- Email: [email protected]
Cell Cycle Regulation of Smooth Muscle Cells—Searching for Inhibitors of Neointima Formation: Is Combretastatin A4 an Alternative to Sirolimus and Paclitaxel?
LABORATORY INVESTIGATION
Cell Cycle Regulation of Smooth Muscle Cells— Searching for Inhibitors of Neointima Formation: Is Combretastatin A4 an Alternative to Sirolimus and Paclitaxel? Daniel Spira, MD, Gerd Grözinger, MD, Nicole Domschke, MD, Rüdiger Bantleon, PhD, Jörg Schmehl, MD, Jakub Wiskirchen, MD, and Benjamin Wiesinger, MD
ABSTRACT Purpose: To compare the effects of sirolimus, paclitaxel, and combretastatin A4 (CA4) on regulatory proteins of the cell cycle in proliferating smooth muscle cells (SMCs). Materials and Methods: Human aortic SMCs were treated with sirolimus, paclitaxel, and CA4 at 5 109 mol/L. After 1 day, half of the cells were harvested (DAY1 group). The treatment medium of the other half was replaced with culture medium on day 4, and those cells were harvested on day 5 (DAY5 group). Cyclins D1, D2, E, and A and cyclin-dependent kinase (CDK) inhibitors p16, p21, and p27 were detected by Western blot technique. Quantification was performed by scanning densitometry of the specific bands. Results: In the DAY1 group, treatment with sirolimus resulted in decreased intracellular levels of cyclins D2 and A (P o .05). Increased D cyclins and reduced levels of cyclins E and A (P o .05) in the DAY5 group indicated a permanent G1/S block by sirolimus. Paclitaxel led to only slight alterations of cyclin and CDK inhibitor expression (P 4 .05). In the DAY1 group, CA4 decreased intracellular levels of cyclins D2, E, and A (P o .05). Despite recovery effects in the DAY5 group (increase of cyclins D1, D2, and A compared with DAY1 group; P o .05), the upregulation of the CDK inhibitor p21, increased D cyclins, and decreased cyclins E and A (P o .05) are compatible with a G1 arrest. Conclusions: CA4 is a stronger inhibitor of the SMC cycle than sirolimus or paclitaxel and may represent an alternative for drug-eluting stents in atherosclerotic luminal stenosis. The effect of CA4 on neointima formation should be evaluated further.
ABBREVIATIONS CA4 = combretastatin A4, CDK = cyclin-dependent kinase, CI = confidence interval, mTOR = mammalian target of rapamycin, PPARγ = peroxisome proliferator activated receptor γ, SMC = smooth muscle cell
From the Department of Diagnostic and Interventional Radiology (D.S.), University Hospital Heidelberg, Im Neuenheimer Feld 110, Heidelberg 69120, Germany; Department of Diagnostic and Interventional Radiology (G.G., R.B., J.S., B.W.), Eberhard-Karls-University, Tübingen, Germany; Department of Internal Medicine (N.D.), SRH Hospital Sigmaringen, Sigmaringen, Germany; and Department of Radiology and Nuclear Medicine (J.W.), Franziskus Hospital, Bielefeld, Germany. Received February 16, 2015; final revision received May 4, 2015; accepted May 10, 2015. Address correspondence to D.S.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2015 J Vasc Interv Radiol 2015; XX:]]]–]]] http://dx.doi.org/10.1016/j.jvir.2015.05.025
In atherosclerosis, restenosis after percutaneous transluminal angioplasty or in-stent restenosis caused by neointima formation is a major issue in clinical situations. Interventional trauma to the vessel wall results in exposure of the subintimal matrix, triggering neointima formation. This process takes 1–3 months and entails proliferation, migration, and apoptosis of medial smooth muscle cells (SMCs) (1). In a 24- to 72-hour time period, 20%–40% of the SMCs enter the cell cycle and begin to proliferate (2). This situation has led to the development of drug-eluting balloons and stents enabling the local application of agents to suppress the inflammatory response and consecutive SMC proliferation. Currently
2
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
available drug-eluting stents convey high levels of the cell cycle inhibitors sirolimus, sirolimus analogues, and paclitaxel to the vessel wall (3). Vascular SMCs normally stay in G0/G1 phase of the cell cycle. After vascular injury, SMCs are stimulated to pass a restriction point and enter S phase, where the DNA is replicated. This entry into S phase is regulated by complexes of cyclins and cyclin-dependent kinases (CDKs) as well as CDK inhibitors. D cyclins bind and activate CDK 4 and CDK 6, which are necessary for progression through G1 phase (4). Cyclin E is selectively expressed in proliferating cells at the transition of G1 to S phase building a complex with its catalytic partner CDK 2 (5). Vascular SMC proliferation is inhibited by the CDK inhibitors p27 and p21 as a result of suppression of CDK 2 activity (6). Sirolimus inhibits mammalian target of rapamycin (mTOR), which prevents the decomposition of the CDK inhibitor p27. The CDK inhibitor p27 suppresses cyclin E/CDK 2 kinase activity leading to G1- to S-phase transition arrest (3). Sirolimus also has inhibitory effects on vascular SMC migration (7). Paclitaxel attaches to β-tubulin and leads to cell cycle arrest during the M (mitosis) phase through tubulin polymerization and suppression of spindle formation and intracellular transport. It inhibits SMC proliferation and migration independently of intracellular signaling mechanisms such as those of sirolimus and its mTOR pathway (8). Despite advances in the drug-eluting technology, some drawbacks remain. First, although sirolimus is efficacious in reducing neointima formation, this effect may be hindered in patients with diabetes who are treated with oral agonists of peroxisome proliferator activated receptor γ (PPARγ), such as rosiglitazone or pioglitazone, as blockade of mTOR by sirolimus was shown to attenuate PPARγ activity (8). Paclitaxel also reduces neointima formation, but treated arteries showed incomplete healing and long-lasting fibrin deposition accompanied by a large number of macrophages (9). Additionally, in the dose range used for clinical applications, paclitaxel induces apoptosis, which hints at a narrow toxic-therapeutic window (10). Furthermore, paclitaxel accumulates in the adventitia of vessels, which assumes an inferior role in the pathophysiology of neointima formation and restenosis (11). Finally, paclitaxel-eluting balloons failed to show their superiority compared with conventional pressure-only balloon angioplasty in the lower limb (12). These drawbacks reopen the question about the best antiproliferative agent for neointima inhibition. Combretastatin A4 (CA4) binds to the colchicine binding site of β-tubulin and inhibits its polymerization, preventing the degradation of the spindle apparatus and hindering mitosis (13). Because CA4 proved to be an effective cytostatic agent, it should inhibit SMC proliferation independently of the mTOR pathway of sirolimus and may suppress neointima formation. The
Spira et al
’
JVIR
purpose of this study was to compare the effects of sirolimus, paclitaxel, and CA4 on regulatory proteins of the cell cycle (cyclin A, cyclin D1, cyclin D2, cyclin E, p16, p21, and p27) in proliferating vascular SMCs.
MATERIALS AND METHODS Cell Culture Human aortic SMCs (ATCC, Rockville, Maryland) were cultured in Weymouth/Ham’s F12 medium (GIBCO, Karlsruhe, Germany) in a 1:1 ratio. The medium was supplemented with 20% fetal calf serum (Boehringer Ingelheim, Mannheim, Germany), 1% glutamine, and 1% penicillin/streptomycin. Cell cultures were incubated in 25 cm2 tissue culture flasks (Falcon; Becton, Dickinson and Company, Franklin Lakes, New Jersey) at 371C in a 5% carbon dioxide atmosphere. Cell density was 2 105 cells per 5 mL culture medium. The culture medium was changed on day 1 and day 4. Cells were harvested using trypsinization every 7 days. Absolute cell numbers were determined with an electronic cell counter (CASY; Schärfe System GmbH, Reutlingen, Germany).
Treatment of Human Aortic SMCs Three cytostatic drugs were compared with untreated controls: sirolimus, paclitaxel, and CA4 (Sigma-Aldrich, Deisenhofen, Germany). Cells were seeded at a density of 400,000 cells in 75 cm2 cell culture flasks with 10 mL medium containing 20% fetal calf serum. After 2 days, the medium was changed, and the medication was added in a concentration of 5 109 mol/L in the treated groups, whereas untreated controls received normal culture medium. Vincent et al (14) described antiproliferative effects of CA4 on endothelial cells at nontoxic doses of 5 nmol/L and 10 nmol/L. Also, Hayashi et al (15) showed inhibitory effects of sirolimus and paclitaxel on the cell cycle in the low nanomolar range. To ensure comparability, we chose 5 109 mol/L as the concentration in all experiments. After 1 day, half of the cells (three out of six 75 cm2 cell culture flasks) were washed with cold phosphate-buffered saline and were harvested mechanically with the help of a cell scraper (DAY1 group). The treatment medium of the other half of the cells was replaced with culture medium on day 4, and those cells were harvested on day 5 (DAY5 group). The rationale for the groups DAY1 (24-h treatment) and DAY5 (4-d treatment and 24-h withdrawal) was to search for effects on the cell cycle early during treatment and early after treatment cessation.
Test Procedures Procedures were performed on ice to inhibit the activity of proteases. Cells were lysed with lysis buffer (0.05 mol/L tris-(hydroxymethyl)-aminomethane, pH 8, 5 mL; 0.12 mol/ L sodium chloride, 2.4 mL; 0.5% Triton X-100, 500 mL; 0.005 mol/L ethylenediamine tetraacetic acid, 1 mL; distilled
Volume XX
’
Number X
’
Month
’
2015
3
water, 91.1 mL) after protease inhibitors aprotinin/leupeptin (1 mL lysis buffer þ 7.5 mL Pefabloc [Roche Diagnostics, Mannheim, Germany; 200 mmol/L] þ 5 mL aprotinin/ leupeptin [aprotinin 0.4 mg/mL, leupeptin 1 mg/mL]) were added and pelleted after centrifugation. To maximize protein extraction, additional sonication was performed using ultrasound (Sonifier B-12; Branson Sonic Power Co., Danbury, Connecticut). The lysed cell solution was centrifuged again, and the supernatant, containing soluble cytoplasmic proteins, was saved. Total protein content was determined by Bradford protein assay, and the solution was stored at 701C until Western blotting. For Western blotting, proteins were denatured by mercaptoethanol 50 mmol/L and heating at 951C for 4 minutes. Then 25 mg of the protein solution was added to each lane of a 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel. After electrophoresis, the protein was transferred onto a Hybond-N nylon membrane (Amersham, Buckinghamshire, United Kingdom) by semidry blotting with help of an electroblotting chamber system (Semi-Phor; Hoefer Pharmacia Biotech, San Francisco, California). For detection of cyclin D1, cyclin D2, cyclin E, cyclin A, p16, p21, and p27, the membrane was incubated with a specific monoclonal antibody in a concentration of 1:2,000 (New England Biolabs, Beverly, Massachusetts). After washing the membrane, a second antibody in a concentration of 1:5,000 labeled with horseradish peroxidase (Amersham) was added; after incubation, LumiGLO (New England Biolabs) was added. Blocking of the polyvinylidene difluoride membrane was performed with milk powder solution. Membrane detection was conducted with a Lumi-Imager Diana III (Raytest GmbH, Straubenhardt, Germany). β-Actin was used to control protein load in the Western blot. Quantification was done by scanning densitometry of the specific bands of each lane and was measured with help of the AIDA BioPackage (Raytest GmbH). The optical density of controls was equated to 100, and the optical density of the other groups was determined accordingly by computer analysis. The experiments were repeated 12 times (n ¼ 12). Four test preparations (sirolimus, paclitaxel, CA4, untreated control) and three Western blots performed for each preparation resulted in 12 measured values for each of the seven regulator proteins (cyclins D1, D2, E, and A and CDK inhibitors p16, p21, and p27) on day 1 and another 12 values on day 5, respectively.
Statistics All data are reported as arithmetic mean and [95% confidence interval (CI)]. Statistical analysis was performed using a multivariate analysis via Tukey-Kramer test in conjunction with an analysis of variance for all pairwise comparisons and Student t test for unpaired data (comparison day 1 vs day 5). P values o .05 were considered statistically significant.
RESULTS Day 1: Effects of Sirolimus, Paclitaxel, and CA4 on Cyclin Expression Differences in cyclin D1 expression on day 1 were not statistically significant after addition of sirolimus (mean, 60.37 [95% CI, 25.07–95.67]), paclitaxel (mean, 113.74 [95% CI, 78.44–149.04]) or CA4 (mean, 38.17 [95% CI, 2.87–73.46]) compared with controls (mean, 100.00 [95% CI, 64.70–135.30]) (all P 4 .05) (Table 1). Mean cyclin D1 levels in cells treated with paclitaxel were slightly elevated, whereas sirolimus and CA4 resulted in decreased expression (Fig 1a). In a direct comparison with paclitaxel, suppression of cyclin D1 by CA4 reached statistical significance (P o .05). Cyclin D2 expression was significantly decreased by sirolimus (mean, 39.53 [95% CI, 13.09–65.96]) and CA4 (mean, 33.19 [95% CI, 6.76–59.63]) compared with either controls (mean, 100.00 [95% CI, 73.57–126.43]) or paclitaxel (mean, 90.898 [95% CI, 64.464–117.33]) (all P o .05) (Fig 1b). Cyclin E levels were significantly reduced after addition of CA4 (mean, 28.22 [95% CI, 3.80 to 60.25]) compared with controls (mean, 100.00 [95% CI, 67.97– 132.03]) (P o .05). Treatment with sirolimus (mean, 74.19 [95% CI, 42.17–106.22]) and paclitaxel (mean, 92.84 [95% CI, 60.82–124.87]) resulted in only a minor, nonsignificant decrease of cyclin E (P 4 .05) (Fig 1c). Cyclin A expression was significantly decreased by sirolimus (mean, 60.59 [95% CI, 39.60–81.58]) and even more so by CA4 (mean, 22.85 [95% CI, 1.86– 43.84]) compared with controls (mean, 100.00 [95% CI, 79.01–120.99]) (P o .05). The reduction of cyclin A levels after treatment with paclitaxel (mean, 72.94 [95% CI, 51.95–93.92]) was not statistically significant (P 4 .05) (Fig 1d).
Table 1 . Synopsis of Cell Cycle Changes on Day 1 Cyclin D1
Cyclin D2
Cyclin E
Cyclin A
p16
p21
Sirolimus
—
↓
—
↓
—
—
p27 —
Paclitaxel Combretastatin A4
— (↓)
— ↓
— ↓
— ↓
— —
(↑) —
— —
↑ ¼ statistically significant increase versus untreated controls; ↓ ¼ statistically significant decrease versus untreated controls; (↑) ¼ not statistically significant but 4 50% increase of mean value versus control; (↓) ¼ not statistically significant but 4 50% decrease of mean value versus control.
4
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
Spira et al
’
JVIR
Figure 1. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of cyclin D1 (a), cyclin D2 (b), cyclin E (c), and cyclin A (d) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 1.
Day 1: Effects of Sirolimus, Paclitaxel, and CA4 on CDK Inhibitor Expression No significant differences in p16 expression occurred after treatment with sirolimus (mean, 74.95 [95% CI, 36.84–113.07]), paclitaxel (mean, 102.77 [95% CI, 64.66– 140.89]), or CA4 (mean, 108.46 [95% CI, 70.34–146.57]) compared with controls (mean, 100 [95% CI, 61.89– 138.11]) (Table 1, Fig 2a). Even the decrease with sirolimus was not statistically significant (P 4 .05). Differences in p21 expression on day 1 were not statistically significant after addition of sirolimus (mean, 79.57 [95% CI, 18.74–140.40]), paclitaxel (mean, 159.96 [95% CI, 99.13–220.79]) or CA4 (mean, 119.70 [95% CI, 58.87–180.53]) compared with controls (mean, 100.00 [95% CI, 39.17–160.83]). Mean p21 levels in cells treated with CA4 and even more so with paclitaxel were elevated (but not statistically significant, P 4 .05) (Fig 2b). The mean expression of p27 after treatment with sirolimus (mean, 142.96 [95% CI, 102.29–183.63]), CA4 (mean, 140.28 [95% CI, 99.61–180.96]), or paclitaxel (mean, 119.97 [95% CI, 79.30–160.64]) was higher than in untreated controls (Fig 2c). Nevertheless, these differences did not reach statistical significance (P 4 .05).
35.19–105.27]), paclitaxel (mean, 100.21 [95% CI, 65.17– 135.25]), or CA4 (mean, 133.39 [95% CI, 98.93–169.01]) compared with controls (mean, 100.00 [95% CI, 64.96– 135.04]) (all P 4 .05) (Table 2). Mean cyclin D1 levels in cells treated with CA4 were slightly elevated, whereas sirolimus resulted in decreased expression (Fig 3a). On day 5, cyclin D2 levels were unaffected by treatment with sirolimus (mean, 97.75 [95% CI, 64.26–131.24]), paclitaxel (mean, 113.59 [95% CI, 80.10–147.08]), or CA4 (mean, 103.92 [95% CI, 70.43–137.41]) compared with controls (mean, 100.00 [95% CI, 66.51–133.49]) (P 4 .05) (Fig 3b). Cyclin E was significantly reduced on day 5 after addition of sirolimus (mean, 51.21 [95% CI, 27.46–74.95]) or CA4 (mean, 55.23 [95% CI, 31.48–78.97]) compared with controls (mean, 100.00 [95% CI, 76.26–123.74]) (P o .05). Treatment with paclitaxel (mean, 93.24 [95% CI, 69.49–116.98]) resulted in only a minor, nonsignificant decrease of cyclin E (P 4 .05) (Fig 3c). Cyclin A expression was significantly decreased by sirolimus (mean, 57.21 [95% CI, 42.15–72.26]) and CA4 (mean, 55.57 [95% CI, 40.27–70.63]) compared with controls (mean, 100.00 [95% CI, 84.94–115.06]) (P o .05). The reduction of cyclin A levels after treatment with paclitaxel (mean, 90.80 [95% CI, 75.75–105.86]) was not statistically significant (P 4 .05) (Fig 3d).
Day 5: Effects of Sirolimus, Paclitaxel, and CA4 on Cyclin Expression
Day 5: Effects of Sirolimus, Paclitaxel, and CA4 on CDK Inhibitor Expression
Cyclin D1 expression on day 5 was not significantly different after addition of sirolimus (mean, 70.23 [95% CI,
On day 5, p16 levels were significantly reduced by sirolimus (mean, 48.83 [95% CI, 27.61–70.04]) compared
Volume XX
’
Number X
’
Month
’
2015
5
Figure 2. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of p16 (a), p21 (b), and p27 (c) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 1.
Table 2 . Synopsis of Cell Cycle Changes on Day 5 Cyclin D1
Cyclin D2
Cyclin E
Cyclin A
p16
p21
Sirolimus
—
—
↓
↓
↓
—
p27 —
Paclitaxel Combretastatin A4
— —
— —
— ↓
— ↓
— —
— (↑)
— —
↓ ¼ statistically significant decrease versus untreated controls; (↑) ¼ not statistically significant but 4 50% increase of mean value versus control.
with controls (mean, 100.00 [95% CI, 78.79–121.21]) (P o .05) (Table 2). The decrease of p16 expression after addition of paclitaxel (mean, 77.38 [95% CI, 56.16– 98.59]) and CA4 (mean, 94.59 [95% CI, 73.37–115.80]) was not statistically significant (P 4 .05) (Fig 4a). Mean expression of p21 after treatment with CA4 (mean, 176.34 [95% CI, 125.61–227.06]) was higher than in untreated controls (mean, 100.00 [95% CI, 49.27– 150.73]) or after addition of paclitaxel (mean, 91.88 [95% CI, 41.16–142.61]) and sirolimus (mean, 83.74 [95% CI, 33.02–134.47]) (Fig 4b). Nevertheless, these differences did not reach statistical significance (P 4 .05). No significant differences in p27 expression occurred on day 5 after treatment with sirolimus (mean, 75.91 [95% CI, 55.81–96.02]), paclitaxel (mean, 84.79 [95% CI, 64.68–104.89]), or CA4 (mean, 85.09 [95% CI, 64.98–105.19]) compared with controls (mean, 100.00 [95% CI, 79.90–120.10]) (P 4 .05) (Fig 4c).
Changes in Cyclin and CDK Inhibitor Expression—Day 1 versus Day 5 Comparing cyclin D1 expression on day 1 versus day 5, no significant differences were found with sirolimus or paclitaxel, whereas a significant increase from 38.17 [95% CI, 2.87–73.46] on day 1 to 133.39 [95% CI, 98.93–169.01] on day 5 was observed with CA4 (P o .05). With sirolimus and CA4, cyclin D2 expression significantly increased from day 1 (sirolimus, mean, 39.53 [95% CI, 13.09–65.96]; CA4, mean, 33.19 [95% CI, 6.76–59.63]) to day 5 (sirolimus, mean, 97.75 [95% CI, 64.26–131.24]; CA4, mean, 103.92 [95% CI, 70.43–137.41]) (P o .05). The increase was not significant in the paclitaxel group. When comparing day 1 with day 5, the cyclin E measurements listed earlier showed a moderate, nonsignificant decrease with sirolimus and increase with CA4 (P 4 .05). Values remained unchanged with paclitaxel.
6
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
Spira et al
’
JVIR
Figure 3. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of cyclin D1 (a), cyclin D2 (b), cyclin E (c), and cyclin A (d) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 5.
Figure 4. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of p16 (a), p21 (b), and p27 (c) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 5.
Volume XX
’
Number X
’
Month
’
2015
With CA4, cyclin A levels significantly increased from day 1 (mean, 22.85 [95% CI, 1.86–43.84]) to day 5 (mean, 55.57 [95% CI, 40.27–70.63]) (P o .05). The increase with paclitaxel and minor decrease with sirolimus were not statistically significant (P 4 .05). From day 1 to day 5, p16 levels as listed earlier decreased with sirolimus, paclitaxel, or CA4. However, these changes were not statistically significant (P 4 .05). When comparing day 1 with day 5, the p21 measurements listed earlier showed a moderate, nonsignificant decrease with paclitaxel and increase with CA4 (P 4 .05). Values remained unchanged with sirolimus. With sirolimus and paclitaxel, p27 expression significantly decreased from day 1 (sirolimus, mean, 142.96 [95% CI, 102.29–183.63]; paclitaxel, mean, 119.97 [95% CI, 79.30–160.64]) to day 5 (sirolimus, mean, 75.91 [95% CI, 55.81–96.02]; paclitaxel, mean, 84.79 [95% CI, 64.68– 104.89]) (P o .05). The decrease in p27 from a mean of 140.28 [95% CI, 99.61–180.96] on day 1 to a mean of 85.09 [95% CI, 64.98–105.19] on day 5 did not reach statistical significance in the CA4 group (P 4 .05).
DISCUSSION Clinical studies have focused on CA4 because, besides its antimitotic effects, it shows selective toxicity toward endothelial cells of tumor vasculature (16). CA4 selectively targets unstable tumor neovessels without periendothelial support through disruption of endothelial cell-cell contact mediated by the vascular endothelial-cadherin/β-catenin complex. However, in the presence of SMCs, endothelial cells were shown to be unaffected by CA4 (14). CA4 has not been investigated with regard to its effects on neointima formation. In this study, CA4 proved to be a strong inhibitor of SMC proliferation decreasing intracellular levels of cyclin D1, cyclin D2, cyclin E, and cyclin A after 24 hours of incubation (DAY1 group) (Table 1). After 4day treatment with CA4 and 24-hour withdrawal (DAY5 group), we observed persistent cell cycle inhibition beyond the treatment period (Table 2). Nevertheless, the increase of cyclin levels from the DAY1 to the DAY5 group indicates recovery effects early after CA4 withdrawal. Khan and Wahl (17) recognized that microtubular damage activates the spindle checkpoint leading to a temporary cell cycle arrest in mitosis. Subsequently, cells may either go into apoptosis or leave mitosis without cell division (mitotic slippage) reaching a tetraploid G1 status. In this case, a p53dependent induction of p21 after activation of the spindle checkpoint holds cells in a cell cycle arrest, preventing them from endoreplication. In the present study, the upregulation of the CDK inhibitor p21, the increase of D cyclins, and decreased cyclin E (and cyclin A) levels are in line with a G1 arrest. Treatment of SMCs for 24 hours with sirolimus (DAY1 group) resulted in decreased intracellular levels of cyclins (Fig 1a–d, Table 1). Although the CDK inhibitor p27 was upregulated, p16 and p21 were expressed slightly below
7
control values. These results parallel findings of other groups (18–21) and point at a sirolimus-induced inhibition of cellular proliferation in early G1 phase after induction of cyclin D1 but before induction of cyclins E and A. The increase we observed in cyclin expression from the DAY1 to the DAY5 group and the decrease of CDK inhibitors p16 and p27 indicate recovery effects early after sirolimus withdrawal. Sirolimus transiently inhibits the cell cycle in G1/S phase but enables cells to reenter the cell cycle after the substance has dissolved. Concerning CA4, our results also speak for a transient inhibition of the SMC cycle with recovery effects early after CA4 withdrawal. Similar to sirolimus, therapeutic exposure of CA4 likely should be permanent via drug-eluting stents instead of coated balloons. CA4 directly binds to β-tubulin inhibiting its polymerization and should suppress SMC proliferation independently of the mTOR pathway of sirolimus. CA4 may be advantageous in patients with diabetes who are treated with PPARγ agonists, which are inhibited by mTOR blockade through sirolimus (8). A significant inhibitory effect of CA4 on human umbilical vein endothelial cell migration was demonstrated even in the low nanomolar range (14). Because SMCs can migrate in G1 phase but not in later phases of the cell cycle, an inhibitory effect similar to sirolimus can also be expected on vascular SMC migration (7). Our results show that in contrast to CA4, paclitaxel in the applied dose of 5 109 mol/L affects SMC cycle progression only to a minor degree. Whether CA4 is also advantageous over other drawbacks of paclitaxel in terms of incomplete vascular healing, fibrin deposition, macrophage recruitment, and its narrow toxic-therapeutic window needs to be investigated in subsequent studies (9,10). Concerning technical applicability, a patent by Chaplin et al (22) described the possibility to load an aluminum oxide coating with a porous surface with CA4 using amounts between 5 mg and 500 mg by spraying or dipping. Analogously, the patent described the wells of a porous polytetrafluoroethylene membrane and various polymers (ie, methacrylate polymers, polyurethane coatings, polytetrafluoroethylene coatings) to be suitable for CA4 drug loading. This study has several limitations. First, a standardized dose of 5 109 mol/L was used for all cytostatic agents. However, differences in cell cycle regulation and cytostatic/cytotoxic effects have been shown for CA4, sirolimus, and paclitaxel depending on the dose applied (23,24), and so the observed findings cannot be extrapolated to higher concentrations. Second, we did not demonstrate additional functional data to support our findings. Such data need to be collected in further follow-up studies. Third, effects of CA4 on endothelial cells may result in paracrine signaling to SMCs, and our experimental design did not simulate the complex interplay between endothelial cells and SMCs. As in every cell experimental study, it is uncertain whether the in vitro findings also apply in vivo. The effect of CA4-induced inhibition of SMC proliferation on
8
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
neointima formation should be evaluated further and compared with the established agents sirolimus and paclitaxel. In conclusion, the spindle toxins CA4 and paclitaxel lead to significantly different expression of SMC cycle regulators. At a standardized dose of 5 109 mol/L, CA4 proved to be a stronger inhibitor of the SMC cycle than sirolimus or paclitaxel. CA4 may represent an alternative for drugeluting stents in atherosclerotic luminal stenosis. The effect of CA4-induced inhibition of SMC proliferation on neointima formation should be evaluated further.
REFERENCES 1. Tasaki T, Yamada S, Guo X, et al. Apoptosis signal-regulating kinase 1 deficiency attenuates vascular injury-induced neointimal hyperplasia by suppressing apoptosis in smooth muscle cells. Am J Pathol 2013; 182:597–609. 2. Hehrlein C, Zimmermann M, Pill J, et al. The role of elastic recoil after balloon angioplasty of rabbit arteries and its prevention by stent implantation. Eur Heart J 1994; 15:277–280. 3. Marx SO, Totary-Jain H, Marks AR. Vascular smooth muscle cell proliferation in restenosis. Circ Cardiovasc Interv 2011; 4:104–111. 4. Sherr CJ. The Pezcoller lecture: cancer cell cycles revisited. Cancer Res 2000; 60:3689–3695. 5. Ohtsubo M, Theodoras AM, Schumacher J, et al. Human cyclin E, a nuclear protein essential for the G1-to-S phase transition. Mol Cell Biol 1995; 15:2612–2624. 6. Tanner FC, Boehm M, Akyürek LM, et al. Differential effects of the cyclindependent kinase inhibitors p27(Kip1), p21(Cip1), and p16(Ink4) on vascular smooth muscle cell proliferation. Circulation 2000; 101:2022–2025. 7. Sun J, Marx SO, Chen HJ, et al. Role for p27(Kip1) in vascular smooth muscle cell migration. Circulation 2001; 103:2967–2972. 8. Finn AV, John M, Nakazawa G, et al. Differential healing after sirolimus, paclitaxel, and bare metal stent placement in combination with peroxisome proliferator-activator receptor gamma agonists: requirement for mTOR/Akt2 in PPARgamma activation. Circ Res 2009; 105:1003–1012. 9. Drachman DE, Edelman ER, Seifert P, et al. Neointimal thickening after stent delivery of paclitaxel: change in composition and arrest of growth over six months. J Am Coll Cardiol 2000; 36:2325–2332. 10. Wang TH, Wang HS, Soong YK. Paclitaxel-induced cell death: where the cell cycle and apoptosis come together. Cancer 2000; 88: 2619–2628.
Spira et al
’
JVIR
11. Levin AD, Vukmirovic N, Hwang CW, et al. Specific binding to intracellular proteins determines arterial transport properties for rapamycin and paclitaxel. Proc Natl Acad Sci U S A 2004; 101:9463–9467. 12. Zeller T, Baumgartner I, Scheinert D, et al. Drug-eluting balloon versus standard balloon angioplasty for infrapopliteal arterial revascularization in critical limb ischemia: 12-month results from the IN.PACT DEEP randomized trial. J Am Coll Cardiol 2014; 64:1568–1576. 13. Cushman M, Nagarathnam D, Gopal D, et al. Synthesis and evaluation of stilbene and dihydrostilbene derivatives as potential anticancer agents that inhibit tubulin polymerization. J Med Chem 1991; 34:2579–2588. 14. Vincent L, Kermani P, Young LM, et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. J Clin Invest 2005; 115:2992–3006. 15. Hayashi S, Yamamoto A, You F, et al. The stent-eluting drugs sirolimus and paclitaxel suppress healing of the endothelium by induction of autophagy. Am J Pathol 2009; 175:2226–2234. 16. Dark GG, Hill SA, Prise VE, et al. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer Res 1997; 57:1829–1834. 17. Khan SH, Wahl GM. p53 and pRb prevent rereplication in response to microtubule inhibitors by mediating a reversible G1 arrest. Cancer Res 1998; 58:396–401. 18. Viñals F, Chambard JC, Pouysségur J. p70 S6 kinase-mediated protein synthesis is a critical step for vascular endothelial cell proliferation. J Biol Chem 1999; 274:26776–26782. 19. Braun-Dullaeus RC, Mann MJ, Seay U, et al. Cell cycle protein expression in vascular smooth muscle cells in vitro and in vivo is regulated through phosphatidylinositol 3-kinase and mammalian target of rapamycin. Arterioscler Thromb Vasc Biol 2001; 21:1152–1158. 20. Kawamata S, Sakaida H, Hori T, et al. The upregulation of p27Kip1 by rapamycin results in G1 arrest in exponentially growing T-cell lines. Blood 1998; 91:561–569. 21. Morice WG, Wiederrecht G, Brunn GJ, et al. Rapamycin inhibition of interleukin-2-dependent p33cdk2 and p34cdc2 kinase activation in T lymphocytes. J Biol Chem 1993; 268:22737–22745. 22. Chaplin D, Wnendt S, Kuttler B, et al. Implants containing combretastatin a-4. Patent DE 20050065595, filed September 2, 2002, and issued March 24, 2005. 23. Tozer GM, Kanthou C, Parkins CS, et al. The biology of the combretastatins as tumour vascular targeting agents. Int J Exp Pathol 2002; 83:21–38. 24. Costa MA, Simon DI. Molecular basis of restenosis and drug-eluting stents. Circulation 2005; 111:2257–2273.
Cell Cycle Regulation of Smooth Muscle Cells— Searching for Inhibitors of Neointima Formation: Is Combretastatin A4 an Alternative to Sirolimus and Paclitaxel? Daniel Spira, MD, Gerd Grözinger, MD, Nicole Domschke, MD, Rüdiger Bantleon, PhD, Jörg Schmehl, MD, Jakub Wiskirchen, MD, and Benjamin Wiesinger, MD
ABSTRACT Purpose: To compare the effects of sirolimus, paclitaxel, and combretastatin A4 (CA4) on regulatory proteins of the cell cycle in proliferating smooth muscle cells (SMCs). Materials and Methods: Human aortic SMCs were treated with sirolimus, paclitaxel, and CA4 at 5 109 mol/L. After 1 day, half of the cells were harvested (DAY1 group). The treatment medium of the other half was replaced with culture medium on day 4, and those cells were harvested on day 5 (DAY5 group). Cyclins D1, D2, E, and A and cyclin-dependent kinase (CDK) inhibitors p16, p21, and p27 were detected by Western blot technique. Quantification was performed by scanning densitometry of the specific bands. Results: In the DAY1 group, treatment with sirolimus resulted in decreased intracellular levels of cyclins D2 and A (P o .05). Increased D cyclins and reduced levels of cyclins E and A (P o .05) in the DAY5 group indicated a permanent G1/S block by sirolimus. Paclitaxel led to only slight alterations of cyclin and CDK inhibitor expression (P 4 .05). In the DAY1 group, CA4 decreased intracellular levels of cyclins D2, E, and A (P o .05). Despite recovery effects in the DAY5 group (increase of cyclins D1, D2, and A compared with DAY1 group; P o .05), the upregulation of the CDK inhibitor p21, increased D cyclins, and decreased cyclins E and A (P o .05) are compatible with a G1 arrest. Conclusions: CA4 is a stronger inhibitor of the SMC cycle than sirolimus or paclitaxel and may represent an alternative for drug-eluting stents in atherosclerotic luminal stenosis. The effect of CA4 on neointima formation should be evaluated further.
ABBREVIATIONS CA4 = combretastatin A4, CDK = cyclin-dependent kinase, CI = confidence interval, mTOR = mammalian target of rapamycin, PPARγ = peroxisome proliferator activated receptor γ, SMC = smooth muscle cell
From the Department of Diagnostic and Interventional Radiology (D.S.), University Hospital Heidelberg, Im Neuenheimer Feld 110, Heidelberg 69120, Germany; Department of Diagnostic and Interventional Radiology (G.G., R.B., J.S., B.W.), Eberhard-Karls-University, Tübingen, Germany; Department of Internal Medicine (N.D.), SRH Hospital Sigmaringen, Sigmaringen, Germany; and Department of Radiology and Nuclear Medicine (J.W.), Franziskus Hospital, Bielefeld, Germany. Received February 16, 2015; final revision received May 4, 2015; accepted May 10, 2015. Address correspondence to D.S.; E-mail: [email protected] None of the authors have identified a conflict of interest. & SIR, 2015 J Vasc Interv Radiol 2015; XX:]]]–]]] http://dx.doi.org/10.1016/j.jvir.2015.05.025
In atherosclerosis, restenosis after percutaneous transluminal angioplasty or in-stent restenosis caused by neointima formation is a major issue in clinical situations. Interventional trauma to the vessel wall results in exposure of the subintimal matrix, triggering neointima formation. This process takes 1–3 months and entails proliferation, migration, and apoptosis of medial smooth muscle cells (SMCs) (1). In a 24- to 72-hour time period, 20%–40% of the SMCs enter the cell cycle and begin to proliferate (2). This situation has led to the development of drug-eluting balloons and stents enabling the local application of agents to suppress the inflammatory response and consecutive SMC proliferation. Currently
2
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
available drug-eluting stents convey high levels of the cell cycle inhibitors sirolimus, sirolimus analogues, and paclitaxel to the vessel wall (3). Vascular SMCs normally stay in G0/G1 phase of the cell cycle. After vascular injury, SMCs are stimulated to pass a restriction point and enter S phase, where the DNA is replicated. This entry into S phase is regulated by complexes of cyclins and cyclin-dependent kinases (CDKs) as well as CDK inhibitors. D cyclins bind and activate CDK 4 and CDK 6, which are necessary for progression through G1 phase (4). Cyclin E is selectively expressed in proliferating cells at the transition of G1 to S phase building a complex with its catalytic partner CDK 2 (5). Vascular SMC proliferation is inhibited by the CDK inhibitors p27 and p21 as a result of suppression of CDK 2 activity (6). Sirolimus inhibits mammalian target of rapamycin (mTOR), which prevents the decomposition of the CDK inhibitor p27. The CDK inhibitor p27 suppresses cyclin E/CDK 2 kinase activity leading to G1- to S-phase transition arrest (3). Sirolimus also has inhibitory effects on vascular SMC migration (7). Paclitaxel attaches to β-tubulin and leads to cell cycle arrest during the M (mitosis) phase through tubulin polymerization and suppression of spindle formation and intracellular transport. It inhibits SMC proliferation and migration independently of intracellular signaling mechanisms such as those of sirolimus and its mTOR pathway (8). Despite advances in the drug-eluting technology, some drawbacks remain. First, although sirolimus is efficacious in reducing neointima formation, this effect may be hindered in patients with diabetes who are treated with oral agonists of peroxisome proliferator activated receptor γ (PPARγ), such as rosiglitazone or pioglitazone, as blockade of mTOR by sirolimus was shown to attenuate PPARγ activity (8). Paclitaxel also reduces neointima formation, but treated arteries showed incomplete healing and long-lasting fibrin deposition accompanied by a large number of macrophages (9). Additionally, in the dose range used for clinical applications, paclitaxel induces apoptosis, which hints at a narrow toxic-therapeutic window (10). Furthermore, paclitaxel accumulates in the adventitia of vessels, which assumes an inferior role in the pathophysiology of neointima formation and restenosis (11). Finally, paclitaxel-eluting balloons failed to show their superiority compared with conventional pressure-only balloon angioplasty in the lower limb (12). These drawbacks reopen the question about the best antiproliferative agent for neointima inhibition. Combretastatin A4 (CA4) binds to the colchicine binding site of β-tubulin and inhibits its polymerization, preventing the degradation of the spindle apparatus and hindering mitosis (13). Because CA4 proved to be an effective cytostatic agent, it should inhibit SMC proliferation independently of the mTOR pathway of sirolimus and may suppress neointima formation. The
Spira et al
’
JVIR
purpose of this study was to compare the effects of sirolimus, paclitaxel, and CA4 on regulatory proteins of the cell cycle (cyclin A, cyclin D1, cyclin D2, cyclin E, p16, p21, and p27) in proliferating vascular SMCs.
MATERIALS AND METHODS Cell Culture Human aortic SMCs (ATCC, Rockville, Maryland) were cultured in Weymouth/Ham’s F12 medium (GIBCO, Karlsruhe, Germany) in a 1:1 ratio. The medium was supplemented with 20% fetal calf serum (Boehringer Ingelheim, Mannheim, Germany), 1% glutamine, and 1% penicillin/streptomycin. Cell cultures were incubated in 25 cm2 tissue culture flasks (Falcon; Becton, Dickinson and Company, Franklin Lakes, New Jersey) at 371C in a 5% carbon dioxide atmosphere. Cell density was 2 105 cells per 5 mL culture medium. The culture medium was changed on day 1 and day 4. Cells were harvested using trypsinization every 7 days. Absolute cell numbers were determined with an electronic cell counter (CASY; Schärfe System GmbH, Reutlingen, Germany).
Treatment of Human Aortic SMCs Three cytostatic drugs were compared with untreated controls: sirolimus, paclitaxel, and CA4 (Sigma-Aldrich, Deisenhofen, Germany). Cells were seeded at a density of 400,000 cells in 75 cm2 cell culture flasks with 10 mL medium containing 20% fetal calf serum. After 2 days, the medium was changed, and the medication was added in a concentration of 5 109 mol/L in the treated groups, whereas untreated controls received normal culture medium. Vincent et al (14) described antiproliferative effects of CA4 on endothelial cells at nontoxic doses of 5 nmol/L and 10 nmol/L. Also, Hayashi et al (15) showed inhibitory effects of sirolimus and paclitaxel on the cell cycle in the low nanomolar range. To ensure comparability, we chose 5 109 mol/L as the concentration in all experiments. After 1 day, half of the cells (three out of six 75 cm2 cell culture flasks) were washed with cold phosphate-buffered saline and were harvested mechanically with the help of a cell scraper (DAY1 group). The treatment medium of the other half of the cells was replaced with culture medium on day 4, and those cells were harvested on day 5 (DAY5 group). The rationale for the groups DAY1 (24-h treatment) and DAY5 (4-d treatment and 24-h withdrawal) was to search for effects on the cell cycle early during treatment and early after treatment cessation.
Test Procedures Procedures were performed on ice to inhibit the activity of proteases. Cells were lysed with lysis buffer (0.05 mol/L tris-(hydroxymethyl)-aminomethane, pH 8, 5 mL; 0.12 mol/ L sodium chloride, 2.4 mL; 0.5% Triton X-100, 500 mL; 0.005 mol/L ethylenediamine tetraacetic acid, 1 mL; distilled
Volume XX
’
Number X
’
Month
’
2015
3
water, 91.1 mL) after protease inhibitors aprotinin/leupeptin (1 mL lysis buffer þ 7.5 mL Pefabloc [Roche Diagnostics, Mannheim, Germany; 200 mmol/L] þ 5 mL aprotinin/ leupeptin [aprotinin 0.4 mg/mL, leupeptin 1 mg/mL]) were added and pelleted after centrifugation. To maximize protein extraction, additional sonication was performed using ultrasound (Sonifier B-12; Branson Sonic Power Co., Danbury, Connecticut). The lysed cell solution was centrifuged again, and the supernatant, containing soluble cytoplasmic proteins, was saved. Total protein content was determined by Bradford protein assay, and the solution was stored at 701C until Western blotting. For Western blotting, proteins were denatured by mercaptoethanol 50 mmol/L and heating at 951C for 4 minutes. Then 25 mg of the protein solution was added to each lane of a 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel. After electrophoresis, the protein was transferred onto a Hybond-N nylon membrane (Amersham, Buckinghamshire, United Kingdom) by semidry blotting with help of an electroblotting chamber system (Semi-Phor; Hoefer Pharmacia Biotech, San Francisco, California). For detection of cyclin D1, cyclin D2, cyclin E, cyclin A, p16, p21, and p27, the membrane was incubated with a specific monoclonal antibody in a concentration of 1:2,000 (New England Biolabs, Beverly, Massachusetts). After washing the membrane, a second antibody in a concentration of 1:5,000 labeled with horseradish peroxidase (Amersham) was added; after incubation, LumiGLO (New England Biolabs) was added. Blocking of the polyvinylidene difluoride membrane was performed with milk powder solution. Membrane detection was conducted with a Lumi-Imager Diana III (Raytest GmbH, Straubenhardt, Germany). β-Actin was used to control protein load in the Western blot. Quantification was done by scanning densitometry of the specific bands of each lane and was measured with help of the AIDA BioPackage (Raytest GmbH). The optical density of controls was equated to 100, and the optical density of the other groups was determined accordingly by computer analysis. The experiments were repeated 12 times (n ¼ 12). Four test preparations (sirolimus, paclitaxel, CA4, untreated control) and three Western blots performed for each preparation resulted in 12 measured values for each of the seven regulator proteins (cyclins D1, D2, E, and A and CDK inhibitors p16, p21, and p27) on day 1 and another 12 values on day 5, respectively.
Statistics All data are reported as arithmetic mean and [95% confidence interval (CI)]. Statistical analysis was performed using a multivariate analysis via Tukey-Kramer test in conjunction with an analysis of variance for all pairwise comparisons and Student t test for unpaired data (comparison day 1 vs day 5). P values o .05 were considered statistically significant.
RESULTS Day 1: Effects of Sirolimus, Paclitaxel, and CA4 on Cyclin Expression Differences in cyclin D1 expression on day 1 were not statistically significant after addition of sirolimus (mean, 60.37 [95% CI, 25.07–95.67]), paclitaxel (mean, 113.74 [95% CI, 78.44–149.04]) or CA4 (mean, 38.17 [95% CI, 2.87–73.46]) compared with controls (mean, 100.00 [95% CI, 64.70–135.30]) (all P 4 .05) (Table 1). Mean cyclin D1 levels in cells treated with paclitaxel were slightly elevated, whereas sirolimus and CA4 resulted in decreased expression (Fig 1a). In a direct comparison with paclitaxel, suppression of cyclin D1 by CA4 reached statistical significance (P o .05). Cyclin D2 expression was significantly decreased by sirolimus (mean, 39.53 [95% CI, 13.09–65.96]) and CA4 (mean, 33.19 [95% CI, 6.76–59.63]) compared with either controls (mean, 100.00 [95% CI, 73.57–126.43]) or paclitaxel (mean, 90.898 [95% CI, 64.464–117.33]) (all P o .05) (Fig 1b). Cyclin E levels were significantly reduced after addition of CA4 (mean, 28.22 [95% CI, 3.80 to 60.25]) compared with controls (mean, 100.00 [95% CI, 67.97– 132.03]) (P o .05). Treatment with sirolimus (mean, 74.19 [95% CI, 42.17–106.22]) and paclitaxel (mean, 92.84 [95% CI, 60.82–124.87]) resulted in only a minor, nonsignificant decrease of cyclin E (P 4 .05) (Fig 1c). Cyclin A expression was significantly decreased by sirolimus (mean, 60.59 [95% CI, 39.60–81.58]) and even more so by CA4 (mean, 22.85 [95% CI, 1.86– 43.84]) compared with controls (mean, 100.00 [95% CI, 79.01–120.99]) (P o .05). The reduction of cyclin A levels after treatment with paclitaxel (mean, 72.94 [95% CI, 51.95–93.92]) was not statistically significant (P 4 .05) (Fig 1d).
Table 1 . Synopsis of Cell Cycle Changes on Day 1 Cyclin D1
Cyclin D2
Cyclin E
Cyclin A
p16
p21
Sirolimus
—
↓
—
↓
—
—
p27 —
Paclitaxel Combretastatin A4
— (↓)
— ↓
— ↓
— ↓
— —
(↑) —
— —
↑ ¼ statistically significant increase versus untreated controls; ↓ ¼ statistically significant decrease versus untreated controls; (↑) ¼ not statistically significant but 4 50% increase of mean value versus control; (↓) ¼ not statistically significant but 4 50% decrease of mean value versus control.
4
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
Spira et al
’
JVIR
Figure 1. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of cyclin D1 (a), cyclin D2 (b), cyclin E (c), and cyclin A (d) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 1.
Day 1: Effects of Sirolimus, Paclitaxel, and CA4 on CDK Inhibitor Expression No significant differences in p16 expression occurred after treatment with sirolimus (mean, 74.95 [95% CI, 36.84–113.07]), paclitaxel (mean, 102.77 [95% CI, 64.66– 140.89]), or CA4 (mean, 108.46 [95% CI, 70.34–146.57]) compared with controls (mean, 100 [95% CI, 61.89– 138.11]) (Table 1, Fig 2a). Even the decrease with sirolimus was not statistically significant (P 4 .05). Differences in p21 expression on day 1 were not statistically significant after addition of sirolimus (mean, 79.57 [95% CI, 18.74–140.40]), paclitaxel (mean, 159.96 [95% CI, 99.13–220.79]) or CA4 (mean, 119.70 [95% CI, 58.87–180.53]) compared with controls (mean, 100.00 [95% CI, 39.17–160.83]). Mean p21 levels in cells treated with CA4 and even more so with paclitaxel were elevated (but not statistically significant, P 4 .05) (Fig 2b). The mean expression of p27 after treatment with sirolimus (mean, 142.96 [95% CI, 102.29–183.63]), CA4 (mean, 140.28 [95% CI, 99.61–180.96]), or paclitaxel (mean, 119.97 [95% CI, 79.30–160.64]) was higher than in untreated controls (Fig 2c). Nevertheless, these differences did not reach statistical significance (P 4 .05).
35.19–105.27]), paclitaxel (mean, 100.21 [95% CI, 65.17– 135.25]), or CA4 (mean, 133.39 [95% CI, 98.93–169.01]) compared with controls (mean, 100.00 [95% CI, 64.96– 135.04]) (all P 4 .05) (Table 2). Mean cyclin D1 levels in cells treated with CA4 were slightly elevated, whereas sirolimus resulted in decreased expression (Fig 3a). On day 5, cyclin D2 levels were unaffected by treatment with sirolimus (mean, 97.75 [95% CI, 64.26–131.24]), paclitaxel (mean, 113.59 [95% CI, 80.10–147.08]), or CA4 (mean, 103.92 [95% CI, 70.43–137.41]) compared with controls (mean, 100.00 [95% CI, 66.51–133.49]) (P 4 .05) (Fig 3b). Cyclin E was significantly reduced on day 5 after addition of sirolimus (mean, 51.21 [95% CI, 27.46–74.95]) or CA4 (mean, 55.23 [95% CI, 31.48–78.97]) compared with controls (mean, 100.00 [95% CI, 76.26–123.74]) (P o .05). Treatment with paclitaxel (mean, 93.24 [95% CI, 69.49–116.98]) resulted in only a minor, nonsignificant decrease of cyclin E (P 4 .05) (Fig 3c). Cyclin A expression was significantly decreased by sirolimus (mean, 57.21 [95% CI, 42.15–72.26]) and CA4 (mean, 55.57 [95% CI, 40.27–70.63]) compared with controls (mean, 100.00 [95% CI, 84.94–115.06]) (P o .05). The reduction of cyclin A levels after treatment with paclitaxel (mean, 90.80 [95% CI, 75.75–105.86]) was not statistically significant (P 4 .05) (Fig 3d).
Day 5: Effects of Sirolimus, Paclitaxel, and CA4 on Cyclin Expression
Day 5: Effects of Sirolimus, Paclitaxel, and CA4 on CDK Inhibitor Expression
Cyclin D1 expression on day 5 was not significantly different after addition of sirolimus (mean, 70.23 [95% CI,
On day 5, p16 levels were significantly reduced by sirolimus (mean, 48.83 [95% CI, 27.61–70.04]) compared
Volume XX
’
Number X
’
Month
’
2015
5
Figure 2. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of p16 (a), p21 (b), and p27 (c) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 1.
Table 2 . Synopsis of Cell Cycle Changes on Day 5 Cyclin D1
Cyclin D2
Cyclin E
Cyclin A
p16
p21
Sirolimus
—
—
↓
↓
↓
—
p27 —
Paclitaxel Combretastatin A4
— —
— —
— ↓
— ↓
— —
— (↑)
— —
↓ ¼ statistically significant decrease versus untreated controls; (↑) ¼ not statistically significant but 4 50% increase of mean value versus control.
with controls (mean, 100.00 [95% CI, 78.79–121.21]) (P o .05) (Table 2). The decrease of p16 expression after addition of paclitaxel (mean, 77.38 [95% CI, 56.16– 98.59]) and CA4 (mean, 94.59 [95% CI, 73.37–115.80]) was not statistically significant (P 4 .05) (Fig 4a). Mean expression of p21 after treatment with CA4 (mean, 176.34 [95% CI, 125.61–227.06]) was higher than in untreated controls (mean, 100.00 [95% CI, 49.27– 150.73]) or after addition of paclitaxel (mean, 91.88 [95% CI, 41.16–142.61]) and sirolimus (mean, 83.74 [95% CI, 33.02–134.47]) (Fig 4b). Nevertheless, these differences did not reach statistical significance (P 4 .05). No significant differences in p27 expression occurred on day 5 after treatment with sirolimus (mean, 75.91 [95% CI, 55.81–96.02]), paclitaxel (mean, 84.79 [95% CI, 64.68–104.89]), or CA4 (mean, 85.09 [95% CI, 64.98–105.19]) compared with controls (mean, 100.00 [95% CI, 79.90–120.10]) (P 4 .05) (Fig 4c).
Changes in Cyclin and CDK Inhibitor Expression—Day 1 versus Day 5 Comparing cyclin D1 expression on day 1 versus day 5, no significant differences were found with sirolimus or paclitaxel, whereas a significant increase from 38.17 [95% CI, 2.87–73.46] on day 1 to 133.39 [95% CI, 98.93–169.01] on day 5 was observed with CA4 (P o .05). With sirolimus and CA4, cyclin D2 expression significantly increased from day 1 (sirolimus, mean, 39.53 [95% CI, 13.09–65.96]; CA4, mean, 33.19 [95% CI, 6.76–59.63]) to day 5 (sirolimus, mean, 97.75 [95% CI, 64.26–131.24]; CA4, mean, 103.92 [95% CI, 70.43–137.41]) (P o .05). The increase was not significant in the paclitaxel group. When comparing day 1 with day 5, the cyclin E measurements listed earlier showed a moderate, nonsignificant decrease with sirolimus and increase with CA4 (P 4 .05). Values remained unchanged with paclitaxel.
6
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
Spira et al
’
JVIR
Figure 3. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of cyclin D1 (a), cyclin D2 (b), cyclin E (c), and cyclin A (d) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 5.
Figure 4. Graphic results of one-way analysis of variance and Tukey-Kramer test on the expression of p16 (a), p21 (b), and p27 (c) after treatment with sirolimus, paclitaxel, and CA4 in the percent of untreated controls on day 5.
Volume XX
’
Number X
’
Month
’
2015
With CA4, cyclin A levels significantly increased from day 1 (mean, 22.85 [95% CI, 1.86–43.84]) to day 5 (mean, 55.57 [95% CI, 40.27–70.63]) (P o .05). The increase with paclitaxel and minor decrease with sirolimus were not statistically significant (P 4 .05). From day 1 to day 5, p16 levels as listed earlier decreased with sirolimus, paclitaxel, or CA4. However, these changes were not statistically significant (P 4 .05). When comparing day 1 with day 5, the p21 measurements listed earlier showed a moderate, nonsignificant decrease with paclitaxel and increase with CA4 (P 4 .05). Values remained unchanged with sirolimus. With sirolimus and paclitaxel, p27 expression significantly decreased from day 1 (sirolimus, mean, 142.96 [95% CI, 102.29–183.63]; paclitaxel, mean, 119.97 [95% CI, 79.30–160.64]) to day 5 (sirolimus, mean, 75.91 [95% CI, 55.81–96.02]; paclitaxel, mean, 84.79 [95% CI, 64.68– 104.89]) (P o .05). The decrease in p27 from a mean of 140.28 [95% CI, 99.61–180.96] on day 1 to a mean of 85.09 [95% CI, 64.98–105.19] on day 5 did not reach statistical significance in the CA4 group (P 4 .05).
DISCUSSION Clinical studies have focused on CA4 because, besides its antimitotic effects, it shows selective toxicity toward endothelial cells of tumor vasculature (16). CA4 selectively targets unstable tumor neovessels without periendothelial support through disruption of endothelial cell-cell contact mediated by the vascular endothelial-cadherin/β-catenin complex. However, in the presence of SMCs, endothelial cells were shown to be unaffected by CA4 (14). CA4 has not been investigated with regard to its effects on neointima formation. In this study, CA4 proved to be a strong inhibitor of SMC proliferation decreasing intracellular levels of cyclin D1, cyclin D2, cyclin E, and cyclin A after 24 hours of incubation (DAY1 group) (Table 1). After 4day treatment with CA4 and 24-hour withdrawal (DAY5 group), we observed persistent cell cycle inhibition beyond the treatment period (Table 2). Nevertheless, the increase of cyclin levels from the DAY1 to the DAY5 group indicates recovery effects early after CA4 withdrawal. Khan and Wahl (17) recognized that microtubular damage activates the spindle checkpoint leading to a temporary cell cycle arrest in mitosis. Subsequently, cells may either go into apoptosis or leave mitosis without cell division (mitotic slippage) reaching a tetraploid G1 status. In this case, a p53dependent induction of p21 after activation of the spindle checkpoint holds cells in a cell cycle arrest, preventing them from endoreplication. In the present study, the upregulation of the CDK inhibitor p21, the increase of D cyclins, and decreased cyclin E (and cyclin A) levels are in line with a G1 arrest. Treatment of SMCs for 24 hours with sirolimus (DAY1 group) resulted in decreased intracellular levels of cyclins (Fig 1a–d, Table 1). Although the CDK inhibitor p27 was upregulated, p16 and p21 were expressed slightly below
7
control values. These results parallel findings of other groups (18–21) and point at a sirolimus-induced inhibition of cellular proliferation in early G1 phase after induction of cyclin D1 but before induction of cyclins E and A. The increase we observed in cyclin expression from the DAY1 to the DAY5 group and the decrease of CDK inhibitors p16 and p27 indicate recovery effects early after sirolimus withdrawal. Sirolimus transiently inhibits the cell cycle in G1/S phase but enables cells to reenter the cell cycle after the substance has dissolved. Concerning CA4, our results also speak for a transient inhibition of the SMC cycle with recovery effects early after CA4 withdrawal. Similar to sirolimus, therapeutic exposure of CA4 likely should be permanent via drug-eluting stents instead of coated balloons. CA4 directly binds to β-tubulin inhibiting its polymerization and should suppress SMC proliferation independently of the mTOR pathway of sirolimus. CA4 may be advantageous in patients with diabetes who are treated with PPARγ agonists, which are inhibited by mTOR blockade through sirolimus (8). A significant inhibitory effect of CA4 on human umbilical vein endothelial cell migration was demonstrated even in the low nanomolar range (14). Because SMCs can migrate in G1 phase but not in later phases of the cell cycle, an inhibitory effect similar to sirolimus can also be expected on vascular SMC migration (7). Our results show that in contrast to CA4, paclitaxel in the applied dose of 5 109 mol/L affects SMC cycle progression only to a minor degree. Whether CA4 is also advantageous over other drawbacks of paclitaxel in terms of incomplete vascular healing, fibrin deposition, macrophage recruitment, and its narrow toxic-therapeutic window needs to be investigated in subsequent studies (9,10). Concerning technical applicability, a patent by Chaplin et al (22) described the possibility to load an aluminum oxide coating with a porous surface with CA4 using amounts between 5 mg and 500 mg by spraying or dipping. Analogously, the patent described the wells of a porous polytetrafluoroethylene membrane and various polymers (ie, methacrylate polymers, polyurethane coatings, polytetrafluoroethylene coatings) to be suitable for CA4 drug loading. This study has several limitations. First, a standardized dose of 5 109 mol/L was used for all cytostatic agents. However, differences in cell cycle regulation and cytostatic/cytotoxic effects have been shown for CA4, sirolimus, and paclitaxel depending on the dose applied (23,24), and so the observed findings cannot be extrapolated to higher concentrations. Second, we did not demonstrate additional functional data to support our findings. Such data need to be collected in further follow-up studies. Third, effects of CA4 on endothelial cells may result in paracrine signaling to SMCs, and our experimental design did not simulate the complex interplay between endothelial cells and SMCs. As in every cell experimental study, it is uncertain whether the in vitro findings also apply in vivo. The effect of CA4-induced inhibition of SMC proliferation on
8
’
Smooth Muscle Cell Cycle Regulation by Combretastatin A4
neointima formation should be evaluated further and compared with the established agents sirolimus and paclitaxel. In conclusion, the spindle toxins CA4 and paclitaxel lead to significantly different expression of SMC cycle regulators. At a standardized dose of 5 109 mol/L, CA4 proved to be a stronger inhibitor of the SMC cycle than sirolimus or paclitaxel. CA4 may represent an alternative for drugeluting stents in atherosclerotic luminal stenosis. The effect of CA4-induced inhibition of SMC proliferation on neointima formation should be evaluated further.
REFERENCES 1. Tasaki T, Yamada S, Guo X, et al. Apoptosis signal-regulating kinase 1 deficiency attenuates vascular injury-induced neointimal hyperplasia by suppressing apoptosis in smooth muscle cells. Am J Pathol 2013; 182:597–609. 2. Hehrlein C, Zimmermann M, Pill J, et al. The role of elastic recoil after balloon angioplasty of rabbit arteries and its prevention by stent implantation. Eur Heart J 1994; 15:277–280. 3. Marx SO, Totary-Jain H, Marks AR. Vascular smooth muscle cell proliferation in restenosis. Circ Cardiovasc Interv 2011; 4:104–111. 4. Sherr CJ. The Pezcoller lecture: cancer cell cycles revisited. Cancer Res 2000; 60:3689–3695. 5. Ohtsubo M, Theodoras AM, Schumacher J, et al. Human cyclin E, a nuclear protein essential for the G1-to-S phase transition. Mol Cell Biol 1995; 15:2612–2624. 6. Tanner FC, Boehm M, Akyürek LM, et al. Differential effects of the cyclindependent kinase inhibitors p27(Kip1), p21(Cip1), and p16(Ink4) on vascular smooth muscle cell proliferation. Circulation 2000; 101:2022–2025. 7. Sun J, Marx SO, Chen HJ, et al. Role for p27(Kip1) in vascular smooth muscle cell migration. Circulation 2001; 103:2967–2972. 8. Finn AV, John M, Nakazawa G, et al. Differential healing after sirolimus, paclitaxel, and bare metal stent placement in combination with peroxisome proliferator-activator receptor gamma agonists: requirement for mTOR/Akt2 in PPARgamma activation. Circ Res 2009; 105:1003–1012. 9. Drachman DE, Edelman ER, Seifert P, et al. Neointimal thickening after stent delivery of paclitaxel: change in composition and arrest of growth over six months. J Am Coll Cardiol 2000; 36:2325–2332. 10. Wang TH, Wang HS, Soong YK. Paclitaxel-induced cell death: where the cell cycle and apoptosis come together. Cancer 2000; 88: 2619–2628.
Spira et al
’
JVIR
11. Levin AD, Vukmirovic N, Hwang CW, et al. Specific binding to intracellular proteins determines arterial transport properties for rapamycin and paclitaxel. Proc Natl Acad Sci U S A 2004; 101:9463–9467. 12. Zeller T, Baumgartner I, Scheinert D, et al. Drug-eluting balloon versus standard balloon angioplasty for infrapopliteal arterial revascularization in critical limb ischemia: 12-month results from the IN.PACT DEEP randomized trial. J Am Coll Cardiol 2014; 64:1568–1576. 13. Cushman M, Nagarathnam D, Gopal D, et al. Synthesis and evaluation of stilbene and dihydrostilbene derivatives as potential anticancer agents that inhibit tubulin polymerization. J Med Chem 1991; 34:2579–2588. 14. Vincent L, Kermani P, Young LM, et al. Combretastatin A4 phosphate induces rapid regression of tumor neovessels and growth through interference with vascular endothelial-cadherin signaling. J Clin Invest 2005; 115:2992–3006. 15. Hayashi S, Yamamoto A, You F, et al. The stent-eluting drugs sirolimus and paclitaxel suppress healing of the endothelium by induction of autophagy. Am J Pathol 2009; 175:2226–2234. 16. Dark GG, Hill SA, Prise VE, et al. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer Res 1997; 57:1829–1834. 17. Khan SH, Wahl GM. p53 and pRb prevent rereplication in response to microtubule inhibitors by mediating a reversible G1 arrest. Cancer Res 1998; 58:396–401. 18. Viñals F, Chambard JC, Pouysségur J. p70 S6 kinase-mediated protein synthesis is a critical step for vascular endothelial cell proliferation. J Biol Chem 1999; 274:26776–26782. 19. Braun-Dullaeus RC, Mann MJ, Seay U, et al. Cell cycle protein expression in vascular smooth muscle cells in vitro and in vivo is regulated through phosphatidylinositol 3-kinase and mammalian target of rapamycin. Arterioscler Thromb Vasc Biol 2001; 21:1152–1158. 20. Kawamata S, Sakaida H, Hori T, et al. The upregulation of p27Kip1 by rapamycin results in G1 arrest in exponentially growing T-cell lines. Blood 1998; 91:561–569. 21. Morice WG, Wiederrecht G, Brunn GJ, et al. Rapamycin inhibition of interleukin-2-dependent p33cdk2 and p34cdc2 kinase activation in T lymphocytes. J Biol Chem 1993; 268:22737–22745. 22. Chaplin D, Wnendt S, Kuttler B, et al. Implants containing combretastatin a-4. Patent DE 20050065595, filed September 2, 2002, and issued March 24, 2005. 23. Tozer GM, Kanthou C, Parkins CS, et al. The biology of the combretastatins as tumour vascular targeting agents. Int J Exp Pathol 2002; 83:21–38. 24. Costa MA, Simon DI. Molecular basis of restenosis and drug-eluting stents. Circulation 2005; 111:2257–2273.