at SciVerse ScienceDirect

Biomaterials 33 (2012) 7984e7992

Contents lists available

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

In vivo biocompatibility of a plasma-activated, coronary stent coating

Anna Waterhouse a,e,1, Steven G. Wise a,e, Yongbai Yin b, Buchu Wu c, Barbara James d, Hala Zreiqat d,David R. McKenzie b, Shisan Bao c, Anthony S. Weiss a, Martin K.C. Ng e, Marcela M.M. Bilek b,*

a School of Molecular Bioscience, University of Sydney, NSW 2006, AustraliabApplied & Plasma Physics Group, Physics School, University of Sydney, NSW 2006, AustraliacDiscipline of Pathology, School of Medical Sciences, University of Sydney, NSW 2006, Australiad School of Aerospace, Mechanical and Mechatronic Engineering, University of Sydney, NSW 2006, AustraliaeHeart Research Institute, Sydney, NSW 2042, Australia

a r t i c l e i n f o

Article history:Received 4 July 2012Accepted 28 July 2012Available online 11 August 2012

Keywords:Coronary stentPlasma-activated coatingDelaminationEndothelializationBiocompatibilityIn vivo rabbit model

* Corresponding author. Tel.: þ61 2 93516079; fax:E-mail address: marcela.bilek@sydney.edu.au (M.M

1 Present address: Wyss Institute, Harvard, Boston

0142-9612/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.biomaterials.2012.07.059

a b s t r a c t

Bare metal and drug-eluting coronary stents suffer an inherent lack of vascular cell and bloodcompatibility resulting in adverse patient responses. We have developed a plasma-activated coating(PAC) for metallic coronary stents that is durable, withstands crimping and expansion, has low throm-bogenicity and can covalently bind proteins, linker-free. This has been shown to enhance endothelial cellinteractions in vitro and has the potential to promote biointegration of stents. Using the rabbit denudediliac artery model, we show for the first time that PAC is a feasible coating for coronary stents in vivo. Thecoating integrity of PAC was maintained following implantation and expansion. The rate of endotheli-alization, strut coverage, neointimal response and the initial immune response were equivalent to baremetal stents. Furthermore, the initial thrombogenicity caused by the PAC stents showed a reduced trendcompared to bare metal stents. This work demonstrates a robust, durable, non-cytotoxic plasma-basedcoating technology that has the ability to covalently immobilize bioactive molecules for surface modi-fication of coronary stents. Improvements in the clinical performance of implantable cardiovasculardevices could be achieved by the immobilization of proteins or peptides that trigger desirable cellularresponses.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Metallic coronary stents now dominate in percutaneous coro-nary interventions [1]. However, the clinical performance of bothbare metal stents (BMS) and drug eluting stents (DES) is less thanideal and causes significant adverse patient outcomes. BMS areprone to high rates of vessel renarrowing known as restenosis [2],while DES suffer from an ongoing risk of late stent thrombosis [3],polymer hypersensitivity [4], polymer delamination [5] anddelayed re-endothelialization [6].

New generation DES aimed at reducing these adverse effects arebeing developed and although promising, reliance on dual anti-platelet therapy and safety outcomes following their cessationremain issues [7]. Additionally, polymer instability is a persistentproblem, with thick polymer coatings delaminating and exposingthrombogenic bare metal stent struts to the vasculature [5]. Simi-larly, newly developed bioresorbable stents could potentially

þ61 2 9351 7725..M. Bilek).

MA 02115, USA.

All rights reserved.

provide benefits by degrading over time, however these are yet tobe fully developed. Their clinical applicability is likely to be limitedby intrinsic problems, such as poor deliverability, flexibility andradial strength. The thicker struts required to compensate forreduced radial strength of the materials [8] are known to causeincreased restenosis and thrombosis [9,10]. There remains a needfor robust, biomimetic coatings for metallic stents.

We have developed a plasma-activated coating (PAC) for surfacemodification of metallic alloys using plasma enhanced chemicalvapour deposition [11]. PAC is smooth and strongly adheres to theunderlyingmetal byanenergetic ion stitchingdepositionprocess. Thecoating iswear resistant under pulsatile flowand is able towithstandcrimping andexpansionwithoutdelaminating invitro [11,12]. PAChasstrikingly low thrombogenicity compared to 316L stainless steel instatic adhesion and flow loops in vitro using whole human blood[12,13]. Furthermore, PAC allows linker-free covalent immobilizationof functional biomolecules to the surface with the potential to allowimproved biointegration of a range of biomedical implants [11,13].

In this proof of principle study, we have evaluated the acuteresponse to PAC stents compared to baremetal stainless steel stentsin a well-established animal model [14]. In a rabbit denuded

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e7992 7985

bilateral iliac artery model we analysed the feasibility of PAC asa stent coating. We evaluated coating integrity, the rate of endo-thelialization, the initial immune response and thrombogenicitycaused by the stent.

2. Methods

2.1. Stent design and treatment

The stent design was based on the dimensions of commercially available stents.The stents were laser cut from slotted tubes of 316L vacuum melted stainless steeland electropolished to remove surface contaminants (Laserage, Waukegan, Illinois,U.S.A). The PAC was deposited without active heating or cooling of the stents. Agraded interface was created using reactive magnetron sputtering from a cathode of316L stainless steel. The substrate holder was rotated to expose all surfaces of thestents to the sputtered flux. Argon and acetylene were injected into the chamberthrough a distributed gas line. To form the graded interface, a pure stainless steelcoating was first deposited onto the stainless steel stents, followed by a coating thatcontained gradually increasing fractions of plasma polymer deposited by plasmaenhanced chemical vapour deposition. To achieve this, the acetylene follow rate wasincreased from zero until a pure plasma polymer layer was formed. The initialsputtering voltage during the deposition of pure metal was 800 V while the cathodecurrent was maintained at 3 A. Increasing the flow rate of acetylene while keepingthe cathode current constant eventually results in the deposition of a pure plasmapolymer material when the cathode is fully covered by a plasma polymer layerdeposited on the cathode surface. In this way, the stainless steel cathode can be fullyprotected from sputtering, resulting in a pure PAC layer at the surface. The exposureto highly energetic reactive species in the plasma ensures that the coated stents aresterile. Transfer of the stents to storage vials was performed under sterile conditions.

2.2. Energy dispersive X-ray microanalysis (EDS)

Samples were imaged with a Philips XL 30 CP scanning electron microscopeusing an acceleration voltage of 10 kV. EDS analysis was carried out using an EDAX

Fig. 1. PAC stent. (A) Photograph of a BMS (left) and a PAC treated stent via the modified dep(blue) and PAC (red) stent detected by EDS. The SEM was operated in spot mode at an accele150 s (C and D) SEM image of a PAC stent surface after crimping and expansion. (C) Scale barthe nanoscale columnar structures is visible. Scale bar indicates 400 nm. (For interpretation oof this article.)

detector and EDX control software at a working distance of 11 mm. The EDS wasoperated in area mode with an accumulated live time of 150 s at 5 kV, producinga penetration depth of 375 nm.

2.3. Animal care and surgery

Study approval was obtained from Sydney South West Area Health AnimalEthics Committee (protocol number 2008/036) and the University of SydneyAnimal Ethics Committee (protocol number K00/12-2009/3/5243). Experimentswere conducted in accordance with the Australian Code of Practice for the Careand Use of Animals for Scientific Purpose. Adult male New Zealand White Rabbitswere housed in individual cages in a temperature and light controlled room andgiven standard rabbit chow ad libitum with free access to sterilized drinkingwater.

A total of 14 animals received surgery, n ¼ 4 controls (denuded on 1 iliac arteryand no manipulation of the bilateral artery) and n ¼ 10 BMS vs PAC. Rabbitsweighing 3e4 kg were sedated with a subcutaneous injection of acetylpromazine(0.5 mg/kg) and anaesthetized with isofluorane (2%) and oxygen (2 l/min) via a facemask. A 5 F sheath was inserted into the right femoral artery by cut down. Endo-thelial denudation of the iliac arteries was carried out using a 3.25 mm � 10 mmangioplasty balloon catheter (Sprinter� Legend RX) passed over a 0.01400 guide wireto the aorta, inflated to nominal pressure (6 atmwith 50% (v/v) contrast/saline) andwithdrawn retrograde to the external iliac artery three times. The four controlanimals received no further manipulation. A BMS or PAC stent was hand crimpedonto a 3.25mmballoon catheter, advanced to the common iliac artery and expandedfor 15 s (6 atm) to a diameter of 3.1 mm, giving a stent to artery ratio of 1.2:1. Thelocation and patency of the stented iliac arteries and surrounding vasculature wasconfirmed by angiography. The procedure was repeated for the left iliac artery. Theorder and side in which each type of stent was implanted was randomized. Rabbitswere anti-coagulated with a single dose of heparin at the time of surgery (100 U/kg)and aspirin (40 mg/day) orally 24 h before surgery and maintained throughout thein-life phase of the study [15,16]. Stented vessels were explanted after 7 days.Anaesthetized animals were exanguinated via left ventricular puncture, perfusedwith lactated Ringer’s solution and given a lethal injection of lethabarb (130 mg/kg).Stents were explanted and flushed with lactated Ringer’s solution before fixation.

osition protocol (right). Scale bar indicates 5 mm. (B) Elemental composition of a BMSration voltage of 5 kV (penetration depth of 357 nm) with an accumulated live time ofindicates 100 mm. (D) The stress accommodation mechanism through the boundaries off the references to colour in this figure legend, the reader is referred to the web version

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e79927986

In total 10 pairs of BMS and PAC stents were implanted into 10 animals. Stentpairs from 4 animals were longitudinally sectioned for scanning electronmicroscopy(SEM) and immunofluorescence analysis. The remaining 6 pairs of stents wereprocessed for histology.

2.4. Scanning electron microscopy

Stented and denuded vessels for SEM processing were longitudinally bisected.Half of each stent was fixed in 2.5% glutaraldehyde for 2 h for SEM. This was followedby osmium fixation, dehydration, mounting and coating with 20 nm gold. Imagingwas done using a Philips XL-30 CP scanning electron microscope.

2.5. Immunofluorescence

Half of a longitudinally bisected stent was fixed in 4% paraformaldehyde for 16 hfor immunofluorescence. Whole aorta was used as controls for immunolabeling.Samples were incubated with 0.2 M glycine for 24 h at 4 �C and blocked with 10% (v/v) goat serum in TBS-T (0.5% Tween in Tris buffered saline) for 2 h at roomtemperature. Samples were then incubated with rabbit anti-CD31 antibody raised inmouse (Abcam, 1:25 in TBS) or with mouse IgG1 k (Abcam, 1:25 in TBS) for 16 h at4 �C. Samples were washed 4 times in TBS for 5 min each and then incubated withmouse anti-IgG-FITC conjugated secondary antibody raised in goat (Sigma, 1:90 inTBS) for 1 h at room temperature. The secondary antibody was removed and thesamples were washed 4 times in TBS for 5 min each. The samples were then stainedwith phalloidin-TRITC (0.1 mg/ml) in TBS for 45 min then washed 3 times with TBS.The samples were then stained with 0.1 mg/ml Hoechst in TBS for 10 min thenwashed 3 times in TBS. Sampleswere imaged using a 10� objective (HC PL FL 10x/0.3PH1, free working distance of 11 mm) on a Leica TCS SPII Multi-Photon Microscope.

2.6. Histology

Stented and denuded vessels for histological processing were incubated in 4%paraformaldehyde for 24 h, dehydrated with ascending grades of ethanol andembedded in methyl methacrylate resin. Transverse sections, 15e20 mm thick, weremade on a diamond knife microtome (EXAKT band saw 300 CL/CP, Norderstedt,

Fig. 2. Stent implantation. (A) Schematic of the location of stent implantation in the lefvasculature following injection of contrast dye (B) and the stented common iliac arteries (C

Germany) and polished to 8e10 mm using a variable speed grinder (Buehler Ecomet 3,Illinois, U.S.A). Sections were de-plasticized with xylene for 2 h and 2% (v/v) hydro-chloric acid in ethanol for 5 min and stained with hemotoxylin and eosin (H&E).

2.7. Statistics

Data are expressed as mean � SE and indicated in figures as (*p < 0.05, **p < 0.01and ***p< 0.001). For in vivo experiments, compiled stent SEM imageswere scored forthe percentage cell coverage of stent struts and intersections and divided by thenumber counted. Stents were also analysed for the area of CD31 positive cell labellingand fibrin deposition, these areas were each divided by the total area visible usingImageJ. The data were analysed by a paired t-test for statistical significance usingGraphPad Prism version 4.00 (Graphpad Software, San Diego, California) for Mac.

3. Results

3.1. PAC stents

Macroscopically, PAC stents were identifiable due to their muchdarker colour (Fig. 1A). The PAC deposited onto stents was detectedusing energy dispersive X-ray microanalysis, which allows detec-tion of different elements present on a surface or in a coating [17].EDS of a BMS (blue trace) showed the presence of carbon (C), iron(Fe), nickel (Ni), sulphur (S), silicon (Si) and chromium (Cr) (Fig. 1B),with Fe observed in the highest relative intensity. In contrast, thepredominant signal in the EDS of a PAC stent (red trace) was carbon,with the relative intensity of all the other elements reduced. Thecontribution of other elements was also decreased when comparedto the BMS. The PAC layer showed no delamination after crimpingand expansion of the stent (Fig. 1C). Examination of the layer athigher magnification (Fig. 1D) showed the presence of a columnar

t and right rabbit common iliac arteries (asterisks). (BeD) Angiograms of the rabbitand D arrows) showing patent vessels following implantation (D).

Fig. 4. Cell coverage and fibrin deposition. (A) Representative images of cell coverage on the lumen of paired stents, BMS (blue), PAC (red). (B) Percentage of cell coverage on stentstruts. (C) Representative images of fibrin thrombus deposition on the lumen of paired stents, BMS (blue), PAC (red). (D) Percent fibrin thrombus formation on stent struts. (Forinterpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. PAC adhesion. (A) Representative SEM images of a BMS (left, blue box) and PAC (right, red box) stent. Scale bar indicates 1 mm. (B) Higher magnification SEM images of PACstents showing strut intersections with no evidence of delamination. Scale bars indicate 100 mm. (For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e7992 7987

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e79927988

microstructure on the nanoscale. The accommodation of strainthrough the generation of small gaps between the columns isevident.

3.2. In vivo assessment of PAC stents

A pilot rabbit endothelialization experiment was carried outwith a total of 10 animals receiving a control BMS (316L SS stent)and a PAC stent implanted into bilateral common iliac arteries(Fig. 2A). Animal variability was accounted for by the internal BMScontrol in each animal and subsequent paired analysis. Vesselposition (Fig. 2B), stent deployment (Fig. 2C) and post-implantpatency (Fig. 2D) were confirmed for each animal via angiog-raphy. The stent to artery diameter ratio was 1.13:1 e1.24:1, withinthe accepted range [18]. All animals survived the in-life phase of thestudy and there was no evidence of infection due to the sterility ofthe stents. Animals were sacrificed 7 days post-procedure, a timepoint chosen such that the control BMS stents would only bepartially endothelialized. BMS have been shown to endothelialize80e90% by 14 days in this model [16].

Control animals underwent balloon denudation without stentimplantation in one iliac artery and nomanipulation of the bilateraliliac artery. Neointima was seen in the denuded iliac artery withmultiple layers of cells on the luminal surface of the internal elastic

Fig. 5. Immunofluorescence of CD31 in the aorta. CD31 immunostaining (green) of rabbit steconjugated secondary antibody (positive staining). Alternatively, an isotype control antibosamples of aorta. Hoechst stained nuclei (blue) and phalloidin-TRITC stained actin cytoskeletin the other images. (For interpretation of the references to colour in this figure legend, th

lamina (IEL). In comparison, a monolayer of endothelium wasobserved in a normal iliac artery (data not shown).

3.3. Acute cellular response

Four pairs of stents were longitudinally bisected and analysedby SEM. 7 days post-implantation, all stents had cells both betweenstruts and covering the struts in areas (Fig. 3A). Out of the 4 PACstents implanted, there was no evidence of delamination at anystrut region visualized (Fig. 3B). Cell coverage of struts on BMS andPAC stents (Fig. 4A) was quantified and determined to be 69 � 6%and 63 � 7% respectively (Fig. 4B). There were no cases of occlusivestent thrombosis in any of the BMS or PAC stents. A small amount ofthrombus adhesion was observed on BMS and PAC stents, both onand between the struts (Fig. 4C). This consisted of fibrin, adherenterythrocytes and platelets. The level of fibrin thrombus formationwas reduced on PAC compared to BMS, although, consistent withbeing a secondary end point in this animal model, the differenceswere not significant (Fig. 4D).

3.4. Detection of endothelium

Immunofluorescent labelling for CD31, an endothelial cellspecific marker, was visualized on stented arteries by confocal

nted iliac arteries (stented artery) or aorta using a rabbit anti-CD31 antibody and a FITCdy (isotype control), or a secondary antibody only (secondary control) was used onon (red) are also shown. Scale bars in stented artery images indicate 200 mm, and 50 mme reader is referred to the web version of this article.)

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e7992 7989

fluorescence microscopy and used to quantify endothelial cellcoverage. The samples were also co-stained with TRITC-phalloidinto detect the actin cytoskeleton and Hoechst for cell nuclei. Samplesof rabbit aorta were used as controls for the immunofluorescentlabelling (Fig. 5). CD31 was present at the periphery of cells on theluminal surface of the aorta (Fig. 5, Positive Control). The isotypecontrol and secondary only antibody showed no defined cellperiphery fluorescence in the emission range of FITC(500e550 nm). Non-specific green staining present in both theisotype control and secondary control samples localized withgrooves of the intima, seen at low power magnification.

Whole stent halves were imaged as z stacks by multipointpositioning. Projections from each channel were automaticallygenerated and the stacks were merged by mosaic stitching (Figs. 5and 6). In all samples, CD31 immunolabeling was observed at cell

Fig. 6. Immunofluorescence of CD31 in stented arteries. (A) CD31 immunostaining(green) of BMS (left) and PAC (right) stented arteries using a rabbit anti-CD31 antibodyand a FITC conjugated secondary antibody, high power images in corresponding boxes.Scale bars indicate 1 mm in low power images and 200 mm in high power images.(B) Percentage area of cell stained positive for CD31 compared to total area imaged.(For interpretation of the references to colour in this figure legend, the reader isreferred to the web version of this article.)

peripheries on the artery between the struts and in some casescovering the struts (Fig. 6A). There was no difference in the amountof fluorescence observed between matched pairs of BMS and PAC(Fig. 6B). For all samples, there were more CD31 positive cellsobserved at the proximal and distal regions of the stents comparedto the middle of the stents. Additionally, there were areas of strutsfrom both BMS and PAC stents that were covered with cells whichstained positive for actin and nuclei, however they did not showimmunolabelling for CD31.

3.5. Neointima formation and strut coverage

Cross sections of paired BMS and PAC stents were analysed afterH&E staining (Fig. 7A). Stent struts were visible as black square orrectangular shapes in cross sections (black arrows in low powermagnifications). Neointima was observed in all samples as a thick-ening of the tissue on the luminal side of the IEL (white arrows inhigh power magnifications). Strut coverage and neointimal areawere equivalent in proximal, middle and distal regions of the BMSand PAC stents (Fig. 7B and C). There was no gross inflammatoryresponse caused by either BMS or PAC stents, with only a smallnumber of multinucleated cells present in both stent types. Addi-tionally, there were areas of orange/brown staining containingerythrocytes occasionally present surrounding stent struts orwithin the neointima in both stent types.

4. Discussion

In vivo analysis of PAC stents has demonstrated the feasibility ofthis coating technology for medical devices, showing that thecoating survived crimping, implantation and expansion with nodelamination. Furthermore, PAC stents displayed equivalentperformance compared to commercial 316L stainless steel BMS, interms of cell coverage, endothelialization, inflammatory and heal-ing response. PAC represents a robust platform coating technologythat has been shown to facilitate the covalent attachment of bio-logical molecules, such as proteins and peptides [11,13,19,20]. Thecovalent immobilization of appropriate biomolecules has greatpotential for directing biological responses to non-biologicalmaterials [21]. In the context of cardiovascular applications,preliminary in vitro studies indicate that covalently immobilisedtropoelastin and its derivative peptides provide benefits inpromoting endothelialisation [11,22] and reducing thromboge-nicity [12,23]. The results presented here are a crucial first steptowards realising this goal in vivo as they demonstrate that the PACcoating we have engineered to enable the covalent immobilisationof key proteins or peptides survives the rigours of surgicalimplantation and remains stable in vivo.

PAC stents were darker in colour and did not delaminate uponrepeated crimping and expansion. The presence of PAC on the stentwas additionally confirmed prior to implantation by detectingsurface elements with EDS which is applicable to 3D objects, incontrast to FTIR and XPS [24]. EDS analysis of BMS showed char-acteristic elements carbon, iron, nickel, sulphur, silicon and chro-mium [25]. EDS of PAC stents confirmed the presence of a carbon-rich coating demonstrated by a substantially increased relativecarbon content and corresponding decrease in iron. The EDS of PACstents correlates with the known carbon based composition of PAC,previously characterized by FTIR [26] and XPS [12]. SEM imaging ofPAC stents after crimping and balloon expansion showed that PAChad adhered well. Delamination was resisted and strain due toplastic deformation of the underlying metal was accommodated bythe formation of nanocracks between nanoscale columns of thecoating. The nanoscale columns are a feature of the deposition

Fig. 7. H&E staining of stented arteries 7 days post-implantation. (A) Representative images of H&E stained BMS (left) and PAC (right) stented arteries 7 days post-implantation atlow (top panel) and high (lower panel) magnification. Black arrows indicate stent struts and white arrows indicate the internal elastic lamina. (B) Percentage strut coverage of BMSand PAC stents. (C) Percent neointimal area on the luminal side of the IEL in BMS and PAC stents.

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e79927990

process and are separated by regions of relative mechanicalweakness where the nanocracks can form [12].

A pilot animal experiment was carried out to establish thefeasibility, safety and deliverability of a PAC stent. The plasma basedmethod of PAC deposition produces a sterile coated stent, thereforeno further sterilization was required. BMS were plasma etched toremove surface contaminants resulting in a sterile homogeneousstainless steel surface. This common method of sterilization doesnot alter surface properties [27]. Paired BMS and PAC stents wereimplanted into bilateral iliac arteries of 10 animals for 7 days toassess the PAC coating integrity and acute response. The denudedrabbit iliac stenting model is the most commonly used animalmodel for assessing the in vivo response of commercially availablestents as the healing response following arterial denudation andstent implantation is comparable to that in humans following stentimplantation [28,29].

Of the 4 PAC stents analysed by SEM, there was no evidence ofdelamination at any of the strut intersections. This is consistentwith previous in vitro studies of crimping and expansion showingresistance to delamination [12]. The ability of PAC stents to resistdelamination puts them in direct contrast with commerciallyavailable DES which frequently delaminate [5,30,31]. The PACcoating combines a number of improved properties to achieve thisbenefit. First, the deposited layer is only 1 mm thick, substantiallythinner than DES coatings which can range from 4 to 19 mm. PACalso has optimized mechanical properties, displaying elasticity andresilience in vitro [12]. This allows the coating to better withstandthe plastic deformation that occurs during stent expansion. Thenanostructured surface has the capacity to accommodate strainbuilt up during deployment through a network of fine nanocracks,whilst allowing the coating to remain fully adherent.

SEM of stented regions showed evidence of organized muralthrombi, macrophages and cells, consistent with the expectedevents observed in this model post-implantation [32]. Cell coverageof stents was quantified from the percentage strut coverage in eachsample. Equivalent cell coverage was observed in paired samples

between BMS and PAC stents. Increased cell coverage of struts wasobserved at the proximal and distal ends of the stents compared tothe middle of the stent in the majority of cases [16]. Fibrinthrombus formation was non-significantly reduced on PAC stentscompared to BMS, but did show a strong trend towards a decreasedamount. This trend correlates with the low thrombogenicity of PACsurfaces extensively demonstrated in vitro [12,13], but additionalanimal models are needed to determine significance in vivo.Although mural thrombus and fibrin deposition are expected tooccur following biomaterial implantation [28], it is a secondary endpoint in the rabbit iliac model. This model does not accuratelyrepresent the acute thrombogenicity of the stent due to high bloodflow in iliac arteries and the aspirin regimen. Further assessment ofthrombogenicity in vivo, for example, in a baboon shunt modelwould be necessary to confirm the reduced thrombogenicity seenin these results.

As the cell type was not discernable via SEM, immunocyto-chemical staining for endothelial cells was performed. Whole-mount immunofluorescence of longitudinally bisected stentedarteries was performed to detect endothelial cells by the expressionof CD31, an endothelial specific intercellular junction protein. Widefield z stack, multipoint positioning confocal microscopy andmosaic stitching were utilized to image the entire length of stentedarteries. This technique offers a powerful tool for generating highquality representative informationwhen analysing stents and otherlarge three-dimensional specimens.

Endothelial cells were identified by positive cell peripherystaining, both between and covering the struts on BMS and PACstents. Equivalent endothelialization between BMS and PAC stentswas detected, with CD31 positive cells covering 30 � 14% and28 � 12% of the total stent area respectively. This degree of endo-thelialization is consistent with previous studies of BMS in thismodel and also reflective of in vitro findings for PAC alone. Incomparative endothelial cell attachment and proliferation studiesPAC has consistently been shown to be equivalent to BMS, with theaddition of a biomolecule such as tropoelastin required to show

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e7992 7991

a benefit for endothelial cell interactions [11,12]. In both stent types,strut coverage by some CD31 negative cells was also observed.These cells could be inflammatory cells, smooth muscle cells ormural thrombus. Alternatively, these areas could be endothelialcells with poorly formed cellecell junctions. Indeed, it has beenshown that endothelial cells down-regulate CD31 expressionduringmigration and proliferation [16,33]. In both stent types thereappeared to be more CD31 positively stained endothelium at theproximal and distal ends of the stents compared to the middlesection of the stents. This was also observed in the SEM analysiswhere the strut coverage was more complete at the ends corre-lating with previous studies showing that endothelialization occursmore quickly at the ends of a stent [16,34].

Neointima formation was observed on both samples along theregion of the stented vessel, with no difference in cell coverage ofthe stent struts or neointimal area observed between BMS and PACstents. This was expected for both stent types, in the absence ofregulatory molecules or drug-elution from these platforms.Encouragingly for PAC, H&E staining showed there was no grossimmune response, however further characterization with immu-nohistochemistry for inflammatory cell markers would helpquantify the initial inflammatory response. The protein depositspresent within the neointima were possibly matrix deposition bycells or plasma protein deposition from the blood or migratedsmooth muscle cells from the media.

Overall, PAC stents showed no sign of delamination followingimplantation. Analysis of cell coverage and immunofluorescentCD31 labelling, revealed there was equivalent cell coverage andhealing response of struts between the commercially used BMS andPAC stents. This is consistent with in vitro data showing thatendothelial cells attach and proliferate comparably on 316L SS andPAC surfaces [12]. PAC has been shown to covalently bind proteinsandmaintain their native conformation and activity [13]. The linkerfree covalent binding capacity of PAC for biomolecules could beused to immobilize a layer of biomolecules, such as tropoelastin orits peptide derivatives, to enhance endothelialization and thehealing response of stented arteries whilst minimizing restenosis.This additional feature of the PAC coating technology has thepotential to dramatically improve stent performance by enhancingbiomimicry and vessel integration.

5. Conclusions

In this proof-of-concept study, the feasibility and safety of PACstents was established by implantation into denuded rabbit iliacarteries. The PAC did not delaminate and maintained its integrityfollowing crimping and in vivo implantation, which representsa significant improvement over commercially available polymercoated DES. PAC stents showed equivalent cell coverage, endothe-lialization, inflammatory response and healing response comparedto paired BMS. This study demonstrates that the PAC coating isa viable platform coating technology for coronary stents. The in vivoperformance of PAC stents was comparable to that of commerciallyavailable 316L stainless steel BMS.We have therefore demonstrateda platform for biological functionalization of metal stents. Covalentbinding of selected bioactive proteins or peptides could be used topromote endothelialization, inhibit smoothmuscle cells and reducethrombogenicity.

Acknowledgements

This work is supported by the University of Sydney MedicalFoundation, University of Sydney Sydnovate Fund and the Austra-lian Research Council. The authors would like to thank Kim Hewitt,HRI Biological Facilities, for her help and dedication in conducting

the rabbit experiments. Additional thanks goes to the AustralianCentre for Microscopy and Microanalysis, especially Dr. ReneeWhan and Dr. Ellie Kable for confocal microscopy expertise, Dr. IanKaplin for SEM expertise and the Australian Microscopy andMicroanalysis Research Facility for access to equipment.

References

[1] Serruys PW, Kutryk MJB, Ong ATL. Coronary-artery stents. N Engl J Med 2006;354(5):483e95.

[2] Serruys PW, Strauss BH, Beatt KJ, Bertrand ME, Puel J, Rickards AF, et al.Angiographic follow-up after placement of a self-expanding coronary-arterystent. N Engl J Med 1991;324(1):13e7.

[3] Cutlip DE, Windecker S, Mehran R, Boam A, Cohen DJ, van Es G-A, et al. Clinicalend points in coronary stent trials: a case for standardized definitions.Circulation 2007;115(17):2344e51.

[4] Pendyala L, Jabara R, Robinson KA, Chronos NA. Passive and active polymercoatings for intracoronary stents: novel devices to promote arterial healing.J Interv Cardiol 2009;22(1):37e48.

[5] Levy Y, Mandler D, Weinberger J, Domb A. J. Evaluation of drug-eluting stents’coating durability-clinical and regulatory implications. J Biomed Mater Res BAppl Biomater 2009;91B(1):441e51.

[6] Joner M, Finn AV, Farb A, Mont EK, Kolodgie FD, Ladich E, et al. Pathology ofdrug-eluting stents in humans: delayed healing and late thrombotic risk. J AmColl Cardiol 2006;48(1):193e202.

[7] Alfonso F, Fernandez C. Second-generation drug-eluting stents: moving thefield forward. J Am Coll Cardiol 2011;58(1):26e9.

[8] Onuma Y, Ormiston J, Serruys PW. Bioresorbable scaffold technologies. Circ J2011;75(3):509e20.

[9] Kastrati A, Mehilli J, Dirschinger J, Dotzer F, Schuhlen H, Neumann F-J,et al. Intracoronary stenting and angiographic results: strut thicknesseffect on restenosis outcome (ISAR-STEREO) trial. Circulation 2001;103(23):2816e21.

[10] Kolandaivelu K, Swaminathan R, Gibson WJ, Kolachalama VB, Nguyen-Ehrenreich K-L, Giddings VL, et al. Stent thrombogenicity early in high-riskinterventional settings is driven by stent design and deployment and pro-tected by polymer-drug coatings/clinical perspective. Circulation 2011;123(13):1400e9.

[11] Yin Y, Wise SG, Nosworthy NJ, Waterhouse A, Bax DV, Youssef H, et al.Covalent immobilisation of tropoelastin on a plasma deposited interface forenhancement of endothelialisation on metal surfaces. Biomaterials 2009;30(9):1675e81.

[12] Waterhouse A, Yin Y, Wise SG, Bax D, McKenzie DR, Bilek MMM, et al. Theimmobilization of recombinant human tropoelastin on metals using a plasma-activated coating to improve biocompatibility of coronary stents. Biomaterials2010;31(32):8332e40.

[13] Bilek MMM, Bax DV, Kondyurin A, Yin Y, Nosworthy NJ, Fisher K, et al.Free radical functionalization of surfaces to prevent adverse responsesto biomedical devices. Proc Natl Acad Sci USA 2011;108(35):14405e10.

[14] Schwartz RSMD, Edelman EMDP, Virmani RMD, Carter ADO, Granada JFMD,Kaluza GLMD, et al. Drug-eluting stents in preclinical studies: updatedconsensus recommendations for preclinical evaluation. Circ Cardiovasc Interv2008;1(2):143e53.

[15] Rogers C, Edelman ER. Endovascular stent design dictates experimentalrestenosis and thrombosis. Circulation 1995;91(12):2995e3001.

[16] Joner M, Nakazawa G, Finn AV, Quee SC, Coleman L, Acampado E, et al.Endothelial cell recovery between comparator polymer-based drug-elutingstents. J Am Coll Cardiol 2008;52(5):333e42.

[17] Porter SC, Felton LA. Techniques to assess film coatings and evaluate film-coated products. Drug Dev Ind Pharm 2010;36(2):128e42.

[18] Schwartz RS, Chronos NA, Virmani R. Preclinical restenosis models and drug-eluting stents: still important, still much to learn. J Am Coll Cardiol 2004;44(7):1373e85.

[19] Yin Y, Fisher K, Nosworthy NJ, Bax D, Rubanov S, Gong B, et al. Covalentlybound biomimetic layers on plasma polymers with graded metallic interfacesfor in vivo implants. Plasma Process Polym 2009;6(10):658e66.

[20] Yin Y, Nosworthy NJ, Youssef H, Gong B, Bilek MMM, McKenzie DR. Acetyleneplasma coated surfaces for covalent immobilization of proteins. Thin SolidFilms 2009;517(17):5343e6.

[21] Hirano Y, Mooney DJ. Peptide and protein presenting materials for tissueengineering. Adv Mater 2004;16(1):17e25.

[22] Blit PH, McClung WG, Brash JL, Woodhouse KA, Santerre JP. Platelet inhibitionand endothelial cell adhesion on elastin-like polypeptide surface modifiedmaterials. Biomaterials 2011;32(25):5790e800.

[23] Srokowski EM, Blit PH, McClung WG, Brash JL, Santerre JP, Woodhouse KA.Platelet adhesion and fibrinogen accretion on a family of elastin-like poly-peptides. J Biomater Sci 2011;22(1e3):41e57.

[24] Ng FL, Wei J, Lai FK, Goh KL. Metallic thin film depth measurements by X-raymicroanalysis. Appl Surf Sci 2006;252(11):3972e6.

[25] Park JB, Kon Kim Y. Metallic biomaterials. Boca Raton, Florida: CRC PressTaylor & Francis Group; 2007.

A. Waterhouse et al. / Biomaterials 33 (2012) 7984e79927992

[26] Yin Y, Bilek MMM, McKenzie DR, Nosworthy NJ, Kondyurin A, Youssef H, et al.Acetylene plasma polymerized surfaces for covalent immobilization of densebioactive protein monolayers. Surf Coating Tech 2009;203(10e11):1310e6.

[27] Cheruthazhekatt S, Cernák M, Slaví�cek P, Havel J. Gas plasmas and plasmamodified materials in medicine. J Appl Biomed 2010;8(2):55e66.

[28] Virmani R, Kolodgie FD, Farb A, Lafont A. Drug eluting stents: are human andanimal studies comparable? Heart 2003;89(2):133e8.

[29] Perkins LE. Preclinical models of restenosis and their application in the evalu-ation of drug-eluting stent systems. Vet Pathol 2010;47(1):58e76. Online.

[30] Wiemer M, Butz T, Schmidt W, Schmitz K-P, Horstkotte D, Langer C. Scanningelectron microscopic analysis of different drug eluting stents after failedimplantation: from nearly undamaged to major damaged polymers. CatheterCardiovasc Interv 2009;75(6):905e11.

[31] Basalus MW, Ankone MJ, van Houwelingen GK, de Man FH, von Birgelen C.Coating irregularities of durable polymer-based drug-eluting stents asassessed by scanning electron microscopy. Eurointervention 2009;5:157e65.

[32] Palmaz JC. Intravascular stents: tissue-stent interactions and design consid-erations. Am J Roentgenol 1993;160(3):613e8.

[33] RayChaudhury A, Elkins M, Kozien D, Nakada MT. Regulation of PECAM-1 inendothelial cells during cell growth and migration. Exp Biol Med 2001;226(7):686e91.

[34] Soucy NV, Feygin JM, Tunstall R, Casey MA, Pennington DE, Huibregtse BA,et al. Strut tissue coverage and endothelial cell coverage: a comparisonbetween bare metal stent platforms and platinum chromium stents withand without everolimus-eluting coating. Eurointervention 2010;6(5):630e7.