Author + information
- Received August 15, 2002
- Revision received November 2, 2002
- Accepted December 18, 2002
- Published online June 4, 2003.
- Craig A Thompson, MD*,‡∥,* (, )
- Boris A Nasseri, MD†,‡,
- Joshua Makower, MD¶,
- Stuart Houser, MD§,
- Michael McGarry, MSc§,
- Theodore Lamson, PhD§,
- Irina Pomerantseva, MD, PhD*,‡,
- John Y Chang, MS ME¶,
- Herman K Gold, MD, FACC*,
- Joseph P Vacanti, MD†,‡ and
- Stephen N Oesterle, MD, FACC*,‡
- ↵*Reprint requests and correspondence:
Dr. Craig A. Thompson, Knight Center for Cardiac Catheterization and Intervention, Massachusetts General Hospital, 55 Fruit Street, Blake 950, Boston, Massachusetts 02114, USA.
Objectives The study evaluated a nonsurgical means of intramyocardial cell introduction using the coronary venous system for direct myocardial access and cell delivery.
Background Direct myocardial cell repopulation has been proposed as a potential method to treat heart failure.
Methods We harvested bone marrow from Yorkshire swine (n = 6; 50 to 60 kg), selected culture-flask adherent cells, labeled them with the gene for green fluorescence protein, expanded them in culture, and resuspended them in a collagen hydrogel. Working through the coronary sinus, a specialized catheter system was easily delivered to the anterior interventricular coronary vein. The composite catheter system (TransAccess) incorporates a phased-array ultrasound tip for guidance and a sheathed, extendable nitinol needle for transvascular myocardial access. A microinfusion (IntraLume) catheter was advanced through the needle, deep into remote myocardium, and the autologous cell–hydrogel suspension was injected into normal heart. Animals were sacrificed at days 0 (n = 2), 14 (n = 1, + 1 control/collagen biogel only), and 28 (n = 2), and the hearts were excised and examined.
Results We gained widespread intramyocardial access to the anterior, lateral, septal, apical, and inferior walls from the anterior interventicular coronary vein. No death, cardiac tamponade, ventricular arrhythmia, or other procedural complications occurred. Gross inspection demonstrated no evidence of myocardial perforation, and biogel/black tissue dye was well localized to sites corresponding to fluoroscopic landmarks for delivery. Histologic analysis demonstrated needle and microcatheter tracts and accurate cell–biogel delivery.
Conclusions Percutaneous intramyocardial access is safe and feasible by a transvenous approach through the coronary venous system. The swine offers an opportunity to refine approaches used for cellular cardiomyoplasty.
Nearly 5 million Americans currently live with congestive heart failure (CHF). This disease accounts for 962,000 hospitalizations and 287,200 deaths annually. In 1998, over $3.6 billion was paid to Medicare beneficiaries for CHF expenditures, and the prevalence of this disease continues to increase (1). Orthotopic cardiac transplantation remains limited by donor supply, and by need for lifelong immunosuppression, invasiveness, and cost. Both cell- and gene-based therapies have been proposed as potential alternatives for treatment of this often recalcitrant patient population (2–7).
Current methods of direct cell and gene delivery have notable limitations. The target patient populations, those with recent myocardial infarction (MI) or advanced heart failure, typically are at higher risk for conventional surgical approaches and for general anesthesia. Furthermore, open surgical approaches have limited septal wall access. Several catheter-based methods, including endoventricular (8–11), intracoronary infusion (12), and coronary vein retroinfusion (13)are being evaluated. Intracoronary and coronary venous infusions are less specific than direct injections and may potentially decrease the therapeutic yield. Endoventricular methods may have limited access to the myocardium in the area of the submitral valve apparatus and may be unstable in the mobile ventricular wall.
We introduce a new concept for direct myocardial access using the coronary venous system as a roadmap to the heart (Fig. 1) and a composite catheter system to allow direct and stable access to remote myocardium (Figs. 2 and 3) ⇓with intravascular ultrasound (IVUS) and fluoroscopic guidance. ⇓This method provides an alternate platform for cell-, gene-, and drug-based cardiovascular bioengineering therapies (14). The objective of this study was to test feasibility, accuracy, and safety (death, cardiac tamponade, sustained arrhythmia) of this transvenous delivery method.
This investigation was approved by the Massachusetts General Hospital and United States Army Subcommittee on Animal Research Care, and was in compliance with the 1996 Guide for the Care and Use of Laboratory Animals and with the Animal Welfare Act. This study was designed primarily to assess the safety and feasibility of percutaneous, trans(coronary)venous, direct myocardial cell delivery using an engineered substrate of an autologous bone marrow subpopulation within a biodegradable polymer.
Bone marrow harvest and cell preparation
Six Yorkshire swine (∼50 kg) received sedation with intramuscular (IM) acepromazine 5 mg/atropine 1 mg and intravenous telazol 110 mg/Xylazine 250 mg. Five of the six animals were subject to bone marrow harvest (the sixth as a negative control). The iliac crest was accessed with a bone marrow aspiration needle (MD Tech, Gainesville, Florida), and ∼50 cc of bone marrow was aspirated into a sterile, heparinized syringe. These cells were injected through a 50-μm filter to exclude large particles and into a 75-cm2uncoated, vented polystyrene Falcon flask (Corning, Inc., Corning, New York). The adherent population was cultured in high glucose Dulbecco’s Modified Eagle Medium (GIBCO, Auckland, New Zealand) to 80% to 90% confluence. This population of adherent cells was intended to be analogous to the putative mesenchymal stem cell population (15,16), but was not further characterized for this study.
These adherent cells were distributed to 30% confluence on six-well plates (Becton Dickson, Franklin Lakes, New Jersey) and transduced with murine green fluorescence protein (GFP) using a vesiculostomatitis vector. These cells were qualitatively assessed under direct fluorescence within three days for GFP expression. All animals had 70% to 75% of cells expressing GFP determined by qualitative assessment of two independent observers. The cells were expanded in culture by serial advancement through 25-cm2, 75-cm2, and 150-cm2Falcon flasks and passaged in 1:2 fashion when 90% confluence was reached.
At the time of procedure, the autologous cells were recombined and resuspended in 0.3% collagen biogel (Cellagen, ICN Biomedicals, Aurora, Ohio), with the addition of tissue dye for gross pathologic identification (n = 5 pigs). An aliquot of cells was evaluated for fluorescence using flow cytometry. This cell–biogel preparation of ∼2.0 × 107cells/ml was divided evenly into 1 cc luer-tipped syringes. Collagen biogel/tissue dye without cells was used in the negative control animal.
Trans(coronary)venous myocardial access procedure
The swine were sedated, ventilated, and monitored (cardiac rhythm, oxygenation, and blood pressure), and then prepped and draped in standard surgical fashion. Six French (F) arterial (Cordis, Miami, Florida) and 14F venous (Transvascular, Menlo Park, California) femoral sheaths were placed percutaneously. The left main coronary artery was selectively engaged with a 6F Hockeystick diagnostic catheter (Medtronic) and angiography performed with emphasis on venous follow-through phases to determine coronary venous anatomy, anomalies, and coronary sinus location.
The coronary sinus (CS) was accessed by placing a 7F Porcine 3 catheter into the right ventricle, withdrawing with clockwise torque across the tricuspid valve. Using this technique, the catheter tends to fall into, or near, the CS. An exchange length, 0.035-inch hydrophilic angled wire (Teruma) with J-tip was advanced into the CS, through the great cardiac vein (GCV), and into the anterior interventricular coronary vein (AIV) (Fig. 4A). The diagnostic catheter was withdrawn, with the guide wire in place, and a 14F CS guiding catheter (Transvascular) and introducer were placed with conventional over-the-wire technique. After removal of introducer, a subselective catheter (Transvascular) was then placed over the wire, through the CS guide, and into the AIV (Fig. 4B). The hydrophilic guide wire was then exchanged for a 0.014-inch HiTorque floppy (Guidant, Temecula, California) guide wire. The TransAccess catheter is a 6F, monorail, composite catheter system combining a phased-array IVUS (compatible with JOMED IVUS, JOMED, NV) and a pre-shaped, sheathed, extendable 24-gauge nitinol needle (Fig. 3A). This TransAccess catheter was advanced over the 0.014-inch guide wire and into the AIV in preparation for myocardial access (Fig. 4C).
Intravascular orientation was performed using the corresponding artery, pericardium, and ventricular chamber as landmarks with IVUS imaging (Figs. 4D and 4E). After confirmation of position within the coronary vein and with respect to surrounding structures, the nitinol needle was extended into the myocardium (Fig. 4F). A 27-gauge microinfusion (IntraLume) catheter was advanced through the needle and into the myocardial tissue (Figs. 4G and 4H). Because the myocardial tissue is a potential space, and without room for prolapse, all of the force for the otherwise floppy IntraLume catheter is forward, essentially allowing this catheter tip to become a drill capable of tunneling through remote myocardium in plane with the needle puncture (Fig. 3B).
The animals (n = 6) were sacrificed at 0- (n = 2), 14- (n = 2), and 28-day (n = 2) timepoints with IV potassium chloride 40 mEq overdose. All animals received cell/collagen biogel/tissue dye except the negative control animal (collagen biogel/tissue dye without cells), which was one of the animals sacrificed at the 14-day timepoint. Accuracy was assessed by correlating the dye stains identified by gross examination with intended anatomic and fluoroscopic landmarks. Histologic sections of the injection sites were subdivided into basal, mid- and apical subsegments of the: 1) anterior, 2) lateral, 3) anteroseptal, and 4) inferoseptal walls to allow for correlation with the respective angiographic counterparts.
Tissue processing and staining
Portions of porcine myocardium acquired from injection sites were frozen in Tissue Tek optimal cutting temperature compound (Sakura Finetek USA Inc., Torrance, California) and stored at −80°C for subsequent histologic analysis. Histopathologic evaluation was performed on 5-μm cryosections of the tissue. Sections were examined with a fluorescence microscope, using an NIB filter, and subsequently stained with hematoxylin–eosin (H&E) and Masson’s trichrome stains for study by light microscopy.
Immunohistochemical study of additional sections was performed using anti-mouse GFP primary antibody (clone-20, Sigma-Aldrich, St. Louis, Missouri) that cross-reacts with pig. Binding was visualized by using a horseradish peroxidase conjugated secondary antibody (Vector, Burlingame, California). Bound antibody was revealed by incubation of cryopreserved sections in aminoethylcarbazole (Dako, Carpinteria, California) or diaminobenzidine (Zymed, San Francisco, California). Sections were counterstained with hematoxylin solution (Sigma). Negative controls were checked by omitting the primary antibody. Cryosections of skin from a GFP transgenic mouse (Jackson Laboratory, Bar Harbor, Maine) were used as a positive control for the immunohistochemical analysis. Given the relatively poor expression of GFP in animals from each timepoint, we focused histologic evaluation on the animal with the highest percentage GFP expression at each timepoint.
Coronary venous access and direct myocardial injection was successful in 100% (6/6) of the animals attempted. Access was gained via the AIV approach to the basal-, mid-, and apical-anterior, lateral, anteroseptal, and inferoseptal left ventricular myocardium. Eighty-eight transvenous punctures were performed (mean 14.6/animal). No death, cardiac tamponade, or sustained arrhythmia occurred during the procedure or the follow-up periods to time of sacrifice (range 0 to 28 days).
Dye correlation on gross examination with angiographic landmarks was 100% (Figs. 5 and 6). ⇓After cell recombination, ⇓GFP expression determined by flow cytometry was relatively poor in several animals (51% and 5% GFP% expression for the two 28-day animals, 5% GFP% expression for the 14-day animal, and 45% and 2% GFP% expression for the two acute sacrifice animals). We suspect that this decrement in GFP% expression is due to a disproportionate expansion of the initial unlabeled cell fraction. The GFP-positive, autologous “donor” bone marrow cells were determined to be present at the 0-, 14-, and 28-day timepoints, based on expression of signal on fluorescence microscopy (Fig. 7) or density of expression above background levels by immunohistochemistry, using light microscopy on sampled sections of targeted myocardium that were demarcated with the biogel/tissue dye (Figs. 8 and 9). ⇓Excessive fibrosis was not seen ⇓in comparison with control areas of normal myocardium (Fig. 8). Determining cell proliferation and transdifferentiation was outside the scope of this study.
Conventional medical and surgical approaches are often inadequate to treat patients with myocardial disease. Both cell- and gene-based bioengineering approaches have been proposed as potential alternate therapies to augment or substitute damaged heart tissue (2,3). We present the initial experience of a novel, catheter-based endovascular platform for direct myocardial cell delivery using the coronary venous system as a “roadmap to the heart” and a composite catheter system to allow a direct, transvenous puncture and access to the myocardium. We hypothesize that this method will have advantages over presently utilized approaches.
Currently available methods for cardiovascular cell and/or gene delivery have potential limitations. Direct surgical injection can provide a high level of substrate per unit area, but this may result in high morbidity, and potential mortality, in the target patient populations with recent MI or advanced CHF. Futhermore, standard, transepicardial surgical approaches do not grant free access to the septal wall and do not afford the opportunity for real-time, contrast-enhanced assessment of the microcirculation. The TransAccess catheter system shares the advantages associated with surgical transepicardial access regarding accuracy and substrate deposition, but it only requires local anesthesia at the femoral access site. In addition, the septal wall is readily accessed, and “online” microcirculatory assessment is readily available. Endoventricular approaches are limited primarily by platform “stability.” Endoventricular catheter systems do not rotate with the heart, which can minimize target accuracy. The pressure of syringe injection with these systems can be destabilizing and cause expulsion of the needle tip from the myocardial interface. Needle withdrawal provides an exit point for cells, gene therapy substrates, and so forth to be released to the systemic circulation. Thinned myocardial tissue (<5 mm) is currently considered a contraindication because of concerns for transepicardial perforation.
The TransAccess system overcomes these hurdles because the needle and Microlume catheter are advanced in a coaxial direction with the myocardial tissue. The catheter is therefore well seated deep within the heart muscle, rotates with the heart, and is not subject to needle expulsion or cell loss directly from the exit tract. Preliminary data suggest that acute retention efficiency with this transvenous method and fluoroscopic guidance is superior to electromechanical mapping-guided endoventricular approaches (17). We hypothesize that the deeper penetration into the myocardium and the stability of the IntraLume microinfusion catheter are largely responsible for this observation. In theory, the risk of stroke from ventricular embolization or excessive catheter manipulations and dwell time is reduced by access and delivery from the venous side. Furthermore, thinned tissue, in concept, is no more a limitation for IntraLume injections than is standard open surgical approaches. We are currently evaluating the second-generation CrossPoint TransAccess catheter, which has a much more flexible shaft. This new catheter system can easily track over a 0.014-inch guide wire, obviating the need for the subselective catheter component. The flexibility of this catheter is well suited for subselective vein access, such as the middle cardiac vein (which parallels the posterior descending artery) for enhanced inferior wall access.
We used a collagen biogel as a cell “delivery vehicle.” The role of biogels and biodegradable polymers is poorly understood in cardiovascular applications. Our bias is that cell packaging and delivery may be as important as the cell sourcing for cardiovascular bioengineering applications (4–7,18–21). We hypothesize that biogels and biodegradable polymers may provide protection from physical compression and lysis of the cells during their harsh transition to the myocardium, provide architectural and nutritional support during the engraftment process, and facilitate vascular ingrowth and/or passive cell nutrition via diffusion processes. Our experience with biodegradable polymers for the application of cellular cardiomyoplasty is that they may create ambiguity with the distinction of donor-host cellular relationships, a factor that must be carefully considered and circumvented when designing investigations in which such relationships are the primary outcome measure. The utility of such biopolymers to facilitate cellular cardiomyoplasty is subject to ongoing investigation.
We identified what we believe to be autologous bone marrow cell implants in normal porcine myocardium at 0-, 2-, and 4-week timepoints, based on GFP expression using a variety of complementary histologic modalities in selected myocardial sites that clearly retained the tissue dye and biopolymer gel. Bone marrow cell subpopulations that are adherent to uncoated flasks may contain mesenchymal stem cells capable of multilineage potential through mesenchymal pathways (muscle, bone, adipose, stroma) (15). Murine investigations suggest that such bone marrow cell subpopulations can be differentiated into a cardiomyocyte phenotype in vitro and, perhaps, in vivo (22,23).
In addition, studies with conditioned media suggest that bone marrow cell subpopulations may secrete high levels of vascular endothelial growth factor and MCP-1, and potentially can recruit vascular supply (8). Our cell population was cultured in a similar fashion (15), but the cells were not characterized prior to injection with cell surface markers, and interspecies differences may certainly exist. This investigation was designed primarily to assess the catheter delivery system, and not transdifferentiation potential of our cell source. We observed labeled cell orientation with the porcine myocardium at the four-week timepoint. It is interesting to speculate about the potential of this cell population to transdifferentiate into myotubules as a milieu-dependent process. However, this is clearly not proof of concept and must be subject to rigorous investigation specifically designed to assess such end points.
In summary, we present the initial experience of intracardiac, autologous bone marrow cell transplantation with a new catheter-based system that allows direct myocardial access with IVUS-guided needle punctures through the coronary venous system and infusion catheter placement into remote myocardium. This method of cell delivery may have potential advantages over currently available approaches, and it provides a new, stable platform for cardiovascular bioengineering therapies. Autologous cell populations from adult bone marrow may have a potential role for in vivo tissue engineering strategies for cardiac angiomyogenesis.
We appreciate the assistance of Tom Aretz, MD, Susan Boucher, Farouc Jaffer, MD, PhD, and Luis Guerrero in the conduct of this investigation.
☆ Joshua Makower, Michael McGarry, Theodore Lamson, and John Y. Chang are employees of TransVascular, Inc.
This work was supported by the Center for Integration of Medicine and Innovative Technology (CIMIT), Cambridge, Massachusetts, and by TransVascular, Inc., Menlo Park, California. Craig A. Thompson received grant support from the Clinical Investigator Training Program: Harvard/MIT Health Sciences and Technology–Beth Israel Deaconess Medical Center, in collaboration with Pfizer Inc.
- anterior interventricular coronary vein
- congestive heart failure
- coronary sinus
- great cardiac vein
- green fluorescence protein
- intravascular ultrasound
- myocardial infarction
- Received August 15, 2002.
- Revision received November 2, 2002.
- Accepted December 18, 2002.
- American College of Cardiology Foundation
- ↵American Heart Association. 2002 Statistical Update. Dallas, TX: American Heart Association
- ↵Thompson C, Oesterle S. Biointerventional cardiology: the future interface of interventional cardiovascular medicine and bioengineering. Vasc Med 2002;7:135–40
- Langer R.,
- Vacanti J.P.
- Fuchs S.,
- Baffour R.,
- Zhou Y.F.,
- et al.
- Losordo D.W.,
- Vale P.R.,
- Hendel R.C.,
- et al.
- Yang X.,
- Atalar E.,
- Li D.,
- et al.
- Boekstegers P.,
- von Degenfeld G.,
- Giehrl W.,
- Kupatt C.,
- Franz W.,
- Steinbeck G.
- Thompson C.,
- Nasser B.,
- Makower J.,
- et al.
- ↵Pittenger MF, Mackay AM, Beck SC. Multilineage potential of adult human mesenchymal stem cells. Science 284;284:143–147
- ↵Smits P, Reijs A, van der Giessen W, et al. Efficiency and retention of a percutaneous transendomyocardial injection of VEGF165 by a fluoroscopy guided transendomyocardial injection catheter. In: XIVth World Congress of Cardiology, Sydney, Australia, 2002
- Hakuno D.,
- Fukuda K.,
- Makino S.,
- et al.