global cardiac-specific transgene expression using cardiopulmonary bypass with cardiac isolation

2002;73:1939-1946 Ann Thorac SurgPatterson and Hansell H. Stedman

Chen, Charles B. Yarnall, Timothy J. Gardner, Alan S. Stewart, Mark M. Stecker, Terry Charles R. Bridges, James M. Burkman, Ramin Malekan, Stephane M. Konig, Haiyan

cardiac isolationGlobal cardiac-specific transgene expression using cardiopulmonary bypass with

http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939on the World Wide Web at:

The online version of this article, along with updated information and services, is located

Print ISSN: 0003-4975; eISSN: 1552-6259. Southern Thoracic Surgical Association. Copyright © 2002 by The Society of Thoracic Surgeons.

is the official journal of The Society of Thoracic Surgeons and theThe Annals of Thoracic Surgery

by on June 3, 2013 ats.ctsnetjournals.orgDownloaded from

http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939

http://ats.ctsnetjournals.org

Global Cardiac-Specific Transgene ExpressionUsing Cardiopulmonary Bypass With CardiacIsolationCharles R. Bridges, MD, ScD, James M. Burkman, MD, Ramin Malekan, MD,Stephane M. Konig, MD, Haiyan Chen, MD, Charles B. Yarnall, BS,Timothy J. Gardner, MD, Alan S. Stewart, MD, Mark M. Stecker, MD, PhD,Terry Patterson, PhD, and Hansell H. Stedman, MDDepartments of Surgery and Neurology, University of Pennsylvania Medical Center, Philadelphia, Pennsylvania

Background. The available techniques for intravasculargene delivery to the heart are inefficient and not organ-specific. Yet, effective treatment of heart failure willlikely require transgene expression by the majority ofcardiac myocytes. To address this problem, we developeda novel cannulation technique that achieves efficientisolation of the heart in situ using separate cardiopulmo-nary bypass (CPB) circuits for the heart and body in dogs.

Methods. The arterial inflow and venous effluent fromthe two circuits were physically isolated. The efficiencyof separation was 98% to 99% in three preliminaryexperiments using Evans Blue dye-labeled albumin. In 6dogs, the cardiac circuit was perfused with oxygenatedcrystalloid cardioplegia at 37°C containing � 4 � 1011

particles of an adenovirus encoding LacZ (AdCMVLacZ)with a perfusion pressure of 170 to 200 mm Hg for 15minutes allowing virus to recirculate through the heart �

15 times. Cross-clamp time was 26 � 2 minutes and CPBtime was 90 � 3 minutes.

Results. Five animals survived and were euthanized at7 days. �-Galactosidase activities measured using achemiluminescent assay were three orders of magnitudehigher in all areas of the heart than in the liver. Histo-logical analyses revealed heterogeneous X-Gal stainingof myocytes in all areas of the myocardium.

Conclusions. Despite using a constitutive promoter,this technique yields relatively cardiac-specific transgeneexpression and is potentially translatable to clinical ap-plications. Future studies will allow for further optimi-zation of the conditions necessary for vector-mediatedgene delivery to the heart.

(Ann Thorac Surg 2002;73:1939–46)© 2002 by The Society of Thoracic Surgeons

Heart failure is a major public health problem with aprevalence of over 2,000,000 in the United States

[1]. The incidence is more than 400,000 new cases peryear [2] with more than 200,000 deaths per year [3]. The5-year survival is less than 50% [3]. Nearly 60,000 patientsper year could benefit from cardiac transplantation orlong-term mechanical circulatory support. However, onlyapproximately 2,500 donor hearts are available for trans-plantation each year in the United States [4]. Genetherapy is emerging as a potential new therapy for thisvexing public health problem. Our hypothesis is that oneimportant reason for the lack of effective gene-basedtherapy to date is the unique obstacle of obtainingefficient gene delivery to the majority of myocytes in situ.Furthermore, experimental validation requires that thesedifficulties be overcome in a clinically translatable largeanimal model. Experimental methods of delivering trans-genes to the myocardium include intramuscular (IM) [5]or intracoronary [6–8] injection, or injection directly into

the left ventricular (LV) cavity with cross-clamping of theaorta [9]. Intramural injection allows for local myocytegene expression only [5]. The other two techniquesresult in a single-pass of vector through the heart withsystemic release of the unabsorbed fraction typicallyresulting in undesirably high-level transduction of theliver [8]. Using adenovirus encoding LacZ (AdCMV-LacZ) or adeno-associated virus encoding delta sarco-glycan (AAVCMV�sarcoglycan) in rodents, our recentstudies have shown that limb isolation with heterotopicgroin transplantation of the heart results in gene expres-sion by essentially 100% of cardiac and skeletal myocyteswhen vectors are coadministered with selected inflam-matory mediators [10]. These observations suggest thatovercoming the endothelial barrier is critical to achievingwidespread myocyte gene expression with Ad or AAVvectors via the intravascular route.

As a first step toward achieving this end in the heart,we present an exciting new technique that allows forefficient isolation of the heart in vivo using separatecardiopulmonary bypass (CPB) circuits for the cardiacand systemic circulations in dogs. We [11], along withDavidson and colleagues [12], simultaneously publisheddescriptions of the use of cardiopulmonary bypass for

Accepted for publication Feb 7, 2002.

Address reprint requests to Dr Bridges, Division of Cardiothoracic Sur-gery, Pennsylvania Hospital, 230 W Washington Sq, 3rd Flr, Philadelphia,PA 19106; e-mail: [email protected].

© 2002 by The Society of Thoracic Surgeons 0003-4975/02/$22.00Published by Elsevier Science Inc PII S0003-4975(02)03509-9



cardiac gene delivery. In contrast to their method whichinvolves standard cold cardioplegic arrest [12, 13], ourtechnique allows for isolation of the heart in situ. Thisnovel configuration is utilized to increase the theoreticalefficiency of transgene delivery using two separate, oxy-genated circuits for the heart and systemic circulations,respectively, allowing for multiple passes of vectorthrough the heart and control of perfusion temperaturewhile minimizing transgene delivery to other organs.

Material and Methods

All animals were treated in compliance with NIH publi-cation No. 82-23 as revised in 1985. Nine mongrel dogs(group 1, labeled albumin, n � 3; group 2, AdCMVLacZ,n � 6) were fasted overnight (12 hours). The animalswere sedated and induced for anesthesia with ketamine(10 mg/kg, IM), diazepam (0.5 mg/kg, IV), atropine(0.05 mg/kg, IM), and the vocal cords sprayed with 2%lidocaine. They underwent endotracheal intubation andmechanical ventilation (Drager anesthesia monitor,North American Drager, Telford, PA) with 100% oxygen.Anesthesia was started with 2% to 3% isoflurane andmaintained with 0.5% to 2% isoflurane. A rectal temper-ature probe was inserted, and a noninvasive pulse oxim-etry transcutaneous monitor placed on the tongue. Theanimals were prepared and draped using sterile tech-nique. All animals received cefazolin (1 gram, IV) prior toanesthetic induction. The surface electrocardiogram wasmonitored throughout the procedure. A percutaneouslyplaced right femoral catheter was used to monitor arte-rial blood pressure and obtain arterial blood gas samples.The right jugular vein was cannulated with an 8.5 Fr.introducer for volume infusion. The right common ca-rotid artery was isolated, and the incision left opentemporarily. The drapes were removed, and the animalwas again prepped and draped using sterile techniqueincluding both the chest, the right jugular, and internalcarotid arteries in the operative field. A sternotomyincision was performed using an oscillating saw. Thepericardium was opened, and the heart was suspendedin a pericardial cradle.

Cardiac Isolation TechniqueTwo separate cardiopulmonary bypass circuits were em-ployed with separate pump-oxygenators and reservoirsfor the systemic and cardiac circuits, respectively. Hepa-rin (3 mg/kg) was given intravenously. Purse stringsutures were placed in the proximal ascending aorta,proximal superior vena cava, and right atrium adjacent tothe inferior vena cava. The left carotid artery was cannu-lated, and the cannula was connected to the arterial limbof the systemic cardiopulmonary bypass circuit. A car-dioplegia cannula was passed through the purse string inthe proximal ascending aorta and connected to the arte-rial limb of the cardiac circuit. Two 26 Fr. right anglevenous cannulae were placed in the superior vena cava(SVC) and inferior vena cava (IVC), respectively. Theircombined outflow was passed via a Y-connector to thevenous limb of the systemic cardiopulmonary bypass

circuit. Snares were placed around the SVC and IVC toallow diversion of all systemic venous return into thevenous limb of the systemic cardiopulmonary bypasscircuit during the cardiac isolation interval. Cardiopul-monary bypass was initiated at 37°C. Ventricular ventcatheters were placed into the right and left ventricles,respectively. These vent catheters were connected via aY-connector to become the venous limb of the cardiaccircuit. The azygous vein was ligated. Systemic cooling to30°C was performed. The ascending aorta was cross-clamped, and 300 cc of crystalloid cardioplegia (Plegisol,Abbott Laboratories, Chicago, IL; pH adjusted with so-dium bicarbonate to pH � 7.4) at 4°C was administeredvia the aortic root. Diphenhydramine 50 mg and methylprednisolone 100 mg were administered intravenously.The SVC and IVC snares were tightened. The pulmonaryartery was cross-clamped proximally. In this way, allblood or crystalloid returning from the heart via thecoronary sinus or the Thebesian veins was then routedvia the right ventricular (RV) vent, and all blood thatregurgitated across the aortic valve was routed via the leftventricular vent to the cardiac venous reservoir. Thecardioplegia solution was oxygenated and returned tothe coronary circulation via the cardioplegia cannula inthe aortic root (cardiac arterial in-flow) and maintained at37°C. All blood returning from the systemic circulationwas routed via the SVC and IVC cannulae to the systemicvenous reservoir. Arterial in-flow via the cardioplegiacannula perfused only the cardiac circulation via the leftand right coronary arteries proximal to the aortic cross-clamp. Systemic arterial in-flow perfused only the sys-temic (noncardiac) circulation.

With the cardiac circulation effectively isolated fromthe systemic circulation, warm, oxygenated crystalloidcardioplegia (Plegisol), alone, (group 2) or containingEvans Blue dye-labeled albumin with histamine (10 mM)and papaverine (10 mg/L) (group 1) was infused into thecardiac circulation and re-circulated continuously forapproximately 15 minutes. In either case, the solution pHwas adjusted to 7.4 using sodium bicarbonate beforestarting the infusion. In group 2, 1 minute after cardiacisolation, a high-flow, four-way stopcock was used torapidly inject AdCMVLacZ, 0.4 ml of 1012 particles/mldirectly into the cardioplegia stream without interrup-tion. The cardioplegia/vector or cardioplegia/albuminsolution was infused at a perfusion pressure of 170 to200 mm Hg resulting in flows of approximately 250 to 400cc/min. The high perfusion pressure was chosen empir-ically to maximize the gradient for transendothelialtransport. This procedure allowed for passage of theentire volume of the cardiac circuit through the coronarycirculation approximately 15 times. When the isolationinterval was completed, 1 liter of Hespan (DuPont Phar-maceuticals, Wilmington, DE) with 100 mg of methylprednisolone and 50 mg of diphenhydramine added wasused to wash out the coronary circulation. The aorticcross-clamp was removed and snares loosened, mergingthe two circulations. Systemic rewarming was initiated.An additional dose of diphenhydramine 50 mg andmethyl prednisolone 100 mg were administered IV. The

1940 BRIDGES ET AL Ann Thorac SurgGLOBAL CARDIAC-SPECIFIC TRANSGENE EXPRESSION 2002;73:1939–46



LV and RV vent catheters were then used as systemicvents to decompress the heart. The animals were rou-tinely weaned from bypass after 30 minutes of reperfu-sion with epinephrine (1 mcg/min) and lidocaine (1mg/min) infusions. After 5 minutes of systemic reperfu-sion, the cross-clamp on the pulmonary artery was re-moved. Once the heat was contracting well, the LV andRV vent cannulae were removed, and inotropic supportwas discontinued. The cardiac isolation procedure isdepicted in Figure 1.

Labeled Albumin StudiesA total of 3 dogs (group 1) underwent labeled albuminstudies. For these animals, the cardiac isolation proce-dure above was followed. At the conclusion of the iso-lated cardiac perfusion interval, blood samples wereobtained from the cardiac and systemic venous reser-voirs for quantification of Evans Blue dye-labeled albu-min concentrations using a spectrophotometric assay.The two circulations were then merged, and the animalsweaned from cardiopulmonary bypass. In these acutestudies, the animals were euthanized 20 minutes aftersuccessful weaning from bypass. Specimens of the rightand left ventricle and the diaphragm were obtained forEvans Blue dye fluorescence histochemistry at 540 nm.The efficiency of separation of the two circuits was

defined as follows: Efficiency � [1 � (systemic concen-tration)/(cardiac concentration)] � 100.

Isolated Perfusion With AdCMVLacZFor the 6 animals in group 2, after weaning from bypass,the sternum was rewired, the sternotomy wound wasclosed, and the animals were allowed to recover. Withinapproximately 4 hours after completion of the procedure,the animals were weaned from mechanical ventilationand inotropic support, and all chest tubes were removed.The animals were returned to an oxygen cage overnightand allowed access to food and water. The next morningthey were transferred to cages without supplementaloxygen and sacrificed after 7 days. A total of 6 dogsunderwent the procedure described. Tissues were pro-cured at necropsy on day 7, following euthanasia of thedogs with an intravenous overdose of sodium pentobar-bital. Cryostat sections of the heart, liver, and lung wereincubated overnight at room temperature in X-Gal solu-tion. �-Galactosidase enzyme activity was quantifiedusing a chemiluminescent reporter assay system (Ga-lacto-Light Plus, Tropix, Inc, Bedford, MA).

Results

Labeled Albumin StudiesThese preliminary experiments were used to develop thesurgical technique utilized in the subsequent experi-ments. The concentrations of histamine (10 mM) andpapaverine (0.3 mg/ml) used were based on the concen-trations in previous isolated hindlimb experiments thatresulted in efficient vector-mediated gene transfer toskeletal muscle and to the heterotopically transplantedheart [10]. All animals were successfully weaned fromcardiopulmonary bypass. Evans Blue dye fluorescencehistochemistry at 540 nm from specimens of the right andleft ventricle and diaphragm obtained for histologicalstudies are demonstrated in Figure 2. There is extensiveand uniform exudation of labeled albumin into the inter-stitium of the myocardium of the left and right ventricles(Fig 2C and D, respectively), affecting the local environ-ment of essentially 100% of the myocytes. The interstitialspace is prominent with increased volume indicative oftissue edema. In contrast, there are only scant amounts oflabeled albumin present in the interstitium of the dia-phragm, and there is no interstitial edema (Fig 2A,B). Inthe diaphragm, the punctate “dots” of fluorescence (Fig2A) represent small amounts of labeled albumin remain-ing in the systemic capillaries surrounding the musclecells. The use of two separate CPB circuits with surgicalisolation of the heart resulted in an average to systemiccardiac-labeled albumin concentration ratio of 66.5. Thisresult is indicative of 98.3% efficiency of separation of thetwo circulations (Table 1). These findings are consistentwith the hypothesis that histamine and papaverine resultin a dramatic increase in endothelial permeability, aneffect that remains localized to the myocardium.

Fig 1. Open chest canine preparation during cardiopulmonary by-pass. Illustrated are all important components of the cardiac isola-tion scheme including the aortic cross-clamp, superior vena cava(SVC) and inferior vena cava (IVC) cannulae, SVC and IVC snares,pulmonary cross-clamp, cardiac arterial in-flow, and right ventricu-lar (RV) and left ventricular (LV) vent catheters.

1941Ann Thorac Surg BRIDGES ET AL2002;73:1939–46 GLOBAL CARDIAC-SPECIFIC TRANSGENE EXPRESSION



Studies With AdCMVLacZTo establish a best case scenario for the effects of vectorrecirculation alone, these animals underwent an other-wise identical cardiac isolation procedure without the useof histamine or papaverine. Photomicrographs obtainedfrom X-Gal staining of representative sections of the rightand left ventricles are presented in Figures 3 and 4,respectively. There is intense but heterogeneous trans-gene expression throughout the myocardium 7 days afterisolated perfusion of the heart in situ with AdCMVLacZ.Careful inspection of cross sections of the heart demon-strate that a significant portion of the observed X-Galstaining in the heart appears to be due to transduction ofendothelial cells rather than myocytes (Fig 3B,D and4A–D). Nonetheless, there is unequivocal evidence oftransgene expression in a number of cardiac myocytes(Fig 3A,C). On other sections, we found little if anyidentifiable staining of skeletal myocytes in the dia-phragm and only rare evidence of transduction of hepa-tocytes. Thus, although transgene expression is relativelycardiac-specific, there appears to be evidence of endo-thelial and myocyte transduction in the heart. In othersections of the right and left ventricles, we observed atendency to see higher numbers of cardiac myocytes

transduced when in close proximity to blood vessels oron the endocardial border of the LV surface (Fig 5A,B). Incontrast, in areas remote from blood vessels and remotefrom the endocardial surface, only scant staining ispresent both in the LV and RV (Fig 5C,D). A summary of�-galactosidase activities is provided in Figure 6. Basedon prior demonstrations that infection of endotheliumand myocardium with LacZ-encoding vectors results inseveral log higher levels of �-galactosidase than seenwith control adenovirus vectors (ie, those lacking theLacZ transgene) [14, 15], we interpret all of our histolog-ical and histochemical findings as evidence of transgene(as distinct from endogenous �-galactosidase) expres-sion. The average �-galactosidase activities in transducedcardiac muscle are also several orders of magnitudehigher than those in the liver.

All 6 animals survived the procedure and were weanedfrom cardiopulmonary bypass without difficulty. Thecardiopulmonary bypass time was 90 � 3 minutes andthe cross-clamp time was 26 � 2 minutes for the animalsin group 2. After weaning from bypass, 1 animal diedfrom an acute aortic dissection after a transient episodeof severe hypertension. The other 5 animals were weanedfrom mechanical ventilation within 4 to 6 hours of com-

Fig 2. Evans blue dye-labeledalbumin fluorescence histochemis-try: cross sections obtained fromthe diaphragm (A, 200�; B,40�), left ventricle (C, 100�),and right ventricle (D, 100�).

Table 1. Summary of Evans Blue Dye-Labeled Albumin Cardiac Isolation Experiments

Experiment No.Mean Systemic Concentration

(�g/mL)Mean Cardiac Concentration

(�g/mL) Cardiac/Systemic Ratio Efficiency (%)

1 1.29 125.9 97.6 99.02 0.31 17.2 46.5 97.93 0.80 37.0 55.4 98.2Mean � SEM 0.80 � 0.28 60.0 � 33.5 66.5 � 15.8 98.3 � 0.3




pletion of the procedure, survived 7 days, and wereeuthanized according to protocol. One of these 5 animalshad incomplete cardiac isolation during the cross-clampinterval due to inadvertent loosening of the snare aroundthe IVC and was excluded from the analysis. The remain-ing 4 animals in group 2 formed the basis for the analysis.Two animals suffered groin wound hematomas that im-proved with conservative treatment during the postop-erative period.

Comment

The results presented here differ in two important waysfrom most previously published studies. Despite the useof a constitutive promoter, relatively cardiac-specificgene expression was achieved. The �-galactosidase activ-ities in the heart were several orders of magnitude higherthan those in the liver. In contrast, Lilly and coworkersfound that intracoronary administration of AdCMVLuc

Fig 3. Gene transfer into rightventricular (RV) cardiomyocytesafter cardiopulmonary bypasswith isolated perfusion of the car-diac circulation in vivo with Ad-CMVLacZ. (A) RV outflow tract,100�; (B) RV free wall, 40�; (C)RV outflow tract, 200�; (D) RVfree wall, 40�, after infusion of4 � 1011 particles of AdCMV-LacZ and staining with X-Gal 1week postoperatively.

Fig 4. Gene transfer into leftventricular (LV) cardiomyocytesafter cardiopulmonary bypasswith isolated perfusion of thecardiac circulation in vivo withAdCMVLacZ. (A) LV apex,100�; (B) LV anterior wall,100�; (C) LV lateral wall, 100�;(D) intraventricular septum,100�, after infusion of 4 � 1011

particles of AdCMVLacZ andstaining with X-Gal 1 weekpostoperatively.




in the adult rabbit left circumflex coronary artery resultedin a threefold to 40-fold higher luciferase activity in theliver compared to the particular region of the heartexpected to have the greatest level of gene expression [8].Davidson and associates also used cardiopulmonary by-pass for delivery of adenovirus encoding the LacZ re-porter gene and the �-adrenoreceptor to the myocardiumin neonatal pigs [13]. However, they used a conventionalconfiguration for bypass with a single pump oxygenator.Their technique resulted in heterogenous gene expres-

sion throughout the myocardium. In contrast to themethod described here, their approach did not includeisolation of the cardiac circulation from the systemiccirculation. As a result, cold (4°C) cardioplegia was re-quired, and only a single pass of the vector through thecardiac circulation was achieved. After the cross-clampinterval, residual vector was washed out into the systemiccirculation. It is, therefore, somewhat surprising thattheir technique did not result in more significant geneexpression in the liver.

Fig 5. Gene transfer into left(LV) and right (RV) ventricularcardiomyocytes after cardiopul-monary bypass with isolated per-fusion of the cardiac circulationin vivo with AdCMVLacZ. (A)LV endocardium, 40�; (B) RV,adjacent to a large artery andvein, 40�; (C) LV region remotefrom blood vessels, 40�; (D) RVregion remote from blood vessels,40�, after infusion of 4 � 1011

particles of AdCMVLacZ andstaining with X-Gal 1 weekpostoperatively.

Fig 6. �-Galactosidase activities in the heart, liver,and lung (relative light units [RLU]/�g/103,mean � S.E.M.). (LV � left ventricle; RV � rightventricle.)




In the technique presented here, separate pump-oxygenators are used for the systemic and cardiac cir-cuits. This novel configuration allows for optimization ofa number of variables that have been shown to beimportant determinants of the efficiency of vector-mediated, intravascular cardiac gene delivery includingtemperature, pressure, ionic composition, flow rate, andexposure time [16]. Moreover, this system allows forrecirculation of the vector through the cardiac circulationmultiple times, theoretically increasing the efficiency ofvector-mediated transduction of cardiac myocytes. At theend of the cross-clamp interval, this technique allows forremoval of the vector from the cardiac circulation, mini-mizing the probability of transgene expression in otherorgans despite the use of a constitutive promoter.

In general, we found the highest �-galactosidase activ-ities in the right atrium and right ventricle (Fig 6). Wespeculate that the higher activities on the right side maybe due to a combination of direct and transvasculartransport of adenoviral vectors on the right side of theheart, while on the left side, primarily transvasculartransport occurs. Note that we retrieve the venous returnfrom the cardiac circuit from the right ventricular lumenvia the right ventricular vent (Fig 1). Therefore, on theendocardial surface, right ventricular and right atrialmyocytes are bathed directly with the cardiac circuitvenous effluent containing adenoviral vector. Since themajority of cardiac venous return occurs via the coronarysinus, the highest concentrations of vector are likely tocontact the endocardial surface of the right atrium. Incontrast, the left ventricular endocardium has directcontact with adenovirus only in those areas bathed byany physiological aortic regurgitation and from Thebes-ian veins draining into the left ventricle or atrium. Asexpected, we found that the vast majority of the venousreturn was from the right ventricular vent and only aminimal amount from the left ventricular vent.

The available literature [5–8] and our own results [10]clearly indicate that the probability of successful trans-duction of any cardiac myocyte, Pt is much less than 1.0when a vector such as AdCMVLacZ is delivered via asingle pass through the cardiac circulation. Without iso-lation of the heart as accomplished here, injected Advectors are most likely to transduce hepatocytes ratherthan cardiac myocytes even when injected directly intothe coronary arteries [8]. Therefore, in general, intracoro-nary injection [6–8] is likely to result in only a single passof a given vector through the cardiac circulation. Basedon these assumptions, the probability of myocyte trans-duction using the technique presented here should beP � nPt where n is the number of times the cardiacreservoir volume recirculates through the heart. Since, inthe experiments presented, the reservoir plus tubingvolume for the cardiac circuit is approximately 300 cc, atflow rates of 300 cc/minute for 15 minutes, n � 15resulting in a theoretical order of magnitude increase intransduction efficiency.

One possible limitation of our technique is the poten-tial for inactivation of the adenovirus due to incompati-bility with components of the bypass circuit including the

tubing and pump-oxgenator surfaces. Marshall and col-leagues showed elegantly that a variety of commonlyused catheter constituents such as stainless steel, nitinol,and polycarbonate rapidly and efficiently inactivate ade-novirus infectivity [17]. The primary component of tubingused for cardiopulmonary bypass circuits is polyvinylchloride (PVC). PVC was not one of the polymers testedin this study. However, polycarbonate is commonly usedin stopcocks and venous reservoirs, and the adenoviruswas usually drawn up through a metal needle. If inaggregate, the components of the bypass circuit lead tosignificant viral inactivation, the theoretical advantagesof virus recirculation would not be realized. Specificmeasures, such as the incorporation of human serumalbumin into the cardioplegia solution then could beused to minimize adenovirus inactivation [17].

We chose to use crystalloid cardioplegia for perfusionof the cardiac circuit rather than blood cardioplegia sinceadenovirus-mediated transduction of isolated myocytesis significantly more efficient in crystalloid media than inmedia containing red blood cells [16]. An oxygenator wasincorporated into the cardiac circuit, as well, to ensureadequate oxygen delivery to meet the metabolic de-mands of the normothermic heart. Oxygenated crystal-loid cardioplegia solutions at 37°C have been shown toprovide adequate myocardial protection for cross-clampintervals greater than 40 minutes with results compara-ble to blood cardioplegia in humans [18]. Temperature isanother important variable determining the efficiency ofvector-mediated gene transfer. Donahue and colleaguesfound that transduction efficiency was approximately tentimes higher at 37°C than at 4°C for isolated myocytesexposed to AdCMVLacZ for 10 to 15 minutes in vitro [16].Therefore, we chose to perfuse the cardiac circuit withoxygenated crystalloid at 37°C to maximize the theoreti-cal efficiency of vector-mediated gene transfer withoutcompromising myocardial protection.

We were encouraged by the relative specificity ofcardiac transgene expression achieved in this model,illustrating one clear advantage of this technique forcardiac gene delivery. Our failure to achieve successfultransduction of the majority of cardiac myocytes in situusing this technique was somewhat disappointing yet notunexpected. Previous work in our laboratories has impli-cated the endothelial barrier as the critical obstacle toobtaining widespread Ad- or AAV-mediated gene trans-fer to skeletal myocytes in situ and to the heterotopicallytransplanted heart in the rat and �sarcoglycan deficienthamster [10]. Overcoming this barrier by coinfusion ofselected inflammatory mediators such as histamine andpapaverine into the isolated hindlimb results in success-ful transduction of essentially 100% of skeletal myocytesin situ and 100% of cardiac myocytes in the heart hetero-topically transplanted into the groin [10]. The resultspresented here support our “restrictive-endothelium”hypothesis that the blood vessel wall is rate limiting forintravascular Ad or AAV-mediated gene transfer to skel-etal or cardiac myocytes in situ.

In summary, a new cardiac surgical technique has beenpresented that represents a novel approach to the goal of




achieving highly efficient vector-mediated gene transferthroughout the myocardium in situ in a large animal.This technique allows for control of temperature, perfu-sion pressure, exposure time, and the ionic compositionof the myocardial perfusion solution, all of which havebeen shown to be important determinants of the effi-ciency of vector-mediated gene transfer to the heart.Using two separate oxygenated bypass circuits for thecardiac and systemic circulations, 98% to 99% efficiencyof separation of the two circulations was achieved. Wefound that gene expression appeared to be most markedin endothelial cells and in myocytes adjacent to bloodvessels or the endocardial surface. Based on these andour own previously published results, we believe that theinclusion of strategies designed to increase endothelialpermeability will be necessary to achieve successfultransduction of the majority of cardiac myocytes in situvia the intravascular route. The technique presentedallows for unique opportunities to investigate systemat-ically a variety of approaches to manipulate the cardiaccirculation to optimize the efficiency of vector-mediatedgene transfer to the myocardium. This method is clini-cally translatable and could be used as an adjunct toother cardiac surgical procedures where cardiopulmo-nary bypass is utilized. The approach presented may alsohave applications to problems in drug delivery whereisolation of the heart is desirable such as organ-specificchemotherapy for neoplastic disorders.

This study was supported by a grant from the NIH (5-P01-AR/NS43648); grants from the Muscular Dystrophy Association ofAmerica, the Association Francaise Contre les Myopahthies, andthe U.S. Veterans Administration; and by funds from the Har-rison Department of Surgical Research, University of Pennsyl-vania Health System. Dr Bridges was supported by an NIHResearch Supplement for Minority Investigators (supplement to5-P01-AR/NS43648).

References

1. Schocken DD, Arrieta MI, Leaverton PE, et al. Prevalenceand mortality rate of congestive heart failure in the UnitedStates. J Am Coll Cardiol 1992;20:301–6.

2. Massie BM, Packer M. Congestive heart failure: currentcontroversies and future prospects. Am J Cardiol 1990;66:429–30.

3. Gillum RF. Epidemiology of heart failure in the UnitedStates. Am Heart J 1993;126:1042–7.

4. Goldstein DJ, Oz MC, Rose EA. Implantable left ventricularassist devices. N Engl J Med 1998;339:1522–33.

5. Svensson EC, Marshall DJ, Woodard K, et al. Efficient andstable transduction of cardiomyocytes after intramyocardialinjection or intracoronary perfusion with recombinant ad-eno-associated virus vectors. Circulation 1999;99:201–5.

6. Kaplitt MG, Xiao X, Samulski RJ, et al. Long-term genetransfer in porcine myocardium after coronary infusion of anadeno-associated virus vector. Ann Thorac Surg 1996;62:1669–76.

7. Shah AS, Lilly RE, Kypson AP, et al. Intracoronary adenovi-rus-mediated delivery and overexpression of the �2-adrenergic receptor in the heart: prospects for molecularventricular assistance. Circulation 2000;101:408–14.

8. Lilly RE, Peppel K, Kypson AP, et al. Intracoronary admin-istration of adenovirus for gene transfer into adult rabbitmyocardium. Surg Forum 1996;47:279–81.

9. Maurice JP, Hata JA, Shah AS, et al. Enhancement of cardiacfunction after adenoviral-mediated in vivo intracoronary�2-adrenergic receptor gene delivery. J Clin Invest 1999;104:21–9.

10. Greelish JP, Su LT, Lankford EB, et al. Stable restoration ofthe sarcoglycan complex in dystrophic muscle perfused withhistamine and a recombinant adeno-associated viral vector.Nature Med 1999;5:439–43.

11. Bridges CR, Burkman JM, Malekan R, Konig SK, et al. Use ofcardiopulmonary bypass with in situ cardiac isolation toachieve global cardiac-specific transgene expression in thecanine heart. Circulation 2000;102(18):II-739.

12. Davidson MJ, Shah AS, Jaggers J, et al. Cardiopulmonarybypass enables cardiac-specific transgene delivery in neo-natal pigs. Circulation 2000;102(18):II-165.

13. Davidson MJ, Jones JM, Emani SM, et al. Cardiac genedelivery with cardiopulmonary bypass. Circulation 2001;104:131–3.

14. Schneider DB, Sassani AB, Vassalli G, et al. Adventitialdelivery minimizes the proinflammatory effects of adenovi-ral vectors. J Vasc Surg 1999;29:543–50.

15. Barr E, Carroll J, Kalynych AM, et al. Efficient catheter-mediated gene transfer into the heart using replication-defective adenovirus. Gene Ther 1994;1:51–8.

16. Donahue JK, Kikkawa K, Johns DC, et al. Ultrarapid, highlyefficient viral gene transfer to the heart. Proc Natl Acad SciUSA 1997;97:4664–8.

17. Marshall DJ, Palasis M, Lepore JJ, et al. Biocompatibility ofcardiovascular gene delivery catheters with adenovirus vec-tors: an important determinant of the efficiency of cardio-vascular gene transfer. Mol Ther 2000;1:423–9.

18. Shanewise JS, Kosinski AS, Coto JA, et al. Prospective,randomized trial comparing blood and oxygenated crystal-loid cardioplegia in reoperative coronary artery bypassgrafting. J Thorac Cardiovasc Surg 1998;115:1166–71.




2002;73:1939-1946 Ann Thorac SurgPatterson and Hansell H. Stedman

Chen, Charles B. Yarnall, Timothy J. Gardner, Alan S. Stewart, Mark M. Stecker, Terry Charles R. Bridges, James M. Burkman, Ramin Malekan, Stephane M. Konig, Haiyan

cardiac isolationGlobal cardiac-specific transgene expression using cardiopulmonary bypass with

& ServicesUpdated Information

http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939including high-resolution figures, can be found at:

References http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939#BIBL

This article cites 16 articles, 5 of which you can access for free at:

Citationshttp://ats.ctsnetjournals.org/cgi/content/full/73/6/1939#otherarticlesThis article has been cited by 7 HighWire-hosted articles:

Permissions & Licensing

[email protected]: orhttp://www.us.elsevierhealth.com/Licensing/permissions.jsp

in its entirety should be submitted to: Requests about reproducing this article in parts (figures, tables) or

Reprints [email protected]

For information about ordering reprints, please email:


http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939

http://ats.ctsnetjournals.org/cgi/content/full/73/6/1939#BIBL

http://www.us.elsevierhealth.com/Licensing/permissions.jsp


global cardiac-specific transgene expression using cardiopulmonary bypass with cardiac isolation

Documents