCs are in a non-proliferative quiescent state and show a well differentiated contractile phenotype. After the vascular injury, there is a loss of differentiated phenotype and a shift to a synthetic phenotype, which is also accompanied by entry into the cell cycle and proliferation (2). Several cytokines, growth factors, vasoregulatory molecules, and extracellular matrix components exert their effects on the proliferation of VSMC. The development of an atherosclerotic lesion can be blocked significantly by effective inhibition of VSMC proliferation (3). Natural glycosaminoglycans such as heparin are also known inhibit VSMC proliferation in vitro in tissue culture (4 -6) and in vivo in animal models (7,8). Despite the well documented antiproliferative effect of heparin on VSMC, the molecular mechanisms responsible for inhibition of cell cycle remain uncharacterized.
Cellular proliferation is regulated primarily by regulation of the cell cycle (9), which consists of four distinct sequential phases (G 0 /G 1 , S, G 2 , and M). This tightly regulated temporal order is controlled by the sequential activation of certain serine/threonine protein kinases known as cyclin-dependent kinases (Cdks) that phosphorylate the Rb protein (10). In quiescent cells, Rb exists in its hypophosphorylated state and is thus able to bind and sequester the members of E2F family of transcription factors (11). Phosphorylation of Rb at multiple sites by Cdks causes the release of E2F because hyperphosphorylated Rb cannot bind and sequester E2F factors, thus enabling them to activate transcription of genes whose products are absolutely essential for further cell cycle progression (12). The activity of Cdks is further regulated negatively by a number of Cdk inhibitors, which are grouped into two classes (13). The members of the INK4 family (p16 INK4a , p15 INK4b , p18 INK4c , and p19 INK4d ) inhibit only Cdk4 and Cdk6 (14), and the members of Cip family (p21 cip1 , p27 kip1 , and p57 kip1 ) inhibit all Cdks (15). It is unknown how heparin affects the activities of these proteins to regulate the cell cycle.
We have shown recently that heparin induces a block in G 1 to S phase transition in VSMC (16). The interferon-induced protein kinase (PKR) activation in heparin-treated VSMC was essential in part for this cell cycle block. In this paper, we have extended our study to determine the mechanism of heparininduced cell cycle block. Our results indicate that heparin causes the p27 kip1 levels to remain elevated even in response to serum, thereby causing an inhibition of Cdk2 activity that results in deficient phosphorylation of Rb, leading to the block in S phase entry. Furthermore, the elevated levels of p27 kip1 in heparin-treated cells were found to result from a posttranscriptional mechanism, mainly by stabilization of p27 kip1 protein. This p27 kip1 stabilization was found to be defective in cells treated with 2-aminopurine, an inhibitor of PKR activity. In addition, the p27 kip1 -null cells were found to be insensitive to heparin-induced cell cycle block, thereby demonstrating p27 kip1 stabilization to be the main cause of the antiproliferative effects of heparin.

EXPERIMENTAL PROCEDURES
Cell Culture-The rat primary aortic vascular smooth muscle cells (RASMCs) were obtained from the thoracic aorta of male Sprague-Dawley rats. The cells were cultured in DMEM (Invitrogen) supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin. RASMCs were made quiescent in DMEM with 0.1% FCS for 72-96 h. For heparin treatment, the cells were shifted to 0.1% serum containing medium with 100 g/ml heparin (Sigma, catalog number H-3149, lot 17H03885) for 2 h. After 2 h, the cells were serumstimulated with DMEM containing 10% serum and 100 g/ml heparin. As a control, RASMCs were subjected to identical treatment in the absence of heparin.
Cdk2 Kinase Assays-Cell extracts were prepared from RASMCs treated with heparin as described in ice-cold lysis buffer (50 mM Hepes, pH 7.6, 150 mM NaCl, 2.5 mM EGTA, 1 mM EDTA, 10 mM ␤-glycerophosphate, 0.2 mM sodium vanadate, 1 mM sodium fluoride, 0.1% Tween 20, 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 2 g/ml aprotinin). Lysates were centrifuged in cold temperature for 10 min in a microcentrifuge at maximum speed, and the supernatant was stored at Ϫ80°C in small aliquots. Cell extracts containing 100 g of the total protein were incubated at 4°C for 2 h under constant rotation in 500 l of lysis buffer containing 250 ng of anti-Cdk2 antibodies (Santa Cruz Biotechnology) and 20 l of protein A-Sepharose beads (Roche Applied Science). Immunoprecipitates were washed twice with 500 l of kinase buffer (40 mM Tris-HCl, pH 7.6, 20 mM MgCl 2 , and 2 mM dithiothreitol). The beads were then suspended in 25 l of kinase buffer containing 2 g of histone H1 (Sigma), 10 M ATP, and 5 Ci of [␥-32 P]ATP. The reaction was incubated at 30°C for 20 min and then stopped by the addition of 10 l of 4ϫ SDS-PAGE sample loading buffer and boiling for 10 min. The proteins were separated by 12% SDS-PAGE and visualized by phosphorimaging analysis of dried gels.
Northern Blot Analysis-Total RNA from RASMC was extracted by using the RNAzol B reagent (TelTest). 30 mg of total RNA from each sample was analyzed by Northern blot analysis according to Church and Gilbert protocol with a 1.5-kb EcoRI fragment of human p27 kip1 cDNA (provided kindly by Sylvain Meloche, University of Montreal, Canada) as a probe. The signals were analyzed by phosphorimaging analysis. The same blot was then stripped and reprobed with a 28 S ribosomal RNA as a probe to ensure equal loading in all of the lanes.
Cloning the p27 kip1 Internal Ribosome Entry Site (IRES) Element in the Bicistronic Vector and Assaying Its Activity in VSMC-The pLL vector is composed of a fragment containing Renilla luciferase from pRL-null (Promega) as the first cistron fused to a fragment encoding the cytosolic firefly luciferase from pGL3-basic (Promega) as the second cistron. This vector and the pLVL construct containing the IRES element from vascular endothelial growth factor (VEGF) mRNA were kindly provided to us by Dr. Orna Elroy-Stein (Tel Aviv University). A plasmid construct containing a 502-bp long murine p27 kip1 5Ј-UTR was kindly provided by Robin Miskimins (University of South Dakota). This 502-bp long murine p27 kip1 5Ј-UTR was PCR-amplified and subcloned at the EcoRV site between the two luciferase reading frames in pLL to generate the M502/pLL construct. 1 g of each of the two plasmids (pLL and M502/pLL) was transfected into rat VSMC using the Lipofectamine (Invitrogen) transfection reagent. 24 h after transfection, the cells were harvested and cell extracts were prepared and assayed for Renilla and firefly luciferase activities using the dual-luciferase reporter assay system (Promega) and a Zylux luminometer.
Pulse-Chase Experiment-The RASMCs were grown in 6-well plates and were made quiescent in DMEM with 0.1% FCS for 72-96 h. The cells were then treated with 100 g/ml heparin in 0.1% serum containing methionine-free medium for 2 h. After 2 h, the cells were serum-stimulated with methionine-free DMEM containing 10% serum and 100 g/ml heparin. As a control, RASMCs were subjected to identical treatment in the absence of heparin. 6 h after serum stimulation, the cells were pulse-labeled for 2 h with [ 35 S]methionine and chased for the times indicated in fresh medium with methionine containing 10% serum. Cell lysates prepared at indicated times were subjected to immunoprecipitation with anti-p27 kip1 monoclonal antibody and protein A-Sepharose, and the labeled immunoprecipitated p27 kip1 protein was analyzed by SDS-PAGE followed by fluorography and phosphorimaging analysis.
Effect of Inhibition of PKR Activity by 2-Aminopurine-RASMCs were made quiescent in DMEM with 0.1% FCS for 72-96 h. The cells were then treated with 10 mM 2-aminopurine (2AP) for 2 h prior to the addition of heparin. After this treatment, 100 g/ml heparin was added to the cells in 0.1% serum containing medium for 2 h. After 2 h, the cells were serum-stimulated with DMEM containing 10% serum and 100 g/ml heparin. As a control, RASMCs were subjected to identical treatment in the absence of 2AP. Cell extracts were prepared 20 h after the start of heparin treatments, and a Western blot analysis was performed with an anti-p27 kip1 antibody. The blot was then stripped and reprobed with anti-␣-actin antibody to ascertain that an equal amount of protein from different cell extracts was analyzed.
Effect of Heparin on p27 kip1 -null Cells-The p27 Ϫ/Ϫ and isogenic p27 ϩ/ϩ mouse embryonic fibroblast (MEF) cell lines were provided by J. M. Roberts (Fred Hutchinson Cancer Research Center). The p27 ϩ/ϩ and p27 Ϫ/Ϫ MEFs were made quiescent in 0.1% serum for 36 -48 h and were stimulated with DMEM containing 10% serum. The cells were pretreated for 2 h with 100 g/ml heparin in low serum medium before the addition of serum-containing medium with 100 g/ml heparin. As a control, the same treatments were performed without any heparin. 24 h after serum stimulation, the cells were permeabilized using the Vindelov's solution (17) and stained with propidium iodide and analyzed by flow cytometry. The data are representative of three separate experiments.

RESULTS
It has been shown previously that heparin induces a block in G 1 to S transition in VSMC (16). To characterize the mechanism of this cell cycle block, we have analyzed the effect of heparin on defined cell cycle events that occur in response to serum. As a first step toward characterizing the effect of heparin on cell cycle, we performed Western blot analysis to examine the expression of cyclin D and E in the presence of heparin. Heparin showed no effect on the expression of cyclins D and E. Cyclin D was barely detectable in quiescent VSMCs and was induced within 4 h after addition of serum, both in presence and absence of heparin ( Supplementary Fig. 1A, lanes 1-8). In the presence of heparin, the phosphorylated form of Rb protein never prevailed completely, even at 24 and 30 h after serum addition, as in the absence of heparin. These results indicate that heparin prevents Rb phosphorylation in response to serum.
Induction of cyclin A synthesis in response to serum occurs as cells enter S phase, and it is abruptly degraded before metaphase during mitosis (18,19). The synthesis of cyclin A is mainly controlled at the transcriptional level, involving E2F and other transcription factors (20). Enzymatic activation of the G 1 Cdk complexes results in the phosphorylation of the Rb protein (10). The hypophosphorylated Rb binds to and inhibits several isoforms of the transcription factor E2F. Hyperphosphorylation of Rb in late G 1 releases these E2Fs, which are responsible for up-regulating gene expression required for progression through the S, G 2 , and M phases (21). Therefore, the expression of cyclin A is regulated by phosphorylation of Rb during late G 1 phase (22). We next examined whether cyclin A expression was affected by reduced Rb phosphorylation in the presence of heparin. As seen in Fig. 1A, in the absence of heparin, there was a robust induction of cyclin A at 30 h after serum addition. However, no induction of cyclin A was detected We next examined whether the expression of Cdk inhibitors p21 and p27 kip1 was regulated by heparin. As shown in Fig. 2A, p27 kip1 protein levels were rapidly down-regulated by serum and fell to undetectable levels at 24 h after the addition of serum. In contrast to this finding, p27 kip1 levels were unaffected in the presence of heparin and, even at 30 h after serum addition, p27 kip1 levels were comparable with those in the quiescent cells. These results indicate that the inhibition of Rb phosphorylation observed may be because of a block in Cdk activities due to high levels of p27 kip1 in the presence of heparin. In contrast to p27 kip1 , the p21 levels showed no changes in the presence of heparin. As seen in Fig. 2B, p21 levels were induced at 6 and 12 h after serum addition before falling later at 18 h. The presence of heparin did not show any changes in this pattern, thereby indicating that p21 is not regulated by the presence of heparin.
Higher levels of p27 kip1 in heparin-treated VSMCs would result in reduced Cdk2 activity, thereby also affecting Rb phosphorylation (12). We next examined whether heparin affected the Cdk2 activity by using the histone H1 phosphorylation assay. As seen in Fig. 3, there was a modest increase in the levels of cyclin E in serum-stimulated cells as compared with quiescent cells (panel B). The levels of Cdk2 (panel A) remained unchanged in quiescent and serum-stimulated VSMC. The total p27 kip1 showed a decrease at 18 h after serum treatment in the absence of heparin but not in its presence (panels C and F). The Cdk2-bound p27 kip1 levels dropped markedly at 18 h after serum treatment in the absence of heparin (panel D and F). However, for the heparin-treated samples, Cdk2-bound p27 kip1 levels were ϳ2-fold higher as compared with those in serumstarved VSMC. Consistent with this finding, the Cdk2 kinase activity assayed by histone H1 phosphorylation was high in serum-stimulated VSMC but remained low (9.7% Cdk2 activity in the absence of heparin) in VSMC treated with heparin and serum (panel E and F). A quantification of these results is shown in Fig. 3F. As seen in Fig. 3F, serum stimulation resulted in a 50% reduction in the p27 kip1 protein levels in 18 h (black bar) compared with the levels in quiescent VSMC (white bar). On the other hand, heparin treatment (diagonally hatched bar) resulted in a moderate rise in p27 kip1 protein levels. Serum-stimulated cells had a 50-fold less p27 kip1 protein that was associated with Cdk2 compared with the quiescent cells (compare the middle bar with the vertically hatched bar).
In the presence of heparin, the Cdk2-bound p27 kip1 levels (horizontally hatched bar) remained high and were 2.1-fold more than the levels in the quiescent cells (vertically hatched bar). These results clearly indicate that the elevated levels of p27 kip1 contribute to the inhibition of Cdk2 activity in the presence of heparin.
It is known that p27 kip1 protein levels are regulated primarily by a posttranscriptional mechanism either at the level of translation or protein stability (23)(24)(25)(26)(27)(28)(29)(30). To examine whether the p27 kip1 levels were regulated at a posttranscriptional level in VSMC, we performed Northern blot analysis of p27 kip1 expression. As seen in Fig. 4A, p27 kip1 mRNA levels were the highest in serum-starved cells and dropped somewhat in serum-stimulated cells, both in presence and absence of heparin to the same extent. These results indicate that heparin treatment seems to regulate p27 kip1 expression at the posttranscriptional level in VSMC. The regulation of p27 kip1 protein levels occurs mainly by regulation of its translation and protein stability (12). The 5Ј-UTR region of p27 kip1 mRNA has been shown to contain an IRES element (25) and is shown to regulate its translation (31). Some IRES elements from VEGF, plateletderived growth factor, and c-Myc mRNAs function to enhance their translation under conditions when eIF2␣ is phosphorylated in cells (32). We have shown previously that heparin treatment of VSMC resulted in PKR activation and eIF2␣ phosphorylation (16). Thus, it is possible that heparin enhances p27 kip1 protein synthesis via the IRES element because it also causes eIF2␣ phosphorylation. To examine whether heparin was enhancing p27 kip1 synthesis via the IRES element in its 5Ј-UTR, we examined the expression of luciferase reporter mRNA, which contained the murine IRES element of p27 kip1 transcript. For this reason, we made use of the vector pLL, which encodes a bicistronic mRNA where Renilla luciferase is the first cistron and firefly luciferase is the second cistron (32). The 502-bp long murine p27 kip1 5Ј-UTR containing the IRES element (33) was cloned between the two reading frames (Fig. 4B). Firefly luciferase is translated at high levels only if FIG. 1. Induction of cyclin A in response to serum is inhibited in the presence of heparin. Quiescent VSMCs (Q) were stimulated with serum either in the absence (A) or in the presence of 100 g/ml heparin (B) for indicated times over a 12-36-h period. Equal amounts of proteins from the cell extracts were resolved by SDS-PAGE on a 12% gel and transferred to polyvinylidene difluoride membranes. The membranes were first probed with antibody specific for cyclin A, and signals were visualized by chemiluminescence detection. The membranes were then stripped and reprobed with antibody against smooth muscle ␣-actin to ensure equal loading in all of the lanes. FIG. 2. Heparin inhibits p27 kip1 down-regulation in response to serum but has no effect on p21 cip1 levels. Quiescent VSMCs (Q) were stimulated with 10% serum containing medium either in the absence or in the presence of 100 g/ml heparin (hep) for indicated times over a 6 -30-h period. Equal amounts of proteins from the cell extracts prepared at the indicated time points were resolved by SDS-PAGE on a 12% gel and transferred to polyvinylidene difluoride membranes. The membranes were first probed with antibody specific for p27 kip1 (A) or p21 cip1 (B), and signals were visualized by chemiluminescence detection. The membranes were then stripped and reprobed with antibody against smooth muscle ␣-actin to ensure equal loading in all of the lanes. the IRES element under study is active. Thus, the ratio of firefly luciferase to Renilla luciferase activity is high when the IRES element is active. As seen in Fig. 4C, the p27 kip1 IRES element was highly active in VSMC because the ratio of firefly/ Renilla luciferase activity is 15-fold higher in cells transfected with M502/pLL than in the cells transfected with empty vector pLL. As a positive control, a previously described (32) construct containing IRES element from VEGF mRNA was used (pLVL), which showed 3-fold higher activity than pLL as expected (data not shown). Thus, the p27 kip1 IRES was more active than VEGF IRES in VSMC under these conditions. However, heparin treatment caused only a modest increase in IRES activity as seen in Fig. 4C. These results indicate that, although the IRES element functions to enhance the translation of p27 kip1 mRNA in VSMC, it is not responsive to heparin treatment. Thus, the increase in p27 kip1 synthesis is unlikely to be a mechanism by which heparin enhances the cellular p27 kip1 protein levels.
To determine whether heparin affected the p27 kip1 protein stability, we performed the pulse-chase experiments. As seen in Fig. 5A (0 h lanes), a 2-h labeling period resulted in the labeling of p27 kip1 to equal levels, both in the absence and presence of heparin. These results are in agreement with the results shown in Fig. 4 and further confirm that the rate of p27 kip1 synthesis is same in the absence or presence of heparin. As seen during the chase, labeled p27 kip1 protein is degraded to undetectable levels within 12 h in the absence of heparin. However, in the presence of heparin, the stability of p27 kip1 protein is significantly enhanced, thereby indicating that heparin causes stabilization of p27 kip1 protein during the G 1 phase. The radioactivity present in the p27 kip1 bands was quantified on a PhosphorImager and was plotted as a percentage of initial amount at 0 h (Fig. 5B). Thus, heparin causes a block in G 1 to S phase transition in VSMC by stabilizing the p27 kip1 protein and a consequent inhibition of Cdk2 activity.
To determine the role of PKR in enhancing p27 kip1 protein stability in the presence of heparin, we examined the effect of inhibiting PKR activity in VSMC by its known inhibitor 2-aminopurine (2AP) (34 -36). As shown in Fig. 6A, VSMCs treated with 2AP and heparin show lower amount of p27 kip1 protein than the VSMC treated only with heparin. These results indicate that PKR activation by heparin is essential at least in part for the observed increase in p27 kip1 stability. The levels of p27 kip1 protein in cells treated with heparin and 2AP are intermediate between those in control and heparin-treated VSMC. Thus, the p27 kip1 stabilization in response to heparin seems to occur both by PKR-dependent and independent pathways. Similar results were obtained with PKR-null fibroblasts where PKR-null MEFs did not show a significant increase of p27 kip1 protein levels in response to heparin (data not shown). We have previously shown (16) that these PKR-null fibroblasts do not show a complete block of G 1 to S phase transition in response to heparin. The results presented here show that the lack of a cell cycle block is due to a defect in p27 kip1 stabilization in response to heparin in the absence of PKR activity.
If heparin-induced block in the cell cycle operates primarily because of a rise in p27 kip1 protein levels, it is expected that cells lacking p27 kip1 would be insensitive to heparin. To establish a direct causal relationship between the higher p27 kip1 levels in heparin-treated cells and the cell cycle block, we examined the effect of heparin on p27 kip1 -null cells (37). We have previously shown that MEFs respond well to the antiproliferative actions of heparin (16). We compared the effect of heparin on MEFs isolated from wild type and p27 kip1 -null mice. Heparin treatment caused a block in G 1 to S phase transition in wild type MEFs. The p27 kip1 -null MEFS were resistant to heparin-induced block of G 1 to S phase transition. As represented in Fig. 6B, 39.4% wild type MEFs were in S phase at 24 h after serum stimulation. This percentage dropped to 19.6% in heparin-treated wild type MEFs, thereby confirming that heparin effectively blocked the S phase entry of MEFs. In addition, there were 15.2% cells in G 2 /M phase in the absence of heparin and this percentage dropped to 6.3% in the presence of heparin. The p27 kip1 -null MEFs seemed to progress through the cell cycle much faster than their wild type counterparts. At 24 h after serum stimulation, p27 kip1 -null cells had progressed through the S phase and thus most of the cells that entered cell cycle were in the G 2 /M phase at this time point. 22.5% p27 kip1 -null cells were in G 2 /M phase at 24 h after serum stimulation in the absence of heparin as compared with only 15.2% in wild type MEFs. Nevertheless, in the case of p27 kip1 -null cells, clearly no effect of heparin treatment was observed on the cell cycle progression. The percentages of p27 kip1 -null cells in S and G 2 /M

FIG. 3. The inactive Cdk2 complexes in heparin (hep)-treated
VSMCs contain the p27 kip1 inhibitory protein. VSMCs were made quiescent by maintaining the confluent cultures in serum-free DMEM for 4 days. They were serum-stimulated by subculturing in 10% serum containing medium at 30% confluency. Heparin treatment was as described before in the previous figure legends. Western blot analysis was performed to determine the total amount of Cdk2 (panel A), cyclin E (panel B), and p27 kip1 (panel C). To determine the association between Cdk2 and p27 kip1 , 500 g of protein extracts were first immunoprecipitated with anti-Cdk antibodies and then the immunoprecipitates were examined by Western blot analysis using anti-p27 kip1 antibodies (panel D). Histone H1 kinase assays were performed on immunoprecipitates obtained with anti-Cdk2 antibody from 100 g of total protein extracts (panel E). The numbers in percentages shown below panel E represent quantification of data in panel E. The data in panels C and D were quantified by PhosphorImager analysis, and the signal in quiescent cells was given an arbitrary value of 1 (white and vertically hatched bars). The values of total and Cdk2-bound p27 kip1 protein at 18 h after serum addition in the absence (black bars) and presence of heparin (diagonally and vertically hatched bars) are expressed relative to the signal in quiescent cells.
phases after serum stimulation either in the absence or presence of heparin were very similar. In the absence of heparin, 22.5% cells were in G 2 /M phase (8.5% in S phase) and in the presence of heparin and 19.5% cells were in G 2 /M phase (6.6% in S phase) at 24 h after serum stimulation. Thus, the cell cycle inhibitory effect of heparin was clearly absent in p27 kip1 -null MEFs.
To examine the effect of heparin on cell cycle in p27 kip1 -null VSMC, experiments were also carried out to induce p27 kip1 down-regulation by small interfering RNA transfections with three different small interfering RNAs. Although this reduced the p27 kip1 mRNA levels to barely detectable levels, it did not completely knock out p27 kip1 protein levels in quiescent VSMC (data not shown). In addition, the transfected small interfering RNA did not persist for the entire duration of the experiment due to the fact that VSMCs need to be maintained in serumstarved conditions for a minimum of 72 h to induce a G 0 /G 1 arrest. We have shown previously that heparin-induced cell cycle block is also present in MEFs and that PKR-null MEFS are insensitive to this block in G 1 to S phase transition (16). The results shown in Fig. 6B with p27 kip1 -null MEFs further show that this heparin-induced block in G 1 to S phase transition is dependent on p27 kip1 .
Thus, we have shown that the cell cycle block imposed by heparin in VSMC occurs by up-regulation of p27 kip1 protein levels. Heparin causes stabilization of p27 kip1 protein without enhancing its transcription or translation. In addition, the p27 kip1 protein stabilization occurs by both PKR-dependent and independent pathways. The p27-null cells are markedly resistant to G 1 to S phase block imposed by heparin, indicating that p27 kip1 is essential for heparin-induced cell cycle block. DISCUSSION We have previously shown that heparin inhibits the VSMC proliferation by blocking the G 1 to S transition. The interferoninduced protein kinase PKR was shown to be involved at least partly in mediating this cell cycle block in response to heparin (16). In PKR-null MEFs and in VSMCs expressing the transdominant negative K296R mutant of PKR, there was a marked but not a complete loss of the antiproliferative effects of heparin, indicating that pathways in addition to PKR activation FIG. 4. A, p27 kip1 mRNA levels are not affected by heparin (hep). Quiescent VSMCs (Q) were stimulated with serum either in absence or presence of 100 g/ml heparin, and total RNA was isolated from quiescent cells and 12-and 18-h posttreatments. The RNA was analyzed by Northern blot analysis using a p27 kip1 cDNA probe. The same blot was stripped and rehybridized to cDNA probe for 28 S rRNA to normalize the loading. The time points are as indicated at the top of the panels. B, a schematic representation of the bicistronic constructs. The 502-bp long p27 kip1 5Ј-UTR was inserted between the Renilla and firefly luciferase coding regions. CMV, cytomegalovirus. C, p27 kip1 5Ј-UTR functions as IRES element to enhance translation but is unresponsive to heparin. Rat VSMCs grown in 6-well plates were transfected in triplicates with 1 g of pLL or M502/pLL using Lipofectamine reagent. 24 h after transfection, the cells were treated with 100 g/ml heparin. 18 -24 h after heparin treatment, the cells were harvested and the luciferase activities in the extract were determined. The data shown are representative of three separate experiments, each one performed in triplicates. The error bars shown represent the mean Ϯ S.D. C, represents untreated cells.

FIG. 5. Heparin treatment of VSMC increases the half-life of p27 kip1 protein.
A, the rat VSMC in G 0 /G 1 phase (6 h after serum stimulation) were pulse-labeled for 2 h with [ 35 S]methionine and chased in fresh medium containing 10% serum for the indicated times. Cell lysates were subjected to immunoprecipitation with anti-p27 kip1 monoclonal antibody, and the labeled immunoprecipitated p27 kip1 protein was analyzed by SDS-PAGE followed by fluorography and PhosphorImager analysis. B, quantification of the data shown in A was plotted as percentage of original amount of radioactivity present in p27 kip1 band at 0 h. may also contribute to growth inhibition. Other documented effects of heparin on VSMCs include inhibition of immediateearly gene expression (38,39), matrix-degrading proteases (40 -42), mitogen-activated protein kinase activation (43), matrix molecules (44,45), and extracellular signal-regulated kinases, ERK1 and ERK2 (46). Thrombin-induced VSMC migration via inhibition of epidermal growth factor receptor transactivation has also been documented (47). The tyrosine kinase receptor EphB2 mRNA is also down-regulated by heparin (48). Cellular proliferation is controlled by multiple Cdks that are regulated by different cyclin proteins during distinct cell cycle phases (49). Cdk activity is also negatively regulated by the interaction with specific Cdk inhibitors, which cause a cell cycle block when overexpressed artificially in several cell lines (50 -52). Two classes of Cdk inhibitors have been defined: the INK proteins (p16 INK4a , p15 INK4b , p18 INK4c , and p19 INK4d ) and the Cip/Kip proteins (p21 cip1 , p27 kip1 , and p57 kip2 ) (53). The INK proteins interact with the Cdk4 and Cdk6 and thus inhibit them. The Cip/Kip proteins are thought to inhibit all Cdks, and p27 kip1 is thought to be the primary modulator of the proliferative status in most cell types where it is thought to cause and maintain the quiescent state (54). p27 kip1 accumulates in serum-starved or density-arrested cultures, and the addition of mitogens to cultures rapidly down-regulates the levels of p27 kip1 (13). Conditions that block p27 kip1 down-regulation prevent quiescent cells from entering the cell cycle in response to mitogens (55). These conditions include treatment of cells with rapamycin (56 -58), hypoxia-inducing agents (59), and cAMP analogs (60). Changes in p27 kip1 protein levels in normal cells usually do not correlate with changes in the transcription of the p27 kip1 gene (for review, see Ref. 12). The p27 kip1 protein levels in normal cells are most commonly regulated by changes in the rate of proteasome-mediated p27 kip1 degradation or the rate of translation of p27 kip1 mRNA (27,61,62). Consistent with the cell cycle regulatory functions of p27 kip1 , the mice lacking p27 kip1 exhibit large size because of enhanced growth and multiple organ hyperplasia (63). p27 kip1 also functions as a tumor suppressor, and reduced expression of p27 kip1 has been correlated with poor cancer patient survival (64,65).
Accumulating evidence implicates p27 kip1 as an important regulator of the phenotypic response of VSMCs to mitogenic or hypertrophic stimuli. At late time points after the balloon angioplasty, p27 kip1 up-regulation has been shown to limit the growth of VSMCs (52,66). In accordance with this finding, overexpression of p27 kip1 can efficiently block mitogen-induced induction of cyclin A promoter activity in cultured VSMCs in vitro (52). Adenovirus-mediated overexpression of p27 kip1 also effectively blocks vascular occlusive lesion formation in balloon-injured arteries in rats (67). In addition, p27 kip1 has also been implicated in determining hypertrophic versus herplastic growth (26,68). The protective role of p27 kip1 against neointimal thickening has been also demonstrated in hypercholesterolemic apolipoprotein-deficient mice in which genetic inactivation of one or both p27 kip1 alleles shows progressively accelerated atherosclerosis (56,69,70). The migratory response exhibited by VSMC in response to mitogens has been shown to be inhibited by overexpression of p27 kip1 (58,67,71,72). Inhibition of VSMC proliferation by salicylate has also been shown to be mediated by up-regulation of p27 kip1 and p21 (73). Contrary to this finding, neointimal hyperplasia after mechanical damage of the artery wall was comparable in both wild-type and p27 kip1 -null mice (56,74). In addition, although rapamycin- FIG. 6. A, PKR kinase activity is essential for p27 kip1 stabilization in response to heparin (hep). Rat VSMCs were made quiescent in DMEM with 0.1% FCS for 72-96 h. The cells were then treated with 10 mM 2AP for 2 h prior to the addition of heparin. After this treatment, 100 g/ml heparin was added to the cells in 0.1% serum containing medium for 2 h. After 2 h, the cells were serum-stimulated with DMEM containing 10% serum and 100 g/ml heparin. As a control, VSMCs were subjected to an identical treatment in the absence of 2AP. C lanes represent serumstimulated cells without any treatment. Cell extracts were prepared 20 h after the start of heparin treatments, and a Western blot analysis was performed with an anti-p27 kip1 antibody. The blot was then stripped and reprobed with anti-␣-actin antibody to ascertain that equal amounts of protein from different cell extracts were analyzed. B, the p27 kip1 -null cells are unresponsive to the antiproliferative effects of heparin. Quiescent (Q) MEFs (0.1% serum for 24 h) were stimulated with 10% serum containing DMEM. A majority of the serum-starved wild-type and p27 kip1null MEFs were in G 0 /G 1 phase as judged by the cell cycle profile. These were pretreated with 100 g/ml heparin in low serum medium for 2 h prior to the addition of 10% serum medium also containing 100 g/ml heparin. As a control, the same treatments were done without any heparin. At 24 h after the serum stimulation, DNA was stained with propidium iodide and the cells were analyzed by flow cytometry. Quantitative analysis of the cell cycle profiles is shown. The data shown are representative of three separate experiments. The error bars represent the mean Ϯ S.D. dependent growth arrest in cultured cells has been shown to result from stabilization of p27 kip1 , rapamycin does not prevent the down-regulation of p27 kip1 seen after balloon injury (57). The neointimal formation after balloon injury was also found to be similar in wild-type and p27 kip1 -null mice in response to rapamycin where rapamycin inhibited VSMC proliferation in both wild-type and p27 kip1 -null mice to similar levels (74). These results indicate that the antiproliferative actions of rapamycin in VSMC are not mediated by p27 kip1 but via other mechanisms.
Effects of heparin on cell cycle in VSMC have been analyzed previously by Castellot et al. (75). It was reported that heparin had a dual effect on G 0 to S transition and that it delayed entry into S phase and also reduced the number of cells entering the cycle from G 0 . Through simple and elegant studies, this work further demonstrated that there were two points during G 1 phase when the cells were sensitive to heparin, one was during early G 1 and the other was during mid-late G 1 . Our data agree well with these early observations, because our results demonstrate that p27 kip1 stability is increased by heparin during mid-late G 1 phases. The early G 1 block observed in this work may result from effects of heparin on cell cycle regulatory molecules other than p27 kip1 . For example, the induction of c-Fos and c-Myc in murine fibroblasts (76) and c-Fos induction in mesangial cells (77,78) is blocked by heparin. In addition, heparin has been shown to inhibit cell cycle by inhibition of multiple pathways such as mitogen activated protein kinase (MAPK), calmodulin-dependent protein kinase II, and protein kinase C (PKC). Although seemingly separate, all of these pathways have been shown to affect cellular p27 kip1 levels in recent years. Ottlinger et al. (79) documented that MAPK activation in response to FCS and phorbol 12-myristate 13-acetate but not in response to epidermal growth factor was inhibited by heparin in VSMC. Although not tested in this work, the MAPK pathway has been shown to have effect on cellular p27 kip1 levels in VSMC. Pharmacological inhibition of MAPKs increased p27 kip1 expression and attenuated VSMC proliferation and migration (72). MAPK cascade has also shown to be required for p27 kip1 down-regulation and S phase entry in fibroblasts and epithelial cells (80). Inhibitory effects of heparin on activation of calmodulin-dependent protein kinase II have been documented by Mishra-Gorur et al. (81,82) in VSMC. Recently, it has been shown that inhibition of calmodulindependent protein kinase II in NIH 3T3 cells enhances association of p27 kip1 with Cdk2 and thus causes a G 1 arrest (83). Our results are in agreement with these findings, because heparintreated VSMC did exhibit a higher amount of p27 kip1 associated with Cdk2 compared with the quiescent VSMC. Herbert et al. (84) showed that heparin blocks VSMC proliferation by interfering with the PKC pathway through a selective direct inhibition of the PKC-␣ isoenzyme. Pukac et al. (85) also showed that heparin inhibits activation of PKC by plateletderived growth factor and serum but not by epidermal growth factor. Any connection between the inhibition of PKC-dependent pathways to the antiproliferative effects of heparin remains intriguing. This is especially true because, in various cell systems, the major effect of PKC activation appears to be an inhibition of cell cycle due to a reduced activity of Cdk2 as a result of increased levels of p27 kip1 (86).
The levels of p27 kip1 protein are regulated by changes in the protein stability during the cell cycle (12). Two different mechanisms have been shown to be responsible for this regulation, one operating during the G 0 /G 1 phase and the other operating during the S phase (87). p27 kip1 protein is much more stable during G 0 /G 1 phase than during S phase. In cells exposed to mitogens, p27 kip1 levels fall before Cdk2 becomes active and the decreases in p27 kip1 levels are required for Cdk2 activation. Thus, Cdk2-independent processes mediate the down-regulation of p27 kip1 levels during G 1 phase (12). The phosphorylation of p27 kip1 at threonine 187 by Cdk2/cyclinE has been shown to destabilize it and cause a rapid decline in its levels via the Skp2-dependent ubiquitin-proteasome pathway in the nucleus during the G 1 -S transition (88 -91). The degradation of p27 kip1 during G 0 /G 1 transition takes place by a distinct pathway in the cytoplasm that is independent of both the threonine 187 phosphorylation and Skp2-mediated proteasome pathway (92). During the G 0 phase, p27 kip1 is phosphorylated by Mirk/ Dyrk1B at serine 10 and this has been shown to increase its stability in the nucleus (93). Mirk/Dyrk1B levels rapidly fall after serum stimulation of G 0 -arrested cells. Another kinase, kinase-interacting stathmin, is induced in response to serum and maintains the serine 10 phosphorylation of p27 kip1 during the G 1 phase (94). Nuclear export of p27 kip1 has been shown to occur during G 1 phase, and phosphorylation at serine 10 has been shown to be required for binding to CRM1 via Jab1/CSN5 and subsequent nuclear export (95). However, because serine 10 is already phosphorylated during G 0 phase, an additional signal must be required for causing the nuclear export of p27 kip1 during G 1 phase (96,97). Evidence has also been presented that, although required for its cytoplasmic localization, serine 10 phosphorylation of p27 kip1 is not a prerequisite for its proteolysis (98). Thus, although the threonine 187 phosphorylation-dependent proteasome pathway has been well established, the exact mechanism(s) regulating the p27 kip1 destabilization during G 1 phase has not been elucidated as yet.
An earlier report showed that heparin inhibits proliferation of myometrial and leiomyomal smooth muscle cells through the induction of ␣-smooth muscle actin, calponin h1, and p27 kip1 (99). However, no mechanistic experiments were presented in this study. Our results presented here indicate that p27 kip1 protein is stabilized in VSMC during G 1 phase by heparin treatment of VSMC via both PKR-dependent and independent pathways. In addition, the p27 kip1 -null cells show no effect of heparin treatment on cell cycle, thereby indicating p27 kip1 stabilization to be a major mechanism for heparin-induced cell cycle arrest.