OUP user menu

Short- and long-term effects of cytochalasin D, paclitaxel and rapamycin on wall thickening in experimental porcine vein grafts

Gavin J. Murphy, Thomas W. Johnson, Martin H. Chamberlain, S. Imran Rizvi, Marcie Wyatt, Sarah J. George, Gianni D. Angelini, Karl R. Karsch, Martin Oberhoff, Andrew C. Newby
DOI: http://dx.doi.org/10.1016/j.cardiores.2006.11.015 607-617 First published online: 1 February 2007


Objectives: Neointima formation and wall thickening caused by smooth muscle cell proliferation compromise long-term patency of human aorto-coronary vein-grafts. We investigated short- and long-term effects of anti-proliferative pharmacological agents on experimental pig vein-grafts with similar dimensions and kinetics to human coronary grafts.

Methods and results: Saphenous veins were treated for 1 h ex vivo with vehicle or concentrations of cytochalasin D, paclitaxel or rapamycin found to be anti-proliferative in preliminary studies. Vehicle and treated veins were implanted contralaterally, end-to-end into the carotid arteries of pigs. Cytochalasin D 2.5 μg/ml non-significantly reduced neointima formation in 4-week vein-grafts (mean±standard error, 2.5±0.6 vs. 3.3±0.6 mm2, n=10, p=NS), whilst paclitaxel 10 μM produced significant inhibition (1.7±0.2 vs. 3.0±0.3 mm2, n=8, p<0.01) as did rapamycin 0.1 mg/ml (0.6±0.3 vs. 1.7±0.5 mm2, n=8, p<0.02). Similar effects were found on total wall cross-sectional area but medial area was unaffected. PCNA staining of 1-week vein grafts confirmed in vivo anti-proliferative effects of paclitaxel (21±2 vs. 36±3%, n=5, p<0.01) and rapamycin (32±1 vs. 57±6%, n=6, p<0.005); neither agent stimulated loss of endothelium at these concentrations. Neointima and total wall area increased significantly between 4- and 12-weeks in all vein-grafts such that there was no longer a significant effect on neointima formation of either paclitaxel (7.5±1.3 vs. 8.9±1.9 mm2 in control, n=5, p=NS) or rapamycin (6.0±0.9 vs. 7.9±1.1 mm2 in control, n=9, p=NS) or on total wall area in 12-week grafts.

Conclusions: Pre-treatment of saphenous vein with anti-proliferative agents paclitaxel or rapamycin reduced neointima and total wall area after 4 weeks but continued growth abolished differences by 12 weeks. These results may help to understand the failure of clinical studies using anti-proliferative treatments in vein-grafts.

  • Vein-grafts
  • Paclitaxel
  • Rapamycin
  • Neointima

1. Introduction

More than 1 million coronary bypass procedures are undertaken each year worldwide and despite increasing use of arterial conduits an average of two saphenous vein-grafts are still deployed per patient [1,2]. Unfortunately, approximately 10–15% of coronary vein-grafts suffer early thrombotic occlusion and another 10–30% occlude in the first 1–5 years because of neointimal hyperplasia [1,2]. Moreover neointima formation in vein-grafts is associated with adverse clinical outcome [1]. Another 30–40% occlude in the next 5–7 years because of accelerated atherosclerosis, superimposed on the thickened neointima. As a result, less than half of vein-grafts remain patent after 12 years and re-intervention is frequently needed to treat recurrent angina pectoris [3].

Successful strategies to reduce vein graft failure have employed early anti-thrombotic treatments (mainly aspirin [4]) followed by aggressive lipid lowering with statins and other agents [5]. Nevertheless late patency rates continue to be much poorer than for arterial conduits [2] because the arterialisation of the vein-graft wall that occurs in response to increased pressure accelerates neointima formation and late atherosclerosis. Placement of an external polyester collar around vein-grafts in pigs prevents neointima formation, graft wall thickening and foam-cell accumulation [6–9]. However, translation of this into a clinically useful strategy is still awaited. In contrast, treatments aimed at inhibiting smooth muscle cell proliferation have been already translated from experimental to clinical studies. Decoy oligodeoxyribonucleotides (ODNS) that target several transcription factors prevent neointima formation and later atherosclerosis in rabbit vein-grafts [10,11], and initial clinical experience with this technique was promising. In the small PREVENT 1 safety trial pre-treatment of human saphenous veins ex vivo with E2F transcription factor decoy reduced proliferation in veins in vitro and marginally significantly decreased peripheral bypass graft failures after 1 year [12]. The unpublished PREVENT II trial apparently showed similar effects in coronary vein grafts. This led to the large Phase 3 PREVENT III/IV trials of edifoligide in peripheral and coronary vein grafts, respectively. Unfortunately, neither trial reached its primary endpoint (improved patency at 1 year) [1] and development of edifoligide has therefore been halted. Following these negative results, the entire anti-proliferative strategy has been called into question and there remains no effective, clinical treatment for neointima formation in vein grafts.

The anti-proliferative agents rapamycin and paclitaxel have been effective clinically in reducing in-stent restenosis. When delivered locally by elution from stents the restenosis rates are reduced by at least three quarters [13,14]. Moreover, effects are sustained for several years [15,16]. Unlike decoy ODNS, which are highly water-soluble, rapamycin and paclitaxel are lipid-soluble and therefore have prolonged effects on smooth muscle cell (SMC) proliferation even when administered for a short period. Spurred on by the success of the clinical trials of drug eluting coronary stents we investigated whether local pre-treatment of veins with paclitaxel, rapamycin or another lipid soluble agent, cytochalasin D would have early and sustained effects on vein-grafts in a pig model. Each of the agents has a very different mode of action. Rapamycin is a macrolide antibiotic with potent antiproliferative properties attributed to the inhibition of mTOR [17]. Cytochalasin disrupts the actin cytoskeleton and destabilises signalling complexes involved in extracellular matrix regulation of the cell cycle [18], whilst paclitaxel stabilises microtubules and hence prevents mitosis [19]. We chose to use the pig model because the wall thickness of pig saphenous vein is more comparable to human saphenous vein than those used in mouse or rabbit models and hence the biomechanical triggers for wall thickening are also more similar to those in human grafts [20]. Moreover the time-course of wall thickening in pig vein-grafts (3 months, [21]) is comparable to that determined by intravascular ultrasound measurements in man [22].

2. Methods

2.1. Materials

Unless stated, all chemicals were obtained from Sigma. Rapamycin, cytochalasin D and paclitaxel were purchased from Calbiochem. Media, antibiotics, and foetal calf serum (FCS) were purchased from Gibco/BRL, and enzymes from Promega. Test agents were dissolved in 100% ethanol vol/vol and then underwent serial dilutions in vehicle – Dulbecco's Modified Eagle Medium (DMEM) – to the desired concentration.

2.2. Animals and tissues

A total of 79 Large white/Landrace cross pigs weighing 29.6±3.9 kg were used. They received care in accordance with the Home Office Guidance on the operation of the Animals (Scientific Procedures) Act 1986, published by HMSO, London. The investigation conforms with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Surplus segments of human saphenous veins were obtained with informed consent from patients undergoing coronary artery bypass grafting. All procedures had local ethical approval.

2.3. In-vitro evaluation of anti-proliferative effects on pig or human saphenous-vein organ cultures

We used the organ culture method of Soyombo and colleagues, as previously described [23]. Briefly, pig or human saphenous veins from either leg were surgically exposed, and the side branches ligated. The veins were placed in sterile wash medium (20 mM Hepes-buffered RPMI 1640 supplemented with 2 mM l-Glutamine, 8 μg/ml gentamycin, 100 IU/ml penicillin, and 100 μg/ml streptomycin). The saphenous veins were then infused for 1 h, at ambient temperature (22–27 °C) with the same medium containing active treatments or vehicle. The veins were then opened longitudinally and cut transversely into 5–10 mm segments. Vein segments were cultured separately, endothelial surface uppermost, in culture medium for 14 days (DMEM containing Sodium pyruvate, Glucose (1000 mg/L), l-Glutamine (5 ml), 20% Foetal Calf Serum (FCS) and 100 IU/ml penicillin, and 100 μg/ml streptomycin). Culture medium was changed every 2 days. Twenty-four hours prior to fixation and embedding of the vein segments for histological analysis, the samples were suspended in culture medium containing 10 μM bromodeoxyuridine (BrdU).

2.4. Evaluation of anti-proliferative agents in porcine saphenous vein-grafts

Following induction with Ketaset (100 mg/ml ketamine hydrochloride), animals were intubated and anaesthetised with halothane and allowed to spontaneously ventilate.

The method of the saphenous vein-carotid interposition-grafting model has been described previously [24]. Approximately 10-cm of the long saphenous vein was harvested from the hind leg by a ‘no-touch’ technique. The vein was cannulated proximally and gently irrigated with heparinised iso-osmotic sodium chloride (NaCl) solution (0.9 g/l). It was divided into two equal sized lengths and bathed in active agent or the appropriate vehicle (DMEM) as described for the organ cultures above. They were then gently irrigated again with heparinised iso-osmotic sodium chloride (NaCl) solution (0.9 g/l). Each common carotid artery was exposed via a longitudinal neck incision. The animal was heparinised by intravenous administration of 100 IU/kg of heparin (CP Pharmaceuticals Ltd, Wrexham, UK). A 4–5 cm portion of carotid artery was isolated between vascular clamps and 10 mm was then removed. The residual ends were bevelled at 45°, and a small longitudinal incision made to lengthen the anastomotic area. The ends of the vein were similarly bevelled and anastomosed as an interposition graft under optical magnification using continuous 7/0 Surgipro (Auto Suture, Dagford, UK) sutures. Animals were recovered, returned to their pen and fed a normal chow diet for the duration of the experiment.

Animals were re-anaesthetised 1–12 weeks later, the grafts were exposed and patency (presence of blood flow) was assessed by transecting the common carotid artery distal to the graft. The grafts were excised and the animal was then euthanised with pentobarbitone. Only patent grafts were used for subsequent analyses (see below).

2.5. Histological methods

After organ culture, segments of pig or human saphenous veins were fixed in 4% Formalin in PBS and then prepared for wax embedding and sectioning into 5 μm transverse sections. Neointima formation was quantified from sections stained with Miller's elastic van Gieson (EVG) and Mayer's haematoxylin and eosin stains using an image analyser (Image-Pro Plus version 4, Media Cybernetics, L.P.) as described previously [25,26]. Neointimal area was quantified in 8 pairs of serial sections. The total number of neointimal cells and neointimal area, 14 days after paclitaxel or rapamycin exposure, were recorded and the mean cell count and cell density calculated for each vein segment. Proliferation was assessed by BrdU labelling, as described previously [26]. In short, paraffin-embedded sections were analysed immunocytochemically (ICC) for the presence of BrdU incorporation using a monocloncal anti-BrdU antibody (Europath, Bude, U.K.). The total number of BrdU-labelled cells was counted for each vein section; a ratio to non-labelled cells was generated by calculation of a total cell number from an averaged cell density (all cells counted in four fields (40× magnification) per vein section) and vein segment area.

Pig vein-grafts were pressure fixed at 100 mm Hg with 4% Formalin in PBS and then processed as for cultured veins (see above). Four transverse sections at equally spaced intervals along their length were stained with EVG. For each section, the neointimal luminal surface, internal elastic lamina, and external elastic lamina were identified and traced from digital images, and neointimal and medial areas were calculated. Vessel wall dimensions were measured by image-analysis (Image-Pro Plus version 4, Media Cybernetic, L.P.).

Cell proliferation was measured by PCNA ICC staining as previously described [27]. Briefly 5 μm sections were de-waxed, dehydrated, and treated with hydrogen peroxide in methanol to remove endogenous peroxidase. Sections were then microwaved in 10 mmol/L citrate buffer, pH 6, quenched in 1 in 3 horse serum in PBS for 30 min, and then incubated with mouse anti-human PCNA antibody diluted 1 in 75 overnight at 4 °C. Sections were then washed, incubated for 30 min with biotinylated goat anti-mouse IgG and then horseradish peroxidase-labeled Extravidin (diluted to 1:800 and 1:200, respectively) in 1% BSA in PBS. In all cases, visualization was performed with 0.05% (wt/vol) 3,3-diaminobenzidine and 0.3% (vol/vol) H2O2. Sections were counterstained with haematoxylin. The total numbers of cells positive for PCNA were counted in 4 fields at ×40 magnifications, which abutted the lumen and included the neointima and inner media. Five sections per graft are assessed. The number of PCNA-positive cells was expressed as a percentage of the total cell number (PCNA index).

Apoptosis was assessed by ICC staining for cleaved-Poly (ADP-Ribose) Polymerase (PARP) (Oncogene Research, Germany). Antigen retrieval was achieved by 15 min incubation with trypsin, following which the protocol for staining was identical to that used for PCNA. Alternatively, in situ end labelling (ISEL) was performed as described previously [26]. Briefly, sections were incubated at room temperature for 15 min with 5 μg/ml proteinase K in 1×TE (10 mm Tris HCl pH 8.0, 1 mm EDTA), washed twice in 1×TE and incubated in labelling mix (50 mm Tris HCl (pH 7.2), 10 mm MgSO4, 0.1 mm dithiothreitol, 0.01 mm dATP, dCTP, dGTP, 0.01 mm biotin-dUTP, and 8 units/ml of DNA polymerase I (Klenow; Promega, Southampton, UK) for 30 min at room temperature. Alternatively, terminal transferase dUTP nick end labelling (TUNEL) was determined over 3 h at 37 °C using the TdT-FragEL kit (Calbiochem catalogue number QIA33), according to the manufacturer's instructions. Irrespective of the labelling method, sections were rinsed in 1×TE, endogenous peroxidase was inhibited by incubation in 2% H2O2 for 5 min and then stained as per the PCNA protocol. Apoptotic activity was assessed using four, randomly selected, fields (×40 magnification) per section; a ratio of positive vs. negative staining cells was quantified in both neointima and media. Field areas were calculated thus enabling assessment of cell density, when combined with the total cell number per field.

Evaluation of endothelial coverage within grafts was achieved using an ICC stain for biotinylated dolichos biflorus agglutinin (DBA) (Vector Laboratories, Peterborough, U.K.). Antigen retrieval was facilitated by incubation in 3% hydrogen peroxide, at room temperature, for 5 min, the subsequent protocol was identical to that used for PCNA. Endothelial coverage was assessed in 4 sections per graft, using a score, and subsequently an average score was generated for each graft. The endothelial score was defined by the proportion of luminal coverage with endothelial cells staining positive (1=<1/3; 2=1/3–2/3; 3=>2/3; 4=100%).

2.6. Statistical analysis

Values are expressed throughout as the mean+/−SEM. Because occluded grafts were lost to follow up we conducted the statistical analysis in two ways. Firstly all patent grafts were compared to the controls conducted in the same series using analysis of variance followed by an unpaired Student's t-test corrected for repeated measures. Secondly, grafts from animals with one treated and one untreated graft were compared by a paired Student's t-test. Values were considered significant if p was less than 0.05. In no case did the type of statistical test influence whether p was less than or greater than 0.05 and so just one specified p value is reported.

3. Results

3.1. Experimental approach

We used preliminary experiments to find concentrations of agents that would inhibit proliferation without major effects on cell viability. To mimic a clinically useful protocol we pre-treated vein segments for 1 h before organ culture or implantation as interposition grafts in vivo. We used computerised morphometry to determine neointima, media and lumen cross-sectional areas; total wall area was computed as the sum of neointima and media areas. We also measured the perimeter of the lumen and the external elastic lamina and used these to compute notional lumen and total vessel areas based on circular profiles. This corrects for any squashing of the veins during processing and yields a coherent set of parameters in area units (e.g. Fig. 1). As expected, calculated lumen area was consistently slightly larger than measured lumen area (Fig. 1). Baseline values for pig saphenous veins prior to grafting were as follows; lumen area 2.5±1.2 mm2, neointima area 0 mm2, media area 1.2±0.9 mm2 and total vessel area 3.7+1.0 mm2.

Fig. 1

Effect of 1-hour ex-vivo treatment with cytochalasin D 2.5 μg/ml on porcine saphenous vein grafts 4 week after grafting.

3.2. Effects of cytochalasin D

Continuous local administration of cytochlasin D 0.1–10 μg/ml has been shown to inhibit proliferation but not viability of rabbit vascular smooth muscle cells in vivo, with effective tissue concentrations for the inhibition of isolated aortic ring contractions in vitro evident after 30 min exposures [28]. To verify that these concentrations were effective when used as a 1-hour pre-treatment, we exposed human saphenous vein segments to 0.25, 0.5, 1.0 and 2.5 μg/ml cytochalasin D for 1 h, and then cultured them for 14 days in the absence of cytochalasin D. Neointima thickness was reduced concentration-dependently from 26±6 μm to 14±6, 15±6, 13±5 and 5±2 μm, respectively. Only the effect of 2.5 μg/ml cytochalasin D was significant (p<0.05, n=6). Pre-treatment of pig saphenous veins with 2.5 μg/ml cytochalasin D reduced neointima cross-sectional area in 28-day vein-grafts by 22% but this was not significant (Fig. 1). There was no effect on medial or total wall dimensions, or on lumen or total vessel size. Based on these results cytochalasin D was not studied further.

3.3. Effects of paclitaxel

We showed previously that pre-treatment of isolated human arterial smooth muscle cells with paclitaxel concentrations of 0.1–10 μM inhibited proliferation without effects on viability. Higher concentrations of paclitaxel (100 μM) were, however, cytotoxic [19]. To confirm that 10 μM paclitaxel was appropriate for our in vivo studies, we pre-treated segments of pig saphenous veins for 1 h, a time feasible in the animal studies, and then cultured them in vitro for 14 days, the longest period we could maintain fully viable organ cultures [23]. Neointima thickness was reduced from 24±6 to 14±4 μm (p<0.004, n=8) and cell proliferation (BrdU index) from 11±2 to 4±1% (p<0.005, n=8). These results showed that pre-treatment with 10 μM paclitaxel had an effect on pig saphenous vein that was sustained for 2 weeks.

We conducted time-course studies with paclitaxel in the pig model in vivo (Fig. 2). As we showed previously [21], the most rapid period of medial growth occurs in the first week after grafting but neointima formation is mainly delayed until after the first week (Fig. 2a, b). The media and neointima (and hence total wall area) enlarge progressively between 1 and 12 weeks after grafting (Fig 2a, b). Pre-treatment of veins with 10 μM paclitaxel had no significant effect on media area in 1-week grafts but it did significantly decrease media SMC proliferation (measured as PCNA index) by 39% (Fig. 3A, Ba, b). No apoptosis of SMC was observed in 1-week control grafts using PARP cleavage or ISEL either with or without paclitaxel pre-treatment, despite excellent staining of positive controls (not shown). TUNEL staining also showed no apoptosis in 1-week control grafts control (Fig. 4Aa), although paclitaxel-pre-treated grafts showed 0.5±0.2% (n=4, p=0.11) of labelled cells (Fig. 4Ab). Endothelial coverage was an average of more than 75% in 1-week grafts and was not influenced by paclitaxel pre-treatment (Table 1, Fig. 4Ba, b). Of 4 graft occlusions observed in this group of experiments only 1 was in the paclitaxel-treated group.

Fig. 4

Effect of paclitaxel and rapamycin on apoptosis and endothelial coverage in 1-week porcine vein grafts. A. Apoptosis marked by TUNEL staining (brown, arrows) of 1-week vein-grafts pre-treated with a) and c) vehicle, b) paclitaxel (10 μM) and d) rapamycin 0.1 mg/ml. Counterstaining (blue) with haematoxylin. Original magnification ×40. B. The endothelium is stained light brown with DBA lectin in a) grafts pre-treated with vehicle b) 10 μM paclitaxel c) 0.1 mg/ml rapamycin or d) 0.5 mg/ml rapamycin where the endothelial layer is fragmented and incomplete. Counterstaining (blue) with haematoxylin. (Original magnification ×40).

Fig. 3

A. Effect of paclitaxel 10 μM and rapamycin 0.1 mg/ml on proliferation (PCNA index) in 1-week porcine saphenous vein grafts. *p<0.01 versus vehicle-treated control. B. PCNA staining (brown) of 1-week vein-grafts pre-treated with a) and b) vehicle; c) paclitaxel (10 μM) and d) rapamycin 0.1 mg/ml. Counterstaining (blue) with haematoxylin. Original magnification ×40.

Fig. 2

Effect of 1-hour ex-vivo treatment with 10 μM paclitaxel on porcine saphenous vein grafts 1, 4 and 12 weeks after grafting. Morphometric parameters are shown for paclitaxel pre-treated (circles, hatched lines) and vehicle pre-treated (squares, continuous lines) veins (n=5–8). *p<0.05 versus control at 4 weeks. † p<0.05 after 12 weeks versus 4 weeks.

View this table:
Table 1

DBA lectin scores as a marker of endothelial coverage

Time1 week4 weeks
TreatmentTreatedPaired controlTreatedPaired control
Paclitaxel 10 μM2.93±0.15 (5)2.67±0.24 (5)3.00±0.17 (8)2.92±0.14 (8)
Rapamycin 0.01 mg/mlNDND3.00±0 (4)3.00±0.12 (4)
Rapamycin 0.1 mg/ml3.33±0.11 (8)3.33+0.20 (8)3.26±0.1 (5)3.26±0.13 (5)
Rapamycin 0.5 mg/mlNDND2.90±0.05 (8)*3.1±0.10 (8)
  • Values are means±SEM with the number of animals in parentheses. *p<0.05 versus paired control.

Pre-treatment of pig saphenous veins with 10 μM paclitaxel reduced neointima area in 4-week vein-grafts by 43% and this was highly significant (p<0.001) (Fig. 2a). Medial area was unaffected (Fig. 2b) but total wall area was significantly reduced by 25% (Fig. 2c). There was no significant effect on measured or calculated lumen area or on total vessel area (Fig. 2d–f). This is important because it demonstrates that the reduction in wall thickness was not accompanied by outward remodelling of the vein-grafts. SMC proliferation was still measurable in 4-week grafts and there was no difference between control and paclitaxel treated grafts either in the neointima (PCNA index 29.6±6.1% control versus 32.9±6.0% paclitaxel, p=NS) or the media (38.0±7.2% control versus 38.7±4.9% paclitaxel p=NS). Apoptosis remained undetectable in 4-week grafts whether or not pre-treated with paclitaxel (results not shown). Endothelial coverage was no more complete after 4 weeks than 1 week and there was still no difference between control and paclitaxel-treated vessels (Table 1).

Not surprisingly given the persistence of SMC proliferation after 4 weeks, the neointima, media and total wall size increased significantly up to 12 weeks (Fig. 2a–c). Lumen and total vessel area also tended to increase progressively, although in this series of pigs differences between 4 and 12 weeks were mostly not significant (Fig. 2d–f). Pre-treatment of vein with 10 μM paclitaxel did not reduce the enlargement of the neointima, media and total wall area between 4 and 12 weeks (Fig. 2a–c). As a result there was no longer a significant effect of paclitaxel pre-treatment on any parameter measured in 12-week vein-grafts (Fig. 2). We concluded that the early benefit of paclitaxel was obliterated by subsequent neointimal and medial enlargement.

3.4. Effects of rapamycin

We were keen to establish whether these transient inhibitory effects were specific to paclitaxel and therefore conducted a similar series of studies with rapamycin. Rapamycin 0.1–0.2 mg/ml has been shown to inhibit neointima formation in mice vein grafts [29]. Since we were unsure about the optimum concentration for our studies we initially compared rapamycin pre-treatment at 0.01, 0.1 and 0.5 mg/ml on 4-week vein-grafts. At 0.5 mg/ml the solubility of rapamycin in vehicle (DMEM) was poor and achieved only after incubation at room temperature for over 1 h; this was therefore the highest feasible concentration. As shown in Fig. 5a, pre-treatment with 0.01 mg/ml rapamycin had no effect on neointima formation in 4-week grafts but 0.1 mg/ml rapamycin significantly reduced neointima formation by 65%. The effect of 0.5 mg/ml of rapamycin on neointima formation was no greater than that of 0.1 mg/ml. None of the concentrations of rapamycin had any effect on medial area (Fig. 5b) but 0.1 mg/ml rapamycin significantly decreased total wall cross-sectional area (Fig. 5c). Rapamycin 0.1 mg/ml also significantly increased lumen area in 4-week grafts (by 28% measured or 24% calculated, Fig 6d, e) but total vessel area was unaffected (Fig. 6f). Increasing concentrations of rapamycin had no effect on parameters measured in the contralateral control vein-grafts (Fig. 5a–c), which makes it unlikely that there were any effects of systemic washout from the pre-treated grafts.

Fig. 6

Effect of 1-hour ex-vivo treatment with 0.1 mg/ml rapamycin on porcine saphenous vein grafts 1, 4 and 12 weeks after grafting. Morphometric parameters are shown for rapamycin pre-treated (circles, hatched lines) and vehicle pre-treated (squares, continuous lines) veins (n=5–8). *p<0.05 versus control at 4 weeks. † p<0.05 1 after 12 weeks versus 4 weeks.

Fig. 5

Dose response to rapamycin in 4-week porcine saphenous vein grafts. Morphometric parameters are shown for rapamycin pre-treated (circles, hatched lines) and vehicle pre-treated (squares, continuous lines) veins (n=5–8). *p<0.02 versus contralateral vehicle-treated controls.

Apoptosis was undetectable in 4-week grafts treated with rapamycin but was present in the positive controls (results not show). Rapamycin 0.01 or 0.1 mg/ml had no effect on endothelial coverage (Table 1, Fig 4Bc) but coverage was slightly but significantly reduced in patent grafts treated with 0.5 mg/ml rapamycin (Table 1, Fig 4Bd). In addition when using 0.5 mg/ml rapamycin thrombosis was observed at harvest in 4 out of an attempted 12 grafts. Of these, 3 grafts had evidence of early neointima formation, which implies that thrombosis occurred late rather in the immediate postoperative period. Together with the evidence for reduced endothelial coverage, these data suggested that 0.5 mg/ml rapamycin increased the thrombogenicity of grafts.

Based on these results we conducted a time course study of the effects on 0.1 mg/ml rapamycin pre-treatment (Fig. 6). Rapamycin tended to reduce media area in 1-week grafts and significantly reduced PCNA index by 45% (n=6, p<0.005) (Fig. 3A, Bc, d) but had no effect on any other parameter in 1-week grafts (Fig. 6). No apoptosis of SMC was detected in 1-week grafts using ISEL staining either in controls or after 0.1 mg/ml rapamycin-treatment, despite good staining of the positive control (results not shown). However TUNEL staining was significantly increased by 0.1 mg/ml rapamycin-treatment from 0.3±0.1% in vehicle-treated controls to 1.5±0.3% (n=10, p<0.0002). Endothelial coverage in 1-week grafts showed no different between the groups (Table 1).

Both rapamycin 0.1 mg/ml and vehicle-treated control grafts demonstrated significant increases in neointima, media, total wall, lumen and vessel areas between 4 and 12 weeks (Fig 6a–f). As with paclitaxel, rapamycin pre-treatment did not prevent any of these increases, and the final areas were indistinguishable in control and rapamycin-treated 12-week grafts.

4. Discussion

4.1. Main findings

We investigated the effect of pre-treating veins with three structurally and mechanistically unrelated anti-proliferative pharmacological agents in a vein-grafting model characterised by similar vessel dimensions and time course to human aortocoronary bypass grafts. We investigated both short-term effects on initial neointimal growth, and longer-term effects to check for catch up. Our study is, therefore, the most thorough test so far of the anti-proliferative pre-treatment strategy in any model of vein-grafting. Importantly, we show for the first time that early inhibition of neointima and wall thickening is feasible with either paclitaxel or rapamycin but subsequent enlargement of the grafts completely abolishes the early benefits. These studies call into question the wisdom of pursuing such early pre-treatment strategies with anti-proliferative agents in human grafts.

Our data showed anti-proliferative effects of paclitaxel pre-treatment was sustained for at least 1 week in pig vein-grafts, which provides a rationale for the reduction in neointima formation and wall thickness seen in 4-week grafts. However, smooth muscle cell proliferation was still ongoing in 4-week grafts and we presented direct evidence that paclitaxel no longer exerted any anti-proliferative effect. Although we did not measure tissue drug levels, tissue half times for paclitaxel after ex vivo application in arteries are about 12 h [30], and it is likely that similar tissue retention characteristics apply to vein-grafts. It seems logical therefore to propose that paclitaxel concentrations had declined in 4-week vein-grafts to below anti-proliferative concentrations. Most importantly, the early anti-proliferative effect of paclitaxel did not result in a sustained reduction of neointima formation or wall thickening in vein-grafts between 4 and 12 weeks. A second series of experiments carried out with rapamycin showed the generally consistent baseline characteristics of this porcine model (Fig. 2 compared to Fig. 6). The biggest differences were seen in the extent of neointima formation (3.0±0.5 vs. 1.7±0.2 mm2) after 4 weeks, a time-point at which this variable is rapidly increasing. Given that the two series of experiments were carried out by different operators, minor differences in surgical skill could impact on the exact timing of neointimal growth. However, the neointimal size after 12 weeks appears consistent (7.5±1.3 vs. 8.2±0.6 mm2). The chosen concentration of rapamycin caused a similar inhibition of smooth muscle cell proliferation in 1-week grafts and neointima formation in 4-week grafts to the chosen concentration of paclitaxel. As with paclitaxel, the early inhibitory effects of rapamycin did not reduce the continued growth of the vein-grafts between 4 and 12 weeks. Hence although the data in Fig. 6 suggest that the early benefit of rapamycin was maintained, this was overwhelmed by subsequent growth so that the final dimensions of the grafts were indistinguishable from vehicle-treated controls.

4.2. Study strengths and limitations

The method of drug administration we used (immersion) would be feasible in the clinic but it exposes all layers of the vein wall and might therefore impair adventitial repair after grafting. However we did not notice any macro- or micro-scopic differences in the adventitia between vehicle and drug treated grafts. To test the ‘anti-proliferative pre-treatment’ concept, we sought to use concentrations of agents that reduced proliferation but were not toxic. In previous in vivo studies continuous infusions of high concentrations (10–100 μM) of cytochalasin D inhibited smooth muscle cell proliferation but not collar-induced neointima formation in rabbits [28]. Cytochalasin D inhibited neointima formation in a porcine drug-eluting stent model, where prolonged exposure was also insured [31]. However, we saw no significant effect of cytochalasin D in pig vein-grafts, despite an 80% inhibitory effect in organ cultures using the same pre-treatment regime. This suggests that cyctochalsin D was not sufficiently retained within the grafts in vivo to maintain anti-proliferative concentrations. We therefore abandoned further studies with this agent.

Previous studies demonstrated that 10 μM paclitaxel inhibits growth of cultured arterial smooth muscle cells by 80% [19] and that the sensitivity of venous smooth muscle cells is comparable or greater [32]. Short-term exposure to 10 μM paclitaxel has also been shown to preserve viability of smooth muscle cells [19] and vasomotor function in porcine coronary arteries [33]. However high concentrations of paclitaxel cause apoptosis in cultured smooth muscle cells [19] and tissue injury around drug-eluting stents [34]. We observed that pre-treatment with 10 μM paclitaxel caused a 40% reduction in neointima formation with pig saphenous veins in organ culture and a comparable 40% inhibition of smooth muscle cell proliferation in 1-week grafts. Hence the concentration of paclitaxel we used was sufficient to inhibit proliferation. Apoptosis was detectable in vein grafts only using the TUNEL assay, perhaps because of its greater sensitivity. Even then levels of apoptosis were less that 1% at any time point. 10 μM paclitaxel did not promote loss of endothelium or thrombosis of vein grafts; neither did it lead to outward remodelling (aneurysm-like dilatation). Hence there was little evidence for toxicity of this concentration of paclitaxel. Pretreatment with rapamycin 0.1 mg/ml produced comparable effects to 10 μM paclitaxel on neointima formation in organ culture and smooth muscle cell proliferation in 1-week grafts. Using the sensitive TUNEL assay a small increase in apoptosis (to 1.5%) was seen at the 1-week time point only. Rapamycin 0.01–0.1 mg/ml also had no effect on re-endothelialisation or graft patency, and it did not cause outward remodelling. In contrast, rapamycin 0.5 mg/ml appeared to inhibit re-endothelialisation and tended to promote thrombosis, although the numbers were too small to achieve statistical significance. The suggested toxicity at the higher doses may relate to the poor solubility of rapamycin 0.5 mg/ml with tissue precipitation producing very high local concentrations, and this may represent a failure of the delivery technique as opposed to a dose dependant effect per se. Nevertheless we can conclude that the concentrations of 10 μM paclitaxel and 0.1 mg/ml rapamycin were appropriate to test the efficacy of anti-proliferative treatments in this model. Whether toxic concentrations of paclitaxel or rapamycin would have produced more sustained effects on graft thickening in our model is unknown. For example the 50% reduction in intima formation observed with perivascular rapamycin administration in a previous study in mouse vein-grafts was associated with high levels of apoptosis [29].

A further question to address is the appropriateness of our pig model. The major differences with human vein-grafts are location, carotid versus coronary, the use of end-to-end rather than end-to-side anastomoses and the fact that the pig model has good run-off into an undiseased distal vascular bed. However the popular rabbit and mouse models share all these disadvantages and also use veins that are at least 10 times thinner than the human saphenous vein. The wall of the pig saphenous vein is approximately half the thickness of human saphenous veins and hence the mechanical loads during to the switch from venous to arterial pressure are more similar than in the rodent models. As a consequence, as we confirm here, pig vein-grafts respond without the initial wave of smooth muscle cell apoptosis seen, for example, in the most commonly used mouse model [35]. Wall thickening in all the vein-grafting models described takes place over at least 12 weeks, although many intervention studies have focussed only on early responses. We conclude that the pig model has distinct advantages for the study we conducted and that prolonging the experiments for at least 12 weeks was necessary to establish sustained effects.

Caution is necessary before extrapolating our findings into the clinical situation. Studies of longitudinal changes in vein-graft morphology in human and animals used largely different methodologies. In pigs we used histomorphometry in groups of animals sacrificed at defined intervals. Human studies used opportunistic autopsy or biopsy material or, in one study, serial intravascular ultrasound measurements. Despite these differences, increased intimal and total wall thickening is evident in both pig and human vein-grafts. Moreover increases in the total number of neointimal and medial smooth muscle cells have been observed. This implies that smooth muscle cell proliferation is a major mechanism underlying vein-graft thickening in man and is a realistic therapeutic target even though the precise kinetics in human grafts remain uncertain. Our experimental findings at least provide a rationale for the disappointing clinical experience with edifoligide [1]. Moreover our studies suggest that we should endeavour to develop methodologies to sustain release of antiproliferative agents, e.g. from a compound applied to the external surface of the vein [36], an external sleeve as a reservoir for slow drug release [37] or by using drug-eluting nanoparticles [38].


This study was supported by a grant from the British Heart Foundation. We thank Dr Ray Bush, Dr Jason Johnson and Dr Chris Rogers for expert advice in animal handling, histology and data analysis, respectively.


  • 1 These authors contributed equally.

  • Time for primary review 48 days


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