These studies report on reversible attachment of adenoviral gene vectors to coatless metal surfaces of stents and model mesh disks. Sustained release of transduction-competent viral particles contingent upon hydrolysis of cross-linkers used for vector immobilization results in a durable site-specific transgene expression in vascular cells and in stented arteries.
In-stent restenosis presents a major complication of stent-based revascularization procedures widely used to re-establish blood flow through critically narrowed segments of coronary and peripheral arteries. Endovascular stents capable of tunable release of genes with anti-restenotic activity may present an alternative strategy to presently used drug-eluting stents. In order to attain clinical translation, gene-eluting stents must exhibit predictable kinetics of stent-immobilized gene vector release and site-specific transduction of vasculature, while avoiding an excessive inflammatory response typically associated with the polymer coatings used for physical entrapment of the vector. This paper describes a detailed methodology for coatless tethering of adenoviral gene vectors to stents based on a reversible binding of the adenoviral particles to polyallylamine bisphosphonate (PABT)-modified stainless steel surface via hydrolysable cross-linkers (HC). A family of bifunctional (amine- and thiol-reactive) HC with an average t1/2 of the in-chain ester hydrolysis ranging between 5 and 50 days were used to link the vector with the stent. The vector immobilization procedure is typically carried out within 9 hr and consists of several steps: 1) incubation of the metal samples in an aqueous solution of PABT (4 hr); 2) deprotection of thiol groups installed in PABT with tris(2-carboxyethyl) phosphine (20 min); 3) expansion of thiol reactive capacity of the metal surface by reacting the samples with polyethyleneimine derivatized with pyridyldithio (PDT) groups (2 hr); 4) conversion of PDT groups to thiols with dithiothreitol (10 min); 5) modification of adenoviruses with HC (1 hr); 6) purification of modified adenoviral particles by size-exclusion column chromatography (15 min) and 7) immobilization of thiol-reactive adenoviral particles on the thiolated steel surface (1 hr). This technique has wide potential applicability beyond stents, by facilitating surface engineering of bioprosthetic devices to enhance their biocompatibility through the substrate-mediated gene delivery to the cells interfacing the implanted foreign material.
The effectiveness of gene therapy as a therapeutic modality is hampered by the poor targeting capacity of gene therapy vectors1,2. The lack of proper targeting results in sub-therapeutic levels of transgene expression at the target location and leads to a wide dissemination of vectors to non-target organs3, including those responsible for mounting immune responses against both the vector and encoded therapeutic product4,5. One potential means to offset the promiscuity of transduction and to promote targeting is to introduce gene vectors at the desired location in a form that precludes their free dissemination via blood and lymph. Typically, such efforts rely on a locally injectable delivery systems comprising of either viral or non-viral vectors admixed with fibrin, collagen or hyaluronic acid hydrogel matrices6-10 that are capable of transiently sustaining gene vectors at the injection site by physically entrapping them in a polymeric network.
Another generally accepted paradigm for localized gene therapy utilizes immobilization of gene vectors on the surface of implanted prosthetic devices11,12. Permanent medical implants (endovascular, bronchial, urological and gastrointestinal stents, pacemakers, artificial joints, surgical and gynecological meshes, etc.) are used yearly in tens of millions of patients13. While generally effective, these devices are prone to complications that are inadequately controlled for by current medical practices14-17. Implantable prosthetic devices present a unique opportunity to serve as proxy platforms for localized gene therapy treatment. From the pharmacokinetic standpoint, surface derivatization of medical implants with relatively low input doses of gene vectors results in achieving both high local concentrations of gene vectors on the implant/tissue interface and slowing the kinetics of their elimination from this location. As a consequence of protracted residence and enhanced uptake by the targeted cell population, vector immobilization minimizes spread of the gene vector. Thus the inadvertent inoculation of non-target tissues is reduced.
Surface tethering of gene vectors on implantable biomaterials (also termed as substrate-mediated gene delivery or solid phase gene delivery) has been implemented in cell culture and animal experiments using both specific (antigen-antibody18-20, avidin-biotin21,22) and non-specific23-26 (charge, van der Waals) interactions. The covalent attachment of vectors to the surface of the implanted device has been previously considered as non-functional since excessively strong bonds with the surface preclude vector internalization by target cells. Recently it was demonstrated that this limitation can be overcome through the use of spontaneously hydrolysable cross-linker used as the tethers between the modified metallic surface of the stent and capsid proteins of the adenoviral vector27,28. Moreover, the vector release rate and time course of transgene expression in vitro and in vivo can be modulated with the use of hydrolysable cross-linkers exhibiting different kinetics of hydrolysis28.
The present paper provides a detailed protocol for the reversible covalent attachment of adenoviral vectors to activated metal surface and introduces a useful experimental setup for studying ensuing transduction events in vitro in cultured smooth muscle and endothelial cells and in vivo in the rat carotid model of stent angioplasty.
1. Preparation of Cy3-labeled Adenovirus for the Release Experiments
2. Activation of Metal Samples
3. Adenovirus Activation and Metal Surface Immobilization
4. Quantification of Surface-associated Ad Vector by PCR
5. Release Kinetics of Hydrolysable Cross-linker-tethered Vector Particles from the Model Steel Mesh
6. Transduction of Cultured Cells by Mesh-immobilized Ad Vectors
7. Validation of Preserved Transduction Capacity at Delayed Time Points
8. Ad-eluting Stent Deployment in the Rat Carotid Model of Stent Angioplasty
9. Bioluminescence Imaging of Arterial Gene Expression
Vector Release Experiments
Tethering of adenoviral vectors to the surface of implants, including interventional devices such as endovascular stents, approximates the vector to the disease site, partially obviating the lack of vectors’ physical targeting. However, to be able to achieve therapeutic effects via the transduction of target tissue, the vector must be released from the surface (Figure 2). The use of hydrolysable cross-linkers was hypothesized to allow 1) effective attachment of vectors to polybisphosphonate-modified, thiol-installed metal implants and 2) sustained release mediated by hydrolysis of the cross-linker.
The number of genomic copies of AdeGFP vector associated with the mesh disks by the end of the derivatization procedure is determined by PCR of viral DNA with eGFP-specific primers (Figure 3). Depending on the specific cross-linker used, the amount of bound Ad vector should vary between 3.5 x 109 and 5.7 x 109 particles/mesh. This amount is in fair correspondence with the estimated maximal capacity (2.5 x 109) of a single mesh disk with a total area of 0.25 cm2 to accommodate 100-nm Ad particles in a monolayer arrangement. Since the starting amount of Ad vector at the step of cross-linker modification is 5 x 1011 particles, only ~1% of the modified vector is eventually immobilized on the PABT/PEIPDT-derivatized stainless steel surface.
Both the rate of in-chain ester hydrolysis and the number of links between the vector and the biomaterial are expected to affect the overall kinetics of Ad vector release from the surface.
To facilitate the release experiments, adenoviral particles can be pre-labeled with Cy3 fluorophore (Protocol; section 1) prior to cross-linker modification and surface immobilization (Protocol; section 3). The decrease of surface-associated fluorescence over time assessed concurrently by fluorometry (Figure 4A) and fluorescence microscopy (Figure 4B) can then be used as an indirect method for monitoring vector release into physiological buffer. Ad vectors attached through RHC and IHC linkages should demonstrate significantly faster release in comparison with their SHC- and NHC-attached counterparts. In the given example, a 45% and a 39% loss of surface-associated vector was observed with RHC and IHC-tethered vector by day 30, respectively, while less than 20% of tethered vector was observed to be released in SHC and NHC-immobilized samples (Figure 4A).
Additionally, Ad vectors immobilized by using different concentrations of the same HC and thus presumably having differing numbers of tethers between the vector and metal substrate should have dissimilar kinetics of vector release. Indeed, vector modification using 0.1 mM RHC resulted in significantly faster release rates than observed for vector immobilized with 0.5 mM RHC (Figure 4A). The fluorescence microscopy data (Figure 4B) correlates well with the fluorometry-based quantitative vector release analysis, demonstrating a faster and more profound decrease of surface-associated fluorescence with mesh samples formulated using IHC-, RHC- and especially low concentration RHC-tethered vectors in comparison with their NHC- and SHC-tethered counterparts.
In vitro Transduction with Mesh Immobilized Ad Vectors
By being compatible with both quantitative expression analysis (fluorometry) and spatial observations (fluorescence microscopy), AdeGFP is the preferred reporter vector to monitor transduction in SMC and endothelial cell cultures treated with both free and mesh-immobilized Ad vectors. Chemical modification of Ad capsid with RHC (Figure 5) as well as with other cross-linkers (not shown) at concentrations exceeding 75 µM significantly impairs transduction effectiveness of non-immobilized vector in vitro, most probably due to masking and reconfiguring of fiber knob domains utilized by Ad for binding to the cells through the Coxsackie-Adenovirus receptors (CAR). However, the role of knob-CAR interactions is less vital for the transduction with substrate immobilized vector since the vector is retained in the vicinity of targeted cells by tethering to substrate. In cell culture experiments regardless of the HC type, mesh-associated Ad vectors are consistently capable of spatially restricting transduction events to cells located within 300-500 µm from the mesh borders (Figures 6A and 6C). Typically the peak levels of transgene expression are achieved 3-5 days after the placement of AdeGFP-loaded meshes into SMC (Figures 6A and 6B) or BAEC (Figures 6C and 6D) cultures. At the same vector load, peak eGFP expression levels increase in the following sequence: NHC-, SHC-, IHC- and RHC-tethered vector, reflecting the differences in their respective release kinetics profiles (Figure 4). Furthermore, more durable transgene expression is observed in cells treated with IHC-formulated meshes in comparison to cells treated with the RHC-formulated counterparts (Figure 6D).
The utilized cell culture system presents a convenient set-up for the investigation of delayed transduction events brought about by vector particles released from the metal surface at or later than 48 hr after mesh placement. In this application, after determining eGFP expression by fluorometry and fluorescence microscopy at 48 hr after commencement of transduction, BAEC are trypsinized and aspirated out of the wells without disturbing the meshes. Freshly passaged non-transduced BAEC are then re-plated over the partially released meshes. “New” transduction events caused by the virus particles released from the mesh carrier after the 2 day time point are then assessed by fluorescence microscopy and fluorometry (Figures 7A and 7B). Robust eGFP expression demonstrating the characteristic spatial restriction to a mesh locale is typically observed in these “new” cultures (Figure 7), confirming a well-preserved transduction capacity of the mesh-bound Ad vectors even after 48 hr exposure to complete cell culture medium at 37 °C.
In vivo Studies
To examine potential effects related to the use of different HC on arterial transduction, animals implanted with AdLuc-eluting stents prepared with RHC-, IHC-, SHC- and NHC-modified vector underwent bioluminescent imaging 1 and 8 days after stent deployment (Figure 8). The imaging was carried out following local perivascular administration of luciferin (2 mg) co-formulated with Pluronic gel. This formulation is a viscous fluid when chilled on ice, but undergoes immediate phase transition to gel when brought in contact with tissue at 37 °C. Preliminary experiments (not shown) demonstrated that perivascular delivery of luciferin results in more stable and reproducible luminescence signals in AdLuc-transduced vascular tissue when compared to systemic (intraperitoneal or intravenous) luciferin administration.
If virus tethering to the stent and the stent deployment surgery were technically sound, a well-defined signal corresponding to the stented segment of rat carotid artery is emitted and recorded. One day after stent implantation, animals treated with RHC-formulated AdLuc stents typically exhibit the highest luminescence signal, followed by the rats receiving IHC- and SHC- formulated stents (Figures 8A and 8B). No perceptible signal is observed in the group of rats implanted with stents prepared using NHC-tethered AdLuc, underscoring the importance of unimpeded vector release from the stent for effective transduction of vascular tissue (Figures 8A and 8B). By day 8, the average intensity of luciferase expression with RHC-formulated stents drops several-fold, while it increases 1.4- and 1.8-fold with IHC- and SHC-formulated stents, respectively (Figures 8A and 8B). This finding is consistent with the hypothesis of the more durable vascular transduction with gene-eluting stents exhibiting a slower kinetics of vector release from the stent surface.
Figure 1. Structural formulas of cross-linkers (NHC, SHC, IHC and RHC) used throughout the described studies (modified with permission28).
Figure 2. A scheme representing Ad vector tethering to a PABT/PEIPDT-modified stainless steel surface via a HC and subsequent release of the vector upon cross-linker hydrolysis (modified with permission28).
Figure 3. PCR-based quantification of AdeGFP immobilized via cross-linker tethering on the surface of PABT/PEIPDT-modified stainless steel meshes. The stainless steel meshes were formulated with AdeGFP immobilized via RHC, IHC, SHC, or NHC (n = 3 for each condition). After proteinase K treatment, viral DNA was eluted and purified using MinElute columns. Amplification of viral DNA using eGFP-specific primers was carried out in a 7500 Real-Time PCR engine and was detected with Sybr Green. Data normalization was based on a calibration curve prepared with a known amount of non-immobilized AdeGFP(modified with permission28).
Figure 4. Release kinetics of Ad vector tethered to metal substrates with HC. Stainless steel meshes were derivatized with ~2 x 109 Cy3-labeled Ad particles attached to the PABT/PEIPDT-modified surface after modification with 500 µM NHC, SHC, IHC, RHC and 100 µM RHC (designated as RHC-L). The release of fluorescently labeled Ad-particles was studied at the indicated time points by (A) well-scan fluorometry at 550/570 nm and (B) fluorescent microscopy (rhodamine filter set; original magnification 200X). Fluorometry results are presented as means ± SEM, n = 8-10; p < 0.001 for all comparisons between the NHC and SHC vs IHC, RHC and RHC-L groups were determined by Anova with a post-hoc Tukey test (modified with permission28).
Figure 5. Transduction effectiveness of non-immobilized AdeGFP following modification with RHC. Rat embryonic aorta-derived SMC (A10 line) were transduced at a MOI of 1,000 with either unmodified AdeGFP or vector modified with RHC as indicated. A reporter expression was determined fluorometrically (485/535 nm) 48 hr after transduction (modified with permission27).
Figure 6. Transduction of cultured SMC and BAEC with Ad vector immobilized to stainless steel meshes with HC. Meshes were derivatized with ~2 x 109 AdeGFP particles appended via NHC, SHC, IHC, and RHC. Meshes were then individually placed on top of sub-confluent A10 (A, B) and BAEC (C, D) monolayers. Transduction (expressed as eGFP expression levels) of cells treated with Ad vector-tethered meshes was assessed by fluorescence microscopy (A, C; FITC filter set; original magnification 100X) and well-scan fluorometry (B, D). Fluorometry results are presented as means ± SEM, n = 4 (modified with permission28). Please click here to view a larger version of this figure.
Figure 7. Transduction competence of substrate immobilized Ad vectors at delayed time points. AdeGFP was immobilized on steel meshes through NHC, SHC, IHC or RHC tethers. The meshes were placed on the subconfluent BAEC monolayers. Two days after placement, BAEC were trypsinized and removed without disturbing the meshes. New, non-transduced BAEC were then seeded over the meshes. AdeGFP transduction competency was measured by eGFP expression using fluorescent microscopy (A; FITC filter set; original magnification 100X) and fluorometry (B). Measurements were taken at indicated days, representative images were used and fluorometry results are means ± SEM, n = 4 (modified from with permission28). Please click here to view a larger version of this figure.
Figure 8.Transduction efficiency of stent-immobilized AdLuc in vivo. AdLuc (1.3 x 1010 particles) was tethered to endovascular stents via NHC, SHC, IHC or RHC (n = 3-4 for all groups). Transduction efficiency of AdLuc eluted from stents was measured by bioluminescence imaging (IVIS Spectrum) 1 and 8 days post stent deployment (A), and plotted using a logarithmic scale (B) (modified from with permission28). Please click here to view a larger version of this figure.
The presented protocol describes an operational method for substrate mediated gene delivery achieved through reversible attachment of adenoviral vectors to coatless stainless steel surfaces. While developed for the specific purpose of stent-based gene therapy of vascular restenosis, this technique has much broader applications in the areas of biomaterials, biomedical implants and gene therapy.
Although presented studies have solely utilized stainless steel as a prototypical metal substrate, PABT binds other metal alloys such as cobalt/chromium and nitinol with comparable affinity (not published). The indiscriminant interaction of PABT with metals significantly expands the number of potential biomedical applications for this technology31. Moreover, the use of HC-tethered gene vectors, albeit requiring other methods of thiol installation onto material surface, is feasible with other types of biomaterials.
While Ad gene vectors achieve robust transduction in many human cell types and tissues, their clinical significance is limited by strong inflammatory and immune responses elicited by Adenoviridae4. To this end, prolonged (4 months) arterial expression with gene eluting stents formulated using HC modification of luciferase-expressing adeno-associated viral vectors was recently shown (not published).
The methodology is robust. Small deviations from the main protocol do not generally lead to significant alterations of gene expression in vitro and in vivo. Nevertheless, several critical issues that may affect outcomes were identified.
First, as the name suggests, hydrolysable cross-linkers are cleavable upon reaction with water due to hydrolysis of the in-chain ester. Moreover, the amine-reactive moiety, N-sulfosuccinimidyl is hydrolysable as well. Cross-linkers should be stored under an argon atmosphere at -20 °C to preserve their activity. For the same reasons, after preparing solutions of HC (protocol section 3.2) proceed immediately with the addition of HC solutions to the virus preps.
Second, thiol groups that are formed on the metal surface during several intermediate steps of the presented bioconjugation sequence are rapidly oxidized by ambient air oxygen. To prevent this, keep the vulnerable metal samples submerged in degassed water at all times.
Third, a perfect alignment of the mesh with the bottom of the well is required for successful transduction. Do not bend mesh disks during handling.
Fourth, avoid excessive rubbing of gene-eluting stents against the Teflon sheath and vessel wall during stent insertion to prevent dislodging of surface-associated Ad vectors.
Covalent attachment of gene vector to the surface of implantable biomaterials has an apparent advantage of better control over the vector release kinetics than realizable with non-covalent tethering of the vectors. In the setting of gene delivery from stent platform most studies report complete release of viral and non-viral vectors within 1-7 days after stent deployment32-34 . On the other hand, Ad vector immobilization via HC demonstrated release of only 20-45% of the vector load in the first month (Figures 4A and 4B). Moreover, a partial modulation of release and transduction kinetics is achievable with the use of different HC exhibiting dissimilar hydrolysis kinetics or by varying the concentration of HC employed for virus surface modification. Additionally, a concurrent co-modification of the vector with different cross-linker molecules, although not explored in the present study, provides another opportunity for fine-tuning vector release and transduction.
The authors have nothing to disclose.
The authors do not have competing financial interests to disclose.
316 stainless steel mesh disks | Electon Microscopy Sciences | E200-SS | |
Generic 304-grade stainless steel stents | Laserage | custom order | |
AdeGFP | University of Pennsylvania Vector Core | AD-5-PV0504 | |
AdLuc | University of Pennsylvania Vector Core | AD-5-PV1028 | |
AdEMPTY | University of Pennsylvania Vector Core | A858 | |
Cy3(NHS)2 | GE Healthcare | PA23000 | |
Sepharose 6B | Sigma-Aldrich | 6B100-500ML | |
UV 96-well plates | Costar | 3635 | |
Fluorometry 96-well plates | Costar | 3915 | |
Cell culture 96-well plates | Falcon | 353072 | |
Tris(2-carboxyethyl)phosphine hydrochloride (TCEP ) | Pierce Thermo Scientific | 20490 | |
dithiothreitol (DTT) | Pierce Thermo Scientific | 20290 | |
sulfo-LC-SPDP | Pierce Thermo Scientific | 21650 | |
Spectrophotometer | Molecular Devices | SpectraMax 190 | |
Spectrofluorometer | Molecular Devices | SpectraMax Gemini EM | |
Orbital shaker incubator | VWR | 1575R | |
Horizontal airflow oven | Shel Lab | 1350 FM | |
Centra-CL2 centrifuge | International Equipment Company | 426 | |
Digital vortex mixerer | Fisher Thermo Scientific | 02-215-370 | |
Eclipse TE300 fluorescence microscope | Nikon | TE300 | |
DC 500 CCD camera | Leica | DC-500 | |
7500 Real-Time PCR system | Applied Biosystems | not available | |
IVIS Spectrum bioluminescence station | Perkins-Elmer | not available | |
EDTA dipotassium salt | Sigma-Aldrich | ED2P | |
Bovine serum albumin fraction V (BSA) | Fisher Thermo Scientific | BP1600-100 | |
Tween-20 | Sigma-Aldrich | P1379 | |
Dumont forceps | Fine Science Tools | 11255-20 | |
A10 cell line | ATCC | CRL-1476 | |
Bovine aortic endothelial cells | Lonza | BW-6002 | |
Luciferin, potassium salt | Gold Biotechnology | LUCK-1Ge | |
Pluronic F-127 | Sigma-Aldrich | P2443-250G | |
PBS without calcium and magnesium | Gibco | 14190-136 | |
Fetal bovine serum | Gemini Bio-Products | 100-106 | |
Penicillin/Streptomycin solution | Gibco | 11540-122 | |
DMEM, high glucose | Corning cellgro | 10-013-CV | |
0.25% Trypsin/EDTA | Gibco | 25200-056 | |
QIAamp DNA micro kit | Qiagen | 56304 | |
Power Sybr Green PCR Master Mix | Applied Biosystems | 4367659 | |
MicroAmp Optical 96-well Reaction Plate | Applied Biosystems | N8010560 | |
MicroAmp Optical Adhesive Film | Applied Biosystems | 4360954 | |
Cephazolin | Apotex | not available | |
Loxicom (Meloxicam) | Norbrook | not available | |
Heparin sodium | APP Pharmaceuticals | not available | |
Ketavet (Ketamine) | VEDCO | not available | |
Anased (Xylazine) | Lloid | not available | |
Forane (Isoflurane) | Baxter | not available | |
Curved Moria iris forceps | Fine Science tools | 11370-31 | |
Curved extra-fine Graefe forceps | Fine Science Tools | 11152-10 | |
Dumont #5 forceps | Fine Science Tools | 11252-20 | |
Vannas spring scissors | Fine Science Tools | 15018-10 | |
Fine scissors – ToughCut | Fine Science Tools | 14058-09 | |
Surgical scissors | Fine Science Tools | 14101-14 | |
Vicryl suture (5-0) | Ethicon | J385 | |
Suture thread (4/0 silk) | Fine Science Tools | 18020-40 | |
Michel suture clips | Fine Science Tools | 12040-02 | |
Wound dilator (Lancaster eye specula) | KLS Martin | 34-149-07 | |
Hot bead sterilizer | Fine Science Tools | 18000-45 | |
Michel suture clip applicator | Fine Science Tools | 112028-12 | |
Insyte Autoguard 24G IV catheter | Beckton-Dickinson | 381412 | |
2F Fogarty catheter | Edwards Lifesciences | 120602F | |
Teflon tubing | Vention | 041100BST | |
PTA catheter | NuMed | custom order | |
Gauze pads | Kendall Healthcare | 9024 | |
Cotton applicators | Solon Manufacturing | WOD1003 | |
Saline | Baxter | 281321 | |
10 ml syringe (Luer-Lok) | Beckton-Dickinson | 309604 | |
1 ml syringe (Luer-Lok) | Beckton-Dickinson | 309628 | |
Clippers with #40 blade | Oster | 78005-314 | |
Transpore surgical tape | 3M | MM 15271 | |
Puralube vet ointment | Pharmaderm | not available |