Bioactive baculovirus nanohybrids for stent based rapid vascular re-endothelialization

Present study, for the first time, reports the development of a nanohybridized baculovirus based stent that can locally promote vascular re-endothelialization by efficient delivery of pro-angiogenic vascular endothelial growth factor (Vegf) genes. In vitro data demonstrated rapid expression of functionally active Vegf by the bioactive stent-transduced vascular cells. In vivo site-specific transgene expression was observed at the stented regions of balloon-denuded canine femoral artery, which eventually lead to significant endothelial recovery at the injured sites. A significant reduction in neointima formation (2.23 ± 0.56 mm2 vs 2.78 ± 0.49 mm2 and 3.11 ± 0.23 mm2, p < 0.05; n = 8) and percent stenosis was observed in treated stent group compared to negative control and bare metal stent groups. These findings collectively implicate the potential of this newly developed baculovirus based biotherapeutic stent to ameliorate damaged vascular biology and attenuate re-narrowing of stented artery by inhibiting neointima formation.

ceases once its job is done 24 . However, the baculoviruses do not have high transduction efficiency compared to mammalian vectors 30 . Thus genetic and surface modifications of Bac have been under intense research to improve their ability to bind on the mammalian cell surface and improve their lack of specificity, thereby enhancing the transduction efficiency, and overcoming drawbacks of baculoviruses as a gene carrier 31 .
Here, in order to enhance the gene delivery efficiency we have surface-functionalized the Bac with cationic Polyamidoamine dendrimer synthetic nanoparticles (PAMAM) and investigated its effect on transduction efficiency 4 . Furthermore, to protect them from serum inactivation and to achieve a controlled release at the target site, the Bac nanohybrids have been efficiently encapsulated within poly(lactic-co-glycolic acid) (PLGA) microspheres (MS) as optimized in in vitro preliminary studies ( Figure S2 and Table S1) 32 . The MS were impregnated in fibrin hydrogel and coated using layer-by-layer assembly on prefabricated stent surface 33 , which eventually acts as the scaffold to deliver the protected viral nanohybrids to the target site. We postulated that this new generation of gene eluting stent device can efficiently deliver angiogenic vascular genes to the affected site and induce favorable therapeutic effect in the local vascular biology as illustrated in Figure S1 to promote local wound healing. Accordingly, the study assessed the hypothesis by investigating the stent device efficacy using Vegf carrying recombinant baculovirus (Bac vegf ) nanohybrids and control Bac carrying no gene (Bac Null ), followed by evaluating its potential to accelerate re-endothelialization and attenuate in-stent restenosis (ISR) in balloon injured dog femoral artery.

Results
Fabrication and characterization of the baculovirus based bioactive stent. For improved gene delivery, as a first step, the anionic Bac particles were surface coated with PAMAM dendrimer to generate the cationic Bac-PAMAM nanocomplexes. The successful formation of Bac-PAMAM nanocomplexes were confirmed by zeta potential analysis of the Bac nanocomplexes as reported in Figure S3a. The data shows that the positively charged PAMAM dendrimers, upon conjugation with the negatively charged baculoviruses, formed nanocomplexes which resulted in an increment of surface charge towards positive. The nanocomplex formation was further confirmed by measuring the particle size of the formed nanocomplexes ( Figure  S3b). The significantly increased size of the Bac-PAMAM (0.05) particles, where value in brackets represent PAMAM concentration in mmol, compared to the free Bac indicates the efficient formation of Bac-PAMAM nanocomplexes, generated by strong electrostatic interactions between baculovirus and PAMAM dendrimers. The excessively bigger size in Bac-PAMAM (0.1) group is probably because of the formation of larger aggregates and clumps of the formed complexes. Transmission Electron Microcope (TEM) images in Figure S3c demonstrates the efficient binding of the rod like baculovirus with PAMAM dendrimers (0.05) leading to the formation of Bac-PAMAM hybrid complex.
In order to encapsulate the baculoviruses, the method of preparation and energy source required for formation of the primary emulsion were studied extensively and optimized in vitro. After optimization (details in supplementary information), we selected double emulsion solvent evaporation method to prepare the virus encapsulated bioactive microparticles 34 . Bac was added to PLGA dissolved in dichloromethane (DCM) and the mixture was either sonicated or homogenized with and without bovine serum albumin (BSA)/glycerol. Hydrophobic (corn oil) solvent was used in the second emulsion in one method (w/o/o) and compared to standard hydrophilic (Polyvinyl alcohol, PVA) solvent method (w/o/w). A significantly higher amount of encapsulation efficiency was noticed in the w/o/o method which confirms that there was less viral loss using oil compared to water in the continuous phase. As evident from Table S1 which summarises the virus encapsulation efficieny and particle diameters from indi-vidual preparation methods, we concluded that a w/o/o double emulsion procedure using mechanical homogenizer can efficiently encapsulate the baculoviruses, and this can be further enhanced by supplementing them with BSA/glycerol to protect the viral activity.
The prepared polymer coated stents containing the PLGA microspheres (by w/o/o method), mixed with Red Nile dye as tracer, were lyophilized and photographed to confirm that the polymeric stent was able to retain the embedded microspheres after stent expansion. Figure 1a shows a stent before and after coating with MS embedded fibrin layers, while the pink color of the stent confirms retention of the loaded MS on the stent surface post stent expansion using balloon catheter. A fluorescence image of the coated stents in Figure 1b further confirms the above findings, where coating of the microspheres was seen on the stent surface. Scanning Electron Microscope (SEM) and Atomic Force Microscope (AFM) images of the MS on stent surface (Figure 1c and S4a) provided further morphological evidences on the surface characteristics of the coated stents demonstrating distributed PLGA MSs embedded on the fibrin multilayered stent struts. Transmission electron microscope (TEM) image of cut section of a virus loaded PLGA microsphere demonstrate the presence of baculovirus inside microspheres which reconfirms the successful microencapsulation.
Preclinical studies have confirmed the safety of using hydrogels, such as fibrin, for coating bomedical devices 35,36 . Although it does not posses the ideal mechanical propoerties and biodegradation features needed for proper stent coating, fibrin has been considered as one of the safest biomaterial for tissue engineering and medical appliactions. With that knowledge, here we aimed to develop a biocompatible stent surface by coating the metallic stent with layers of optimized fibrin hydrogel, impregnated with baculovirus loaded biodegradable PLGA MS. The topmost genipin/fibrin layer serves as the barrier to external damages during crimping on the balloon catheter as well as protects the inner layers from premature virus release during its passage through the lumen of the artery at time of implantation at the desired site. The addition of genipin, as a natural cross-linker to fibrin, also works towards reducing the chance of any inflammatory reactions 37 . Through the stent coating method presented here, the inner surface of the stent (i.e., the mandrel contact surface) was not coated with the polymer. Consequently, there was no fibrin coating on the blood contact side of the stent. In this way we tried to reduce the loss of loaded viruses into the blood stream.
To determine the release kinetics of the encapsulated baculoviruses from the stent surface and modulate their release behavior, two different PLGA concentrations (50 and 100 mg/ml of dichloromethne, DCM) were used using w/o/o emulsion method and compared with standard w/o/w method. In both the methods the stent group with MS prepared from lower PLGA concentration showed much faster and higher percentage of viral release with time compared to the higher concentration ( Figure S4b). Moreover, w/o/o group showed significantly higher release of active virus at lower PLGA concentration than w/o/w group, 24 h post-incubation in Phosphate Buffered Saline (PBS). A further decrease in PLGA concentration led to non-uniform MS formations and hence we restricted to 50 mg/ml DCM PLGA MS formulation protocol using w/o/o method for further studies.
To investigate whether the released Bac MGFP , recombinant baculovirus carrying Monster Green Fluorescent Protein (MGFP) reporter gene, were bioactive and can transduce the human aortic smooth muscle cells (HASMC), the incubation buffers from different release time points were collected, diluted in same proportions and added to HASMC for transduction. After 72 h, the MGFP expressions were quantified in terms of percentage cells transduced. For the study, four different baculovirus formulations using varied concentrations of PAMAM dendrimers were prepared before encapsulating and loading them on the stent surface. The data in Figure 1d showing consistently higher transduction efficiency than other groups, including free Bac MGFP (control). Further increasing the PAMAM concentration in Bac MGFP -PAMAM (1.0) however did not show any significant improvement. Moreover, the percentage of transduction was proportional to hour of incubation buffer. This is because of the higher amount of baculovirus released in the incubation buffer with higher incubation time, although there were no significant differences in transduction efficiencies between 18 h and 24 h in all groups. This indicates that the active viruses were mostly released within 12 h to 18 h from the stent. Furthermore, 72 h cytotoxicity studies on the effect of these different formulations on HASMC show that Bac MGFP -PAMAM can be safely used for cell transduction, although Bac MGFP -PAMAM (1.0) group with high PAMAM concentration showed significantly high cytotoxicity compared to the control free Bac and other experimental groups ( Figure S5).
Importantly, the stent group with microencapsulated Bac-PAMAM also showed significantly higher protection of bioactive baculoviruses against serum inactivation compared to stents with microencapsulated Bac, and non-encapsulated Bac and Bac-PAMAM stents. Our data confirms that even PAMAM coating can also protect the Bac against serum inactivation to some extent, which is further enhanced by encapsulating them inside PLGA microspheres ( Figure S6a). Additionally, the preservation of bioactivity of the stent for at least 3 months stored at 280uC, as illustrated in Figure  S6b, can be of significant logistic advantages under real life clinical settings as the stored stent can be of immediate off-the-shelf use for any patient undergoing angioplasty and stenting without delay.
In vitro biofunctionality of the developed bioactive stent. To study the functionality of the Vegf proteins, expressed by the treated HASMC, different functionality assays were performed using in the conditioned media (CM) on human umbilical vein endothelial cells (HUVEC) which have high mitotic and chemotactic response to Vegf. Thus, CM were collected from Bac Vegf -PAMAM (0.5) and Bac Vegf transduced HASMCs at regular intervals, and used to quantify the Vegf release profile using Vegf ELISA analysis. CM from non-transduced cells was taken as the negative control. The data, as shown in Figure 2a, demonstrates rapid expressions of hVegf in Bac Vegf -PAMAM and Bac Vegf groups in the first 4 days which gradually decreased over the week. Although Vegf expression decreased considerably in Bac Vegf group after three weeks, Bac Vegf -PAMAM group maintained significantly higher Vegf expression. The bioactivities of the released Vegf in media from the Bac Vegf -PAMAM (0.5) transduced HASMC were evaluated in vitro by observing the proliferative capacity of the HUVEC. Cell Proliferation assay was used to assess the proliferation capabilities of the HUVEC treated with CM (containing secreted Vegf) from experimental samples, collected on day 4 post transduction. As shown in Figure 2b, groups Bac Vegf and Bac Vegf -PAMAM showed significantly high HUVEC proliferation on day 3 compared to unstimulated control, with Bac Vegf -PAMAM exhibiting the highest proliferation (3.98 6 0.19 3 10 4 cells). As expected, group treated with antibodies against Vegf showed no proliferative effects proving that it was because of the bioactivity of released Vegf from genetically modified SMCs that contributed to the drastic HUVEC proliferations. This proliferation rate was directly dependant on the amount of the Vegf released which explains why Bac Vegf -PAMAM demonstrated better results than Bac Vegf and unstimulated control groups.
Similar results were noticed in the wound healing assay where the abilities of CM from different experimental groups, collected on day 4 post transduction, to promote HUVEC migrations were measured. As depicted in Figure 2c-d, stimulation of wounded HUVEC monolayer with CM from Bac Vegf -PAMAM exhibited significant healing of wounded area (89.4 6 7.5%) compared to CM from unstimulated control (12.4 6 2.4%) and Bac Vegf (61.7 6 5.5%). Pre-incubation of CM with the neutralizing anti-Vegf antibodies completely hindered Bac Vegf -PAMAM CM induced wound healing, clearly suggesting that chemotactic signals from Vegf are required for proper wound healing effect.
In addition, the biologic activities of the CM were also evaluated by HUVEC tube formation assay. As illustrated in Figure 2e-g, cells treated with CM from Bac Vegf -PAMAM induced significantly enhanced effect on HUVEC capillary network formation as compared to the cells treated with CM from Bac Vegf and unstimulated groups. The relative tubule length was significantly enhanced in Bac Vegf -PAMAM and Bac Vegf group compared to control (141.5 6 7.9 and 100 6 13.6 vs 47 6 2.1). Also, to assess the extent of impact of Vegf present in the CM on HUVEC tube formation, anti-Vegf antibody was added to the supernatant with released CM. A significantly reduced tube formation was observed which provides evidence of the strong pro-angiogenic nature of Vegf present in the CM. Similar results were also noticed when total capillary tubule numbers were quantified in Bac Vegf -PAMAM and Bac Vegf group and compared to unstimulated control (15. In vivo efficacy of the bioactive stent in balloon denuded canine femoral artery. The in vivo stent implantation procedure has been demonstrated in Figure 3a, where the dogs underwent bilateral femoral artery denudation by balloon angioplasty, followed by stent deployment at the injured site 29,38 . Stented femoral arteries transduced with Coated (1) stent, carrying Bac vegf -PAMAM, Coated (2) stent carrying Bac Null -PAMAM, and uncoated bare metal stents (negative control) were harvested 14 days after stent placement. Following RT-PCR analysis, Vegf signal expression could be observed in all the proximal, middle and distal sections of the Bac Vegf -PAMAM [Coated(1)] transduced arteries examined (n 5 3 stents), while nothing was detected in Bac Null -PAMAM [Coated (2)] transduced arteries (Figure 3b). Although, the RT-PCR product from 2 weeks Bac Vegf -transduced tissue samples demonstrated the presence of the appropriate-sized band for Vegf in the stented artery sections, no bands were detected from 16 weeks samples proving that the transgene expression is transient and disappears over time. Moreover, artery sections from 1 cm proximal and distal to the stented artery did not show any transgene expression, thus indicating that the gene delivery was localized and restricted to the stented region where the virus containing polymer coated strut touches the inner wall of the artery. The result was reconfirmed by immunostaining for expressed Vegf in the stented artery. This baculovirus nanohybrid mediated Vegf transgene expression was localized to the areas around the stent struts in the intimal and medial layers (Figure 3c), where the polymer of the stent touches the arterial surface; in contrast, there was no Vegf staining observed in the stented arteries treated with Coated (2) stents.
Endothelial regeneration was determined using three independent methodologies: Evan's blue staining, scanning electron microscopy and histology. Two weeks after balloon injury and stent placement, endothelial regeneration was assessed in the animals treated with Coated (1) and Coated (2) stents using Evans blue staining (n 5 3 stent). Luminal staining for Evans blue demonstrated that the balloon angioplasty and stenting procedure completely denuded the femoral artery in Control (2) stent while Coated (1) showed marked recovery at week 2 ( Figure S7a). As evident from Figure S7b, re-endothelialization was significantly greater in the Coated (1) vessels (55.36 6 4.64%) in comparison with Coated (2) control vessels (37.5 6 6.51%, P , 0.05).
Furthermore, at 2 weeks and 16 weeks after stent deployment, SEM pictures of the inner surface of the stented arteries were taken. Cells consistent with endothelial morphology were noted on the surface of Bac Vegf treated stent struts on week 2, which completely covered the stent surface in a uniform way after 16 weeks (Figure 3d). In contrast, an irregular rough surface, mainly comprising of the exposed stent strut and partially covered neointima tissues, was noticed in the arteries with Bac Vegf treated stent on both week 2 and week 16. Histological assessment of endothelial regeneration, sixteen weeks post stent placement, demonstrated a significant difference in endothelial regeneration between the three groups, Coated (1), Coated (2) and Uncoated bare metal negative control stents. The percentage of endothelial cells observed in the Coated (1) vessels was significantly higher than in the control vessels (93.5 6 5.2% vs 75.4 6 9.3% and 72.4 6 4.1; P , 0.05) as shown in Figure 3e.
All vessels were angiographically and histologically patent throughout the period of study. Stent malapposition was also not detected in any animal. 16 weeks after site-specific Vegf gene transfer, the stenotic area was significantly reduced in Coated (1) compared to control uncoated group stents (54.58 6 14.1% vs 85.4 6 10.14%; P , 0.05) as analyzed by angiography (Figure 4a and S7d) using standard methods 39 . The extent of vessel injury at the stent site, as analysed by histological examinations 20 , was similar in all the groups as determined by injury score (Coated1 1.13 6 0.14 vs Coated-1.  Discussion To the best of our knowledge, this study for the first time demonstrates that insect cell generated recombinant baculoviruses can be efficiently used for stent based gene therapy applications, such as inhibition of post angioplasty ISR. This work has focused on design-ing a new anti-restenosis strategy using a baculovirus based geneeluting stent which ameliorates the biology of the vessel wall, in contrary to drug-based vascular disruption, by accelerating endothelial recovery after stent deployment. The data substantiate our hypothesis by demonstrating significantly improved endothelial recovery assessed by Evan's blue staining and scanning electron microscopy, and restricted ISR illustrated by histomorphometric and angiographic analysis over a study period of 4 months post implantation.
So far, baculoviruses have long been considered as a biologically safe gene delivery vector for mammalian cells 27 . Although it has been shown to effectively transduce mammalian cells in vitro, its full potential for in vivo gene therapy is not yet proven. Through this work, we have developed methods to protect it from host immune system as well as enhance its transduction ability which are its main limitations for in vivo applications. Moreover, its transient nature of recombinant gene expression is ideal in such a case where the expression ceases once the job is done. This also confirms the safety of the system to avoid unnecessary side effects in the post healing period. The availability of easily grown, suspension-ready host insect cells also facilitates the scale up process. As such, it is easier to obtain high titre of recombinant baculoviruses than other mammalian viruses. This makes the baculovirus expression system logistically beneficial and economically viable for process development and large production capacity.
This technology can be a potential alternative to the currently available DESs, which mainly use anti-proliferative drugs. These drugs, while preventing smooth muscle cell overgrowth, also lead to incomplete regeneration of the damaged endothelial layer. The implication of the newly formulated bioactive stent will enhance the Vegf based rapid re-endothelialization mechanism 40 leading to better recovery process and reduced chance of restenosis. Moreover, the bioactivity of stent can be retained for at least 3 months once stored in the freezer, which further extends its commercial viability. In addi- tion, the invention of the method documented here to effectively encapsulate functionally active baculoviruses also opens up the potential applications of recombinant baculoviruses for wider gene therapy applications. However, the present work does not demonstrate whether this system can inhibit long term in-stent thrombosis. An important area which needs further research is to characterize the in vivo dose dependant safety and efficacy analysis of the stent. In the present study we used a comparatively high dose of baculovirus nanohybrids. This is because, as a first study of its kind with no prior literature data, we aimed to achieve maximum transgene delivery at the stented site to achieve high growth factor expression, which can be a key determinant for the degree of vascular healing. Baculoviruses, with surface functionalized targeting ligands, can be the next step for this technology where the released viruses can site-specifically repair the denuded regions with minimal dosage. Furthermore, this advanced baculovirus based gene delivery system can be integrated with existing DES technology for combined stent based therapy of restenosis and thrombosis.
Taken together, the major findings of the current work capitalize on the development of this novel approach where local delivery of Bac Vegf -PAMAM nanobiohybrid vector, entrapped within PLGA MSs embedded layer-by-layer on fibrin-coated stent, can significantly enhance endothelial recovery and thereby block subsequent neointimal proliferation and ISR. Importantly, this transient nature of insect cell-originated baculovirus nanohybrid gene delivery system makes it a prospective biologically safe material for advanced translational research.

Methods
Generation of recombinant baculoviruses. The recombinant Vegf baculoviruses (Bac Vegf ) were generated by cotransfection of Sf9 insect cells with linearized baculovirus DNA (BD Baculogold) along with purified recombinant transfer vectors using cellfectin (Invitrogen Life Technologies, Carlsbad, CA) transfection reagent as mentioned in earlier study 24 . The viral titre [plaque forming unit (pfu)/mL] of the amplified viral stock was then determined using the Baculovirus Fast Plax Titer Kit (Novagen, Madison, WI) according to the manufacturer's protocol. Similarly, baculoviruses carrying MGFP gene (Bac MGFP ) and no transgene (Bac Null ) were also generated.
Formulation of bac-PAMAM nanohybrids. PAMAM dendrimer (generation 0) with ethylenediamine core (MW: 516.68) containing 4 surface primary amino groups was procured from Sigma Chemicals and resuspended in phosphate buffered saline (PBS). The concentration of PAMAM peptide stock solution was initially adjusted to 10 mmol in PBS solution. In order to form the PAMAM-baculovirus nanocomplex, the solutions of the PAMAM and baculoviruses were first brought to room temperature and adjusted to desired concentrations. Then the PAMAM solution and baculovirus solution were mixed according to a desired PAMAM/virus ratio (0, 0.01 mmol, 0.1 mmol, 0.5 mmol and 1.0 mmol PAMAM molecules per 10 8 baculovirus). The mixture was incubated at room temperature for 30 min to form complexes, with gentle vortexing from time to time. The mixture was further centrifuged at 24,000 rpm for 45 min, and the pellet containing the heavier Bac-PAMAM nanocomplex was collected, leaving the unreacted excess dendrimer in the supernatant. The collected Bac-PAMAM was washed twice using PBS using the same centrifugation process. For every experiment, the preparation was freshly made.
Preparation of bioactive microparticles containing baculovirus nanohybrids. In order to encapsulate the the viral particles by water-oil-oil (w/o/o) double emulsion and solvent evaporation method 34 , 5 3 10 13 pfu was resuspended in 100 ml of PBS containing 10% glycerol and 50 mg/ml BSA. The primary waterr in oil emulsion solution was prepared by homogenizing it in 1 ml of DCM containing 50 or 100 mg of PLGA [poly(D,L-lactic-co-glycolic acid)] and 1 ml acetonitrile for 1 min at 10,000 rpm. The resulting primary emulsion was added to 5 ml of corn oil containing 2% span80 and homogenised for 3 min at 14000 rpm to form the secondary w/o/o emulsion. This solution was further agitated with a magnetic stirring bar in 10 ml of of corn oil containing 2% span80 for 4 h to evaporate the the DCM. The hardened PLGA microspheres were centirfuged at 9000 g for 10 min, washed thrice with PBS and stored temporarily at 4uC. In a similar way, PLGA microparticles were prepared by water-oil-water (w/o/w) method 34 . The method of PLGA microencapsulation of virus particles are demonstrated schematically in Figure S2.
Preparation of bioactive stent using layer-by-layer coating. Firstly, the baculovirus was coated with PAMAM (0.5 mmol) and formulated in PLGA microspheres by w/o/ o method. The prepared PLGA microspheres were resuspended in 5 mg/ml of bovine plasma derived fibrinogen, supplemented with aprotinin (20 mg/ml) to reduce fibrin degradation, and loaded in a 3 ml syringe with a 0.2-mm nozzle. The balloon expandable bare metal stainless steel stents with basic dimensions of 16 mm 3 3.5 mm (Liberte Monorail stent, Boston Scientific, Mississauga, Ontario) was first mounted on a PTFE (polytetrafluoroethylene) mandrel that was driven by a rotator. The loaded aqueous fibrinogen mixture in the syringe was then gradually drip-coated on the surface of the mounted stent layer by layer (0.2 ml of Bac loaded aqueous fibrinogen per layer). In between every layer, 0.05 ml of thrombin solution (20 U/ml) was added all over the stent surface using a micropipette with 200 ml micropipette tip and waited for 15 min to form thin fibrin gel layer. Polymerization of the fibrin occurred around the stent which completely encased the stent ( Figure S1). The stent was subsequently coated with a top layer of fibrinogen (2.5 mg/ml) cross-linked with genipin (0.045 mg/mL final) followed by polymerization with thrombin. The process resulted in a microsphere impregnated fibrin coated stent loaded with 5 3 10 12 pfu Bac. The device was then crimped onto standard collapsed angioplasty balloon and delivered to the femoral artery. [The detailed methods for in vitro transduction and biofunctional studies are present in the Supplementary information.] In vivo surgical procedures for arterial injury and stent implantation. All procedures were in compliance with the Canadian Council on Animal Care and McGill University animal use protocol, following all the ethical guidelines for experimental animals. Adult beagle dogs (Marshall Farms, North Rose, NY) weighing 9.5 to 11 kg were used in this study. A total of 30 stents were implanted bilaterally in deep femoral arteries of 15 dogs in a randomized, blinded fashion following denudation of the arterial endothelial wall, with contralateral arteries receiving stents from different groups. The animals were divided into three groups: fibrin coated stent loaded with microencapsulated PAMAM-Bac Vegf virus [Coated (1); n 5 11 stents per group)], fibrin coated stent loaded with microencapsulated PAMAM-Bac Null virus [Coated (2); n 5 11 stents per group] and bare metal stent (Uncoated; n 5 8 stents per group). To confirm the transgene delivery and re-endothelization, animals from first two groups were sacrificed (n 5 3) after 2 weeks and arteries were harvested. The remaining animals were sacrificed after 16 weeks for further analysis.
Denudation of the femoral arterial endothelial layer and stent implantation was performed according to the procedure performed in earlier studies 38 . The selected portion of artery lumens were flushed with PBS to avoid mixture with arterial blood. On the basis of angiograms, femoral segments with comparable diameters were selected on both legs so that the stent-to-artery diameter ratio remains approximately 1.3. After an arteriotomy, a fogarty arterial embolectomy balloon catheter (Edwards Lifesciences Canada Inc, Ontario) was infused through the saphenous arteries and advanced to the two preselected femoral segments and secured with a tie as mentioned elsewhere 29 . This was followed by severe balloon injury of the inner lumen of the artery with inflated balloon (balloon/artery ratio 1.251) to induce endothelial abrasion. Eventually the balloon catheter mounted stents were deployed at the site. 3 animals were euthanized at week 2, while the remaining at week 16.
Analysis of transgene expression. RT-PCR analysis. At week 2 post stent placement, the stented femoral arteries (n 5 3 stents/group) were harvested from Coated (1) and Coated (2) groups and divided laterally into three sections (proximal, mid and distal) after removing the stents. A part of the sections were used to detect hVegf transcript by RT-PCR, while the remaining parts were used for to detect the hVegf protein by immunostaining and reendothelization by staining. Similarly, sections were also collected at week 16. PCR amplification was performed on the reverse transcribed product using Taq DNA Polymerase (Invitrogen) and forward primer (59CTTGCCTTGCTGCTCTACCTCC39) and reverse primer (59GCTGCGCTGATAGACATCCATG39) for hVegf gene (112-bp product). Amplifications were carried out for 25 cycles at 94uC for 35 s (denaturation), 57uC for 35 s (annealing), and 72uC for 25 s (extension).
Immunohistochemical studies. For immunostaining rabbit anti-hVegf (Santa Cruz Biotechnology Inc., Santa Cruz, California) primary antibodies were used. Donkey anti-rabbit IgG-FITC was used as secondary antibody. The proportions and intensities of FITC-positive regions in the tissue sections, as seen under fluorescence microscope, gave a qualitative idea of the relative amount of hVegf expressed in the stented vascular tissue regions due to the transgene delivery from stent platform.
Assessment of stent re-endothelialization. Evans blue staining. Two weeks after stent deployment animals were anesthetized [n 5 3 , Coated (1) and (2)] and a portion of the harvested artery was incubated in 1% Evans blue (Sigma-Aldrich) for 15 min, followed by washing ,fixing and examining longitudinally using image software as mentioned elsewhere 20 .
Scanning electron microscope. Re-endothelialization was also assessed at 2 week and 16 week post implantation (n 5 3 each). Retrieved stents were washed with saline, fixed in 4% paraformaldehyde and longitudinally incised as above and examined using scanning electron microscopy (SEM).
Histological assessment. Stents were retrieved after twelve weeks and harvested vessels were embedded in methylmethacrylate plastic (Accel Lab Inc, Quebec, Canada and McGill SAIL Lab). After polymerization, the proximal, mid and distal sections of the stents were cut into 5 mm sections and stained with hematoxylin and eosin. Reendothelialization was assessed directly under the microscope by a histologist blinded to treatment. Endothelial coverage was expressed as the percentage of the average lumen circumference covered by the endothelial cells 20 .
Assessment of ISR. Angiogram analysis. Initial and follow-up angiograms were performed with fluoroscopic angiography (GE Stenoscop) using Omnipaque Iohexol contrast dye in anterior oblique projection and percent diameter stenosis at follow-up was calculated by (minimal stent diameter at follow-up/the mean diameter of the stent at full expansion) 3 100, using standard procedure as described elsewhere 39 . Sections were also used to evaluate the presence of inflammation and vessel wall injury at the stented sites of all the groups by injury and inflammation score method 20,39 .
Morphometric analysis. Retrieved vessels were embedded in methylmethacrylate plastic, cut into thin sections as mentioned earlier, and stained with elastic Van Gieson stain. All sections were examined by light microscopy and photographed to quantify the neointima formation and stenotic area using methods mentioned elsewhere 20 .
Statistical analysis. Quantitative variables are presented as mean 6 Standard Deviation (SD) from independent experiments as described in the figure legends. Statistics were performed using student's t-test or one-way ANOVA by Bonferroni's multiple comparison post-hoc test. All statistical analyses were performed with Prism 5 (GraphPad Software). P value ,0.05 was considered significant.