Two-dimensional schematic of the structure of the adenovirus, depicting the penton base proteins from which the fiber shaft, fiber knob, and trimeric fiber coat proteins project.
Schematic of the adenoviral virion penton base, fiber shaft, fiber knob, and trimeric fiber proteins. The area encircled is a schematic representation of the site on the fiber protein that is genetically modified with the sequences listed. These genetic modifications to the fiber protein alter the adenoviral tropism for receptor binding. AdZ.F(RGD) is redirected toward Iv-integrin receptors, and AdZ.F(pK7) is redirected toward heparan sulfate–containing receptors.
The transfer efficiency of AdZ.F(pK7) is much greater than that of AdZ.F(RGD) and AdZ.F in vitro. A, Rat aortic smooth muscle cells (RASMCs) were infected with 1000 particles per cell of the 3 different adenoviruses. The β-galactosidase activity was determined using a β-galactosidase kit (Promega, Madison, Wis). B, The RASMCs were exposed to increasing concentrations of heparin for 15 minutes, rinsed with phosphate-buffered saline, and then infected with the 3 different adenoviruses. The β-galactosidase activity was determined using a β-galactosidase kit (Promega). Data represent the mean ± SE (n = 3, each experiment repeated 4 times).
The transfer efficiency of AdZ.F(pK7) is greater than that of AdZ.F(RGD) and AdZ.F following heparin administration using the pig iliac artery injury model. Three days following infection, iliac arteries were harvested and β-galactosidase activity was determined. Results were normalized with protein concentration. A, The transfer efficiency of the different adenoviral vectors, AdZ.F(pK7) and AdZ.F, was compared by instillation of AdZ.F(pK7) into the left and AdZ.F into the right iliac arteries of the same animal. B, To compare AdZ.F(pK7) directly with AdZ.F(RGD) within the same animal, another set of experiments was performed in a similar manner. All animals received heparin intravenously, 100 U/kg, and the lumen of the vessels was irrigated with heparinized saline, 10 U/mL, before infection with the adenovirus. Data represent the mean ± SE (n = 4, separate vessels for each group).
Qualitatively, the transfer efficiency of AdZ.F(pK7) is greater than that of AdZ.F(RGD) and AdZ.F following heparin administration using the pig iliac artery injury model. Porcine iliac arteries were infected with the 3 different adenoviral vectors. All animals received heparin intravenously, 100 U/kg, and the lumen of the vessels was irrigated with heparinized saline, 10 U/mL, before infection with the adenovirus. Three days following infection, the iliac arteries were harvested and fixed in 2% paraformaldehyde and then stained with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside. A, The control, uninjured, uninfected arteries (n = 4). B, The AdZ.F-infected arteries (n = 4). C, The AdZ.F(RGD)-infected arteries (n = 4). D, The AdZ.F(pK7)-infected arteries (n = 8).
Kibbe MR, Murdock A, Wickham T, Lizonova A, Kovesdi I, Nie S, Shears L, Billiar TR, Tzeng E. Optimizing Cardiovascular Gene TherapyIncreased Vascular Gene Transfer With Modified Adenoviral Vectors. Arch Surg. 2000;135(2):191-197. doi:10.1001/archsurg.135.2.191
Copyright 2000 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2000
Adenovirus is widely used as a vector for gene transfer to the vasculature. However, the efficiency of these vectors can be limited by ineffective viral-target cell interactions. Viral attachment, which largely determines adenoviral tropism, is mediated through binding of the adenoviral fiber coat protein to the Coxsackievirus and adenovirus receptor, while internalization follows binding of the adenoviral RGD motif to αv-integrin receptors. Modifications of the fiber coat protein sequence have been successful for targeting the adenovirus to more prevalent receptors in the vasculature, including heparan sulfate–containing receptors and αv-integrin receptors.
Modified adenoviral vectors targeted to receptors more prevalent in the vasculature result in an increased transfer efficiency of the virus in vitro and in vivo even in the presence of clinically relevant doses of heparin.
We tested 2 modified E1- and E3-deleted Ad5 type adenoviral vectors containing the β-galactosidase gene. AdZ.F(pK7) contains multiple positively charged lysines in the fiber coat protein that target the adenovirus to heparan sulfate receptors, while AdZ.F(RGD) contains an RGD integrin-binding sequence in the fiber coat protein that allows binding to αv-integrin receptors. The gene transfer efficiency of these modified viruses was compared in rat aortic smooth muscle cells in vitro and in an in vivo porcine model of balloon-induced arterial injury. Because of the use of heparin during most vascular surgical procedures and the concern that heparin might interfere with the binding of AdZ.F(pK7) to heparan sulfate receptors, the effect of heparin on the in vitro and in vivo transfer efficiency of these 2 modified adenoviruses was evaluated.
In vitro infection of rat aortic smooth muscle cells with AdZ.F(pK7) and AdZ.F(RGD) resulted in significantly higher levels of β-galactosidase expression compared with the unmodified adenovirus (mean ± SEM, 1766.3 ± 89.1 and 44.8 ± 3.4 vs 10.1 ± 0.7 mU per milligram of protein; P<.001). Following heparin administration, the gene transfer efficiency achieved with AdZ.F(pK7) diminished slightly in a concentration-dependent manner. However, the transfer efficiency was still greater than with the unmodified virus (mean ± SEM, 1342.3 ± 101.8 vs 4.8 ± 0.4 mU per milligram of protein; P<.001). In vivo, following injury to the pig iliac artery with a 4F Fogarty balloon catheter, we found that AdZ.F(pK7) transduced the artery approximately 35-fold more efficiently than AdZ.F and 3-fold more efficiently than AdZ.F(RGD) following the administration of intravenous heparin, 100 U/kg body weight, and heparinized saline irrigation.
Modifications of the adenovirus that lead to receptor targeting resulted in significantly improved gene transfer efficiencies. These improvements in transfer efficiencies observed with the modified vectors decreased slightly in the presence of heparin. However, AdZ.F(pK7) was still superior to AdZ.F(RGD) and AdZ.F despite heparin administration. These data demonstrate that modifications of adenoviral vectors that enhance binding to heparan sulfate receptors significantly improve gene transfer efficiency even in the presence of heparin and suggest an approach to optimize gene transfer into blood vessels.
THE ADENOVIRUS has gained popularity as the vector of choice for cardiovascular gene therapy because it exhibits high gene transfer efficiency compared with other vectors. It has been used successfully in animal models to prevent restenosis following balloon-induced arterial injury, vein graft intimal hyperplasia, arterial thrombosis, and transplantation vasculopathy.1 Nevertheless, gene delivery can still be quite inefficient because of limitations in methods of delivery to blood vessels and the need to minimize vessel occlusion times. Approaches that increase the efficiency of gene transfer could broaden the potential applications of gene therapy and enhance the therapeutic benefit.
The efficiency with which the adenovirus can infect cells is determined by the receptors present on the target cells that mediate viral attachment and internalization. The adenoviral virion is comprised of hexon polypeptides, penton base polypeptides, and trimeric fiber protein polypeptides (Figure 1).2 Viral attachment is mediated by binding of the trimeric fiber protein to the Coxsackievirus and adenovirus receptor, while internalization is mediated via binding of an RGD sequence in the penton base of the adenovirus to αv-integrin receptors.3- 6 In the vasculature, the presence of these receptors is low, which accounts for the low rates of transfection observed in cardiovascular gene transfer studies.
One method to increase the transfer efficiency of the adenovirus to the vasculature is to target the adenovirus to receptors that are more prevalent on endothelial or smooth muscle cells. We tested 2 E1- and E3-deleted Ad5 type modified adenoviruses carrying the β-galactosidase gene. The genetic modifications of the adenoviral genome target the adenoviruses to either αv-integrin receptors, AdZ.F(RGD), or heparan sulfate–containing receptors, AdZ.F(pK7), both of which are abundant in the cells of the vasculature (Figure 2). AdZ.F(pK7) has been successfully used in the absence of heparin to increase transfer efficiency in porcine vessels.6 The present study was undertaken to compare the gene transfer efficiency achieved with AdZ.F(RGD) and AdZ.F(pK7) with that achieved with the control adenovirus in cultured smooth muscle cells and in porcine iliac arteries in vivo, with a special emphasis placed on the effect of heparin on adenovirus-mediated gene transfer. AdZ.F(RGD) and AdZ.F(pK7) demonstrated improved transduction efficiency in rat aortic smooth muscle cells (RASMCs) in vitro (4-fold and 200-fold, respectively; P<.001) compared with the unmodified adenoviral vectors. Although gene transfer with AdZ.F(pK7) was diminished by heparin in a concentration-dependent manner, the efficiency was still greater (>200-fold) than that achieved with the unmodified adenovirus, AdZ.F. Similarly, delivery of these vectors to pig iliac arteries in vivo following balloon-induced arterial injury and intravenous and intralumenal heparin administration resulted in increased gene transfer efficiency compared with the unmodified vector. Therefore, these modified adenoviral vectors have tremendous potential for improving the gene transfer efficiency of adenovirus-mediated cardiovascular gene therapy and may therefore have an impact on therapeutic efficacy.
An E1- and E3-deleted adenoviral vector carrying the β-galactosidase gene under the control of a cytomegalovirus promoter (AdZ.F) was designed and constructed as previously described.6,7 AdZ.F(pK7) and AdZ.F(RGD) were derived from the E4-deleted vector, AdZ.11A,8 to incorporate the additional amino acids present on the C termini of their fiber protein, as previously described.6 AdZ.11A contains a complete deletion of E4 and inclusion of the E4 spacer element. Briefly, the plasmid DNAs from pNS(F5)pK7 and pNS(F5)RGD were linearized6 with SalI, purified, and transfected by using calcium phosphate into 293 cells. The transfected 293 cells had been preincubated with the E1-, E3-, and E4-deleted construct AdZ.11A8 at a multiplicity of infection of 1 fluorescent-forming unit per cell, 1 hour before transfection with the pNS plasmids. Recombination of the E4+ pNS plasmid with the E4-deleted vector resulted in the rescue of an E1-deleted, E3-deleted, E4+ vector capable of replication in 293 cells. The resultant vectors, AdZ.F(pK7) and AdZ.F(RGD), were isolated in 2 successive rounds of plaque purification on 293 cells. Each vector was verified to contain the correct insert by sequencing polymerase chain reaction products derived from virus DNA template by using primers that span the region of the insert DNA. Restriction analysis of adenoviral DNA from each of the viruses showed that the viruses were pure and contained the BamHI restriction site unique to the correctly constructed virus.
All viruses were purified from the infected 293 cells at 2 days after infection by 3 freeze-thaw cycles followed by 3 successive bandings on cesium chloride gradients. Purified virus was dialyzed into a combination of Tris, 10 mmol/L, and sodium chloride, 150 mmol/L (pH 7.8), containing magnesium chloride, 10 mmol/L, and 3% sucrose and frozen at −80°C until required for use. The viruses were titrated by plaque assay and fluorescent focus assay to determine the plaque-forming units and fluorescent-forming units, respectively.
Aortic smooth muscle cells from Sprague-Dawley rats (Harlan, Indianapolis, Ind) were cultured from explanted thoracic aortae, as previously described.7 Cultured cells had the characteristic hills-and-valleys appearance and were routinely more than 95% pure by smooth muscle cell α-actin staining. Cells were grown in a combination of Dulbecco modified Eagle medium (low glucose)–Ham's F12 (1:1, vol/vol) (BioWhittaker, Walkersville, Md) supplemented with 10% fetal bovine serum; penicillin, 100 U/mL; streptomycin sulfate, 100 µg/mL; and L-glutamine, 4 mmol/L; and maintained in a 37°C, 95% air–5% carbon dioxide incubator.
Rat aortic smooth muscle cells (passages 3-6) were plated onto 6-cm tissue culture dishes for 24 hours. Cells not being exposed to heparin were rinsed once with Hanks' balanced salt solution (Gibco BRL, Rockville, Md) followed by infection with AdZ.F, AdZ.F(RGD), or AdZ.F(pK7), using 1000 particles per cell for 2 hours at 37°C. Cells exposed to heparin were incubated with increasing concentrations of heparin in Hanks' balanced salt solution (1.5, 5.0, and 10.0 U/mL) for 15 minutes, rinsed once with phosphate-buffered saline, and then infected with the modified adenoviruses. Following infection, the cells were placed in growth media containing 10% fetal bovine serum for 48 hours. The cells were collected by scraping, and the β-galactosidase activity was determined by using a standard β-galactosidase assay kit (Promega, Madison, Wis). Total cellular protein was quantified using the bicinchonic acid protein assay (Pierce, Rockford, Ill).
All animal procedures were performed using an aseptic technique in accordance with the Institutional Animal Care and Use Committee of the University of Pittsburgh, Pittsburgh, Pa. Male domestic pigs weighing between 12 and 15 kg (Walter Whippo, Enon Valley, Pa) were placed under general anesthesia, and the bilateral common iliac arteries were exposed through a low midline laparotomy incision. Five minutes before obtaining proximal and distal vascular control with noncrushing vascular clamps, intravenous heparin, 100 U/kg body weight, was administered to the pigs. Common iliac arterial injury was then created with a 4F embolectomy catheter inserted through a side branch and inflated to 2 atm for 5 minutes. The lumen of the artery was then irrigated with 1 mL of heparinized saline, 10 U/mL, followed by 1 mL of isotonic sodium chloride solution. The adenoviral solution (2 × 1010 particles per milliliter, ≈500 µL per artery) was instilled into the common iliac artery and allowed to incubate for 30 minutes. After the incubation period, the adenoviral solution was evacuated, the side branch ligated, and blood flow reestablished.
Three days after viral infection, the animals were euthanatized with potassium chloride and thiopental sodium over-doses, and the iliac arteries were collected. For qualitative β-galactosidase staining, the arteries were opened on their long axes, fixed in 2% paraformaldehyde in phosphate-buffered saline for 2 hours, and then stained with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside overnight. Vessels were then imaged by light microscopy (Nikon, Melville, NY) for the amount of blue-staining cells on the lumenal surface. In addition, after 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside staining, portions of the vessels were quick-frozen with a disposable cryosurgical device (HistoFreeze-2000; Fisher, Pittsburgh, Pa), and 7-µm cryosections were cut and counterstained with eosin before imaging. For quantitative evaluation of the β-galactosidase activity using a standard β-galactosidase assay kit (Promega), the arterial segments were immediately frozen upon harvesting in liquid nitrogen and ground into a powder using mortar and pestle. Approximately one third of the sample was added to 0.5 µL of protein lysis buffer (Promega) containing phenylmethylsulfonyl, 1 mmol/L; leupeptin, 10 µg/mL; aprotinin, 10 mg/mL; and EDTA, 1 mmol/L. The sample was rocked for 30 minutes at 4°C and then the cellular debris was pelleted by centrifugation. The clear lysate was assayed using a fluorometric assay kit (Galactolight; Tropix, Bedford, Mass). Protein concentration in the lysate was measured using a Bradford assay. β-Galactosidase activity in the samples was then normalized by the protein concentration.
Results are expressed as the mean ± SEM. Differences between groups were analyzed using a 1-way analysis of variance with the Student-Newman-Keuls post hoc test for all pairwise comparisons (SigmaStat; SPSS Inc, Chicago, Ill). Statistical significance was assumed when P<.05.
Rat aortic smooth muscle cells were infected with AdZ.F, AdZ.F(RGD), and AdZ.F(pK7) using 1000 particles per cell. Vascular smooth muscle cells were chosen as a target because these cells are the predominant target following surgical procedures or angioplasty, during which the endothelial lining is lost. Forty-eight hours following infection, β-galactosidase activity, a marker of gene transfer, was the highest with AdZ.F(pK7), followed by AdZ.F(RGD) and then AdZ.F (1766.3 ± 89.2, 44.8 ± 3.4, and 10.0 ± 0.7 mU per milligram of protein, respectively) (Figure 3, A). To determine if heparin would interfere with the receptor-mediated infection process of either the control or the modified adenoviruses, specifically AdZ.F(pK7), RASMCs were exposed to increasing concentrations of heparin before infection with the virus. The concentrations of heparin chosen were based on heparin doses administered in vivo. The lowest concentration of heparin (1.5 U/mL) was calculated as the equivalent dose for the intravenous administration to the pigs based on weight and total blood volume. The highest dose (10 U/mL) reflects the concentration used for the intralumenal heparinized saline irrigant. Exposure of the heparin to RASMCs resulted in a concentration-dependent decrease in AdZ.F(pK7) gene transfer efficiency (no heparin, 1766.3 ± 89.1 mU per milligram of protein; heparin, 1.5 U/mL, 1843.4 ± 231.4 mU per milligram of protein; heparin, 5.0 U/mL, 1475.4 ± 97.9 mU per milligram of protein; and heparin, 10.0 U/mL, 1342.3 ± 101.8 mU per milligram of protein). The difference between no heparin and heparin, 10.0 U/mL, was significant (P = .04) (Figure 3, B). However, AdZ.F(pK7) was still much more effective than either AdZ.F(RGD) or AdZ.F at the highest heparin concentration (1342.3 ± 101.8, 35.1 ± 0.5, and 4.7 ± 0.4 mU per milligram of protein, respectively; P<.001). Interestingly, while the transfer efficiencies of AdZ.F(RGD) and AdZ.F were low, there was a significant (P = .02 and P = .03, respectively) reduction in gene transfer efficiency with these viruses in the presence of the highest concentration of heparin.
To determine if the improved gene transfer efficiency of AdZ.F(pK7) observed in vitro would also occur in vivo, the porcine iliac artery injury model was implemented. Following the administration of heparin intravenously (100 U/kg) and intralumenally (10 U/mL of isotonic sodium chloride solution), the vascular gene transfer efficiency of the different modified adenoviruses was evaluated qualitatively and quantitatively. Animals treated with AdZ.F(pK7) expressed a 35-fold increase in β-galactosidase activity compared with those treated with AdZ.F, even in the presence of heparin (199.6 ± 89.4 vs 5.7 ± 1.3 reactive light units per microgram of protein, respectively; P = .04) (Figure 4, A). In addition, animals treated with AdZ.F(pK7) expressed approximately 3-fold more β-galactosidase activity compared with those treated with AdZ.F(RGD) (81.7 ± 24.9 vs 24.1 ± 8.6 reactive light units per microgram of protein, respectively; P = .05) (Figure 4, B). Background β-galactosidase activity from the uninfected control artery was negligible. On histological review, we found that the transfer efficiency of AdZ.F(pK7) was too great to count individual cells. However, the qualitative pattern of expression of the different adenoviruses supported the quantitative β-galactosidase activity assay results (Figure 5). Also, most of the staining was detected in the vascular smooth muscle cells.
Adenoviral gene therapy directed toward preventing neointima formation following arterial injury or vein grafting has limitations. One of the most formidable limitations is gene transfer efficiency. For gene therapies aimed at arterial injury, the target cell is predominately the vascular smooth muscle cell and, to a lesser extent, the adventitial fibroblast. Because these cell types are relatively deficient in the receptors necessary for adenoviral attachment and internalization, adenoviral uptake is poor. Furthermore, gene transfer efficiency may be impaired by vector-vessel exposure times as might be encountered during a vascular procedure. To circumvent these limitations, modifications of the adenovirus have been made to target the virus to certain cell types via the use of cell-specific promoters or to redirect viral binding to more prevalent receptors.9- 13 We show in this study that by modifying the adenoviral genome we can target binding of the virus to more prevalent receptors in the vasculature, such as αv-integrin and heparan sulfate–containing receptors. Most important, the enhanced gene transfer efficiency achieved with these modified vectors vs the parent virus persists even in the presence of heparin.
In vitro, we showed that AdZ.F(pK7), which is the modification that redirects binding to heparan sulfate–containing receptors, resulted in a significantly higher gene transfer efficiency in RASMCs compared with AdZ.F(RGD) and the unmodified adenovirus, AdZ.F (Figure 3, A). The approximately 200-fold greater gene transfer efficiency achieved with this virus can be explained by the prevalence of the ubiquitously expressed heparan sulfate receptors on smooth muscle cells. However, given that essentially all vascular and cardiovascular procedures are performed in the presence of heparin, the impact of heparin administration on the binding of the modified vectors had to be determined. Using clinically relevant concentrations of heparin, we exposed RASMCs to increasing concentrations of heparin immediately before infection with the modified adenoviruses. The highest concentration of heparin decreased the β-galactosidase expression after infection with all 3 of these vectors. Whether this is due to persistent binding of the heparin to specific receptors is unclear. Alternatively, high concentrations of heparin could have impaired other steps in the expression of the transgene such as transcription or translation. Our studies do not address this issue. However, it is important that AdZ.F(pK7), the vector that specifically uses the heparan sulfate receptor, still exhibited a greater than 200-fold increase in gene transfer efficiency compared with the unmodified vector in the presence of heparin.
Studying the effect of heparin on gene transfer efficiency is important because heparin is being administered to most patients undergoing cardiovascular gene therapy. Phase 1 clinical trials of vascular and cardiovascular gene therapy demonstrate the point. For example, 2 phase 1 trials use vascular endothelial growth factor to induce angiogenesis. The models used in these trials include intralumenal delivery of the gene to coronary arteries following percutaneous coronary artery angioplasty14 or direct injection of the gene into the myocardium in combination with coronary artery bypass grafting.15 During these procedures, patients are routinely administered intravenous heparin to prevent the incidence of thrombotic events. Vascular gene therapy models directed toward preventing neointimal hyperplasia in vein grafts are also in phase 1 clinical trials.16 Harvested vein grafts are exposed to the vector of choice in a solution for a given period. Following infection, the vein is then grafted into the circulation during which time the patient receives intravenous heparin and copious heparinized saline irrigation of the vein graft. Hence, it is apparent that heparin is administered systemically to patients undergoing vascular or cardiovascular gene therapy procedures. It is, therefore, imperative to determine if heparin would affect the gene transfer efficiency of one of the more frequently used vectors, namely, the adenovirus. Since we found that the levels of heparin administered systemically did not have a significant effect on gene transfer efficiency, while the higher concentrations of heparin routinely used in heparinized saline irrigant did have a significant impact on gene transfer efficiency, the amount of heparin being administered must be taken into consideration when designing and implementing adenoviral gene therapy protocols.
In conclusion, the AdZ.F(pK7) modification that redirects binding of the adenovirus to heparan sulfate–containing receptors appears to be quite promising for future vascular gene therapy. Not only would this modified adenovirus result in improved gene transfer efficiencies and greater expression of the gene, it would also theoretically allow much lower titers of the adenovirus to be used. This has further implications in that there would be less of a host immune response evoked by using the lower titers of the adenovirus. In addition, there would be less of a demand for the production of clinical grade adenovirus in quantities necessary for clinical trials or routine use. Therefore, by increasing the gene transfer efficiency of the adenovirus and potentially lowering the titer needed for optimal gene transfer, many limitations of vascular gene therapy can be overcome.
Gene therapy as an approach to therapeutically affect cardiovascular disease processes has gained tremendous popularity. However, the biological impact resulting from most studies has been limited by the transfer efficiency of the vector being used. This study was performed to determine whether modified adenoviral vectors would show markedly improved transfer efficiencies in clinically relevant vascular and cardiovascular models. Two different modifications of the adenovirus were tested. AdZ.F(pK7) is an adenovirus that is redirected toward binding with heparin receptors, while AdZ.F(RGD) is an adenovirus redirected toward binding to αv-integrin receptors. The concern of whether these vectors could be used for vascular or cardiovascular procedures was questioned, especially for the AdZ.F(pK7) vector, since most all procedures of this nature require heparin. Therefore, in this model, we show that these modified vectors do result in a dramatically increased transfer efficiency in vitro and during administration of heparin in vivo, intravenously and intralumenally. This has significant implications for all vascular and cardiovascular gene therapy trials limited by the transfer efficiency of vectors. Improving this variable will dramatically affect the biological outcome following these therapeutic genetic interventions in the future management of cardiovascular disease.
This study was supported by grant R29-HL-57854 from the National Institutes of Health, Bethesda, Md (Dr Tzeng), Scientist Development grant 9630257N from the American Heart Association, Dallas, Tex (Dr Shears), and the Ethicon–Society of University Surgeons Surgical Award, Toronto, Ontario (Dr Kibbe).
Corresponding author: Melina Kibbe, MD, Department of Surgery, University of Pittsburgh, 677 Scaife Hall, Pittsburgh, PA 15261 (e-mail: email@example.com).