This method describes the placement of interposition vein grafts in rabbits, the transduction of the grafts, and the achievement of durable transgene expression. This allows the investigation of physiological and pathological roles of transgenes and their protein products in grafted veins, and testing of gene therapies for vein graft disease.
Vein graft bypass surgery is a common treatment for occlusive arterial disease; however, long-term success is limited by graft failure due to thrombosis, intimal hyperplasia, and atherosclerosis. The goal of this article is to demonstrate a method for placing bilateral venous interposition grafts in a rabbit, then transducing the grafts with a gene transfer vector that achieves durable transgene expression. The method allows the investigation of the biological roles of genes and their protein products in normal vein graft homeostasis. It also allows the testing of transgenes for the activities that could prevent vein graft failure, e.g., whether the expression of a transgene prevents the neointimal growth, reduces the vascular inflammation, or reduces atherosclerosis in rabbits fed with a high-fat diet. During an initial survival surgery, the segments of right and left external jugular vein are excised and placed bilaterally as reversed end-to-side common carotid artery interposition grafts. During a second survival surgery, performed 28 days later, each of the grafts is isolated from the circulation with vascular clips and the lumens are filled (via an arteriotomy) with a solution containing a helper-dependent adenoviral (HDAd) vector. After a 20-min incubation, the vector solution is aspirated, the arteriotomy is repaired, and flow is restored. The veins are harvested at time points dictated by individual experimental protocols. The 28-day delay between the graft placement and the transduction is necessary to ensure the adaptation of the vein graft to the arterial circulation. This adaptation avoids rapid loss of transgene expression that occurs in vein grafts transduced before or immediately after grafting. The method is unique in its ability to achieve durable, stable transgene expression in grafted veins. Compared to other large animal vein graft models, rabbits have advantages of low cost and easy handling. Compared to rodent vein graft models, rabbits have larger and easier-to-manipulate blood vessels that provide abundant tissue for analysis.
Atherosclerosis is a chronic inflammatory disease in which lipid accumulation and inflammation in the blood vessel wall lead to narrowing of the vessel lumen, heart attacks, strokes, and loss of limbs1,2. Percutaneous interventions (e.g., angioplasty and stenting) and medical therapy (e.g., statins and antiplatelet agents) are useful treatments for atherosclerosis; however, they are often ineffective in treating severe obstructive disease both in the coronary and peripheral circulations. Bypass grafting, using autogenous vein segments, remains a common procedure for treating patients with severe, diffuse coronary and peripheral vascular disease3,4. However, vein grafts placed in both the coronary and peripheral circulations have poor long-term patency rates. In the coronary circulation, approximately 10-20% of vein grafts are occluded at 1 year and 50% are occluded by 10 years5,6.In the peripheral circulation, vein graft failure rates are 30-50% at 5 years7.
Gene therapy is an attractive approach for the prevention of vein graft failure because it can deliver a therapeutic gene product precisely at the site of the disease. Accordingly, numerous preclinical studies have tested vein graft gene therapy8,9. However, essentially all of these studies have examined the efficacy at early time points (2-12 weeks)10,11,12,13,14,15,16,17. We are aware of no evidence that gene-therapy interventions can provide durable (years) protection against late vein graft failure that typically results from neointimal hyperplasia and atherosclerosis4. We developed a method that allows durable transgene expression in grafted veins, and thereby allows the testing of gene-therapy interventions at late as well as early time points. To achieve durable transgene expression, the method incorporates HDAd vectors and a delayed transduction strategy. HDAd vectors provide prolonged transgene expression because they lack viral genes, preventing the recognition (and rejection) of transduced cells by the immune system18,19,20,21. Delayed transduction (performed 28 days after the graft placement) prevents the loss of transduced cells during the arterialization process that occurs early after the grafting22.
Other methods that achieve therapeutic transgene expression in the vein graft wall rely on the transduction of the vein graft at the time of the graft placement10,11,12,15,16,17. When measured serially, transgene expression using this approach declines quickly after the transduction22,23. Accordingly, studies using this approach have not examined the efficacy beyond 12 weeks after the vein grafting, with most not assessing efficacy beyond 4 weeks. In contrast, our method achieves vein graft transgene expression that persists stably for at least 24 weeks and-based on similar studies performed in arteries-likely continues far longer22,24. We are aware of no other vein graft gene therapy intervention that achieves stable transgene expression of this duration.
We used a rabbit model to develop our method. Others have used rodents, rabbits, or larger animals to test vein graft gene therapy10,11,12,15,16,17,25,26. Compared to rodent models, rabbits are more expensive and are subject to more stringent regulatory requirements. However, compared to larger animals (e.g., pigs and dogs), rabbits are far less expensive to purchase and house and much easier to handle. Moreover, rabbit vessels resemble human vessels physiologically27, they are sufficiently large that they can be used for testing percutaneous interventions28,29, and they provide sufficient tissue that multiple endpoints (e.g., histology, protein, RNA) can be examined using a single blood vessel specimen22,30. In addition, when the rabbits with vein grafts are fed with a high-fat diet, they develop vein graft atherosclerosis31,32, which is a common cause of coronary artery bypass vein graft failure4,5. These atherosclerotic rabbit vein grafts can serve as a substrate for testing gene-therapy interventions delivered with this method. The provided protocol can help investigators to master the technical skills required to achieve durable transgene expression in rabbit vein grafts.
All animal protocols and studies were approved by the University of Washington Office of Animal Welfare.
1. Pre-operation (for All Surgeries)
2. Vein Graft Surgery (Survival)
3. Transcutaneous Ultrasound
4. Gene Transfer Surgery Performed ~ 28 Days After the Graft Placement (Survival Surgery)
5. Harvest Surgery (Terminal)
The technical proficiency of a new operator must be validated before the operator can use this method to generate experimental data. The first milestone that a new operator must achieve is consistent vein graft patency after both the initial vein-grafting surgery and the subsequent delayed-transduction surgery. Over 90% patency after each of the surgeries is desirable and achievable. The patency can be assessed non-invasively using transcutaneous ultrasound, which we typically perform on postoperative day 5-7. In the rabbits with patent veins, the ultrasound exam will reveal a large blood vessel with brisk caudal-to-cranial blood flow on both sides of the anterior neck (Figure 3A). If either of the vein grafts is occluded, there will be no detectable caudal-to-cranial blood flow on the side(s) with the occluded graft. However, cranial-to-caudal blood flow will still be detected in the internal jugular vein (Figure 3B).
The second milestone that a new operator must achieve is efficient transduction of vein-graft endothelial cells. Here, an adenoviral vector that expresses β-galactosidase is used to evaluate the efficiency of endothelial gene transfer. A new operator in our lab transduced vein grafts with an adenovirus expressing β-galactosidase (AdnLacZ) using the methods described. The grafts were harvested 3 days after the transduction and cut into segments. Some segments were stained with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal). En face imaging of the luminal surfaces showed vein grafts with efficient transduction (Figure 4A) and poor transduction (Figure 4B). To further evaluate the proficiency of a new operator, β-galactosidase mRNA in the extracts of the transduced vein graft is also measured using quantitative reverse transcriptase-mediated PCR and normalizing the signal to the level of glyceraldehyde phosphate dehydrogenase (GAPDH) mRNA measured in the same vein grafts.
The levels of β-galactosidase mRNA in vein grafts transduced by the new operator were not significantly different from the levels of β-galactosidase mRNA in vein grafts transduced by an experienced operator (Figure 5). If mRNA samples from the vein grafts transduced by an experienced operator are not available for comparison, negative control vein graft mRNA can be generated by transducing vein grafts with a non-expressing control vector (AdNull). We have found that the mean levels of β-galactosidase mRNA in vein grafts transduced by an experienced operator are approximately 1,000-fold higher than the background PCR signal for β-galactosidase mRNA measured in AdNull-transduced grafts (Figure 5).
Figure 1. The placement of a reversed right external jugular vein-to-right common carotid artery graft with end-to-side anastomoses. The vein graft (blue) is sutured to the common carotid artery (red) at two anastomoses. At each anastomosis, stitches (white X's) are numbered in the order in which they are placed. For both of the anastomoses, stitch 1 sutures the caudal end of the arteriotomy to the external jugular vein. Stitch 2 sutures the cranial end of the arteriotomy to the external jugular vein. Stitch 3 is placed on the lateral side of the anastomosis at a point equidistant from the first two stitches. Stitch 4 is placed on the medial side of the anastomosis, equidistant from the first two stitches and across from stitch 3. Four additional stitches (stitches 5-8) are placed midway between the four existing stitches. The carotid artery segment between the anastomoses is ligated at both ends with 3-0 silk sutures (arrows) and divided (dashed line) halfway between the ligatures. Left-sided grafts are performed in a similar manner. Please click here to view a larger version of this figure.
Figure 2. The creation and closure of a carotid arteriotomy. (A) The common carotid artery adventitia is grasped near the vein graft with fine forceps and upward traction is applied to the artery wall. This creates a vertical surface, allowing the bent 21 G needle to be inserted approximately parallel to the vessel lumen, decreasing the risk of puncturing the back wall of the artery. (B) The arteriotomy is sutured using 7-0 polypropylene in an X-pattern. The first suture pass enters the artery lumen at the bottom right of the arteriotomy (site 1) and exits the lumen at the bottom left (site 2). The suture then crosses the arteriotomy. The next pass re-enters the lumen at the top right (site 3), and exits the lumen at the top left (site 4). Gentle traction on the ends of the suture (the ends exit from site 1 and site 4) closes the arteriotomy. The suture ends are tied with 2 square knots. The solid blue circles indicate the locations where the suture penetrates the arterial wall. The solid blue lines indicate where the suture passes outside of the arterial wall; dotted blue lines indicate that the suture is located inside the lumen. Please click here to view a larger version of this figure.
Figure 3. Transcutaneous ultrasound images of rabbit necks, evaluating vein graft patency 5-7 days after grafting. (A) Patent vein graft (red) with blood flow in caudal-to-cranial direction is indicated (arrow). (B) Occluded vein graft. No caudal-to-cranial blood flow (which would appear red) is detected. The internal jugular vein (blue) with blood flow in cranial-to-caudal direction is indicated (asterisk). Please click here to view a larger version of this figure.
Figure 4. Beta-galactosidase transgene expression in vein grafts. 28 days after grafting, rabbit vein grafts were transduced by a 20-min incubation of adenoviral vectors expressing β-galactosidase within the lumens of surgically isolated vein grafts. Vein grafts were harvested 3 days after transduction and stained with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside. (A) En face image of the luminal surface of a vein segment showing efficient transduction. (B) En face image of the luminal surface of a vein segment showing poor transduction. Scale bar = 1.0 mm. Please click here to view a larger version of this figure.
Figure 5. Expression of beta-galactosidase (β-gal) transgene mRNA in vein graft segments. 28 days after the placement of jugular vein interposition grafts in rabbit carotids, the grafts were transduced either with an adenoviral vector expressing beta-galactosidase (AdnLacZ) or with a control adenoviral vector that does not express a transgene (AdNull). Surgeries were completed by either an experienced operator (Surgeon 1) or a new operator (Surgeon 2). Vein grafts transduced with AdnLacZ were all harvested 3 days after transduction. RNA extracted from grafts transduced with AdNull and harvested 14 days after transduction was also analyzed, as a negative control. The expression of beta-galactosidase mRNA was quantified by quantitative reverse transcriptase-mediated PCR, with values normalized to glyceraldehyde phosphate dehydrogenase mRNA measured in the same extracts, and expressed as arbitrary units (AU). Bars indicate mean values; p values are from rank-sum tests. Please click here to view a larger version of this figure.
Critical steps in this protocol include the management of anesthesia, anticoagulation, surgical manipulation of the artery/grafted vein, and hemodynamic measurements of the grafted vein. Proper management of anesthesia is critical in this multiple survival surgery model that includes two relatively long operations (typically 3-3.5 h for bilateral vein grafting and 1.5-2.5 h for bilateral graft transduction). We have administered anesthesia both via a nosecone and by endotracheal intubation and found that the intubation improved the survival after the graft transduction surgery, possibly because positive pressure ventilation prevents the atelectasis and associated respiratory failure. Rabbits are challenging to intubate, and intubation-related airway trauma can cause post-operative stridor. However, the challenges and complications of the intubation are a small price to pay for the benefit of eliminating most of the perioperative morbidity and mortality associated with the vein grafting surgery.
Anticoagulation is essential in order to prevent the thrombosis of grafted veins, which can occur after either of the survival surgeries. The thrombosis appears to be an early postoperative event, because the vein grafts that are patent on transcutaneous ultrasound examination 5-7 days postoperatively are nearly always patent at harvest. To prevent the thrombosis after the first surgery (i.e., vein grafting), IV heparin is administered before harvesting the first external jugular vein. Based on the plasma half-life of heparin in rabbits (1-2 h), an additional half dose of IV heparin is administered before harvesting the contralateral external jugular vein. In addition, until both arteriovenous anastomoses of an individual vein graft are complete, we flush the lumens of the carotid artery and the vein graft approximately every 8 min with heparinized normal saline. To avoid graft thrombosis, care should be taken not to damage the vein graft-especially the endothelium-during harvesting and grafting. This includes gentle handling of the vein with surgical instruments and leaving a thin layer of adventitial adipose tissue attached to the vein. We also administer a single dose of topical papaverine to the grafted carotid artery, to prevent or reverse vasospasm, which could contribute to thrombosis.
Meticulous attention to surgical technique is required during the vein grafting surgery. To avoid bleeding at the anastomoses, graft thrombosis, and the introduction of uncontrolled hemodynamic variables, the two arteriotomies must be of the proper length. If an arteriotomy is too short, the vein graft ostial wall will be redundant at the site of anastomosis, creating gaps that allow bleeding. If the arteriotomy is too long, stretching the vein to cover the arteriotomy will bring the vein graft walls in close proximity, making it difficult for the surgeon to avoid suturing the opposing vein graft walls together (essentially guaranteeing postoperative thrombosis). Moreover, if the vein graft lumen is narrowed because the vein was stretched to cover an excessively long arteriotomy, a stenosis will be created that increases shear stress, predisposes to thrombosis33, and potentially alters outcome variables such as neointimal growth and vascular remodeling34,35. In addition, it is important to avoid introducing twists into the vein graft, which could result in lumen narrowing or abnormal flow. To help avoid twisting of the graft, we always align the vein segment along the ventral surface of the common carotid artery and ensure there are no twists in the vein. Because of the potential for variable surgical technique to alter vein graft hemodynamics, and because altered hemodynamics can affect outcome variables independently of treatment, we routinely measure blood flow and graft diameter before vector infusion and at the time of harvest. We use these values to calculate shear stress and pulsatility. Vein grafts with hemodynamic measurements outside of predetermined ranges can be excluded objectively from further analysis, decreasing experimental variability and increasing statistical power.
As we developed this method, we made several modifications. Initially, we attempted vein graft gene transfer during the grafting operation. However, under these circumstances, transgene expression was nearly completely lost by 3 days after gene transfer22. Transgene expression was rapidly lost whether the vein segment was transduced in situ and then grafted or whether the vein was grafted first, and then transduced. Only by delaying transduction until after the vein graft had adapted to the arterial circulation (we waited until 28 days after grafting to perform gene transfer, although a shorter delay might be possible), was the rapid loss of transgene expression avoided22. We also converted from nosecone-delivered anesthesia to endotracheal tube-delivered anesthesia for the second (gene transfer) surgery after experiencing intraoperative and postoperative deaths. In addition, after encountering apparent aspiration pneumonia in a postoperative rabbit, we began to place a small plastic table over the rabbit's abdomen (under the sterile drapes). The table was positioned to prevent the surgeon from inadvertently leaning on the rabbit's abdomen, potentially stimulating gastroesophageal reflux and aspiration. We had no more aspiration events after the placement of the table. However, because this placement coincided with the shift to endotracheal intubation, we cannot be sure which intervention is responsible for eliminating aspiration events.
This method also has limitations. In the United States, the use of rabbits instead of rodents necessitates compliance with the requirements of the Laboratory Animal Welfare Act of 1966. Accordingly, investigators using rabbits (or other animals covered by this Act) must have well-equipped operating rooms and expert anesthesiologists and provide a high level of postoperative monitoring and care. The second limitation is that 2 surgeries are required in order to ensure persistent transgene expression22. The second surgery increases the cost of experiments and exposes each rabbit to an additional set of potential operative and perioperative complications. However, we are not aware of any other gene transfer approach that achieves durable transgene expression in vein grafts. The third limitation of this method is that it requires expertise in construction of HDAd vectors and preparation of high-titer HDAd stocks. Many laboratories have expertise with first-generation adenoviral, lentiviral, and adeno-associated viral (AAV) vectors, but relatively few groups have experience with HDAd and would typically need to establish a collaboration to obtain this expertise. The clinical applicability of the method may also be limited. For humans, a second surgery would increase both cost and complication risk. In addition, compared to human saphenous veins (used for coronary and peripheral bypass surgeries), rabbit external jugular veins have very thin walls. It is possible that the wall thickening and remodeling that occurs after grafting a rabbit external jugular vein into the arterial circulation would be attenuated if the vein graft already has a thicker wall. Finally, in experiments in this model lasting up to 6 months, none of the grafted veins developed a stenosis22. Therefore, this method does not appear to model all of the biological processes that contribute to vein graft failure.
In conclusion, the method uses a robust animal model of human vascular disease (rabbits) to generate grafted veins in which transgenes are stably expressed for prolonged periods of time (at least 6 months)22. Stable expression of transgenes in vein grafts will reveal their roles in normal vein graft homeostasis and will clarify their potential for vein graft gene therapy. The method could be improved and made more clinically applicable by the development of techniques that allow percutaneous delivery of vectors to the vein graft, thereby avoiding the second surgery. These techniques might be catheter-based36,37, or they could incorporate specially prepared vectors that are injected into a peripheral vein and then targeted to the vein graft wall, for example by magnetic fields38. The method would also allow testing of vectors other than HDAd for vein graft gene therapy, including lentiviral and AAV vectors as well as non-viral vectors.39,40,41,42
The authors have nothing to disclose.
Disposables | |||
3mL syringe with 24G needle | Becton Dickinson | 309571 | 2x for vein graft surgery; 2x for gene transfer surgery |
1 mL syringe with 27G needle | Becton Dickinson | 309623 | 2x for vein graft surgery, 5x for gene transfer surgery |
19G needle | Becton Dickinson | 305187 | Gene transfer surgery |
20 mL syringe, luer lock | Nipro Medical Corp | JD+20L | |
Catheters, 24Ga x 3/4” | Terumo Medical Products | SROX2419V | |
21G needle | Becton Dickinson | 305165 | Gene transfer surgery and for 20 mL syringe of saline |
Gauze 4” x 4” | Dynarex | 3242 | ~10-15 per surgery |
3-0 silk suture | Covidien Ltd. | S-244 | |
5-0 silk suture | Covidien Ltd. | S-182 | |
7-0 polypropylene suture | CP Medical | 8703P | Vein graft surgery |
7-0 polypropylene suture | CP Medical | 8648P | Gene transfer surgery |
5-0 polyglycolic acid suture | CP Medical | 421A | |
3-0 polyglycolic acid suture | CP Medical | 398A | |
Alcohol swabs | Covidien Ltd. | 6818 | For the placement of I.V. line |
Catheter plug | Vetoquinol | 411498 | |
Ketamine HCl, 100 mg/mL | Vedco Inc. | 5098916106 | |
Xylazine, 100 mg/mL | Akorn Inc. | 4821 | |
Lidocaine, 20 mg/mL | Pfizer | 409427702 | |
Marcaine 0.5% | Pfizer | 409161050 | |
Beuthanasia D-Special | Intervet Inc. | NDC 00061047305 | Harvest surgery only |
Buprenorphine HCl, 0.3 mg/mL | Patterson Veterinary | 12496075705 | |
Saline IV bag, 0.9% sodium chloride | Baxter | 2B1309 | |
Heparin (5000 U/mL) | APP Pharmaceuticals | NDC 63323-047-10 | |
Papaverine (3.5 mg/ml) | American Reagent Inc. | NDC 0517-4002-25 | Diluted from 30mg/mL stock; Use 1 mL maximum |
Fentanyl patch, 25 mcg/h | Apotex Corp. | NDC 60505-7006-2 | |
Isoflurane | Multiple vendors | Catalog number not available | |
Viral vector | Gene transfer surgery only | ||
Surgical Instruments | |||
Metzenbaum needle holder 7" straight | Roboz | RS-7900 | |
Operating scissors 6.5" straight blunt/blunt | Roboz | RS-6828 | |
Needle holder /w suture scissors | Miltex | 8-14-IMC | |
Castroviejo scissors | Roboz | RS-5658 | |
Castroviejo needle holder, 5.75" straight with lock | Roboz | RS-6412 | |
Stevens scissors 4.25" curved blunt/blunt | Roboz | RS-5943 | |
Alm retractor 4" 4X4 5mm blunt prongs | Roboz | RS-6514 | 2x |
Backhaus towel clamp 3.5" | Roboz | 4x | |
Micro clip setting forceps 4.75" | Roboz | RS-6496 | |
Micro vascular clips, 11 mm | Roboz | ||
Surg-I-Loop | Scanlan International | 1001-81M | 5 cm length |
Bonaccolto forceps, 4” (10 cm) long longitudinal serrations, cross serrated tip, 1.2mm tip width | Roboz | RS-5210 | |
Dumont #3 forceps Inox tip size .17 × .10 mm | Roboz | RS-5042 | |
Graefe forceps, 4” (10 cm) long serrated straight, 0.8 mm tip | Roboz | RS-5280 | |
Halstead mosquito forceps, 5" straight, 1.3 mm tips | Roboz | RS-7110 | 2x |
Halstead mosquito forceps, 5" curved, 1.3mm tips | Roboz | RS-7111 | |
Jacobson mosquito forceps 5" curved extra delicate, 0.9 mm tips | Roboz | RS-7117 | |
Kantrowitz forceps, 7.25" 90 degree delicate, 1.7 mm tips | Roboz | RS-7305 | |
Tissue forceps 5", 1X2 teeth, 2 mm tip width | Roboz | RS-8162 | |
Allis-Baby forceps, 12 cm, 4×5 teeth, 3 mm tip width | Fine Science Tools | 11092-12 | 2x |
Adson forceps, 12 cm, serrated, straight | Fine Science Tools | 11006-12 | |
Veterinary electrosurgery handpiece and electrode | MACAN Manufacturing | HPAC-1; R-F11 | |
Surgical Suite Equipment | |||
Circulating warm water blanket and pump | Multiple vendors | Catalog number not available | |
Bair hugger warming unit | 3M | Model 505 | |
IV infusion pump | Heska | Vet IV 2.2 | |
Isoflurane vaporizer and scavenger | Multiple vendors | Catalog number not available | |
Veterinary multi-parameter monitor | Surgivet | Surgivet Advisor | |
Veterinary electrosurgery unit | MACAN Manufacturing | MV-9 | |
Surgical microscope | D.F. Vasconcellos | M900 | 25X magnification for vein graft surgery; 16X magnification for gene transfer surgery |