European Journal of Pharmaceutical Sciences 37 (2009) 351–362

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences

journa l homepage: www.e lsev ier .com/ locate /e jps

Novel farnesylthiosalicylate (FTS)-eluting composite structures

Amir Kraitzera, Yoel Kloogb, Meital Zilbermana,∗

a Department of Biomedical Engineering, Faculty of Engineering, Tel-Aviv University, Tel-Aviv 69978, Israelb Department of Neurobiochemistry, Faculty of Life Sciences, Tel-Aviv University, Tel-Aviv 69978, Israel

a r t i c l e i n f o

Article history:Received 18 November 2008Received in revised form 19 February 2009Accepted 9 March 2009Available online 21 March 2009

Keywords:Farnesylthiosalicylate (FTS)Drug-eluting stentsDrug-eluting fibersStent coatingsDegradation

a b s t r a c t

Farnesylthiosalicylate (FTS) is a new specific nontoxic drug with a mild hydrophobic nature, which acts asa Ras antagonist and can therefore be used for stent applications as well as for local cancer treatment. FTS-loaded bioresorbable core/shell fiber structures were developed and studied in order to investigate theFTS release mechanism. These structures were composed of a polyglyconate core and a porous poly(d,l-lactic-glycolic acid) shell loaded with FTS, prepared using freeze drying of inverted emulsions. The effectsof the emulsion’s composition (formulation) and process kinetics on the FTS release from the coatingswere studied with reference to the shell morphology and degradation profile. The FTS release profilesexhibited a burst effect accompanied by a release rate which decreased with time and lasted for 15–40days. The process was found to affect the drug release profile via two routes: (1) Direct, through wateruptake and swelling of the structure, leading to a FTS burst release. Degradation of the host polymer affectsthe FTS release rate at a later stage. (2) Indirect effect of the microstructure on the release profile, which

occurs via an emulsion stability mechanism. The copolymer composition is the most important parameteraffecting the release behavior in our system. Other parameters, including polymer content, O:A phase ratioand homogenization rate exhibited only minor effects on the FTS release profile. The controlled releaseof the new drug FTS is reported here for the first time.

1

oaLpw2sotJictT(a(

ap

trscwsgasso

t

0d

. Introduction

Drug-eluting stents (DES) significantly reduce the incidencef in-stent restenosis (ISR), which was once considered a majordverse outcome of percutanous coronary stent implantation.ocalized release of antiproliferative drugs interferes with theathological proliferation of vascular smooth muscle cells (VSMC)hich is the main cause of in-stent restenosis (Heldman et al.,

001). Drug uptake into the vessel wall occurs by passive diffu-ion and convection and is facilitated by the hydrophobic naturef these antiproliferative drugs that establish substantial parti-ioning and spatial gradients across the tissue (Creel et al., 2000;ackson et al., 2004). In the USA, use of drug-eluting stents hasncreased from 19.7% at the time of their approval to 78.2% of per-utaneous procedures at the end of 2004 (Rao et al., 2006). Today,he most extensively studied DES are the commercially available

AXUSTM (Boston Scientific, paclitaxel-eluting stent) and CypherTM

Cordis, J&J, sirolimus-eluting stent). However, drug-eluting stentsre associated with increased rates of late stent thrombosis (LST)Lagerqvist et al., 2007) and hypersensitivity reactions (Virmani et

∗ Corresponding author. Tel.: +972 3 6405842; fax: +972 3 6407939.E-mail address: meitalz@eng.tau.ac.il (M. Zilberman).

saio2s2T

928-0987/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.ejps.2009.03.004

© 2009 Elsevier B.V. All rights reserved.

l., 2004), which are serious low-frequency life-threatening com-lications.

Three factors affect the efficacy and safety of current DES:he stent platform, the drug release matrix, and the type of drugeleased. Cypher and Taxus are based on a 316 L stainless steeltrut stent platform. The coating of the Cypher and Taxus stents isomposed of durable dense polymers (Moreno and Macaya, 2005),hich may trigger hypersensitivity reactions as described in a case

tudy by Virmani et al. (2004). The stent coating should also haveood mechanical properties in terms of flexibility and long-lastingdherence to the stent surface (Rogers, 2004), especially when thetent is expanded to the required size during surgery. This expan-ion can seriously affect the release of the drug from the polymer,r worse, can embolize the polymer (Finkelstein et al., 2003).

The drug release kinetics of current DES is far from optimal inerms of safety and efficacy, despite the accumulated knowledgeupporting its importance. The hydrophobic nature of the releasedntiproliferative drugs and the non-degradable nature of the coat-ng matrix along with the small thickness of current DES coatings

ffer relatively poor control over the release period (Wang et al.,006) and a relatively low drug load. Cypher contains 70–300 �girolimus (140 �g/mm2) and 80% of the drug is released within8 days after stent implantation (Venkatraman and Boey, 2007).axus contains 50–200 �g (1 �g/mm2) paclitaxel, where ∼2 �g

352 A. Kraitzer et al. / European Journal of Pharm

a9eairdappoaamebk2

SccsrfPp((ort

nviKatmteFa2eaieFls

taia

cmbn9msbbto

blwsuthwbwwefictcfiqctf

2

2

dTb

o

•

•

C

2.1.1. Reagents1,1,1,3,3,3-Hexafluoro-2-propanol (H1008) was purchased from

Scheme 1. The chemical structure of farnesylthiosalicylate (FTS, Salirasib).

re released within 15 days (Venkatraman and Boey, 2007) and2.5% of the drug remains in the matrix for a long period (Serruyst al., 2005). The drug release kinetics from these stents may beltered via the drug/polymer ratio (Halkin and Stone, 2004), coat-ng thickness, and/or coating a drug-free top layer onto the drugeservoir layer, in order to create the slow-release formulation asemonstrated for the Cypher sirolimus-eluting stent (Venkatramannd Boey, 2007). Drachman et al. (2000) studied bioresorbableoly(lactide-co-�-caprolactone) coated metal stents that releaseaclitaxel and reported that the in vitro release profile followed firstrder kinetics with almost 90% of the drug being released withinbout 2 months, and a burst effect of about 36%. Finkelstein etl. (2003) used alternating layers of 50/50 PLGA and PLLA insideetal struts to control the paclitaxel release profile and burst

ffect. More recent studies changed the hydrophobic/hydrophilicalance in the polymer matrices in order to affect the releaseinetics from the stents (Breitenbach et al., 2000; Westedt et al.,006).

The released drug also affects the efficacy and safety of DES.irolimus is a nonspecific immunosuppressive agent that inhibitsell cycle progression at the early stage (leading to reversion of theell to the quiescent state) (Venkatraman and Boey, 2007). Thus,irolimus-coated DES prevent not only VSMC hyperplasia but alsoe-endothelialization (Li et al., 2006), or may cause endothelial dys-unction (Hofma et al., 2006) on the inner side of the metal stent.aclitaxel is a microtubule stabilizing agent that inhibits cell cyclerogression at the late stage (leading to cell death and apoptosis)Venkatraman and Boey, 2007), and has a low therapeutic indexFonseca et al., 2002). It is therefore very effective in the treatmentf neointimal hyperplasia, which is known as the main cause ofestenosis (Feng and Huang, 2001). However, it may also be veryoxic.

Farnesylthiosalicylate (FTS, Salirasib) is a new rather specificontoxic drug which was recently developed at the Tel-Aviv Uni-ersity (Marom et al., 1995). Its chemical structure is presentedn Scheme 1. FTS acts as a Ras antagonist (George et al., 2004;loog and Cox, 2004) and has a mild hydrophobic nature. In itsctive form (GTP-bound), Ras promotes enhanced cell prolifera-ion, tumor cell resistance to drug-induced cell death, enhanced

igration and invasion. Ras is therefore considered an importantarget for cancer therapy as well as for therapy of other prolif-ration diseases, including restenosis. The apparent selectivity ofTS towards active (GTP-bound) Ras and the absence of toxic ordverse side effects was proven in animal models (George et al.,004) and in humans (Concordia Pharmaceuticals, Inc., Ft. Laud-rdale, FL). In the rat carotid artery injury model, which serves asmodel for restenosis, FTS was found to be a potent inhibitor of

ntimal thickening while not interfering with endothelial prolif-ration (George et al., 2004). The incorporation of the new drug,TS, in a stent coating may overcome the incomplete healing andack of endothelial coverage associated with current drug-elutingtents.

Current drug-eluting biodegradable or biostable stent coatings

hat contain antiproliferative drugs exhibit serious side effects andre consequently far from optimal in terms of their drug load-ng and release profiles. These coatings cannot carry sufficientmounts of drug because of the trade-off between the mechani-

S

H

aceutical Sciences 37 (2009) 351–362

al properties and drug loading. Completely biodegradable stentsay eliminate late complications of stent implantation simply

y degrading into nontoxic substances after maintaining lumi-al integrity during the period of high risk restenosis in the first–12 months (Commandeur et al., 2006). Biodegradable stentsay also eliminate current DES endothelial related limitations and

uggest a larger drug reservoir. Although both drug-eluting andioresorbable stents have long been studied, there is to date noiodegradable or stable drug release device that can provide con-rolled release of drugs within the therapeutic dose and safe healingf the tissue.

We have recently developed and studied paclitaxel-elutingioresorbable core/shell fiber structures for stent applications and

ocal cancer treatment (Kraitzer et al., 2008). These structuresere composed of a polyglyconate dense core and a porous PDLGA

hell loaded with the antiproliferative agent paclitaxel, preparedsing freeze drying of inverted emulsions. Paclitaxel’s release fromhese porous structures was relatively slow due to its extremelyydrophobic nature. In the current study such composite fibersere loaded with the new drug, FTS. We hypothesized that the com-ination of the highly adjustable porous shell structure togetherith the relatively less hydrophobic nature of the FTS moleculesill afford release profiles more suitable for stents. The drug’s

ffectiveness and safety may also increase the suitability of thesebers as basic elements of drug-eluting stents. The FTS-elutingoating (shell) can also be applied on metal stents. The objec-ive of the current study was to characterize the new FTS-elutingore/shell composite fiber structures in terms of the FTS release pro-le, degradation, morphology and tensile mechanical properties. Aualitative model for the FTS release mechanism is suggested andompared to the paclitaxel release model. It is important to notehat the controlled release of the new drug, FTS, is reported hereor the first time.

. Materials and methods

.1. Materials

MaxonTM polyglyconate monofilament (3-0) sutures, with aiameter of 0.20–0.25 mm, Syneture, USA were used as core fibers.his polymer contains 67% glycolic acid and 33% trimethylene car-onate.

Bioresorbable porous structures (the shell coating) were madef the following polymers:

75/25 poly(d,l-lactic-co-glycolic acid), inherent viscosity(i.v.) = 0.65 dl/g (in CHCl3 at 30 ◦C, approximately 97,100 g/mole),Absorbable Polymer Technologies Inc., USA. This polymer istermed 75/25 PDLGA.50/50 poly(d,l-lactic-co-glycolic acid), inherent viscosity(i.v.) = 0.56 dl/g (in CHCl3 at 30 ◦C, approximately 31,300 g/mole),Absorbable Polymer Technologies Inc., USA. This polymer istermed 50/50 PDLGA.

The drug farnesylthiosalicylate (FTS, Salirasib) was a gift fromoncordia Pharmaceuticals (Ft. Lauderdale, FL).

pectrum Chemical Mfg. Corp.Chloroform, CHCl3, HPLC grade and methylene chloride, CH2Cl2,

PLC grade were purchased from Frutarom, Israel.Acetonitrile, CH3CN, HPLC grade was purchased from J.T. Baker.

harm

2

2

sfi1Tp

2

tDafrifta2aprp

2

piiaifis1otT

2

(6vva

mpmggS

uawtws

2

ptlitwaiist

2

3a1oMtAwctorm

JUiup(rwa1i

2

pm6tt(tsau

i

A. Kraitzer et al. / European Journal of P

.2. Preparation of core/shell fiber structures

.2.1. Fiber surface treatmentThe sutures were surface-treated in order to enhance the adhe-

ion between the core fiber and the coating. The polyglyconatebers were slightly stretched on special holders and dipped in,1,1,3,3,3-hexafluoro-2-propanol (hexafluorisopropanol) for 40 s.he fibers were then washed with ethanol and dried at room tem-erature.

.2.2. Emulsion formationA known amount of 50/50 PDLGA was dissolved in chloroform

o form an organic solution and FTS was added to the solution.ouble-distilled water was then poured into the organic phase (inscintillation vial) and homogenization of the emulsion was per-

ormed using a homogenizer (Polytron PT3100 Kinematica, 12 mmotor) operating at 16,500 rpm (medium rate) for 2 min, for mostnvestigated samples. As a reference sample we chose an emulsionormulation containing 12.5% (w/v) polymer in the organic solu-ion, 2% (w/w) FTS (relative to the polymer load), and an organic toqueous (O:A) phase ratio of 4:1 (v/v). Other formulations included0% (w/v) polymer, 1% and 4% (w/w) FTS, O:A phase ratios of 2:1nd 8:1 and copolymer composition of 75/25 PDLGA. Certain sam-les were prepared using homogenization rates of 8500 rpm (lowate) or 22,500 rpm (high rate) in order to investigate the effect ofrocessing kinetics on the porous shell structure.

.2.3. Core/shell fiber structure formationThe treated core polyglyconate fibers were dip-coated (while

laced on holders) in fresh emulsions and then frozen immediatelyn a liquid nitrogen bath. The holders + samples were then placedn a pre-cooled (−105 ◦C) freeze dryer (Virtis 101 equipped with

nitrogen trap) capable of working with organic solvents (freez-ng temperature of the condenser was approximately −105 ◦C) andreeze dried in order to preserve the temporal state of the emulsionn a solid form. The shell’s microstructure thus reflects the emul-ion’s stability. The freeze dryer chamber pressure was reduced to00 mTorr while the temperature remained constant (−105 ◦C) inrder to sublimate the water and solvents. Room temperature washen slowly restored in order to evaporate residual solvent vapors.he samples were then stored in desiccators until use.

.3. Morphological characterization

The morphology of the composite core/shell fiber structurescryogenically fractured surfaces) was observed using a Jeol JSM-300 scanning electron microscope (SEM) using an acceleratingoltage of 5 kV. The SEM samples were Au sputtered prior to obser-ation. The mean pore diameter of the observed morphologies wasnalyzed using the Sigma Scan Pro software.

All quantitative measurements were summarized aseans ± standard deviations, unless otherwise specified. Com-

arisons of the mean pore diameter were carried out for theicrostructure analysis using the unpaired Student’s t-test for two

roup comparisons, or ANOVA (post hoc Tukey-Kramer) for threeroup comparisons. SPSS was used for all statistical calculations.tatistical significance was determined at p < 0.05.

In vitro morphology of wet shell structures was characterized

sing an environmental SEM (Quanta 200 FEG ESEM) using anccelerating voltage of 10 kV in a pressure of 4.5 Torr. Fractographsere taken at three time points while the specimens were main-

ained in double-distilled water outside the ESEM. The specimensere fixed to a special base, and initial coordinates were recorded

o as to enable good return to these coordinates.

iemtpc

aceutical Sciences 37 (2009) 351–362 353

.4. In vitro FTS release studies

The composite core/shell fiber structures were immersed inhosphate buffered saline (PBS) at 37 ◦C and pH = 7.4 for 35 days, inriplicates, in order to determine the release kinetics from the FTS-oaded composite structures. The release studies were conductedn closed glass tubes containing 3 ml PBS medium, using a horizon-al bath shaker operated at a constant rate of 130 rpm. The mediumas removed (completely) periodically, at certain sampling times,

nd measured as described in Section 2.5. Fresh medium was thenntroduced. At the end of the experiment the fibers were immersedn methylene chloride and the residual amount of drug was mea-ured. This experiment enabled simulating conditions similar tohose which exist in a blood vessel.

.5. FTS extraction procedure

FTS extraction from the medium was performed as follows: theml PBS/FTS medium was completely removed at each time pointnd placed in a scintillation vial. Three milliliters acetonitrile andml methylene chloride were added and methylene chloride evap-ration was performed under a nitrogen stream (99.999% grade).edium (50/50, v/v acetonitrile/PBS) was added up to 4 ml in each

est tube. The FTS concentration was then estimated using HPLC.n extraction factor was used for correction. Known weights of FTSere dissolved in 3 ml acetonitrile and 3 ml PBS and 1 ml methylene

hloride was added. The known concentrations were subjected tohe same extraction procedure as the unknown concentrations inrder to determine the efficiency of the extraction procedure. Theecovery efficiency of the method was 88.4% and the value of theeasured drug was corrected accordingly.The FTS content of the medium samples was determined using

asco High Performance Liquid Chromatography (HPLC) with aV 2075 plus detector and a reverse phase column (ACE 5 C18,

nner diameter = 4.6 mm, length = 250 mm), equipped with a col-mn guard, and was kept at room temperature (25 ◦C). The mobilehase consisted of a mixture of acetonitrile and phosphate buffer30 mM, pH = 4.5) at a ratio of 70/30 (v/v), respectively, at a flowate of 1 ml/min with a quaternary gradient pump (PU 2089 plus)ithout gradient. Hundred microliter samples were injected with

n autosampler (AS 2057 Plus). The column effluent was eluted for5 min and detected at 322 nm. The area of each eluted peak wasntegrated using the EZstart software version 3.1.7.

.6. Residual drug recovery from the composite fibers

On the final day of the in vitro trial, residual FTS from the com-osite fibers was measured as follows: the fibers were placed in 1 mlethylene chloride for 10 min and the coating shell was dissolved.ml of a 50/50 acetonitrile/water solution were then added and

he polyglyconate core was removed. Methylene chloride evapora-ion was performed under a nitrogen (99.999%) stream. Medium50/50, v/v acetonitrile/water) was added until 4 ml in each testube and the FTS concentration was estimated by HPLC using theame method as above. An overall calibration curve for both HPLCnd method recovery was calculated using known amounts of FTSnder the same conditions.

The cumulative release profiles were determined relative to thenitial amount of FTS in the composite fibers (quantity released dur-ng the incubation period + the residue remaining in the fibers). All

xperiments were performed in triplicates. Results are presented aseans ± standard deviations. The effects of the emulsion’s formula-

ion on the release profile were studied by examining the followingarameters: polymer content in the organic phase (%, w/v), drugontent relative to polymer content (%, w/w), organic:aqueous

3 Pharm

(cs

2

ttramtatbtop

E

2

1ewop

atbiwuts

2

rwea1tso4a

mwa2u2ladS

Wt1m

2

asaciwo

TaE(

N

2

dsasasipai

W

2d

(wsot(wTtbsd

stm

54 A. Kraitzer et al. / European Journal of

O:A) phase ratio and copolymer composition. The effect of the pro-ess kinetics (homogenization rate) on the release profile was alsotudied.

.7. Encapsulation efficiency calculations

The encapsulation efficiency (EE) of the fibers was calculated ashe actual amount (Ma) of drug measured in each fiber divided byhe theoretical amount of drug (Mt) encapsulated during the fab-ication process, presented in percentage as shown in Eq. (1). Thectual amount of drug encapsulated within each fiber is the accu-ulated amount of FTS released at each measurement point of the

rial plus the residual amount measured on day 35. The theoreticalmount of FTS is the formulation drug concentration multiplied byhe weight of the coating. The weight of the coating is the differenceetween the coated fiber (weighed at the beginning of the in vitrorial) and the bare fiber (weighed at the end of the trial, denudedf the coating by immersion in methylene chloride). The results areresented as means ± standard deviations (n = 3).

E = Ma

Mt× 100 (1)

.8. FTS chemical stability

Two milligrams of FTS were placed in 3 ml PBS at 37 ◦C for00 days in order to determine its stability throughout the releasexperiment. The closed scintillation vials containing the FTS in PBSere placed in a horizontal bath shaker operated at a constant rate

f 130 rpm. The FTS content of a single vial was measured at eachoint of time.

FTS was extracted as follows: 3 ml methylene chloride weredded to the 3 ml PBS/FTS mixture and stirred for 10 min in ordero dissolve the drug. One milliliter was carefully removed from theottom of the vial (the organic phase) using a pipette and placed

n a new vial. Five milliliters of 50/50 acetonitrile/double-distilledater (the medium) were added to the vial and then evaporatednder nitrogen conditions. The content of each vial was transferredo a test tube and diluted to 20 ml. The FTS content of the mediumamples was determined using HPLC, as described above.

.9. In vitro degradation of the porous PDLGA structure

Porous 50/50 PDLGA and 75/25 PDLGA film structures were fab-icated using the emulsion formulation of the reference sample butithout drug. The inverted emulsion was prepared as described

arlier, poured into an aluminum plate, quenched in liquid nitrogen,nd freeze dried. Each sample (three repetitions), approximatelycm2, was incubated in 40 ml phosphate buffered saline (PBS) con-

aining 0.05% (v/v) sodium azide (as preservative) at 37 ◦C undertatic conditions for 5 weeks. PBS was added when the pH was outf range (between 7 and 8) or when the PBS volume dropped below0 ml. The samples were taken out at weekly intervals, dried usingvacuum oven (35 ◦C for 2 h), and stored in a dessicator.

The weight-average molecular weight of the samples was deter-ined by gel permeation chromatography (GPC). The dried sampleas dissolved in methylene chloride before elution to achieveminimal concentration of about 0.13% (w/v). The GPC (Waters

1515 isocratic pump, operating temperature 40 ◦C using a col-mn oven) was equipped with a refractive index detector (Waters

414, operating temperature 40 ◦C) and calibrated with poly-l-actic acid MW kit standards (Polysciences Inc., USA). Data werenalyzed using the Breeze version 3.3 software. The samples wereissolved in methylene chloride, filtered, and eluted through 4tyragel columns (model WAT044234 HR1 THF, WAT044237 HR2,

wr1sI

aceutical Sciences 37 (2009) 351–362

AT044225 HR4, WAT054460 HR5, 300 mm × 7.8 mm, 5 �m par-icle diameter) equipped with a guard column at a flow rate ofml/min. The elution medium was Baker analyzed HPLC gradeethylene chloride.

.10. In vitro weight loss profile of the porous PDLGA structure

Porous 50/50 PDLGA and 75/25 film structures were fabricateds described earlier, in the paragraph about the in vitro degradationtudy. Each film sample was cut into pieces of approximately 1 cm2

nd then incubated in 15 ml PBS at 37 ◦C and pH = 7.4 under staticonditions. Samples (three repetitions) were taken out at weeklyntervals, filtered using a 70 mm porcelain Büchner funnel equipped

ith a Whatman size 2 �m filtration paper and dried in a vacuumven (35 ◦C for 2 h).

Mass loss was measured using a Mettler-Toledo microbalance.he normalized mass loss was calculated by comparing the masst a given time point (wt) with the initial mass (w0) as shown inq. (2). The results are presented as means ± standard deviationsn = 3).

ormalized weight (%) = wt

w0× 100 (2)

.11. Measurements of water uptake

Porous 50/50 PDLGA film structures were fabricated asescribed earlier, in the paragraph about the in vitro degradationtudy. Each film sample was cut into pieces of approximately 1 cm2

nd then incubated in 15 ml double-distilled water at 37 ◦C undertatic conditions. Samples (triplicates) were taken out periodicallynd immediately subjected to measurement of wet weight, afterurface water was removed with a clean-wipe tissue. Water uptake,.e. adsorption and absorption of each sample during the swellingeriod, was determined according to Eq. (3) (w is the wet weightt each time point and w0 is the dry weight measured before thencubation):

ater uptake (%) = w − w0

w0× 100 (3)

.12. Measurements of tensile mechanical properties and theiregradation

Core shell fiber structures were fabricated as described abovein the paragraph about core/shell fiber structure formation butithout drug). The fibers’ tensile mechanical properties were mea-

ured at room temperature, under unidirectional tension at a ratef 50 mm/min, using a 5544 Instron uniaxel machine. Three fiberypes (Maxon sutures, surface-treated fibers and coated fiberswithout the drug), n = 5 for each sample, 14 cm in length) wererapped around a thick paper and inserted between the jigs.

he tensile strength was defined as the maximum strength inhe stress–strain curve. The maximal strain was defined as thereaking strain. Young’s modulus was defined as the slope of thetress–strain curve in the elastic (linear) region. Means ± standardeviations are presented.

In order to evaluate the degradation of mechanical properties,amples were immersed in 14 ml PBS filled tubes for 84 days. Eachube contained five specimens of 14 cm coated fibers. The pH was

aintained between 7.3 and 7.5 and the medium was changed

hen the pH was out of the range. Each week a single tube was

etrieved and the fibers were dried using a vacuum oven (35 ◦C for.5 h) and kept in a dessicator. Each specimen’s diameter was mea-ured using a caliper and a tension test was carried out using annstron machine as described above.

A. Kraitzer et al. / European Journal of Pharmaceutical Sciences 37 (2009) 351–362 355

Table 1The examined specimens’ shell structural characteristics and encapsulation efficiency.

Process parameters Mean pore size ± SD (nm) Mean porosity ± SD (%) Mean EE ± SD (%)

Reference a 2.89 ± 1.07 84.24 ± 4.54 56.75 ± 17.93

Pofymer content (%, w/v) 7.5 2.71 ± 0.65** 83 ± 1.94** 44.54 ± 10.720 4.75 ± 1.23*,** 74.8 ± 1.55*,** 59.25 ± 2.83

Organic to aqueousphase ratio (v/v)

8:1 3.01 ± 0.78** 78.07 ± 5.9** 47.49 ± 242:1 2.64 ± 0.62** 86.42 ± 1.8** 62.06 ± 13.42

Homogenization rate(rpm)

8500 6.24 ± 2.33* 74.8 ± 3.23** 61.96 ± 10.4122,500 2.24 ± 0.47* 84.19 ± 4.32** 69.96 ± 10.89

PDLGA copolymer composition (w/w) 75/25 4.53 ± 1.36* 76.25 ± 2.27* 51.83 ± 4.35

*

sed on 50/50 PDLGA, 17.5% (w/v) polymer, 2% (w/w) FTS, O:A = 4:1 (v/v), homogenizationr

3

tradTtrtFdsw

3

redppiotafhtpt((

2mrboppcwata

Significant compared to the reference.** Significant compared to another specimen, which is not the reference.a The reference sample was prepared using the following conditions: emulsion ba

ate = 16,500 rpm.

. Results and discussion

Farnesylthiosalicylate (FTS, Salirasib) is a new targeted non-oxic drug which can be used for prevention of restenosis wheneleased from a vascular stent as well as for treatment of cancer,s already mentioned in Section 1. In our composite fibers, theense core enables obtaining the desired mechanical properties.he drug is located in a porous shell and should therefore not affecthe mechanical properties. The shell is highly porous so as to enableelease of relatively hydrophobic drugs in a desired manner. In ordero characterize our FTS-eluting core/shell fiber platform, we studiedTS release from the fibers, and the shells’ morphology and degra-ation profiles. Since mechanical properties are important for thetent application, degradation of the tensile properties of the fibersas also studied.

.1. Drug release and degradation

The freeze drying technique that was used for the shells’ prepa-ation is unique in being able to preserve the temporal state of themulsion in a solid form. We used this technique in order to pro-uce inverted emulsions in which the continuous phase containedolymer and drug dissolved in a solvent, with water being the dis-ersed phase. In the current study we investigated the effects of the

nverted emulsion’s parameters, i.e. polymer content, drug content,rganic:aqueous (O:A) phase ratio and copolymer composition onhe shells’ microstructure and on the drug release profile. FTS is

relatively hydrophobic drug. Therefore, most of the emulsionormulations we chose were based on 50/50 PDLGA—the poly(�ydroxyl acid) with the highest degradation rate in order to be ableo release it at an appropriate rate. Based on our experience withaclitaxel-loaded composite fibers, we chose an emulsion formula-ion containing 12.5% (w/v) 50/50 PDLGA in the organic solution, 2%w/w) FTS (relative to the polymer load), and an organic to aqueousO:A) phase ratio of 4:1 (v/v) as the reference sample.

The diameter of the treated core fibers was in the range of00–250 �m and a shell thickness of 30–70 �m was obtained forost emulsion formulations. The shell’s porous structure contained

ound-shaped pores in all the specimens that were based on sta-le emulsions, usually within the 2–7 �m range, with a porosityf 75–92%. The shell microstructure was uniform in each sample,robably due to rapid quenching of the emulsion, which enabledreservation of its microstructure. The pores were partially inter-

onnected by smaller inner pores. The emulsions were stable onlyithin a certain formulation range. The encapsulation efficiency of

ll studied samples was in the 40–70% range and all specimens con-ained 33.3 ± 4.8 mg FTS. The structural characteristics of the shellnd encapsulation efficiency values of all examined specimens are

Fig. 1. The effect of copolymer composition on: (a) FTS cumulative release profilefrom the fibers, (b) degradation profile (expressed as log 10 weight-average molecu-lar weight) of the porous shell, (c) weight loss profile (expressed as wt/w0) of porousshell. ( )—50/50 PDLGA 50/50 and ( )—75/25 PDLGA.

356 A. Kraitzer et al. / European Journal of Pharmaceutical Sciences 37 (2009) 351–362

F ing th(

sto1i

3

tofFeuFarwi

nttPtirt2tdtcrt

ig. 2. SEM fractographs of 50/50 PDLGA and 75/25 PDLGA shell structures showc—50/50 PDLGA, d—75/25 PDLGA); day 14 (e—50/50 PDLGA, f—75/25 PDLGA).

ummarized in Table 1. As mentioned above, FTS is a new drug. Weested its chemical stability using HPLC and found no minor peaksther than the FTS peak nor any changes in elution time even after00 days in an aqueous medium at 37 ◦C, indicating that the drugs stable under these conditions.

.2. Effect of copolymer composition

The effect of the copolymer composition, i.e. the relative quan-ities of lactic acid (LA) and glycolic acid (GA) in the copolymer,n the drug release profile and on the shell microstructure wasound to be the most pronounced of all parameters tested. TheTS release profile from a shell based on 50/50 PDLGA (the ref-rence sample) and from a shell based on 75/25 PDLGA preparedsing similar formulation and process parameters are presented in

ig. 1a. Both release profiles exhibited a burst effect accompanied byrelease rate which decreased with time. The 50/50 PDLGA sample

eleased 62% of the encapsulated drug during the first day of release,hereas the 75/25 PDLGA sample released only 30%. This difference

s attributed mainly to differences in the hydrophilic/hydrophobic

Pstlp

eir microstructural features at: day 0 (a—50/50 PDLGA, b—75/25 PDLGA); day 7

ature of these two copolymers. The 50/50 PDLGA copolymer con-ains more glycolic acid groups and fewer lactic acid groups alonghe polymer chain and is therefore less hydrophobic than the 75/25DLGA and probably exhibits higher water uptake during the ini-ial phase of release. Consequently, this enables more rapid waternflow which results in a higher burst release. Furthermore, theate of release from the 50/50 PDLGA formulation is slightly higherhan the rate obtained with the 75/25 PDLGA formulation and after

weeks of degradation the 50/50 formulation released 100% ofhe drug, whereas the 75/25 formulation released only 79%. Thisifference is attributed to difference in the degradation rate ofhese two copolymers, which assists the drug’s diffusion. The 50/50opolymer degrades faster than the 75/25 copolymer and thereforeeleases the drug at a faster rate. The degradation profiles of thewo copolymers are presented in Fig. 1b. The slope of the 50/50

DLGA’s log 10 of the weight-average molecular weight is indeedignificantly higher than that of the 75/25 PDLGA. It should be notedhat both polymers exhibited an exponential decrease in molecu-ar weight with time. An exponential decay degradation profile oforous PDLGAs was also reported by Wu and Ding (2004) and Lu

A. Kraitzer et al. / European Journal of Pharmaceutical Sciences 37 (2009) 351–362 357

F ral fea

eew5nTp

tis57et

suososiisf

ig. 3. (a) ESEM fractograph of a wet 50/50 PDLGA sample showing its microstructu

t al. (2000). The weight loss (erosion) profiles of both polymersxhibited a small weight loss of less than 10% during the first 3eeks of degradation, whereas after 3 weeks of degradation the0/50 PDLGA exhibited a fast weight loss while the 75/25 PDLGA didot erode during the measured time period (Fig. 1c), as expected.hese results indicate that most of the FTS is released from ourorous coatings before they undergo massive weight loss.

The changes of the shell structures of both copolymers duringhe first 2 weeks of exposure to the aqueous medium are presentedn Fig. 2. The starting point (day 0) shows highly porous delicate

tructures with round pores for both samples. The pore size of the0/50 PDLGA shell (Fig. 2a) is significantly lower than that of the5/25 PDLGA (Fig. 2b) and its porosity is higher (Table 1), whichnables more surface area for diffusion. After 7 days of degrada-ion in the aqueous medium the 50/50 PDLGA exhibited a rougher

uila

tures at three time points. (b) The water uptake profile of the 50/50 PDLGA sample.

tructure (Fig. 2c). Most pores disappeared, probably due to waterptake and swelling. At this point of time the 75/25 PDLGA showednly slight changes in morphology compared to the 50/50 PDLGAample (Fig. 2d), probably due to a lower water uptake. After 14 daysf degradation the 50/50 PDLGA exhibited a totally rough poroustructure, while at the same time the 75/25 PDLGA began exhibit-ng a rough structure but the porous features remained (Fig. 2f). Its therefore suggested that structural changes of our highly poroustructures resulting from water uptake affect the FTS release profilerom the coated fibers.

The shells’ structural changes were further investigated by ESEMsing a special technique which enabled us to take consecutive

mages at different in vitro time points at approximately the sameocations. Observations of the 50/50 PDLGA samples indicated

decrease of approximately 20% in the film’s thickness during

3 Pharmaceutical Sciences 37 (2009) 351–362

ta(pfsvwsopus4bsoiar

caawram

3

demhriaAoafcwsdii

3

tfscadppmtoi

Fig. 4. The effect of process parameters on the cumulative release profile of FTSf)))

oiad

3

dsrr

58 A. Kraitzer et al. / European Journal of

he first week of exposure to the aqueous medium, followed byn increase in the film’s thickness during the following 2 weeksFig. 3a). We hypothesized that although water penetrates theorous structure, it shrinks during the first week due to the changerom a porous to a bulk structure and then expands, probably due towelling. It should be noted that the porous structure gradually con-erts from a porous to a dense structure and the round pores vanishhile the dimensions of the polymer walls increase. The dimen-

ions of polymers usually increase due to swelling. However, mostf our structure contains voids rather than polymer. Thus, as theolymer swells, pores vanish, and the overall thickness decreasesntil reaching a bulk structure and then increases due to furtherwelling. The water uptake measurements indicate an increase of0% in the sample’s weight during the first 48 h, after which it sta-ilizes (Fig. 3b). This further supports our hypothesis that the earlytructural changes are due to water uptake rather than degradationr erosion. Since the drug diffusion through the fiber’s shell occursn an aqueous swollen phase, a relatively high water uptake, suchs that of our 50/50 PDLGA porous shell, enables faster diffusionate of the drug molecules.

It can be concluded that a higher glycolic acid content in theopolymer, i.e. a less hydrophobic copolymer, enables more surfacerea for diffusion and higher FTS release from the fibers due tocombination of early swelling followed by degradation. Higherater uptake affects the microstructure and results in a higher burst

elease and a higher degradation rate of the host polymer, whichssists diffusion. The contribution of the swelling effect is probablyore significant in our system.

.3. Effect of polymer content

The effect of the polymer content of the organic phase on therug release from the porous shell is presented in Fig. 4a. Burstffects of 75%, 62% and 44% were obtained for 50/50 PDLGA for-ulations containing low (7.5%, w/v), medium (12.5%, w/v) and

igh (20%, w/v) polymer contents, respectively. All formulationseleased most of the encapsulated drug within 15 days. Pore sizencreased with polymer content while porosity decreased (Fig. 5and b and Table 1), resulting in less overall surface area for diffusion.s more polymer is incorporated, a more hydrophobic emulsion isbtained, leading to emulsion instability which results in largerqueous domains that eventually convert into large pores afterreeze drying. The decrease in burst effect by increased polymerontent may also have resulted due to denser “polymeric walls” thatere created between adjacent pores which slows down the diffu-

ion rate. Some specific interactions may also exist between therug and the host polymer. A relatively high polymer content thus

ncreases the interactions between the polymer and FTS, resultingn a lower diffusion coefficient which results in slower drug release.

.4. Effect of the O:A phase ratio

The effect of the O:A phase ratio on the FTS release profile fromhe composite fibers is presented in Fig. 4b. The release profilerom samples derived from emulsions with an O:A = 2:1 is veryimilar to the profile derived from emulsions with an O:A = 4:1. Inontradistinction, an increase in the O:A ratio to 8:1 resulted indecrease in the burst release and prolonged FTS release for 30

ays. These phenomena are attributed to changes in the shell mor-hology. The relatively high O:A phase ratio of 8:1 (or low aqueous

hase content) enables lower porosity (Table 1) and larger poly-eric domains between pores (Fig. 5c), which create a barrier to

he diffusion of drug molecules. An increased O:A phase ratio inur FTS-eluting systems is therefore an effective way for decreas-ng the burst release and increasing the release period. The porosity

dr6Ts

rom core/shell fiber structures: (a) effect of polymer content (( )—7.5%, w/v; (—12.5%, w/v; ( )—20%, w/v), (b) effect of O: A phase ratio (( )—8:1, v/v; (—4:1, v/v; ( )—2:1, v/v), (c) effect of homogenization rate (( )—8500 rpm; (—16,500 rpm).

f samples derived from emulsions with O:A ratios higher than 4:1s not high enough to enable effective release of the water-insolublegents in systems that contain more hydrophobic antiproliferativerugs, such as paclitaxel (Kraitzer et al., 2008).

.5. Effect of process kinetics

The effects of the emulsion’s homogenization rate (prepareduring a 180 s homogenization) on the drug release from the poroushell and on the shell structure are presented in Figs. 4c and 5d,espectively. The homogenization rate had some effect on the FTSelease profile due to changes in the shell’s porous structure. A

ecrease in homogenization rate from 16,500 rpm to 8500 rpmesulted in a significant increase in mean pore size from 2.89 �m to.24 �m with a decrease in porosity from 84.24% to 74.8% (Table 1).he low surface area for diffusion of the low homogenization ratepecimen enabled some decrease in the burst release.

A. Kraitzer et al. / European Journal of Pharmaceutical Sciences 37 (2009) 351–362 359

F ell micF h 20%(

3p

sAmdadartta

actep

pctd

ig. 5. SEM fractographs showing the effect of the processing parameters on the shTS, phase ratio of 4:1 and homogenization rate of 16,500 rpm, (b) formulation wit8500 rpm).

.6. Qualitative model of FTS-eluting systems and comparison toaclitaxel-eluting systems

Our results show that the most important parameter in ourystem affecting release behavior is the copolymer composition.n increase in the glycolic acid content of the PDLGA copoly-er resulted in an increase in the burst effect and release rate

uring the first 2 weeks, mainly due to higher water uptakend changes in microstructure, but also due to a higher degra-ation rate of the host polymer (Figs. 1–3). Other parameters

ffecting the FTS release profile are polymer content, O:A phaseatio, surfactants and homogenization rate. However, the effects ofhese parameters are less prominent in the tested specimens. Inhese cases the microstructure determines the surface area avail-ble for diffusion, which affects the drug release rate. In general,

b

Fig. 6. Schematic representation of the qualitative model describing the dependence

rostructure: (a) the reference specimen containing 12.5% (w/v) polymer, 2% (w/w)(w/v) polymer, (c) formulation with O:A ratio of 8:1, (d) low homogenization rate

higher diffusion rate can be achieved when high porosity isombined with a fine structure of lower pore size. The drug con-ent, whose effect was also studied, but not presented, did notxhibit any effect on either the shell microstructure or the releaserofile.

A qualitative model describing the process → structure →roperty (drug release profile) effects in our FTS-eluting systeman be described as follows (Fig. 6): there are two routes by whichhe process affects the drug release profile: direct and indirect. Therug release is controlled by diffusion, which may be accelerated

y early and late mechanisms:

Direct (process → release profile): the emulsion formulation (espe-cially the host polymer) may affect the water uptake and swellingof the structure and therefore also the FTS burst release. Degra-

of the FTS release profile on the emulsion’s formulation and process kinetics.

360 A. Kraitzer et al. / European Journal of Pharmaceutical Sciences 37 (2009) 351–362

Table 2The tensile mechanical properties of the FTS-eluting fibers.

Core polyglyconate fiber Surface-treated fiber Core/shell structurea

Tensile strength (MPa) 288 ± 32 271 ± 21 268 ± 48YM

mfd

Pdfwrmm

toem2ipiwffelptifi

3

aitotewTti

pfiacemOur drug-eluting core/shell fiber structures can therefore be usedfor biomedical applications which require good initial mechanicalproperties, such as stents for blood vessel support.

The core/shell fiber samples were immersed in aqueous mediumat 37 ◦C in order to evaluate the deterioration of the tensile mechan-

oung’s modulus (GPa) 1.424 ± 0.085aximal strain (%) 37 ± 5

a The effective diameter (of the treated core fiber) was considered.

dation of the host polymer affects the FTS release rate at a laterstage.Indirect (process → structure → release profile): the process effect onthe microstructure occurs via an emulsion stability mechanism.The emulsion formulation affects the stability much more thanthe process parameters (i.e. homogenization rate). The emulsionstability determines the surface area through the microstructure,e.g. the surface area increases when porosity is high and pore sizeis low. These enables enhanced diffusion rate of the drug from theporous shell and affect both the burst release and later the release.

Diffusion of the drug molecules is thus controlled chemically,ainly by the water uptake and swelling (affected by the emulsion

ormulation) and also geometrically, through the surface area. Theirect effect is more significant than the indirect effect.

It is important to note that most formulations based on 50/50DLGA released the FTS molecules within 2 weeks and that theifferences between their release profiles resulted mainly from dif-erences in the burst release. Formulations based on host polymersith a higher lactic acid content should be used if FTS release is

equired for a longer period of time. Furthermore, a higher poly-er content and O:A phase ratio or a lower homogenization rateay change the release profile to a lesser extent.Our experience with paclitaxel-eluting composite fiber struc-

ures (Kraitzer et al., 2008) shows that the drug release behaviorf this system is different from that of the currently reported FTS-luting system. Paclitaxel is more hydrophobic than FTS and createsore specific interactions with the host polyesters (Ranade et al.,

004). Therefore, paclitaxel’s diffusion through the host polymers much slower and massive degradation and erosion of the hostolymer must occur in order to enable it. Consequently, changes

n formulation parameters affect its release profile mainly after 10eeks of degradation. In paclitaxel-eluting systems the emulsion

ormulation influences the diffusion by producing binding regionsor the drug as more hydrophobic materials are introduced into themulsion, thus delaying the drug molecules. The emulsion’s formu-ation also significantly affects the stability of the emulsion duringrocessing, and as a result determines the freeze dried microstruc-ure. In general, the release profile of FTS from our composite fiberss faster than the paclitaxel release and more adjustable, and there-ore more suitable for the stent application. Furthermore, since FTSs less toxic, some burst release may be tolerated.

.7. Tensile mechanical properties and their degradation

The polyglyconate monofilament sutures were surface-treated,s described in Section 2.2.1, in order to dispose of the fiber’s orig-nal coating and enhance the adhesion between the core fiber andhe coating. The tensile strength, modulus and ultimate strengthf the studied samples are presented in Table 2. It should be notedhat in practice, the highly porous shell cannot carry the load. The

ffective diameter (of the treated core fiber) was therefore usedhen evaluating the mechanical properties of the core/shell fibers.

he tensile strength, Young’s modulus and ultimate strain of thereated and core/shell (coated) fibers are similar to those of the orig-nal polyglyconate fibers. There is some decrease in all three tensile

Fdcs

1.418 ± 0.03 1.353 ± 0.1732 ± 1 32.7 ± 3

roperties, probably due to cracks and imperfections caused at theber surface as a result of the treatment. However, the changesre not significant (Table 2). This small deterioration in mechani-al properties can be avoided in the future if the carrying fiber isxtruded specifically for this application instead of using a com-ercial suture material, evoking the need for surface treatment.

ig. 7. Mechanical properties of the composite core/shell fibers as a function ofegradation time in PBS medium at 37 ◦C (cross-section was calculated using theore’s effective diameter): (a) tensile strength, (b) Young’s modulus, (c) maximaltrain.

harm

iacadtwoaca

4

dnoawttpwowmosfpw

sTelttrtcrmpiebo

taFt

A

gv

eSv

Do

R

B

C

C

D

F

F

F

G

H

H

H

J

K

K

L

L

L

M

M

R

R

R

S

A. Kraitzer et al. / European Journal of P

cal properties. The changes in tensile strength, Young’s modulusnd ultimate tensile strain with time are presented in Fig. 7a, b and, respectively. Our results show that the composite fibers exhibiteddecrease of approximately 50% in both modulus and tensile strainuring 10 weeks of degradation, whereas a decrease of 50% in theensile strength was exhibited after 8 weeks of degradation, afterhich it was relatively low. Therefore, it is suggested that in terms

f mechanical properties, use of our core/shell fibers is relevant forduration of 2 months. If support is needed for extended periods, aore fiber based on a polymer with a slower degradation rate, suchs poly(l-lactic acid), is recommended.

. Summary and conclusions

Farnesylthiosalicylate (FTS, Salirasib) is a new specific nontoxicrug with a mild hydrophobic nature, which acts as a Ras antago-ist. Thus, the incorporation of this new drug in a stent coating mayvercome the incomplete healing and lack of endothelial coveragessociated with current drug-eluting stents. In the current studye developed and studied FTS-loaded bioresorbable shell struc-

ures in order to investigate the FTS release profile from a platformhat can be used as a stent coating. These structures were com-osed of a polyglyconate core and a porous PDLGA shell loadedith the antiproliferative agent FTS, prepared using freeze drying

f inverted emulsions. In addition to the usage of the coating, thehole unique core/shell fiber platform can be used as basic ele-ents of bioresorbable vascular stents as well as for local treatment

f cancer. Our study focused on the effects of the emulsion’s compo-ition (formulation) and process kinetics on the FTS release profilerom the fibers, in light of the shells’ morphology and degradationrofiles. The tensile mechanical properties and their degradationere also studied.

In general, porous shell structures (porosity of 75–92% and poreize of 2–7 �m) were obtained with good adhesion to the core fiber.he FTS release profiles of all studied specimens exhibited a burstffect accompanied by a release rate which decreased with time andasted for 15–40 days. The copolymer composition was found to behe most important parameter affecting release behavior in our sys-em. A decrease in the lactic acid content of the PDLGA copolymeresulted in an increase in the burst effect and release rate duringhe first 2 weeks, mainly due to higher water uptake, swelling andhanges in microstructure, but also due to a higher degradationate of the host polymer. Other parameters, demonstrating only ainor effect on the FTS release profile, are: polymer content, O:A

hase ratio and homogenization rate. In all these studied cases anndirect effect of microstructure on release profile occurs via anmulsion stability mechanism. In general, a lower diffusion rate cane achieved when low porosity is combined with a coarser structuref higher pore size.

The release profile of FTS from our new coatings is fasterhan paclitaxel release (previously studied by us) and moredjustable, and therefore more suitable for the stent application.urthermore, since FTS is less toxic, some burst release may beolerated.

cknowledgements

The authors are grateful to the Israel Science Foundation (ISF,rant number 1312/07) and to the Slezak Foundation, Tel-Aviv Uni-

ersity, for supporting this research.

We would like to thank Concordia Pharmaceuticals (Ft. Laud-rdale, FL) for kindly providing us the farnesylthiosalicylate (FTS,alirasib), Ms. Tami Shpiro and Ms. Tanya Rosenblit, Tel-Aviv Uni-ersity, for their assistance with the drug release studies and Mr.

V

aceutical Sciences 37 (2009) 351–362 361

aniel Giler, Tel-Aviv University, for his assistance with the studyf mechanical properties.

eferences

reitenbach, A., Mohr, D., Kissel, T., 2000. Biodegradable semi-crystalline combpolyesters influence the microsphere production by means of a supercriticalfluid extraction technique (ASES). J. Control Release 63, 53–68.

ommandeur, S., van Beusekom, H.M., van der Giessen, W.J., 2006. Polymers, drugrelease, and drug-eluting stents. J. Interv. Cardiol. 19, 500–506.

reel, C.J., Lovich, M.A., Edelman, E.R., 2000. Arterial paclitaxel distribution anddeposition. Circ. Res. 86, 879–884.

rachman, D.E., Edelman, E.R., Seifert, P., Groothuis, A.R., Bornstein, D.A., Kamath,K.R., Palasis, M., Yang, D., Nott, S.H., Rogers, C., 2000. Neointimal thickening afterstent delivery of paclitaxel: change in composition and arrest of growth over sixmonths. J. Am. Coll. Cardiol. 36, 2325–2332.

eng, S., Huang, G., 2001. Effects of emulsifiers on the controlled release of pacli-taxel (Taxol) from nanospheres of biodegradable polymers. J. Control Release71, 53–69.

inkelstein, A., McClean, D., Kar, S., Takizawa, K., Varghese, K., Baek, N., Park, K.,Fishbein, M.C., Makkar, R., Litvack, F., Eigler, N.L., 2003. Local drug delivery viaa coronary stent with programmable release pharmacokinetics. Circulation 107,777–784.

onseca, C., Simoes, S., Gaspar, R., 2002. Paclitaxel-loaded PLGA nanoparticles:preparation, physicochemical characterization and in vitro anti-tumoral activity.J. Control Release 83, 273–286.

eorge, J., Sack, J., Barshack, I., Keren, P., Goldberg, I., Haklai, R., Elad-Sfadia, G., Kloog,Y., Keren, G., 2004. Inhibition of intimal thickening in the rat carotid artery injurymodel by a nontoxic Ras inhibitor. Arterioscler. Thromb. Vasc. Biol. 24, 363–368.

alkin, A., Stone, G.W., 2004. Polymer-based paclitaxel-eluting stents in percuta-neous coronary intervention: a review of the TAXUS trials. J. Interv. Cardiol. 17,271–282.

eldman, A.W., Cheng, L., Jenkins, G.M., Heller, P.F., Kim, D.W., Ware Jr., M., Nater,C., Hruban, R.H., Rezai, B., Abella, B.S., Bunge, K.E., Kinsella, J.L., Sollott, S.J.,Lakatta, E.G., Brinker, J.A., Hunter, W.L., Froehlich, J.P., 2001. Paclitaxel stent coat-ing inhibits neointimal hyperplasia at 4 weeks in a porcine model of coronaryrestenosis. Circulation 103, 2289–2295.

ofma, S.H., van der Giessen, W.J., van Dalen, B.M., Lemos, P.A., McFadden, E.P.,Sianos, G., Ligthart, J.M., van Essen, D., de Feyter, P.J., Serruys, P.W., 2006.Indication of long-term endothelial dysfunction after sirolimus-eluting stentimplantation. Eur. Heart J. 27, 166–170.

ackson, J.K., Smith, J., Letchford, K., Babiuk, K.A., Machan, L., Signore, P., Hunter,W.L., Wang, K., Burt, H.M., 2004. Characterization of perivascular poly(lactic-co-glycolic acid) films containing paclitaxel. Int. J. Pharm. 283, 97–109.

loog, Y., Cox, A.D., 2004. Prenyl-binding domains: potential targets for Rasinhibitors and anti-cancer drugs. Semin. Cancer Biol. 14, 253–261.

raitzer, A., Ofek, L., Schreiber, R., Zilberman, M., 2008. Long-term in vitro studyof paclitaxel-eluting bioresorbable core/shell fiber structures. J. Control Release126, 139–148.

agerqvist, B., James, S.K., Stenestrand, U., Lindback, J., Nilsson, T., Wallentin, L.,2007. Long-term outcomes with drug-eluting stents versus bare-metal stentsin Sweden. N. Engl. J. Med. 356, 1009–1019.

i, Y., Fukuda, N., Kunimoto, S., Yokoyama, S., Hagikura, K., Kawano, T., Takayama,T., Honye, J., Kobayashi, N., Mugishima, H., Saito, S., Serie, K., 2006. Stent-based delivery of antisense oligodeoxynucleotides targeted to the PDGF A-chaindecreases in-stent restenosis of the coronary artery. J. Cardiovasc. Pharmacol. 48,184–190.

u, L., Peter, S.J., Lyman, M.D., Lai, H.L., Leite, S.M., Tamada, J.A., Uyama, S., Vacanti,J.P., Langer, R., Mikos, A.G., 2000. In vitro and in vivo degradation of porouspoly(dl-lactic-co-glycolic acid) foams. Biomaterials 21, 1837–1845.

arom, M., Haklai, R., Ben-Baruch, G., Marciano, D., Egozi, Y., Kloog, Y., 1995. Selec-tive inhibition of Ras-dependent cell growth by farnesylthiosalisylic acid. J. Biol.Chem. 270, 22263–22270.

oreno, R., Macaya, C., 2005. Stent-based delivered anti-proliferative drugs in theprevention of coronary stent restenosis. Curr. Med. Chem. Cardiovasc. Hematol.Agents 3, 221–229.

anade, S.V., Miller, K.M., Richard, R.E., Chan, A.K., Allen, M.J., Helmus, M.N., 2004.Physical characterization of controlled release of paclitaxel from the TAXUSexpress 2 drug-eluting stent. J. Biomed. Mater. Res. A 71, 625–634.

ao, S.V., Shaw, R.E., Brindis, R.G., Klein, L.W., Weintraub, W.S., Peterson, E.D., 2006.On-versus off-label use of drug-eluting coronary stents in clinical practice(report from the American College of Cardiology National Cardiovascular DataRegistry [NCDR]). Am. J. Cardiol. 97, 1478–1481.

ogers, C.D., 2004. Optimal stent design for drug delivery. Rev. Cardiovasc. Med. 5(Suppl. 2), S9–S15.

erruys, P.W., Sianos, G., Abizaid, A., Aoki, J., den Heijer, P., Bonnier, H., Smits, P.,McClean, D., Verheye, S., Belardi, J., Condado, J., Pieper, M., Gambone, L., Bressers,

M., Symons, J., Sousa, E., Litvack, F., 2005. The effect of variable dose and releasekinetics on neointimal hyperplasia using a novel paclitaxel-eluting stent plat-form: the Paclitaxel In-Stent Controlled Elution Study (PISCES). J. Am. Coll.Cardiol. 46, 253–260.

enkatraman, S., Boey, F., 2007. Release profiles in drug-eluting stents: issues anduncertainties. J. Control Release 120, 149–160.

3 Pharm

V

W

W

62 A. Kraitzer et al. / European Journal of

irmani, R., Guagliumi, G., Farb, A., Musumeci, G., Grieco, N., Motta, T., Mihalcsik,

L., Tespili, M., Valsecchi, O., Kolodgie, F.D., 2004. Localized hypersensitivity andlate coronary thrombosis secondary to a sirolimus-eluting stent: should we becautious? Circulation 109, 701–705.

ang, X., Venkatraman, S.S., Boey, F.Y., Loo, J.S., Tan, L.P., 2006. Controlled releaseof sirolimus from a multilayered PLGA stent matrix. Biomaterials 27, 5588–5595.

W

aceutical Sciences 37 (2009) 351–362

estedt, U., Wittmar, M., Hellwig, M., Hanefeld, P., Greiner, A., Schaper, A.K.,

Kissel, T., 2006. Paclitaxel releasing films consisting of poly(vinyl alcohol)-graft-poly(lactide-co-glycolide) and their potential as biodegradable stent coatings. J.Control Release 111, 235–246.

u, L., Ding, J., 2004. In vitro degradation of three-dimensional porouspoly(d,l-lactide-co-glycolide) scaffolds for tissue engineering. Biomaterials 25,5821–5830.