A

fagF(aaswrFt©

K

1

itmrt

o

A

0d

Atherosclerosis 197 (2008) 515–523

Inhibition of non-Ras protein farnesylationreduces in-stent restenosis

Paul Coats, Simon Kennedy 1, Susan Pyne,Cherry L. Wainwright 2, Roger M. Wadsworth ∗

Division of Physiology and Pharmacology, Strathclyde Institute of Pharmacy and Biomedical Sciences,University of Strathclyde, 27 Taylor Street, Glasgow G4 0NR, Scotland, UK

Received 28 March 2007; received in revised form 11 June 2007; accepted 19 June 2007Available online 30 July 2007

bstract

Ras has a key role in relation to cell proliferation, survival and migration and requires farnesylation for full activity. The effects of a Rasarnesyl transferase inhibitor, FPT III on human atherosclerotic vascular smooth muscle (VSM) cells proliferation and p42/p44 mitogen-ctivated protein kinase (p42/p44 MAPK) activity was measured. In addition the ability of FPT III to modify the development of neointimalrowth was tested in cultured human arteries and in a rabbit model of in-stent restenosis. In human VSM cells FPT III (25 �M) inhibitedCS-stimulated cell proliferation through a ras-dependent mechanism (after 18 h exposure) and also a novel ras-independent mechanismfollowing 15 min exposure). FPT III incubation (18 h) inhibited platelet-derived growth factor (PDGF)-stimulated p42/p44 MAPK activationnd p21 Ras membrane localization, whereas 15 min incubation had no effect on the activation of p42/p44 MAPK in response to PDGF (addedt 18 h) or on membrane p21 Ras localization (measured at 18 h). In cultured human atherosclerotic arteries, the presence of 25 �M FPT IIIignificantly reduced neointimal growth. In vivo, 15 min local infusion of 25 �M FPT III significantly reduced in-stent restenosis 28 days laterithout affecting vascular function in normal rabbit artery. This study demonstrates that brief administration of a farnesyl transferase inhibitor

educed in-stent restenosis in a rabbit model without deleterious effects on vascular function or endothelial regrowth. Acute application ofPT III was found to act through a novel mechanism to inhibit smooth muscle cell proliferation via a non-ras pathway, which may contribute

o the prevention of in-stent restenosis.2007 Elsevier Ireland Ltd. All rights reserved.

nsducti

dapahp

eywords: Coronary intervention; Coronary disease; Restenosis; Signal tra

. Introduction

Targeting the mitogenic signalling protein Ras by admin-stration of a dominant negative ras mutant has been showno inhibit neointimal hyperplasia in balloon injury animal

odels [1,2]. Members of the Ras subfamily have key

oles in relation to cell proliferation, survival and migra-ion [3] which are considered to be essential steps in the

∗ Corresponding author. Tel.: +44 141 548 2154; fax: +44 141 552 2562.E-mail address: r.m.wadsworth@strath.ac.uk (R.M. Wadsworth).

1 Present address: Institute of Biomedical and Life Sciences, Universityf Glasgow, Glasgow G12 8QQ, UK.2 Present address: School of Pharmacy, The Robert Gordon University,berdeen AB10 1FR, UK.

[

stiiibs

021-9150/$ – see front matter © 2007 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.atherosclerosis.2007.06.007

on; Stents

evelopment of neointimal hyperplasia. Expression of fullctivity in Ras subfamily members requires a series ofosttranslational modifications including covalent isoprenoidttachment [4–7]. Thus inhibition of Ras farnesyl transferaseas proved an effective approach to inhibit neointimal hyper-lasia following experimental angioplasty balloon dilation8,9].

Ras exists as a membrane associated pool with a relativelylow turnover (ca. 5–12 h) in cultured cells [10]. Howeverhe farnesyl transferase inhibitor FPT III [11] is effectiven vivo following a brief (15 min) application [8]. Since it

s unlikely that Ras would be depleted from the membranen a short period, other farnesylated signalling proteins maye important in the development of neointimal hyperpla-ia.

5 sclerosi

ciaahpih[sr

iprpwmslatieo[

2

2

Grtfapsi

nawd(22ctbwps

ceits

2

[VpstTFaouw1pb

2

wb1bwc1fnt5t(NSbR4mw1(i

2

16 P. Coats et al. / Athero

Coronary stent insertion is a highly effective treatment fororonary artery disease and has been greatly improved by thencorporation of antiproliferative compounds (e.g. paclitaxelnd sirolimus) to limit neointimal hyperplasia occurring asresponse to injury. Nevertheless these drug eluting stents

ave limited effectiveness in multivessel disease, in diabeticatients and in narrow vessel lesions [12]. Moreover theres significant concern that the drug eluting stents may delayealing [13] which may be linked to cases of late thrombosis14]. Thus there is a need for a new approach to resteno-is prevention that will be effective rapidly, allowing swiftecovery of the endothelium.

Although our previous studies demonstrated that localnhibition of farnesyl transferase reduced restenosis after sim-le balloon injury, their effect on human tissues and in-stentestenosis is currently untested. In this study we used humanrimary vascular smooth muscle cells derived from subjectsith ischaemic vascular disease to study the efficacy andechanisms through which FPT III acts. We also employed

hort-term local administration of FPT III immediately fol-owing stent implantation in the rabbit iliac artery in vivond addition to cultured human arteries in vitro to measurehe effect on neointimal area and composition. The rabbitliac model is particularly suited to stenting studies aimed atxamining neointima formation and normocholesterolaemicr diet-induced hypercholesterolaemic animals can be used15].

. Methods

.1. Human tissue culture and biochemistry

This study was approved by the University and thelasgow West Hospitals ethics committees. Each subject

ecruited to the study gave informed written consent prior toheir participation. The characteristics of the 8 patients are asollows: male:female = 3:5, age = 65 ± 3 years, plasma cre-tinine = 102 ± 9.2 �M, plasma cholesterol = 5.2 ± 0.2 mM,lasma glucose = 5.7 ± 0.3 mM, arterial blood pres-ure = 108 ± 7 /60 ± 4 mmHg, and drug therapy: ACEnhibitor 3, diuretic 4, nitrates 3, aspirin 5, and statin 3.

Tibial arteries were acquired from subjects clinically diag-osed with critical limb ischaemia who were undergoingmputation of the affected leg. The artery was removedithin theatre at the time of surgery. Arteries were imme-iately placed in ice cold sterile physiological saline solutionPSS; composition in mm—NaCl 119, KCl 4.5, NaHCO35, KH2PO4 1.0, MgSO47H2O 1.0, glucose 11.0, and CaCl2.5) and transferred to the vascular laboratory. Within thelass two culture hood the artery was cleaned of any adherentissue, then cut into 3–4 mm ring sections which were num-

ered in sequence. From each artery, three adjacent ringsere treated in parallel as fresh fixed, cultured or culturedlus FPT III. Tibial artery rings were also used for vascularmooth muscle explants. The vascular smooth muscle (VSM)

Cw

s 197 (2008) 515–523

ells were used for assay between passage 3 and 4. In somexperiments, human tibial artery rings were organ culturedn the presence or absence of FPT III (25 �M), after whichhey were fixed in formal saline, embedded in wax and 4 �mection cut using a microtome [8].

.2. 3H-thymidine cell proliferation assay

Cell proliferation was determined using serum-induced3H]-thymidine incorporation as described previously [8].SM cells were seeded onto the wells of a 24 well-culturelate in the presence of culture medium with 15% foetal calferum (FCS) and grown to 60–70% confluency. Cells werehen quiesced for 24 h in media containing 0.1% (v/v) FCS.he quiesced cells were stimulated for 24 h using 15% (v/v)CS with thymidine addition for the final 5 h. FPT III wasdded acutely at concentrations of 1–25 �M for 15–120 minr chronically from the beginning of the quiescing stagentil the end of the stimulation period. Finally, cells wereashed with ice cold phosphate buffered saline followed by0% trichloroacetic acid and then by lauryl sulphate (10%)lus sodium hydroxide (0.2 M) and radioactivity quantifiedy liquid scintillation.

.3. Western blotting

VSM cells close to confluency were quiesced, then treatedith FPT III either for 18 h or for 15–120 min followedy removal and replacement with quiescing medium alone8 h prior to stimulation with PDGF-AB (10 ng/ml), or phor-ol 12-myristate 13-acetate (PMA, 1 �M). The cells wereashed with ice cold PBS and lysed by addition of ice

old isotonic sucrose buffer (0.25 M sucrose, 1 mM EDTA,0 mM Tris (pH 7.4), 2 mM benzamidine, 0.1 mM PMSF)ollowed by repeated aspiration through an 18 gauge syringeeedle. Total well protein was measured using Bio-Rad pro-ein assay from a 10 �l aliquot (absorbance measured at95 nm). Samples were then diluted to a total protein con-ent of 1 �g/�l and mixed with sample buffer containingfinal concentrations) 0.125 M Tris–HCl (pH 6.7), 0.5 mMa4P2O7, 1.25 mM EDTA, 1.25% (v/v) glycerol, 0.5% (w/v)DS, 25 mM dithiothreitol and 0.02% (w/v) bromophenollue. For experiments involving membrane-associated p21as, lysates were centrifuged at 15,000 × g for 15 min at◦C to separate membrane and cytosolic fractions and theembrane fractions mixed with sample buffer. Samplesere Western blotted for total p42 MAPK (anti-p42 MAPK;:1000) and phospho-p42/p44 MAPK (anti phospho-p42/p44Tyr204/Thr202); 1:1000) or p21 Ras (anti-Ras; 1:400) byncubation overnight at 4 ◦C as described previously [16].

.4. Caspase 3 activity

Caspase 3 activity was measured using Promega’saspACETM colorimetric assay kit. VSM cells were treatedith either FPT III or staurosporine 10 �M as a positive

sclerosi

ccitiaa3

2

m1frHoatSdggaaasr(toolp

2

wiaht

2

sseiesat

sf

2

obfctcN2

2

ptCvtpbE

2

ttjtgctuics

3

3l

1di

P. Coats et al. / Athero

ontrol for up to 4 h and thereafter Caspase 3 activity wasompared with untreated cells (background control). Follow-ng treatment, cells were harvested and cell density adjustedo 106 cells/ml. A 20 �l aliquot of the cell suspension wasncubated with Caspase 3 substrate for 4 h at 37 ◦C. There-fter, absorbance in each well was measured on a plate readert 405 nm and plotted against a standard curve to give Caspasespecific activity.

.5. Stent implantation in vivo

All surgery and animal care conformed to the require-ents of the U.K. Animals (Scientific Procedures) Act

986. A previously characterised rabbit model was usedor stent implantation [17,18]. Male New Zealand Whiteabbits (2.5–3.0 kg) were premedicated with 1.2 ml i.m. ofypnorm® (fluanisone/fentanyl citrate mixture), 100 mg i.m.f ampicillin suspension, and 500 units of heparin i.v. Thenimals were maintained throughout the procedure on a mix-ure of 2% nitrous oxide and 1–1.5% isoflurane in oxygen.tents (2.5 mm diameter, BeStent 2, Medtronic) were intro-uced through the femoral artery using a 0.014 in. steerableuidewire and implanted in the iliac artery under fluoroscopicuidance. The stent was deployed by inflating the balloon topressure of 10 atmospheres for a period of 60 s then deflatednd withdrawn. Although it was not possible to performngiography, measurements of perfusion fixed iliac arteriesuggest an internal diameter of 2 mm, giving an artery:stentatio of 1:1.25. After stent deployment, an infusion balloon2.5 mm Scimed Dispatch®, Boston Scientific) was advancedo the stented region and used to deliver 25 �M FPT III (n = 7)r vehicle (0.9% saline containing 0.25% of DMSO; n = 10)ver a 15 min period. Total volume delivered was 5 ml. Fol-owing withdrawal of the catheter, the femoral artery wasermanently ligated and the wound closed and sutured.

.6. Rabbit artery removal and processing

Rabbits were sacrificed at 28 days and their iliac arteriesere removed. The entire stented artery was fixed overnight

n formal saline solution, embedded in resin and cut usingdiamond saw and grinder [19]. Sections were stained withaematoxylin and eosin and at least three sections (locatedhroughout the length of the stented section) were measured.

.7. Artery staining and analysis

Sections of the organ cultured human arteries and thetented rabbit arteries were digitally photographed and mea-ured with commercially available software (Image-Proxpress). The severity of injury induced by stent implantationn the rabbit arteries was gauged using the method of Gunn

t al. [20]. Each visible stent strut was assigned an injurycore from 0 to 4 dependent on the penetration and damage,nd the overall injury score was calculated as the average ofhe individual strut scores. All interpretation of strut injury

tmc1

s 197 (2008) 515–523 517

cores and histological staining was performed in a blindedashion by two independent observers.

.8. Vessel function study

Isometric function studies were used to assess the effectf FPT III incubation on the function of normal rab-it iliac arteries [21]. FPT III was incubated for 15 minollowed by washout and concentration–response curvesonstructed to KCl (5–80 mM) and phenylephrine (10−9

o 10−6 M), the endothelium-dependent vasodilator carba-hol (10−9 to 2 × 10−5 M) and the endothelium-independentO donor drug 3-morpholinosydnonimine (SIN-1, 10−9 to× 10−5 M).

.9. Materials

Unless otherwise stated, all drugs and chemicals wereurchased from Sigma–Aldrich, Dorset, UK. Farnesyl pro-ein transferase inhibitor III (FPT III) was obtained fromalbiochem, UK. FPT III was dissolved in water for initro experiments (for in vivo experiments the infusion solu-ion contained 0.25% DMSO). Cell culture materials wereurchased from Gibco BRL Life Technologies Ltd. The anti-odies for Western blotting were all obtained from Newngland Biolabs, UK.

.10. Statistics

All results are shown as mean ± S.E.M. The n used in sta-istical analyses and in graphs is the number of rabbits forhe in vivo studies and the number of different human sub-ects for cell studies. A Student’s unpaired t-test was usedo compare stented vessels in vehicle and FPT III-infusedroups. Two-way analysis of variance (ANOVA) was used toompare concentration–response curves for artery contrac-ion and relaxation studies. ANOVA or Student’s t-test wassed to compare the effect of FPT III in biochemical stud-es. Student’s t-test was used to compare effect of FPT III onultured vascular rings. p < 0.05 was taken to be indicative oftatistical significance.

. Results

.1. Effect of FPT III on cell proliferation after short orong exposure

Incubation of human vascular smooth muscle cells for8 h with FPT III (1–25 �M) resulted in a concentration-ependent inhibition of serum-induced [3H]-thymidinencorporation into DNA (Fig. 1A). Under these conditions

he IC50 was 9.6 ± 1 �M. Cell viability was checked by

easuring the protein content of the wells and was not signifi-antly different from control at any concentration (12.5 ± 2.8,8.2 ± 2.9, 19.1 ± 2.2, 16.0 ± 2.3, 15.3 ± 1.9 �g at 0, 1, 5, 10

518 P. Coats et al. / Atherosclerosi

Fig. 1. Both acute and chronic exposure to FPT III inhibited proliferation ofhuman vascular smooth muscle cells. (A) Concentration-dependent effect of18 h exposure to FPT III on FCS stimulated [3H]-thymidine incorporation(DPM = disintegrations per minute) in cultured human vascular smooth mus-cle cells. US = unstimulated cells. *p < 0.05, comparing 5, 10, and 25 �M vs.0 �M, one-way ANOVA with Dunnett multiple comparison test, n = 6. (B)Human vascular smooth muscle cells were exposed to FPT III (25 �M) forvarious times (15–120 min) prior to FCS stimulation, followed by measure-ment of [3H]-thymidine incorporation (DPM = disintegrations per minute).UID

atth35tbCgIw

3s

n(s

Pikbthctao

fsktap1oseceepo

bbari0r

3h

coc6nir

3s

d

S = unstimulated cells. MAX = FCS stimulation in the absence of FPTII. *p < 0.05 comparing each time point vs. MAX, one-way ANOVA withunnett multiple comparison test, n = 6.

nd 25 �M FPT III, respectively). Briefer periods of exposureo FPT III also significantly reduced serum-induced [3H]-hymidine incorporation into DNA. Thus acute exposure ofuman vascular smooth muscle cells to 25 �M FPT III for 15,0, 60, and 120 min reduced [3H]-thymidine incorporation by6.6 ± 3.6, 59.9 ± 3.1, 59.9 ± 2.9, and 64.4 ± 3.9%, respec-ively (all are p < 0.05 compared to the FCS-stimulated value,ut not significantly different from each other) (see Fig. 1B).hronic (18 h) treatment with FPT III produced significantlyreater inhibition when compared with acute exposure (FPTII 25 �M gave 76.2 ± 5.8% inhibition after 18 h, p < 0.05hen compared with acute exposure).

.2. Effect of FPT III on p42/p44 MAPK mitogenicignal transduction

Prolonged (18 h) exposure to FPT III (25 �M) sig-ificantly reduced basal p42/p44 MAPK phosphorylationFig. 2A). FPT III had a differential action on mitogen-timulated p42/p44 MAPK phosphorylation. The response to

d(dp

s 197 (2008) 515–523

DGF, which activates the Gi/ras pathway, was substantiallympaired by FPT III, whereas PMA, which acts via proteininase C, was unaffected (Fig. 2A). In contrast to the inhi-ition of p42/p44 MAPK phosphorylation found after 18 hreatment with FPT III, acute exposure (15 min) to FPT IIIad no effect upon p42/p44 MAPK phosphorylation whenells were stimulated (18 h later) for 10 min (Fig. 2B). Thushe inhibition of smooth muscle cell proliferation followingcute exposure to FPT III (Fig. 1B) occurred independentlyf p42/p44 MAP kinase activation.

Farnesylation of p21 Ras is necessary for its translocationrom the cytosol to the membrane, a required step for down-tream signaling including activation of the p42/p44 MAPinase cascade. Cell membrane p21 Ras levels were unal-ered after 6 or 18 h incubation in control cells (Fig. 3). Ancute 15 min exposure to FPT III 25 �M had no effect on21 Ras membrane localization, whether measured after 6 or8 h (Fig. 3). However, when FPT III was present through-ut the incubation period, membrane associated p21 Ras fellignificantly at 18 h (93 ± 3.4%, p < 0.01, Fig. 3). These West-rn blot data for membrane associated p21 Ras suggest thathronic exposure to FPT III in excess of 6 h is required tolicit a significant effect upon membrane associated Ras lev-ls. This may account for the differential effect of FPT III on42/p44 MAPK phosphorylation depending on the durationf exposure.

Caspase 3 activity of VSM cells was unaltered by incu-ation with FPT III (25 �M) for 1 or 4 h when compared toackgrounds. Basal absorbance at 405 nm was 0.029 ± 0.01,bsorbance at 1 and 4 h was 0.027 ± 0.02 and 0.053 ± 0.03,espectively (n = 3). Staurosporine elicited a significantncrease in Caspase 3 activity (absorbance = 1.13 ± 0.27 and.84 ± 0.01 at 1 and 4 h, respectively; n = 3, p < 0.05 stau-osporine versus background; paired Student’s t-test).

.3. Effect of FPT III on neointimal growth in cultureduman arteries

Rings of atherosclerotic human artery maintained in organulture for 28 days developed significant neointima. Inclusionf FPT III 25 �M in the organ culture medium signifi-antly reduced neo-intimal growth in arterial cultures by4.4 ± 3.7% when compared with controls (Fig. 4A, p < 0.05,= 6 pairs). Despite significant reduction in neointimal mass

n tissues cultured with FPT III for 28 days the endotheliumemained intact (Fig. 4B and C).

.4. Effect of FPT III on neointimal formation followingtent insertion in vivo

Average artery injury scores using the five-point scaleevised by Gunn et al. [20] revealed that animals in the

rug-infused group had a slightly increased injury scoreaverage was 1.26 ± 0.07 in control and 1.65 ± 0.01 inrug-infused; n = 10 for control and n = 7 for drug-treated;< 0.05). This represents a relatively modest degree of injury.

P. Coats et al. / Atherosclerosis 197 (2008) 515–523 519

Fig. 2. Differential effect of FPT III on PDGF and PMA stimulation of p42/p44 MAPK in human vascular smooth muscle cells. Samples were analysed bySDS-PAGE and Western blotted for phospho-p42/p44 MAPK. Subsequently blots were stripped and re-probed with pan-p42 MAP kinase antibody to ensureequal protein loading. Histograms show the ratio of phospho-p42/p44 MAPK to p42 MAPK (relative to that of quiesced control samples) determined fromd lms ares < 0.05w B (10 np 60 and

AncimI

imcstia

3

n

v1v(vFsptt

4

ensitometry measurements (mean + S.E.M., n = 6). Typical examples of fitimulation with PDGF-AB (10 ng/ml, 10 min) or PMA (1 �M, 10 min). *pith FPT III (+, 25 �M, 15 min) prior to stimulation for 18 h with PDGF-Aaired t-test. Similar results were obtained with exposure to FPT III for 30,

rtery size, measured as the circumference of the exter-al elastic lamina, was not significantly different betweenontrol (8.99 ± 0.14 mm) and drug-infused (9.13 ± 0.26 mm)liac arteries. Neointimal area, calculated as a percentage of

edial + neointimal area was significantly reduced in the FPTII group compared to the control group (Fig. 5A–C).

Endothelial regrowth was observed 28 days after injuryn all sections studied (Fig. 5D), with no difference in gross

orphology between the two treatment groups. Inflammatoryells were observed in the area immediately adjacent to thetent struts in some sections from both control and FPT III-reated groups (Fig. 5E). There was no apparent difference innflammatory cell involvement in control or FPT III-infusedrteries.

.5. Vessel function study

Short-term (15 min) incubation with FPT III (25 �M) inon-stented arteries had no effect on endothelium-dependent

tHtd

shown. (A) Cells were pretreated with FPT III (+, 25 �M, 18 h) prior to, +FPT III vs. −FPT III, Student’s paired t-test. (B) Cells were pretreatedg/mL) or PMA (1 �M, 10 min). FPT III had no significant effect, Student’s120 min.

ascular relaxation to carbachol (maximum relaxation was00 ± 0% in control versus 98.8 ± 0.8% in FPT III treatedessels) nor on endothelium-independent relaxation to SIN-198.3 ± 1.1% in control versus 99.4 ± 0.4% in FPT III treatedessels). Contractility in response to KCl was unchanged byPT III (maximum contraction was 5.9 ± 0.7 g in control ver-us 5.4 ± 0.6 g in FPT III treated vessels) but the response tohenylephrine was significantly reduced following exposureo FPT III (6.5 ± 0.8 g in control versus 4.6 ± 0.3 g in FPT IIIreated vessels; n = 6; p < 0.05).

. Discussion

A number of studies have demonstrated the clinical poten-

ial of the FTI group of drugs for proliferative diseases.owever the majority of these studies have focused on cul-

ured cell lines with a few studying models of neoplasticisease. Here we study the FTI, FPT III on intact human vas-

520 P. Coats et al. / Atherosclerosi

Fig. 3. Effect of FPT III on membrane associated p21 Ras in human vascularsmooth muscle cells. The histogram shows densitometric analysis of blotsfor membrane associated p21 Ras (mean + S.E.M., n = 3). A typical exampleof a blot is shown. FPT III (25 �M) was added for 15 min and washed outor was continually present, as indicated. BG = background in the absenceof any additions. **p < 0.01 (ANOVA), FPT III+ at 18 h vs. BG; **p < 0.01FPT III at 18 h vs. FPT III+ at 6 h.

caiat

citfwtoatuttet

wr

Fig. 4. FPT III treatment inhibits neo-intimal proliferation in organ culture. (A) Hudays with or without FPT III 25 �M. ••p < 0.01 fresh fixed tissue vs. FPT III treatt-test, n = 6 pairs. (B and C) Photomicrograph of human atherosclerotic arteries cu(magnification 15×). (D) Shows the presence of endothelial cells (EC) stained for f

s 197 (2008) 515–523

ular tissues and cells derived from the same tissues as wells the effect of in vivo administration on in-stent restenosisn rabbits. Where possible, we have chosen to use diseasedtherosclerotic arteries as such diseased tissues would be thearget of any potential clinical intervention.

The exact mechanism through which the FTIs preventellular proliferation [7] has not been established, howevernhibition of Ras farnesylation has gained foremost atten-ion [10]. The therapeutic application of FTIs has largelyocused on preventing cellular proliferation within tumourshere often mutations of the Ras protein are responsible for

he aberrant growth [22]. A previous study from this lab-ratory was the first to show the potential novel therapeuticpplication of FTIs in a vasculoproliferative condition, neoin-imal hyperplasia [8]. This present study has extended ournderstanding of this therapeutic strategy by demonstratinghat FTIs can also inhibit in-stent restenosis. Significantlyhe present study is the first, to our knowledge, to study thefficacy of FTIs on diseased human atherosclerotic vascular

issues.

A number of important findings were made. First, FPT IIIas shown to inhibit the proliferation of human atheroscle-

otic artery VSM cells and reduced neointimal hyperplasia in

man atherosclerotic arterial rings were maintained in organ culture for 28ed; **p < 0.01 control vs. FPT III treated and fresh fixed, Student’s pairedltured for 28 days B with FPT III and C in the absence of FPT III 25 �Mactor VIII by immumohistochemistry.

P. Coats et al. / Atherosclerosis 197 (2008) 515–523 521

Fig. 5. FPT III treatment inhibits neo-intimal proliferation in vivo. (A) Stents were deployed into the iliac arteries of rabbits with analysis 28 days later.Neointima was calculated as a percentage of the total neointima + media area to correct for any differences in artery size. Three sections from each artery wereanalysed and the mean value used. Infusion of FPT III (5 ml of 25 �M solution) at the time of stent implantation significantly reduced in-stent restenosis 28 dayslater (n = 7–10; **P < 0.005 vs. control). (B–E) Photomicrographs of iliac arteries removed from rabbits 28 days after stent insertion, stained with haematoxylinand eosin. L = lumen, NI = neointima, S = stent strut, I = inflammatory cells, and magnification = 400×. (B) Vehicle-infused control artery showing neointimaf dotheliuw ferenceF

htrwwItalpvpF

Cc

tfpbop

ormation. (C) FPT III-infused artery has much less neointima. (D) The enere frequently observed adjacent to the metal stent struts. There was no difPT III-infused animals.

uman arterial explants. Second, we established that short-erm in vivo application of FPT III is sufficient to significantlyeduce neointimal hyperplasia following stent deploymentith no detrimental effect upon endothelial regrowth. Finallye have shown using in vitro studies that exposure to FPT

II does not compromise endothelial function. Significantly,he mechanisms of inhibition of VSM cell proliferation uponcute exposure to FPT III appear to be independent of theevel of membrane-associated Ras and acute stimulation of

42/p44 MAPK. Since Caspase 3 was shown to be acti-ated at 1 h in the presence of a known activator of thisathway, but not by FPT III, the antiproliferative action ofPT III is highly unlikely to be attributed to apoptosis as

Mtio

m had regrown fully in all FPT III treated vessels. (E) Inflammatory cellsin the frequency or severity of inflammatory cell involvement in control or

aspase 3 is a key mediator of apoptosis of mammalianells.

Ras plays a fundamental role in cell signalling transduc-ion pathways that regulate cellular proliferation. Proteinarnesylation of Ras enables its association with thelasma membrane. Inhibition of farnesylation processeslocks membrane anchoring and consequent transductionf Ras-dependent proliferative mechanisms. Ras-dependentroliferative mechanisms are predominantly through the

AP kinase pathway, Raf/MEK/p42/p44 MAPK [23]. Ini-

ially FTIs were designed to block Ras farnesylation, howevert has been shown that FTIs can also inhibit prenylationf CENP-F [6] and lamin-B [22]. In this present study

5 sclerosi

whR

Ittaa1llnsRrecateiaastPMwasfadRt

r(tFctptto

cawAspse

tiwaofoeov

anpcitmilsrar

A

(

R

22 P. Coats et al. / Athero

e have clearly demonstrated that inhibition of neointimalyperplasia by FTI may involve both Ras-dependent andas-independent mechanisms.

Our data strongly supports the notion that the action of FPTII on human atherosclerotic smooth muscle cells involveswo distinct pathways. Previous data has shown that theurnover of the Ras membrane pool takes several hours [10]nd our present data shows that this is also true in humantherosclerotic artery smooth muscle cells, since a brief5 min treatment with FPT III inhibited FCS-induced pro-iferation but did not induce a change in either ras membraneocalisation or MAP kinase phosphorylation. This implies aon-ras target for this FTI. Moreover, prolonged 18 h expo-ure to FPT III almost abolished membrane localisation ofas and inhibited PDGF-induced p42/p44 MAPK phospho-

ylation. This implies that FPT III also exerts antiproliferativeffects through a Ras-dependent pathway. PDGF is a well-haracterised mitogen in smooth muscle acting through Rasnd p42/p44 MAPK. Thus these experiments demonstratehat depletion of Ras from the membrane prevents prolif-ration of human atherosclerotic smooth muscle cells. Thenhibition of cell proliferation by FPT III after the longer 18 hdministration was significantly greater than that observedfter only 15–120 min treatment with FPT III. The use ofpecific activators, PDGF (which acts through Ras, leadingo phosphorylation of p42/p44 MAP kinase (ERK 1/2)) andMA (which activates protein kinase C leading to p42/p44APK phosphorylation) suggests that both of these path-ays may be functionally important in human atherosclerotic

rtery smooth muscle cells. Neointimal hyperplasia is con-idered to be caused by the release of a mix of several growthactors within the artery wall consequent upon the injurynd inflammation caused by stent deployment. These con-itions are ideal for activation of both Ras-dependent andas-independent pathways, and both are likely to contribute

o neointimal hyperplasia and restenosis.We found that in vitro organ culture of human atheroscle-

otic artery rings with a prolonged application of FPT IIIpresumably acting through the ras pathway) inhibits neoin-imal hyperplasia. With the in vivo protocol the application ofPT III was restricted to a short 15 min interval. This was alsoapable of inhibiting neointima formation and may involvehe novel ras-independent pathway although a ras-dependentathway cannot be excluded. Taken together with the cul-ured VSM cell results, this data shows that both the Ras andhe non-ras pathways may be effective targets for inhibitionf neointimal hyperplasia

In vivo, evidence of inflammatory cell recruitment, espe-ially around stent struts was still apparent in FPT III-treatednimals, suggesting that a general anti-inflammatory effectas not responsible for reduced neointimal hyperplasia [24].lthough endothelial regrowth was unaffected 28 days after

tent implantation, it has recently been reported that FTIs mayrotect endothelial cells against apoptosis induced by oxidanttress [25]. Thus, in addition to preventing the vasculoprolif-rative response, FPT III may also protect endothelial cells,

s 197 (2008) 515–523

hus promoting rapid regrowth which would help to reducen-stent restenosis. Arteries removed at an earlier time point,here endothelial regrowth is still underway may help to

nswer this question. In vitro, FPT III also had little effectn contractile or relaxant function of normal iliac arteriesollowing 15 min incubation, which is in agreement withur previous studies in the pig coronary artery [8]. How-ver, a small reduction in contractility to phenylephrine wasbserved which may reduce the propensity of the injuredessel to undergo vasospasm in vivo.

In conclusion, acute FPT III treatment, which could bechieved with rapid release from a drug-eluting stent, inhibitseointimal hyperplasia, possibly via a non-ras pathway. Therecise mechanism of action remains to be determined butould include the inhibition of farnesylation of proteinsnvolved in cell division, e.g. the centromere associated pro-eins CENPE and CENPF [6]. This mechanism of action

ay have advantages in preventing restenosis whilst allow-ng endothelial regrowth and thereby protecting againstate thrombosis. Our data demonstrate that FPT III inhibitsmooth muscle cell proliferation without exerting a delete-ious effect on vessel function or endothelial regrowth ands such, may be a promising compound to treat in-stentestenosis.

cknowledgements

This work was supported by the British Heart FoundationPG/01/169) and by the University of Strathclyde R&D Fund.

eferences

[1] Indolfi C, Avvedimento EV, Rapacciuolo A, et al. Inhibition of cellularras prevents smooth muscle cell proliferation after vascular injury invivo. Nat Med 1995;1:541–5.

[2] Wu CH, Lin CS, Hung JS, et al. Inhibition of neointimal formation inporcine coronary artery by a Ras mutant. J Surg Res 2001;99:100–6.

[3] Ehrhardt A, Ehrhardt GRA, Guo X, Schrader J.W. Ras andrelatives—job sharing and networking keep an old family together.Exp Hematol 2002;30:1089–106.

[4] Lane KT, Beese L.S. Structural biology of protein farnesyltransferaseand geranylgeranlytransferase type 1. J Lipid Res 2006;47:681–99.

[5] Basso AD, Mirza A, Liu G, et al. The farnesyl transferase inhibitor(FTI) SCH66336 (lonafarnib) inhibits Rheb farnesylation and mTorsignalling. J Biol Chem 2005;280:31101–8.

[6] Hussein D, Taylor SS. Farnesylation of Cenp-F is required forG2/M progression and degradation after mitosis. J Cell Sci2002;115:3403–14.

[7] Basso AD, Kirschmeier P, Bishop WR. Farnesyl transferase inhibitors.J Lipid Res 2006;47:15–31.

[8] Work LM, McPhaden AR, Pyne NJ, et al. Short-term local deliveryof an inhibitor of Ras farnesyltransferase prevents neointima forma-tion in vivo after porcine coronary balloon angioplasty. Circulation

2001;104:1538–43.

[9] Chemla E, Castier Y, Julia P, et al. Inhibition of intimal hyperplasiaby an antisense oligonucleotide if farnesyl transferase delivered endo-luminally during iliac angioplasty in a rabbit model. Ann Vasc Surg1997;11:581–7.

sclerosi

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[by leukocyte depletion of neointima formation after balloon angioplasty

P. Coats et al. / Athero

10] Crul M, de Klerk GJ, Beijnen JH, Schellens JHM. Ras biochemistryand farnesyl transferase inhibitors: a literature survey. Anticancer Drugs2001;12:163–84.

11] Manne V, Ricca CS, Brown JG, et al. Ras farnesylation as a target fornovel antitumor agents: potent and selective farnesyl diphosphate ana-log inhibitors of farnesyltransferase. Drug Dev Res 1995;34:121–37.

12] Flaherty JD, Davidson CJ. Diabetes and coronary revascularisation.JAMA 2005;293:1501–8.

13] Togni M, Windecker S, Cocchia R, et al. Sirolimus-eluting stents asso-ciated with paradoxic coronaray vasoconstriction. J Am Coll Cardiol2005;46:231–6.

14] Iakovou I, Schmidt T, Bonizzoni E, et al. Incidence, predictors andoutcomes of thrombosis after successful implantation of drug-elutingstents. JAMA 2005;293:2126–30.

15] Schwartz RS, Edelman ER. Drug-eluting stents in preclinical stud-ies: recommended evaluation from a Consensus Group. Circulation2002;106:1867–73.

16] Sambi BS, Hains MD, Waters CM, et al. The effect of RGS12 onPDGFbeta receptor signalling to p42/p44 mitogen activated protein

kinase in mammalian cells. Cell Signal 2006;18:971–81.

17] Kennedy S, McPhaden AR, Wadsworth RM, Wainwright CL. Corre-lation of leukocyte adhesiveness, adhesion molecule expression andleukocyte-induced contraction following balloon angioplasty. Br JPharmacol 2000;130:95–108.

[

s 197 (2008) 515–523 523

18] Van Belle E, Maillard L, Tio FO, Isner JM. Accelerated endothelial-ization by local delivery of recombinant human vascular endothelialgrowth factor reduces in-stent intimal formation. Biochem BiophysRes Commun 1997;235:311–6.

19] Malik N, Gunn J, Holt CM, et al. Intravascular stents: a new tech-nique for tissue processing for histology, immunohistochemistry, andtransmission electron microscopy. Heart 1998;80:509–16.

20] Gunn J, Arnold N, Chan KH, et al. Coronary artery stretch versus deepinjury in the development of in-stent neointima. Heart 2002;88:401–5.

21] Kennedy S, Wadsworth RM, Wainwright CL. Effect of antiprolifeta-tive agents on vascular function in normal and in vitro balloon injuredporcine coronary arteries. Eur J Pharmacol 2003;481:101–7.

22] End DW, Smets G, Todd AV, et al. Characterization of the antitumoreffects of the selective farnesyl protein transferase inhibitor R115777in vivo and in vitro. Cancer Res 2001;61:131–7.

23] Vojtek AB, Der CJ. Increasing complexity of the Ras signaling pathway.J Biol Chem 1998;273:19925–8.

24] Miller AM, McPhaden AR, Wadsworth RM, Wainwright CL. Inhibition

in a rabbit model of restenosis. Cardiovasc Res 2001;49:838–50.25] Cuda G, Paterno R, Ceravolo R, et al. Protection of human endothelial

cells from oxidative stress—role of Rasd-ERK1/2 signalling. Circula-tion 2002;105:968–74.