Author + information
- Received October 5, 1998
- Revision received February 2, 1999
- Accepted March 15, 1999
- Published online July 1, 1999.
- Peter R Vale, MDa,
- Douglas W Losordo, MD, FACCa,
- Tengiz Tkebuchava, MDa,
- Donghui Chen, MDa,
- Charles E Milliken, MAa and
- Jeffrey M Isner, MD, FACCa,* ()
- ↵*Reprint requests and correspondence: Dr. Jeffrey M. Isner, St. Elizabeth’s Medical Center, 736 Cambridge Street, Boston, Massachusetts 02135
This study investigated the feasibility and safety of percutaneous, catheter-based myocardial gene transfer.
Direct myocardial gene transfer has, to date, required direct injection via an open thoracotomy.
Electroanatomical mapping was performed to establish the site of left ventricular (LV) gene transfer. A steerable, deformable 7F catheter with a 27G needle, which can be advanced 3 to 5 mm beyond its distal tip, was then directed to previously acquired map sites, the needle was advanced, and injections were made into the LV myocardium.
In two pigs in which methylene blue dye was injected, discretely stained LV sites were observed at necropsy in each pig, corresponding to the injection sites indicated prospectively by the endocardial map. In six pigs in which the injection catheter was used to deliver plasmid using cytomegalovirus promoter/enhancer, encoding nuclear-specific LacZgene (pCMV-nlsLacZ) (50 μg/ml) to a single LV myocardial region, peak beta-galactosidase activity after five days (relative light units [RLU], mean 135,333 ± 28,239, range = 31,508 to 192,748) was documented in the target area of myocardial injection in each pig. Percutaneous gene transfer of pCMV-nlsLacZ(50 μg/ml) was also performed in two pigs with an ameroid constrictor applied to the left circumflex coronary artery; in each pig, peak beta-galactosidase activity after five days (214,851 and 23,140 RLU) was documented at the injection site. All pigs survived until sacrifice, and no complications were observed with either the mapping or the injection procedures.
Percutaneous myocardial gene transfer can be successfully achieved in normal and ischemic myocardium without significant morbidity or mortality. These findings establish the potential for minimally invasive cardiovascular gene transfer.
From the time of the initial description of direct gene transfer to the artery wall by Nabel et al. (1), gene therapy has been proposed or investigated for a number of cardiovascular applications including prevention of restenosis, treatment of congestive heart failure and therapeutic angiogenesis (2). Each of these potential applications requires consideration of the appropriate gene, vector and delivery technique. To date, successful transfection with a variety of genes, using both viral and nonviral vectors, has been accomplished by intravascular gene transfer (3,4), direct transcutaneous intramuscular injection (5–9), transcutaneous pericardial injection (10), and direct intramyocardial injection (11–13).
Although intravascular, pericardial, and intramuscular gene transfer have all been performed using minimally invasive delivery techniques, intramyocardial gene transfer has to date required an operative thoracotomy. Such an approach clearly implies additional morbidity and limits the feasibility of repeat administrations. Successful execution of percutaneous, catheter-based myocardial gene transfer has not been previously reported.
Accordingly, we sought to investigate the safety and feasibility of a novel delivery catheter for percutaneous myocardial gene transfer. To determine whether the delivery catheter could be used in a site-specific manner, myocardial gene transfer was integrated with a previously described catheter mapping technology (14). The results of this preliminary study indeed establish that percutaneous myocardial gene transfer can be successfully achieved in normal and ischemic myocardium in a relatively site-specific fashion without significant morbidity or mortality. These findings thus establish the potential to replace currently used operative approaches with a minimally invasive technique for applications of cardiovascular gene therapy designed to target myocardial function.
Electromechanical left ventricular mapping
The NOGA system (Biosense, Johnson & Johnson) is designed to acquire, analyze, and display electroanatomical maps of the human heart. The maps are constructed by combining and integrating information from intracardiac electrograms acquired at multiple endocardial locations. Catheters designed for use with the NOGA system are equipped with an electromagnetic sensor, which provides real-time location of the catheter. As the catheter is moved along the endocardium, local endocardial electrograms, together with the catheter tip location, are reported simultaneously. The system then uses this information to construct a three-dimensional (3D) electroanatomical map that constitutes a geometrical representation of the left ventricle (LV).
The NOGA system analyzes both global and local parameters that characterize mechanical, dynamic, and electrical LV function. Functional analysis is based on local shortening as an index of local mechanical function, whereas measurement of local intracardiac signals determines viability based upon preserved electrical function. The combination of these data permits assessment of electromechanical coupling (15).
The mapping and navigation system consists of a locator pad, a reference catheter, a mapping catheter and a processing unit with a graphics computer (Silicon Graphics, Mountain View, California). A similar system, CARTO (Biosense), has been previously described in detail (14,16).
The locator pad consists of a triangular arrangement of three magnetic coils, which generate an ultra-low intensity magnetic field (0.02 G to 0.5 G). The pad attaches to the undersurface of the catheterization table.
The reference catheter (Cordis-Webster, Baldwin Park, California) is taped directly to the skin overlying the anterior or posterior chest wall within the frame of reference created by the three coils of the locator pad. It is used to detect small changes in intracardiac position due to respiration or movement of the subject. These small changes are analyzed by the computer to correct spatial information generated by the mapping catheter.
The mapping catheter (Cordis-Webster) is a 7F fused tip catheter with a miniature passive magnetic field sensor embedded within its distal tip. On the basis of the strength of the magnetic field emitted from the locator pad coils, this sensor maps the distance from each coil; these distances determine the radii of theoretical spheres around each coil. The intersection of these three spheres determines the location and orientation of the sensor in 6 degrees of freedom (x, y, z, roll, pitch and yaw), which thus indicates the position and rotation of the distal catheter segment. The accuracy of the sensor position in this low magnetic field is 0.8 mm and 5 degrees (14,16). Unipolar or bipolar signals (timing related to a reference signal) are also obtained from the distal tip of the catheter, allowing generation of activation times in relation to the position of the catheter in the heart.
The mapping procedure has been previously described in detail (14,16). Briefly, the reference catheter was placed within the field of reference. The mapping catheter was introduced via a femoral arteriotomy and advanced to the LV. Three points (high septum, high lateral wall, and apex) were obtained with fluoroscopic guidance to generate the initial 3D image of the LV. The location of the mapping catheter was gated to a reliable point in the cardiac cycle (recorded relative to the location of the fixed reference catheter at that time) and its location was continuously shown on the screen of the mapping computer. An icon of the mapping catheter is displayed superimposed on the 3D map, thus enabling catheter manipulation in relation to the 3D map.
At each site, three parameters are calculated to determine the stability of endocardial contact with the catheter tip: location, cycle length (CL) and local activation time (LAT). Location is a measure of how stable the tip location is between beats. The CL indicates the difference between the current CL and the median CL of the last 100 acquired points. The LAT is calculated as the interval between a reliable point on the body-surface electrocardiogram (ECG) and the steepest negative intrinsic deflection from the mapping-catheter unipolar recording, as determined from the intracardiac electrogram. This electrophysiologic information is color-coded (red being the shortest LAT and purple being the longest) and superimposed on the 3D chamber geometry. The reconstruction was updated in real time with the acquisition of each new site. Validation of both intracardiac signal recording and location accuracy, both in vitro and in vivo, and correlation between electromechanical characteristics and pathology, have been previously established (14–16). In addition, early reports suggest that LV mapping may allow the detection of on-line myocardial viability (17).
The injection catheter (Cordis-Webster) is a modified 7F mapping catheter, the distal tip of which incorporates a 27G needle that can be protruded 3 to 5 mm. The injection catheter was manipulated to acquire stable points within the target region based on the parameters described above and superimposed upon the previously acquired 3D electroanatomic map. Once a stable point was attained, the needle was advanced 4 to 5 mm into the myocardium; the intracardiac electrogram detected transient myocardial injury or premature ventricular contractions as evidence of needle penetration into the myocardium. Injectate was delivered according to one of three protocols as outlined below. Each injection consisted of 1 ml of solution (total volume = 6 ml/animal) delivered from a 1-ml syringe. The lumen was prefilled with 0.1 ml of sterile saline prior to entry into the circulation, and following each injection the lumen was again flushed with 0.1 ml of sterile saline. Following completion of the injection, the needle was retracted and the catheter was moved to another endocardial site.
A total of 10 swine weighing 30 to 50 kg each were studied under protocols approved by the Animal Care and Use Committee of the St. Elizabeth’s Medical Center. All procedures were performed under anesthesia using a combination of intramuscular (IM) ketamine (15 mg/kg), acepromazine (0.2 mg/kg), and atropine (0.05 mg/kg). All animals received inhalation ventilation with 2% isofluorane to ensure adequate anesthesia throughout the experiment, and supplemental oxygen @ 3 l/min. Before introduction of the mapping catheter, all pigs received intravenous (IV) cefazolin 500 mg (Kefzol, Eli Lilly, Indianapolis, Indiana), heparin 5,000 IU, bretylium tosylate (50 mg/kg), and lidocaine (0.5 mg/kg). Levodromaron (2 mg subcutaneous [s.c.]) was given for analgesia at the beginning of the procedure with supplemental doses as required; on completion of each procedure, animals received an additional injection of levodromaron. A further dose of lidocaine (0.5 mg/kg IV) was given before the injection procedure.
Following the mapping and injection procedures, the arteriotomy was closed and the pig was allowed to recover (except in Protocol 1 in which the animals were immediately sacrificed). Animals were observed during recovery until fully conscious, returned to housing 24 h later, and given cefazolin 2 × 500 mg s.c. daily for three to five days. At the end of the study period, the animals were returned to the laboratory, sacrificed with euthanasia solution (sodium pentobarbital), and the heart excised for macroscopic and microscopic evaluation.
Two healthy and nonischemic pigs each received six 1-ml injections of methylene blue dye. Injections were made in three LV myocardial regions in each pig. Immediately postoperatively, each animal was sacrificed and the heart removed. Both the success and the accuracy of myocardial injections were assessed by correlating the number and location of apparent injection sites on the in vivo 3D electroanatomical map with those identified at necropsy. The extent of myocardial staining was measured to identify the spread of dye from the injection site.
To determine the feasibility of using the injection catheter to perform gene transfer, six healthy, nonischemic pigs each received six injections of 1 ml of a reporter gene (see below) to a single area of LV myocardium. These pigs were sacrificed three to five days later and tissue obtained for quantitative analysis of gene expression.
Under general anesthesia administered via an endotracheal tube, ischemia was induced in two pigs by placing an ameroid constrictor around the proximal left circumflex (LCx) artery as previously described (13,18,19). Three weeks later, selective coronary angiography was performed to determine the maturity of the constrictor, and endocardial mapping was used to identify the area of ischemic myocardium (as evidenced by uncoupling of mechanical function and electrical activity). Each pig received six injections of 1.0 ml of LacZ(50 μg/ml) to a single LV area. Pigs were sacrificed after three to five days and tissue obtained for macroscopic and microscopic evaluation.
Naked plasmid deoxyribonucleic acid (DNA) encoding for nuclear-specific beta-galactosidase transcriptionally regulated by the CMV promoter/enhancer (pCMV-nlsLacZ) was used as a reporter gene to evaluate percutaneous myocardial gene transfer. Gene expression was evaluated with a chemoiluminescence assay (20)(Galacto-Light, Tropix, Bedford, Massachusetts) designed to measure beta-galactosidase activity. Before measuring beta-galactosidase activity, tissue homogenates were pretreated with Chelex 100 to inactivate a natural inhibitor of the enzyme (21).
During the mapping procedure, heart rate (premapping = 118 ± 8/min vs. postmapping = 120 ± 10/min), systolic blood pressure (BP, 92 ± 3 vs. 87 ± 4 mm Hg) and O2saturation (98 ± 0.5 vs. 99 ± 0.4%) remained stable. Mapping was associated with transient ventricular ectopic activity but no sustained ventricular arrhythmias. No other complications were associated with the mapping procedure.
Activation (electroanatomic) maps of the LV during sinus rhythm were created in all pigs. All maps were completed in <20 min. The mean number of points acquired per map was 93 ± 6 (45 to 127). The site of earliest activation was in each case at the superior part of the septum (red/orange); the latest site of activation was on the left lateral wall close to the mitral valve apparatus (purple). Before injection, electromechanic interrogation was performed, consisting of maximum voltage (electrical activity) and linear log shortening (mechanical function) maps. Electrically viable tissue produced maximum unipolar voltage >10 mV, and mechanically functional myocardium produced linear log shortening >5%. In all nonischemic pigs, both mechanical function and electrical activity were within normal limits. In the two pigs with an ameroid constrictor, evidence of myocardial ischemia was detected in the lateral wall as evidenced by electromechanical uncoupling (high electrical voltage but low linear log shortening).
Percutaneous LV gene transfer
Percutaneous catheter-based myocardial injections caused no significant changes in heart rate (preinjection = 120 ± 10/min vs. postinjection = 128 ± 11/min), systolic BP (87 ± 3 vs. 89 ± 4 mm Hg), or O2saturation (99 ± 0.4 vs. 98 ± 0.7%). Transient unifocal ventricular ectopic activity was observed at the time the needle was extended into the myocardium. In all pigs, sporadic premature ventricular contractions occurred during injection. No episodes of sustained ventricular (or atrial) arrhythmias occurred. No sustained injury pattern was observed during the injections as recorded by the endocardial electrogram. Likewise, the surface ECG showed no evidence of myocardial infarction in any pig. All pigs survived until sacrifice; complications, including pericardial effusion or cardiac tamponade, were not observed in any animals.
Six discrete sites of methylene blue staining, located in three LV myocardial areas (anteroapical [n = 2], septum [n = 2] and posterolateral wall [n = 2]), were identified at necropsy in each heart; these sites corresponded to the injection sites indicated prospectively in vivo on the endocardial map (Fig. 1). Myocardial staining was 5.2 ± 1.7 mm in depth and 6.4 ± 0.7 mm in width. No epicardial staining was demonstrated. In addition, X-gal staining produced no evidence of nuclear-specific beta-galactosidase activity in the myocardium at these sites (Table 1); these two hearts thus constituted negative controls (because no gene transfer was performed in either case) for Protocol 2 below.
The injection catheter was used to deliver pCMV-nlsLacZ(50 μg/ml) to a single LV myocardial region (Fig. 2)in six pigs (apex [n = 2], septum [n = 2], posterolateral wall [n = 1] and the anterior wall [n = 1]). In each of the six pigs, peak beta-galactosidase activity after five days (relative light units [RLU], mean = 135,333 ± 28,239, [31,508 to 192,748]) was documented in the target area of myocardial injection (Table 1; Fig. 3). ⇓Adjacent myocardial areas demonstrated low-level activity, and areas remote from the injection sites had negligible activity. Thus, percutaneous LV myocardial gene transfer was directed in a relatively localized fashion to those sites indicated by preinjection electroanatomical mapping.
Percutaneous gene transfer of pCMV-nlsLacZ(50 μg/ml) was also performed in two pigs in which an ameroid constrictor had applied to the left circumflex coronary artery (Fig. 4). Myocardial ischemia was demonstrated in the lateral wall of both pigs (Fig. 5). In each of these two pigs, peak beta-galactosidase activity after five days (214,851 and 23,140 RLU) was documented in the target area of myocardial injection (Table 1; Fig. 6). ⇓As in the nonischemic hearts, beta-galactosidase activity was markedly diminished in tissue sections retrieved from adjacent myocardium and was negligible at remote sites.
Although intra-arterial delivery was used to establish the precedent for cardiovascular gene transfer (1)and has been subsequently exploited to accomplish gene transfer successfully in a variety of animal models (2)as well as patients (3), this route of administration has several inherent limitations for myocardial gene transfer. In the case of naked DNA (i.e., DNA unassociated with viral or other adjunctive vectors), cellular uptake is virtually nil when the transgene is directly injected into the coronary artery lumen, presumably due to prompt degradation by circulating nucleases. Gene transfer performed to the coronary arterial wall itself requires access to a satisfactory donor site; in patients with chronic myocardial ischemia—particularly those patients whose anatomy is not amenable to angioplasty or bypass surgery—diffuse distribution of neointimal thickening or extensive calcific deposits (22)may limit gene transfer to the smooth muscle cells of the arterial media (20).
Ischemic muscle as a target for naked DNA
Ischemic muscle represents an alternative target for gene transfer. Striated (5–8)and cardiac (11,12,23)muscle have been shown to take up and express naked plasmid DNA as well as transgenes incorporated into viral vectors (9,13). Moreover, previous studies (6,24)have shown that the transfection efficiency of intramuscular (IM) gene transfer is augmented more than fivefold when the injected muscle is ischemic. This finding may be the result of the skeletal muscle regeneration, including stem cell (myoblast) proliferation. Vitadello et al. (25)reported an 80-fold increase in chloramphenicol acetyltransferase (CAT) activity following transfection of regenerating versus control muscle. Consistent with this concept, Danko et al. (26)found that bupivacaine, which produces myonecrosis followed by satellite cell (muscle stem cell) proliferation and myotube formation one to three days later, may be used to enhance the expression of naked DNA injected IM into striated muscles.
Therapeutic angiogenesis as a model for myocardial gene therapy
We have previously exploited these features of skeletal and cardiac muscle to perform gene transfer of naked DNA encoding for angiogenic growth factors. Preclinical animal studies from our laboratory established that IM gene transfer could be utilized to accomplish successful therapeutic angiogenesis in animals with hindlimb ischemia (6). Subsequently, Phase 1 clinical studies from our institution have established that IM gene transfer of naked DNA encoding for vascular endothelial growth factor (VEGF) may be utilized to accomplish safe and successful therapeutic angiogenesis in patients with critical limb ischemia (7).
The notion that this concept could be extrapolated to the treatment of chronic myocardial ischemia was implied by experiments performed in both our laboratory (19)and in others’ (13,27,28)in which recombinant human VEGF was administered to a porcine animal model of chronic myocardial ischemia. This same animal model has been utilized to demonstrate that therapeutic angiogenesis can also be successfully achieved by direct myocardial administration of VEGF, as naked DNA (29)or using an adenoviral vector (13). Preliminary results utilizing this approach as sole therapy (no bypass) for patients with chronic myocardial ischemia suggest that direct injection of VEGF into cardiac myocytes improves collateral blood flow and markedly reduces the frequency of and threshold for myocardial ischemia (30,31).
Percutaneous myocardial gene transfer
To date, all of the aforementioned work involving myocardial gene transfer has been achieved via a mini-thoracotomy used for less invasive coronary artery bypass surgery. Although this approach has clearly reduced the length of hospital stay and morbidity associated with conventional bypass surgery, it nevertheless requires general anesthesia and is not easily repeatable. The capability to perform myocardial gene transfer percutaneously could thus further reduce the morbidity and facilitate repeat use of myocardial gene transfer.
The experiments described in this report suggest that percutaneous myocardial gene transfer is indeed feasible and can be safely performed in normal and ischemic myocardium of swine (Table 1). Injection of methylene blue was successfully achieved at 6/6 (100%) sites in two pigs. The extent of transmural distribution was limited to the LV wall, as evidenced by the absence of epicardial staining. Maximum gene expression was localized to the injection site in all pigs, ischemic as well as normal, injected with pCMV-nlsLacZ. Myocardium adjacent to the target sites of injection demonstrated low-level beta-galactosidase activity, indicating limited distribution following gene transfer. All remote noninjected areas of myocardium were essentially devoid of beta-galactosidase activity.
Role of adjunctive electromechanical mapping
Although the mapping capabilities of the NOGA system utilized in this study were useful for demonstrating that gene expression could be directed to specific LV sites, it must be acknowledged that these findings do not establish that LV endocardial mapping is required for percutaneous myocardial gene transfer. Electroanatomic mapping clearly may be advantageous both for avoiding gene transfer to sites of myocardial scar and for relocating with accuracy the tip of the injection catheter to areas of myocardial ischemia (or hibernating myocardium) where gene transfer may be potentially optimized (6,32). Theoretically, adjunctive mapping could be employed in a serial fashion to gauge the success of certain gene therapy strategies (e.g., therapeutic angiogenesis). These potential advantages, however, will require further documentation and assessment of physiologic improvement following delivery of a nonreporter gene to establish the value of adjunctive mapping.
☆ This study was supported in part by NIH grants HL-53354, HL-57516 and HL-60911 (Dr. Isner), a grant from the E.L. Weigand Foundation, Reno, Nevada, and the Peter Lewis Educational Foundation. Dr. Vale is the recipient of a travelling fellowship from the St. Vincent’s Clinic Foundation, Sydney, Australia.
- left ventricle
- left circumflex coronary artery
- plasmid using cytomegalovirus promoter/enhancer, encoding nuclear-specific LacZgene
- relative light units
- vascular endothelial growth factor
- Received October 5, 1998.
- Revision received February 2, 1999.
- Accepted March 15, 1999.
- American College of Cardiology
- Nabel E.G,
- Plautz G,
- Nabel G.J
- Nabel E.G
- Wolff J.A,
- Malone R.W,
- Williams P,
- et al.
- Tsurumi Y,
- Takeshita S,
- Chen D,
- et al.
- Baumgartner I,
- Pieczek A,
- Manor O,
- et al.
- Tripathy S.K,
- Svensson E.C,
- Black H.B,
- et al.
- Tripathy S.K,
- Goldwasser E,
- Lu M.-M,
- Bar E,
- Leiden J.M
- Lin H,
- Parmacek M.S,
- Morle G,
- Bolling S,
- Leiden J.M
- Mack C.A,
- Patel S.R,
- Schwarz E.A,
- et al.
- Gepstein L,
- Hayam G,
- Ben-Haim S.A
- Gepstein L,
- Goldin A,
- Lessick J,
- et al.
- Kornowski R,
- Hong M.K,
- Leon M.B
- White F.C,
- Carroll S.M,
- Magnet A,
- Bloor C.M
- Dietz W,
- Tobis J.A,
- Isner J.M
- Banai S,
- Jaklitsch M.T,
- Shou M,
- et al.
- Tio R.A,
- Lebherz C,
- Scheuermann T.H Jr.,
- Tkebuchava T,
- Magner M,
- Symes J
- Vale P.R,
- Losordo D.W,
- Symes J.F,
- Isner J.M
- Losordo D.W,
- Vale P.R,
- Symes J,
- et al.
- Wijns W,
- Vatner S.F,
- Camici P.G