Materials, chemicals, and biologicals
316L grade stainless steel foil and mesh disks were from Goodfellow (Coraopolis, PA, USA) and EMS (Hatfield, PA, USA), respectively. Clinical grade L605 cobalt-chromium alloy stents of open cell design were mounted on 3 mm and 3.5 mm balloon catheters. Polyallylamine bisphosphonate with latent thiol groups (PABT) and branched polyethyleneimine with installed pyridyldithio groups PEI(PDT) were synthesized in our laboratory as described before58. Tris(2-carboxyethyl)phosphine hydrochloride (TCEP), 8Br-cAMP, CM-H2DCFDA, and CellROX Orange were purchased from Thermo Fisher Scientific (Waltham, MA, USA). Thiolated protein G was acquired from Protein Mods (Madison, WI, USA). Anti-AAV2 (clone A20) and anti-β tubulin antibody were were from GeneTex (Irvine, CA, USA). Anti-human apoA1 antibody was from Abcam (Cambridge, MA, USA). Anti-eGFP antibody was from Rockland Immunochemicals (Limerick, PA, USA). Anti-Ki67 antibody was from Novus Biologicals (Centennial, CO). The secondary AlexaFluor 488 and AlexaFluor 555-labelled antibodies were from Invitrogen (Waltham, MA, USA). Human recombinant apoA1 and tert-butyl hydroperoxide (tBHP) were from Sigma-Aldrich (St. Louis, MO, USA). AAV2-human apoA1(WT), AAV2-human apoA1(4WF), and AAV2-eGFP were manufactured by the University of Pennsylvania Vector Core using the triple plasmid transfection protocol59. Human apoA1 ELISA assay was purchased from R&D Systems (Minneapolis, MN, USA). WST-8 assay kit was purchased from Cayman Chemical, Ann Arbor, MI, USA). 3H-cholesterol was obtained from Perkin Elmer (Waltham, MA, USA). Primary rat aortic smooth muscle (RASMC) and endothelial cells (RAEC) were from Lonza (Rochester, NY, USA) and Gelantis (San Diego), respectively. Human embryonic kidney (HEK-293) and Raw 264.7 macrophage cell lines were from ATCC (Manassas, VA, USA). Rat blood outgrowth endothelial cells (BOEC) were isolated and characterized as described60.
Vector immobilization on metal substrates
A protocol for reversible immobilization of AAV2 vectors to the bare metal surfaces was previously published by our group32. Briefly, cobalt/chromium stents and stainless steel mesh disks were cleaned with isopropanol and incubated in a 0.5% aqueous solution of PABT at 72ºC with shaking for 1 h followed by deprotection of thiol groups in the side chains of PABT with TCEP (12 mg/ml in 0.1 M acetate buffer at room temperature (RT) with shaking for 15 min). The samples were then extensively washed with degassed DDW and reacted under argon atmosphere with an 0.5% aqueous solution of PEI(PDT) at 28ºC with shaking for 1 h, washed and incubated with thiolated protein G (PrG-SH; 125 µg/ml in degassed PBS) at 28ºC with shaking for 1 h. The specimens were washed with PBS and reacted with 50 µg/ml of the anti-AAV2 antibody with shaking at RT for 45 min. The samples were then washed in PBS and incubated with AAV2 suspension (1010 VG/ml in PBS) at RT with mild shaking for 45 min to immobilize the vector particles on the metal surface.
Cell culture
All primary cells were used in passages 3 to 7. RAEC and rat BOEC were grown in EGM-2 medium. All other cells were maintained in DMEM supplemented with 10% FBS (Gemini) and 1% antibiotic/antimycotic mixture (Gibco). When appropriate, the medium was switched to respective basal medium supplemented with lipoprotein-free FBS (Kalen Biomedical) to avoid the effects of HDL and apoA1 derived from serum on experimental endpoints.
Transduction experiments
Rat BOEC, RAEC, RASMC and Raw264.7 cells were transduced with AAV2-apoA1(WT) and AAV2-apoA1(4WF) at MOI of 105–5 × 105 (1010–3.5 × 1010 VG/ml). The respective culture media were formulated using a lipoprotein-depleted FBS. The media were collected after 3 days of culture, and the concentration of human apoA1 in the conditioned media was assayed with ELISA. For immunofluorescence studies, RAEC and RASMC were transduced with AAV2-apoA1(WT) and AAV2-apoA1(4WF) as above. Three days post-transduction the cells were fixed in cold methanol and were stained with anti-apoA1 antibody (Abcam, ab52945), followed by either Alexa-488 or Alexa-549 labelled secondary antibodies. Properly stained untransduced RAEC and apoA1-transduced RASMC not exposed to the primary antibody served as controls.
Cholesterol efflux experiments
HEK 293 cells were seeded in a 24-well plate. Upon reaching 75–85% confluence, the triplicate wells were transduced with AAV2-apoA1(WT), AAV2-apoA1(4WF) (both at MOI of 105) or were left untransduced. Forty-eight hours after transduction the medium was changed for unsupplemented DMEM. After 24 h, the medium was collected and concentrated 20-fold using centrifugal concentration devices with a 3 kDa cut-off membrane. The concentration of human apoA1 in each preparation was then determined by human apoA1 ELISA (R&D Systems) per manufacturer instructions. ApoA1 concentration in the conditioned media originated from the AAV2-apoA1(WT)- and AAV2-apoA1(4WF)-transduced cells was then adjusted with unsupplemented DMEM to 10 µg/ml. Likewise, the media from untransduced cells was spiked with the commercial recombinant human apoA1 to 10 µg/ml concentration.
Murine Raw 264.7 macrophages in a 12-well plate format were treated with 0.5 µCi 3H-cholesterol (Perkin-Elmer) formulated in sterile culture media with the addition of 0.15 µM acetylated human LDL and incubated for 48 h to allow ample cholesterol uptake by macrophages. Twenty-four hours before testing, 8Br-cAMP was added to the medium at the final 0.15 mM concentration to activate the ABCA1 production. Immediately before starting the cholesterol efflux phase of the experiment, concentrated conditioned media from HEK-293 cells containing 10 µg/mL of apoA1(WT), 10 µg/mL apoA1(4WF), or 10 µg/ml of recombinant human apoA1 spiked to the media from untransduced cells were mixed with increasing concentrations of hypochlorous acid (0, 2, 8, and 16:1 molar ratio to apoA1). ApoA1 oxidation reaction was run in triplicate aliquots at 37 °C for 1 h. Raw 264.7 cells were then carefully washed with PBS and incubated with the differently oxidized HEK 293-conditioned DMEM in the cell culture incubator for 4 h. The media were then collected and centrifuged to exclude cell debris. Cells were washed with PBS, scraped, resuspended in PBS. 5 mL of Ecolite + scintillation fluid was added to each media or cell containing scintillation vial and vortexed. 3H count rates were then determined by using a scintillation counter (Beckman-Coulter LS6500). The 3H-cholesterol efflux was calculated as a percentage of 3H in the media over the total 3H-cholesterol content in both the media and cells.
Proliferation assay
RASMC and rat BOEC grown to 70–80% confluence in T-75 flasks were transduced at MOI 105 (1010 VG/ml) with AAV2-apoA1(WT), AAV2-apoA1(4WF), AAV2-eGFP or left untransduced. Four days after transduction, the cells were trypsinized, collected by centrifugation, mixed with FBS/DMSO, aliquoted and frozen. The single aliquots of non-transduced (NT), AAV2-eGFP-transduced, AAV2-apoA1(WT)-transduced, and AAV2-apoA1(4WF)-transduced cells of both types were then reseeded into the 96-well plates (N = 4 wells per group; 5 × 103 cells/well across all experimental conditions). A group of wells in each plate was seeded at a higher density (105 cells/well) to achieve immediate confluence. The cells were maintained in DMEM supplemented with 10% lipoprotein-free FBS (Kalen). 20 ng/ml of rat TNFα was added to RASMC cultures to emulate the atherosclerotic milieu. The relative cell number for each transduction type was determined at days 2 and 4 with WST-8 assay by normalizing the optical density values for each well of growing cells to that of 100% confluent reference wells. Cell growth kinetics were expressed as the percent of a monolayer confluency. At the completion of the WST-8 assay at day 4 post-seeding, the cells were fixed with 10% formalin and immunostained with rabbit anti-Ki67 antibody/goat anti-rabbit AlexaFluor 555-labelled antibody and counterstained with Hoechst 33,342.
Migration assay
RASMC and rat BOEC pre-transduced en masse with AAV2 apoA1(WT), AAV2-apoA1(4WF), and frozen as detailed above for the proliferation assay, were seeded in a 48-well plate and cultured for 3–4 days in the respective media supplemented with 10% lipoprotein-depleted FBS until confluent. The cells were then starved for 36 h in media containing 0.5% lipoprotein-depleted FBS. A linear scratch injury was then inflicted to each well with a 200 µl pipette tip. The cells were washed with PBS to remove debris and imaged at 40 × magnification immediately after the scratch injury and 24 h after. The closure of the gap by inwardly migrating cells was quantitated using Image J (v1.53a).
ROS assay
Rat aortic endothelial cells (RAEC) grown to 60–70% confluence in a 96-wells plate were either transduced with AAV2-apoA1(WT), AAV2-apoA1(4WF) at MOI of 5 × 105 (3.5 × 1010 VG/ml) or left untransduced (N = 12 wells for each treatment). Seventy-two hours after transduction, the cells were washed with PBS and added DMEM supplemented with 10% lipoprotein-depleted FBS. Five hours after the medium change, half of the wells were treated with 40 ng/ml rat TNFα. Twenty-four hours after boosting ROS production with TNFα, the cells were washed with PBS and treated with 10 µm CM-H2DCFDA for 30 min, followed by fluorimetry at 485/538 nm. Additionally, ROS mitigation with apoA1(4WF) transduction (MOI of 5 × 105; 3.5 × 1010 VG/ml) was assessed in tBHP-stimulated RAEC using CellROX Orange reagent (N = 5 wells per condition).
Monocyte adhesion assay
Rat aortic endothelial cells (RAEC) grown to 60–70% confluence in a 96-wells plate were either transduced with AAV2-apoA1(WT), AAV2-apoA1(4WF) at MOI of 5 × 105 or left untransduced (N = 4 wells for each condition). Seventy-two hours after transduction, the cells were washed with PBS, administered DMEM supplemented with 10% lipoprotein-depleted FBS and treated with 40 ng/ml rat TNFα for 24 h.
Monocytes were isolated from 10 ml of heparinized blood harvested from naïve male Sprague–Dawley rats by Ficoll-Paque gradient centrifugation with subsequent magnetic immunoseparation using a cocktail of anti-CD8, anti-CD5, anti-CD45RA, and anti-pan T cell antibodies61. Isolated monocytes were then fluorescently labeled with PKH-26 dye (Millipore-Sigma, St. Louis, MO, USA) as directed by a manufacturer. 5 × 104 fluorescently-labeled monocytes were then added to each well with differently transduced, TNFα-activated rat endothelial cells. Following 30 min incubation in the cell culture incubator, monocyte adhesion was examined by fluorescence microscopy. The number of monocytes attached to the activated RAEC monolayers was derived from the 100 × magnification images of the central area of each well.
To eliminate the possibility that cell signaling events triggered by RAEC transduction with AAV2 vector decrease monocyte adhesion to the endothelial monolayer, in a separate experiment PKH-26 labeled rat monocytes were added to the wells of AAV2-Egfp—transduced and non-transduced RAEC stimulated with 40 ng/ml rat TNFα (N = 4 per condition).
Scanning electron microscopy
The mesh disks formulated with 1010 VG/ml of AAV2-eGFP at the vector incubation step were washed in PBS and cacodylate buffer (pH 7.4) and fixed with 2% glutaraldehyde/cacodylate buffer. The samples were dehydrated with graded ethanol and hexamethyldisalazane, sputter-coated with gold/palladium alloy and imaged using a Quanta250 scanning electron microscope (FEI, Hillsboro, OR).
Quantification of AAV2 load on stents and in the arterial wall
The extra AAV2-carrying stents not used in the animal experiments (N = 3), as well as stents (N = 3) harvested from the animal that died within 1 h of the stenting surgery, were carefully cut into 1–2 mm fragments. Arterial tissue segments (N = 3) underlying the stents harvested from the pig coronaries were separated from the metal struts and processed separately. Specifically, four fragments were excised from the central part of each arterial segment and pooled. The wet weight of the pooled specimens was ~ 30 mg. The tissue was minced using stainless steel beads and a Bullet Blender (both from Next Advance, Troy, NY). All specimens were then individually processed using QIAamp DNA minikit (Qiagen) to isolate the viral DNA. Calibration curve samples spanning 104–1010 AAV2 genomes were prepared from the stock solution of the vector. An RT-PCR reaction was then carried out with AAV2-specific primers, thus providing direct quantification of viral genomes associated with each specimen. The fractions of the AAV2 load associated with the retrieved stents and the underlying arterial tissue were then calculated, assuming the viral load of undeployed stents as 100%. To verify completeness of viral DNA extraction from the stents, a second round of DNA extraction from the already processed samples was attempted in the preliminary experiments, yielding no DNA.
AAV2-GFP stent reporter study
To study transduction of arterial tissue with AAV2-carrying stents in a pig model, 4 Yorkshire domestic pigs (22–28 kg) of both genders received AAV2-eGFP stents in the left anterior descending (LAD) and the circumflex (Cx) coronary arteries. The animals were euthanized 7 days after the surgery, and the harvested stented arteries (N = 8) were snap-frozen in liquid N2, pulverized under liquid N2, and the tissue powder was suspended in 500 µl of T-Per buffer (Thermo Scientific) supplemented with protease inhibitors (Roche), incubated on ice for 30 min, and centrifuged at 10,000 G for 10 min. Protein concentration in the supernatant of each sample was determined by the BCA assay. Fifty µg protein samples were resolved on a 4–12% NuPAGE™ Bis–Tris polyacrylamide gel, blotted to a nitrocellulose membrane, blocked with 5% dry milk/PBS and consecutively reacted with anti-eGFP antibody (Rockland, 1:7000 dilution) and anti-β tubulin antibody (GeneTex, 1:5000 dilution), peroxidase-conjugated goat anti-rabbit antibody (Santa Cruz; 1:2000 dilution) and SuperSignal™ Pico Plus luminescent substrate (Thermo Scientific). The signal was detected using a Luminescence detection station (IVIS Spectrum) and analyzed with Image J software (v1.53a).
Hypercholesterolemic/diabetic pig model and pig stenting experiments
All animal experiments were pre-approved by the Institutional Animal Care and Use Committee of the University of Pennsylvania, carried out in accordance with federal regulations and reported in accordance with ARRIVE guidelines. To test the therapeutic effectiveness of stents formulated with AAV2-apoA1(4WF) vector immobilized on the bare metal stent struts, a hypercholesterolemic/diabetic pig model62 was used. Before instituting model-related pharmacological and dietary interventions, all animals were screened for the presence of preformed anti-AAV2 antibodies, and the animals tested positive were either excluded from the study or were ascribed to the group treated with BMS (Supplemental Fig. 6). Totally eleven 10–12 week-old female Yorkshire domestic pigs (22–28 kg) were made diabetic by an intravenous injection of Streptozotocin (125 mg/kg). After verification of persistent hyperglycemia (> 250 mg/dl) the animals were administered a hypercholesterolemic diet (0.5% cholesterol, 5% lard, 1.5% sodium cholate) for 24 weeks. Total blood cholesterol and blood glucose levels were monitored throughout the study. If blood glucose exceeded 400 mg/dl, insulin was administered as needed.
All study animals underwent cardiac catheterization with coronary angiography. Bare metal stents, AAV2-eGFP stents, and AAV2-apoA1(4WF) stents (all 18-mm length, mounted onto 3 mm or 3.5 mm balloon catheters) were implanted in proximal and/or distal locations of each animal’s left anterior descending (LAD) and the circumflex (Cx) coronary arteries (2–4 stents per animal). An inflation pressure of 10–14 atm was applied to deploy stents to achieve a 1.1–1.2 stent/artery diameter ratio. Two animals died because of ventricular fibrillation within 1 h of stenting (Supplemental Fig. 6). The harvested arteries of one of the deceased animals were used to quantify the vector load on the stents and the vector loss during deployment and the initial phase of the release. All survived animals were euthanized 4 weeks after stent deployment by IV injection of KCl (125 mg/kg). The hearts were harvested. The stented arterial segments were excised preserving the 2–3 mm non-stented flaps of arterial tissue on both sides and flushed with heparinized saline. The stent-free overhangs were then dissected and snap-frozen in liquid nitrogen for RT-PCR studies. The stented portions of the arteries were fixed in 10% buffered formalin, methyl methacrylate-embedded, sectioned, deplastified and stained according to the Verhoeff-van Giesson method. Five sections cut 3 mm apart from each other through the entire length of the stented segment were stained and analyzed. Ends of the stented segments were excluded. Digital images of the stained arterial sections were captured at 20 × magnification, and the areas of the lumen, internal, and external elastic laminas were measured to derive the extent of restenosis, expressed as a neointima-to-media area ratio, % of luminal stenosis and neointimal thickness. These restenosis indices from 5 individual sections were averaged and the mean values were used for comparison between the different treatment groups.
Human apoA1(4WF) expression in the GDS-treated porcine arteries
Snap-frozen arterial samples adjacent to stents were stored at −80 °C until processing. The samples were homogenized with steel beads using a Bullet Blender (Next Advance, Troy, NY). RNA was extracted using a RNeasy Fibrous Tissue Mini Kit (Qiagen, Germantown, MD). Five hundred µg of isolated RNA were converted into cDNA with TaqMan reverse transcription reagents (Applied Biosystems, Waltham, MA). RT-PCR reaction was carried out in a 7500 fast Real-Time PCR engine using the following primer sequences for amplification of apoA1(4WF) cDNA: (fwd: 5’-TTT-GAT-CGA-GTG-AAG-GAC-CTG-3’ and rvs: 5’-GGT-TAT-CAA-AGA-ACT-CCT-GGG-T-3’). A ddCt algorithm was applied for the interpretation of RT-PCR results. Porcine β2 microglobulin housekeeping gene amplification with the respective primers (fwd: 5’-CGC-CCC-AGA-TTG-AAA-TTG-ATT-TGC-3’ and rvs: 5’-GCT-ATA-CTG-ATC-CAC-AGC-GTT-AGG-3’) was used for the normalization, and the arterial tissue harvested adjacent to the BMS was used as a reference sample.
AAV2-neutralizing antibodies titering
To determine the prevalence of AAV2 neutralizing antibodies in the serum of experimental animals, blood sampled at the beginning of the study, 1 week before the intervention, and at the sacrifice was allowed to clot and centrifuged at 1500G for 15 min. Obtained sera were stored at −80 °C prior to analysis. HEK-293 cells grown in the 96-well plates to 80% confluence were transduced with AAV2-eGFP at MOI of 105 in the presence of 1:20 diluted serum samples or without the addition of the sera. The ensuing transgene expression was determined by fluorimetry at 485/538 nm 3 days after transduction. 50% or higher inhibition of transgene expression manifested as reduction of eGFP fluorescence intensity, denoted presence of neutralizing antibodies in the serum. The animals that exhibited the presence of neutralizing antibodies at the beginning of the study were excluded from the study or slated to the BMS treatment group.
Statistical methodology
Data are presented as means ± SD, unless specified otherwise. Differences between the groups were analyzed by ANOVA followed by a post-hoc Tukey’s test, and were termed statistically significant at p < 0.05.
