Cardiovascular Research 64 (2004) 12 – 23 www.elsevier.com/locate/cardiores

Review

Genes, stem cells and biological pacemakers Michael R. Rosen a,b,c,*, Peter R. Brink c, Ira S. Cohen a,c, Richard B. Robinson a a

Center for Molecular Therapeutics, Deparment of Pharmacology, Columbia University, New York, NY, USA b Department of Pediatrics, Columbia University, New York, NY, USA c Departments of Physiology and Biophysics, Institute of Molecular Cardiology, SUNY Stony Brook, Stony Brook, NY, USA Received 16 April 2004; received in revised form 13 May 2004; accepted 18 May 2004 Time for primary review 17 days Available online

Abstract

Keywords: Arrhythmia therapy; Ion channels; Gene therapy; ECG; Gap junctions

1. Introduction With few exceptions, the therapy of cardiac disease focuses on prevention or palliation; only rarely (as in some forms of congenital heart disease) is complete cure a realistic goal. Yet now there is greater promise of cure, deriving in part from attempts to revascularize the myocardium, generate new myocytes, and engineer novel functions into existing or newly generated cells. Most of this research shares the techniques of gene therapy and/or cell therapy. Among the strategies employed have been insertion of genes to replace those which are malfunctioning (e.g., cystic fibrosis [1]), creation of new vasculature (e.g., vascular endothelial growth factor [2]), and more recently creation of a cellular substrate to supplant abnormally functioning myocardium (e.g., cell implants to seed poorly contracting myocardium) [3]. These strategies have * Corresponding author. Center for Molecular Therapeutics, Deparment of Pharmacology, Columbia University, 630 West 168 Street, PH 7 West321, New York, NY 10032, USA. Tel.: +1-212-305-8754; fax: +1-212-3058351. E-mail address: [email protected] (M.R. Rosen).

been met with great fanfare and variable success. Certainly, revascularization and replacement of malfunctioning myocardium carry the promise of improved contractile function. As fringe benefits, they might also reduce the incidence of arrhythmias. There is less knowledge of the direct applicability of gene/cell therapy to arrhythmia prevention or termination. Yet, gene and cell transfer techniques offer the possibility to design and create both a trigger to initiate the heartbeat and an excitable substrate capable of propagating it. An important step in demonstrating proof of concept for gene- and cell-based antiarrhythmic therapies has been the study of potential triggers of the heartbeat, that is, the generation of biological pacemakers. There have been at least two stimuli for designing and fabricating biological pacemakers: (1) the desire to improve on electronic pacemakers that are currently the state of the art for treating many rhythm disorders; (2) to use this paradigm as a template for developing other gene/cell-based antiarrhythmic strategies. In this review, we will define a biological pacemaker, consider the needs we envision for biological pacemakers as therapeutic tools, and review strategies in use for their development.

0008-6363/$ - see front matter D 2004 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cardiores.2004.05.012

Downloaded from by guest on January 20, 2017

The advent of gene therapy and cell therapy has led to reconsideration of standard therapies for cardiac disease. One such area of reconsideration is that of the cardiac pacemaker, which has been the mainstay of treatment for high-degree heart block and sinoatrial node dysfunction. Over the past five years, gene therapy has been used to explore the overexpression of h2-adrenergic receptors, the downregulation of inward rectifier current, and the overexpression of pacemaker current as potential sources of biological pacemakers. Cell therapy approaches have explored the ‘‘forcing’’ of embryonic stem cells to evolve along cardiac (and specifically pacemaker) cell lines and the use of adult mesenchymal stem cells as platforms for delivery of specific gene therapies. This review considers the strengths and weaknesses of each of the approaches used to date and attempts to look to the future of biological alternatives to electronic pacemakers. D 2004 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

1.1. What is a biological pacemaker? The sinoatrial node is the primary biological pacemaker in the heart and a potential template for any biological pacemaker to be fabricated. All channels and transporters necessary to generate the heartbeat in mammals are present in the sinoatrial node myocyte membrane and in other pacemaker tissues of the conducting system and myocardium. The sinoatrial node has the following characteristics:

The sinus node generates phase 4 depolarization by activating an inward current, If, which opens over hundreds of milliseconds-to-seconds, a time course well-attuned to physiologic heart rates (Fig. 1) [4,5]. However, other inward

and outward currents contribute to the pacemaker potential. Inward currents include the L- and T-type calcium currents [6,7] and tetrodotoxin-sensitive Na current [8]. Outward currents include the Na/K pump [9] and delayed rectifier, IK [10]. Still, other currents (e.g., the Na/Ca exchanger) can operate in inward or outward directions [11]. Given the ensemble function of these channels, any net increase in inward current and/or decrease in outward current would tend to increase heart rate. Importantly, the sinoatrial node is not the only biological pacemaker in the mammalian heart. Other tissues including atrium, atrioventricular node, His – Purkinje system, and even the ventricular muscle can initiate pacemaker potentials in specific experimental or clinical settings. Such pacemaker function is readily seen as atrioventricular junctional escape rhythms and idioventricular escape rhythms manifested during high-degree atrioventricular block. These escape rhythms are progressively slower as sites distal to the sinus node are involved [12]. This redundancy is an exquisite ‘‘fail safe’’ mechanism providing cardiac output in lifethreatening situations. A major reason for the slower rates at sites distal to the sinoatrial node is the presence of a background potassium current, IK1, that is minimal-to-absent in the sinoatrial node [13]. IK1 is an inwardly rectifying current, which is large at diastolic potentials. As an outward current, it opposes the depolarizing influence of If. Because IK1 is larger in ventricular muscle than the His – Purkinje system and because If activates at more positive voltages in the His – Purkinje system than myocardium [14], the initiation of escape rhythms by the His –Purkinje system is favored. The biophysical properties of IK1 have suggested a strategy to increase ventricular pacemaker expression and idioventricular rates, that is, to reduce the magnitude of myocardial IK1. In principle, this would depolarize the membrane, an effect augmented by unopposed inward currents (as has, in fact, been demonstrated [15]). Hence, the heart incorporates a reservoir of potential biological pacemakers. However, based on the rate it generates, its modulation by autonomic influences and its position with regard to orderly activation of the myocardium, the sinoatrial node, which is ideally suited to drive the heart, provides the optimal template for a fabricated biological pacemaker.

2. What is the need for extrinsic pacemakers? Fig. 1. A family of If currents recorded from a myocyte isolated from the rabbit sinoatrial node. Note that time-dependent inward current was readily observable at 55 mV within the diastolic range of potentials. Also note that the current activates over hundreds of milliseconds to seconds, ideally tuned to pacemaker depolarization. The holding potential was 35 mV, and the temperature was 32 jC. The permeabilized patch-clamp technique was employed to avoid rundown of the current (reprinted by permission from Shi W, Wymore R, Yu H, et al. Distribution and prevalence of hyperpolarization-activated cation channel (HCN) mRNA expression in cardiac tissues. Circ Res 1999; 85:e1 – e6.).

As recently as the mid-20th century, many patients with complete heart block were at risk of death [16,17]. Therapy in adults was largely limited to positive chronotropic interventions, typically sublingual isoproterenol administration, which required repetition every two hours. The first mass-produced implantable pacemakers were fixed rate units featuring the attractiveness and dimensions of a sterile hockey puck, but they saved lives.

Downloaded from by guest on January 20, 2017

1. It is a specialized tissue well-integrated electrically and structurally into the cardiac substrate, whose pacemaker current is optimally suited (see below) to initiate the cardiac impulse. 2. Its operation is determined by interactions among ion channels and pumps to initiate phase 4 depolarization and carry membrane potential towards the threshold voltage at which action potentials are elicited. 3. It initiates an impulse whose rhythm tends to be regular while the rate varies in nearly instantaneous response to the physiological and emotional demands of the body. The modulatory checks and balances on rate and rhythm are largely the province of the sympathetic (excitatory) and parasympathetic (inhibitory) nervous systems. 4. The propagation of its impulse maximizes efficiency of cardiac contraction and output.

13

14

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

Improvements in design and manufacture, insightful adaptation of computer technologies to provide programming and microcircuitry, and imaginative approaches to a variety of cardiac pathologies have ultimately seen pacing used epicardially as well as endocardially to treat not only disorders of heart rate and rhythm but heart failure as well [17]. The development of cardioverters/defibrillators and their incorporation in the pacemaker industry represented a further major advance [18]. Hence, the hardware and the methods initially applied to a very limited spectrum of heart rhythm disorders developed into the medical device industry and into one of the most successful and effective palliative therapies devised in the 20th century.

3. Why replace electrical with biological pacemakers? Given the extraordinary effectiveness of electronic pacemakers in reducing mortality associated with complete heart block and in reducing the morbidity of sinoatrial node dysfunction, they are clearly a formidable medical therapy. Yet they have shortcomings:

Given these concerns, then as good as electronic pacemakers are, a biological alternative that might last for the life of the patient, respond to physiologic demands for different heart rates at different times, and activate the heart via a pathway tailored to the anatomy of disease in any individual is an exciting possibility.

We propose that the ideal biological pacemaker would (1) create a stable physiologic rhythm for the life of the individual; (2) require no battery or electrode, and no replacement; (3) compete effectively in direct comparison with electronic pacemakers; (4) confer no risk of inflammation/infection; (5) confer no risk of neoplasia; (6) adapt to changes in physical activity and/or emotion with appropriate and rapid changes in heart rate; (7) propagate through an optimal pathway of activation to maximize efficiency of contraction and cardiac output; (8) have limited and preferably no arrhythmic potential; (9) represent cure, not palliation. Whereas characteristics 1– 5 are self-evident, 6 – 9 require some explanation as follows: clearly, if it is to fulfill the physiologic requisites of a primary initiator of cardiac rhythm, the biological pacemaker must not only develop a stable pacemaker potential but this must respond to autonomic regulation. With gene therapy, the transfected myocyte likely incorporates adrenergic and muscarinic receptor coupling pathways and will generate hormone-responsive rhythms if the transfected gene that regulates current is susceptible to autonomic regulation. In contrast, for cell therapy receptor expression and appropriate coupling to effectors may be more formidable challenges, depending on the gene and the functional signaling cascade in the cell system used. Depending on the extent of intercurrent cardiac disease, the optimal site for impulse initiation will likely vary among patients. Hence, in most young, otherwise healthy individuals with heart block, His – Purkinje activation would be optimal. In contrast, in the presence of extensive atherosclerosis, old infarct, or fibrosis accompanying cardiomyopathy or aging, the optimal pathway may require more individualization. An additional benefit of identifying an ideal activation path for each patient might be a reduction of arrhythmogenesis. In other words, by testing whether stimulation at a given site and over a range of physiologic rates in individual patients initiates reentrant or triggered rhythms, one might prioritize regions for pacemaker placement in any patient who harbors an arrhythmogenic cardiac substrate. Another way to limit arrhythmogenicity is to create a biological pacemaker that neither prolongs nor shortens the action potential duration, nor increases transmural dispersion of repolarization. One means to do so would create the pacemaker by altering diastolic membrane currents while not altering any membrane currents active in the voltage ranges that characterize action potential repolarization. If, the primary pacemaker current, fulfills this requisite [4,5].

Downloaded from by guest on January 20, 2017

(1) Electronic pacemakers impose limitations on the exercise tolerance and cardiac rate-response to emotion. Despite the use of paradigms to improve heart rate response during increased physical activity, there is no substitute currently available for the autonomic modulation of heart rate. (2) In pediatric patients, the age and size of the patient, the mass of the power pack, and the size and length of the electrode catheter become important considerations. Moreover, the hardware must be tailored to the growth of the patient. (3) The placement site of the stimulating electrode in the ventricle and the resultant activation pathway may have beneficial or deleterious effects on electrophysiologic or contractile function. Although epicardium and endocardium can be accessed, one must identify a site at which the electrode remains stably affixed. (4) The battery has a long-but-limited life expectancy, requiring testing and replacement at periodic intervals. (5) Intercurrent infection may require removal and/or replacement of the pacemaker. (6) Various devices including neural stimulators [19], metal detectors, and magnetic resonance imaging equipment [20] have been reported to interfere at times with electronic pacemaker function.

4. What are the ideal characteristics of a biological pacemaker?

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

5. What are the strategies for building a biological pacemaker?

We will now discuss gene therapy, as this has been employed in various models to explore all three strategies described above, and then will continue with cell therapy. 5.1. Gene therapy 5.1.1. Overexpression of b-adrenergic receptors In these experiments, a plasmid incorporating the gene encoding the h2-adrenergic receptor was injected into the atria of pigs [21,22]. Forty-eight hours later, atrial rhythms occurred, whose rate-response to h-adrenergic stimulation significantly exceeded control. Although these experiments demonstrated that biological pacemaking was not just a concept but a potential reality, concerns exist regarding this strategy, as follows: (1) the duration of effect was brief (only 24 h); (2) there was no testing of whether such a pacemaker could drive the heart during sinoatrial or atrioventricular block; (3) modulating h-adrenergic responsiveness, even in a localized area of the heart, will not provide new pacemaker channels; rather, it will only modulate the rate of those present and to the extent that they are autonomically responsive; (4) if there is sinoatrial node dysfunction, h receptor overexpression may incorporate the risk of worsening some manifestations of the disease. 5.1.2. Down-regulation of outward, hyperpolarizing current The logic here was that diminishing the hyperpolarizing currents that clamp myocardial membrane potentials at negative voltages would permit inward currents to contribute to membrane depolarization and a pacemaker potential [15,32]. Proof of concept was provided in experiments targeting Kir2.1, the gene encoding one of the a subunits of the channel carrying the inward rectifier current, IK1 [15]. The strategy was to create a dominant negative construct by replacing three amino acid residues in the pore of Kir2.1. This construct and green fluorescent protein (GFP) were packaged in an adenoviral vector and injected into the left ventricular cavity of guinea pigs. After 3– 4 days, there was about 80% suppression of IK1, as well as the electrocardiographic demonstration of idioventricular rhythms in the hearts of the treated guinea pigs. Moreover, action potential recordings from myocytes disaggregated from these hearts demonstrated phase 4 depolarization and rapid automatic rates (Fig. 2). Concerns regarding this strategy are the following: (1) IK1 was down-regulated at sites throughout the ventricle, which could create competing idioventricular foci; however, focal injection of the construct would likely obviate this concern; (2) reduction of repolarizing current in this fashion results in about a 15% prolongation of repolarization [32]; this raises concerns regarding the potential for proarrhythmia via increased dispersion of repolarization and/or occurrence of early after-depolarizations; (3) the ionic nature of this pacemaker is uncertain in that several inward currents would contribute to driving the heart in the setting of

Downloaded from by guest on January 20, 2017

Three strategies have been reported to create biological pacemaker activity: (1) up-regulation of neurohumoral (specifically h-adrenergic) actions on heart rate [21,22]; (2) reduction of repolarizing current [15]; and (3) increasing inward current during diastole [23]. All three strategies had their foundations in 20th century pharmacology and physiology. In studies of autonomic modulation, h-adrenergic catecholamines or sympathetic stimulation increased heart rate via an increase in pacemaker current in the sinus node and in accessory pacemakers [24], whereas increasing vagal tone or stimulating muscarinic receptors decreased heart rate [25]. In studies of ionic determinants of pacemaker activity, augmentation of hyperpolarizing, outward currents decreased pacemaker rate [26], suggesting that the opposite intervention, i.e., decreasing hyperpolarizing, outward currents, would increase rate [15]. Finally, other pharmacological experiments demonstrated that suppressing inward current carried by If [27] or by the T-type or L-type Ca channel [28] slows pacemaker rate. Given that the theoretical basis for building a biological pacemaker had its foundation in 20th century pharmacology, we needed only the tools to apply this knowledge to the molecular/genetic determinants of the pacemaker potential. The necessary information was provided in part via the identification and cloning of the gene products that determine the h-adrenergic receptors, the inward rectifier current, and the pacemaker current, If, itself. Also of central importance was the development of tools for (1) gene therapy, wherein genes encoding the molecular subunits of interest are inserted via plasmids or viral vectors into cells of the myocardium; (2) cell therapy via the use of embryonic stem cells, whose differentiation is directed into myocardial precursors manifesting pacemaker activity, or mesenchymal stem cells used as platforms to implant channels into cardiac myocytes. A critical factor is the development of models in which to test pacemaker constructs. In vitro models of cells in culture are a standard for testing a variety of gene therapies. We have found that infecting neonatal rat ventricular myocytes with replication-deficient adenoviral constructs incorporating the gene of interest (with or without coexpression of GFP) provides a cost-effective and reproducible assay [29]. We use a variation on this model for testing the ability of stem cells to transmit the electrical signal of interest [30]. We consider a 100 times or more overexpression of current and a statistically significant effect on beating rate as standards that discriminate efficacy, but more research is required to establish uniform guidelines permitting reliable correlation of in vitro and in vivo effectiveness. As an intact animal screen, the use of guinea pig [15], swine [22], and dog [23,30,31] has been reported. Our use of dog is based on its cardiac size, tractability as a chronic model, and similar electrophysiologic properties to those of man.

15

16

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

Fig. 2. Expression of pacemaker function in guinea pig ventricular myocytes and on ECG following suppression of Kir2.1 channels. (a) Control action potential evoked by depolarizing external stimuli in ventricular myocyte from untreated animal. (b) Spontaneous action potentials in Kir2.1AAA-transduced myocytes with suppressed IK1. (c) Baseline ECG from control animal in normal sinus rhythm (control). (d) Ventricular rhythm 72 h after transduction of Kir2.1AAA. P waves (A, arrow) and wide QRS complexes (V, arrow) are dissociated from one another (reprinted from Ref. [15] by permission).

5.1.3. Overexpression of inward depolarizing current Our own studies have focused on If as the primary candidate for a biological pacemaker for two major reasons: (1) it is the only current that flows only at diastolic potential ranges and hence would have no effect to prolong action potential duration, and (2) it is well-regulated by the autonomic nervous system. The logic driving these experiments derived from the identification and cloning of the HCN (hyperpolarization-activated, cyclic nucleotide-gated) family of gene products that determine the primary pacemaker current, If [36,37]. There are four isoforms of HCN, and three of these (1, 2, and 4) are present in the heart. They are all activated upon hyperpolarization and directly bind cAMP at the C-terminus. The cAMP binding shifts activation towards more positive voltages, effectively increasing inward current at a given voltage [38]. There are only small differences in the voltage dependence of activation for the three cardiac isoforms [38]; hence, the choice of HCN gene to serve as a biological pacemaker was based on cAMP sensitivity and channel kinetics. HCN1 activation exhibits rapid kinetics but is only minimally affected by cAMP binding so it was not chosen. HCN2 and HCN4 both respond strongly to cAMP, but HCN2 has faster kinetics than HCN4 [38] and thus was the isoform of choice for our initial studies. For our proof-of-concept studies, we hypothesized that HCN2 overexpression in myocytes might create an If -like pacemaker current capable of initiating the heartbeat. Proof of concept was first obtained in cell culture experiments, in

which rat myocytes were incubated with an adenoviral construct that incorporated the HCN2 isoform [33]. These myocytes manifested IHCN2 (the analog of If) over 100-fold greater in magnitude than If in wild-type myocytes. The infected cultures beat significantly faster than controls, and the cells responded to the positive chronotropic effects of isoproterenol and to the negative chronotropic effects of carbamylcholine (Fig. 3). A variation on this experiment was then performed in the hearts of intact dogs. Animals received a left atrial injection of adenovirus containing HCN2 as well as GFP, while controls received adenovirus incorporating GFP alone [21]. About 4 days later, the animals were anesthetized and right vagal stimulation was used to suppress sinus rhythm. Animals receiving HCN2 generated atrial beats originating from near the injection site; this did not occur in animals receiving GFP alone. Moreover, myocytes disaggregated from the site of injection showed IHCN2 current over 100-fold greater than native If . To test whether such a pacemaker could drive the ventricle during transient heart block, other dogs were subjected to catheter injection into the left bundle branch of adenovirus containing HCN2 and GFP [31]. One negative aspect of this experiment was that the catheter injection itself resulted in ecchymosis formation and ventricular ectopic rhythms arising from near the injection site for the first 48 h after injection (regardless of whether the injection incorporated HCN2 + GFP, GFP alone, or simply saline). However, these arrhythmias then ceased. The vagal stimulation protocol described above was then performed to induce atrioventricular block. Those animals receiving HCN2 developed a stable idioventricular escape rhythm (rate approximating 60 bpm) originating at a site near the

Downloaded from by guest on January 20, 2017

reduced IK1. This might give rise to interindividual and regional variability of inward currents, creating rhythms of unpredictable behavior [33 –35].

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

17

Fig. 3. Effect of autonomic agonists on neonatal rat ventricle cultures expressing mHCN2 delivered via an adenovirus. Spontaneous action potentials were recorded at room temperature from monolayer cultures with a patch electrode. Left: spontaneous action potentials recorded in control solution (top) and during superfusion of the same culture with 10 7 M isoproterenol (bottom). Right: recording from another culture in control solution (top) and during superfusion with 10 7 M carbamylcholine (bottom). Records provided by Dr. Andrea Barbuti.

activity of the construct used, which was anticipated to be a matter of days to weeks. Summarizing gene therapy approaches, only If has provided evidence of stable idioventricular escape rhythms having physiologically acceptable rates as well as evidence of autonomic responsiveness. Testing of the other approaches with regard to these criteria is awaited. Moreover, for all gene therapy approaches attempted to date, there has been no ‘‘head-to-head’’ comparison with the efficacy and duration of effect of electronic pacemakers,

Fig. 4. Effect on ECG (leads I, II, III) of injecting adenovirus containing GFP alone (upper panel) and HCN2 plus GFP (lower panel) on canine ECG 4 days after implantation. Upper panel: sinus rhythm interrupted by vagal stimulation (arrow) is followed by a very slow idioventricular rhythm. Lower panel: vagal stimulation (arrow) is followed by slowing of sinus rate and then by sinus arrest and an idioventricular rhythm faster than that with GFP alone.

Downloaded from by guest on January 20, 2017

HCN2 injection (Fig. 4). This rate was significantly faster than that of controls; the latter also manifested diverse right and left ventricular sites of impulse initiation. Isolated left bundle branches also revealed automatic rates significantly greater than those of control and GFP-injected animals. Finally, immunohistochemical and biophysical evidence of HCN2 overexpression were demonstrated [31]. Concerns regarding this strategy included (1) a 5- to 30-s period of quiescence after cessation of sinus rhythm before expression of the idioventricular rhythm; (2) the question of duration of

18

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

evidence that is important to our understanding of how to proceed in this field. Finally, gene therapy that is dependent on viral constructs is itself suspect. The replication-deficient adenoviruses used in some studies represent an episomal strategy in which the genetic material is not incorporated into the genome. Such therapy is not permanent. Other viral vectors introduce questions of infectivity and neoplasia that await exploration and solution. Hence, there are important concerns that must be addressed as this field moves forward. 5.2. Embryonic stem cells

5.3. Adult mesenchymal stem cells Mesenchymal stem cells are multipotent, having the property to differentiate into a number of cell lines including musculoskeletal and connective tissues, such as bone, cartilage, muscle, tendon, and fat. These cells are readily available from various tissues of the body, although the usual means of harvest is currently via the bone marrow. A property of these cells that makes them potentially attractive for clinical use is that they may be immunopriviliged; that is, in limited studies they do not appear to have elicited major immune responses [44]. Such a property raises the possibility that banked sources of adult mesenchymal stem cells may serve as a reservoir for human administration, whereas individuals who are not viewed as recipients for banked cells can provide a source of autologous stem cells via their own bone marrow. However, the entire question of ‘‘immunoprivilege’’ must receive more concerted study before solid conclusions can be reached. The ion channel properties of adult human mesenchymal stem cells were reported by Heubach et al. [45]. Outward currents were recorded in almost all cells. These included a Ca-activated K current active at potentials positive to + 20 mV and a clofilium-sensitive outward current at potentials positive to 30 mV. In a few cells L-type calcium currents were recorded. We found robust expression of two specific gap junctional proteins, connexins (Cx) 40 and 43 [46]. This conclusion is based on immunhistochemistry, Western blots demonstrating protein expression and electrophysiologic recordings, and dye-transfer experiments showing cell – cell coupling between stem cell pairs and between stem cells and cardiac myocytes [30,46]. Given their stability in cell lines and their potential for low antigenicity, we viewed adult human mesenchymal stem cells as a potential platform for administering gene therapy and for delivering small molecules. Their potential role as such a platform was supported by their ability to transfer dye and to transmit current to one another, to other cell lines, and to myocytes [30,46]. Moreover, immunostaining demonstrated that adult human mesenchymal stem

Downloaded from by guest on January 20, 2017

Extensive research is now being performed on animal and human embryonic stem cells. These are pluripotent and can potentially differentiate into most of the cell types of the body. Hence, it is theoretically possible to use embryonic stem cells to create/replace any organ, a possibility that one day may be realized. The major problems encountered to date with respect to human embryonic stem cells have been sociopolitical (seen as governmental restrictions on human embryonic stem cell research) and technical. The former problem is beyond the scope of this review, although in urgent need of thoughtful, unemotional resolution. The latter problem holds for both animal and human embryonic stem cell lines; scientists have to learn how to identify cells that are precursors of a specific lineage early after cell isolation and how to ensure that cells can be brought to differentiate along and to maintain identity within that specific lineage. Those subsets of embryonic stem cells that are of a cardiac lineage are most actively being investigated with regard to replacing damaged myocardium [39,40]. The potential to use cells for this purpose holds exciting promise for the treatment of conditions such as intractable heart failure. Because subsets of embryonic stem cells initiate impulses similarly to pacemaker cells, their use as biological pacemakers is being explored [41]. Several problems have been identified to date. One is the property of immature embryonic stem cells to terminally differentiate and in so doing to lose their pacemaker characteristics. Hence, driving these cells down a cardiac lineage and then stopping them precisely and uniformly at the sinoatrial node stage will be a true tour de force. Another issue is the uncertainty regarding whether those embryonic stem cells that have pacemaker function will generate spontaneous rhythms as a result of the operation of the same ion channels that control pacemaker function in the human heart. Also uncertain is their autonomic responsiveness. A different type of problem arises from the report of arrhythmic events in mouse embryonic stem cells [42]. Although heterogeneity of repolarization with occurrence of after-depolarizations and triggered activity was demonstrated in these cells, the same phenomena are seen with a variety of mature mammalian myocytes when they are disaggregated and studied as single cells or are studied in culture [43]. Hence, while this observation of arrhythmogenicity certainly was reported accurately, we would be

surprised if it were as important a problem when embryonic stem cells were permitted to differentiate into mature myocytes. Certainly, if this remained the case, the concern for arrhythmogenicity would be great. Finally, the immunogenicity of the terminally differentiated cells may present a problem to any particular host. A suggested strategy to counteract this is development and banking of multiple cell lines with tissue typing performed to determine compatibility for individual patients. Although these concerns are formidable, research in this field is in its early stages, and the technical problems that need be met are—we believe—ultimately solvable. Hence, we view this resource as a potential source of biological pacemakers on which we can comment little at present, but whose promise is great.

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

cells injected into canine myocardium form Cx43 junctions among themselves and with ventricular myocytes [46]. Given this information, we reasoned that adult human mesenchymal stem cells loaded with the appropriate genes might serve as a platform for a biological pacemaker. The philosophy was as follows (Fig. 5): in a normal sinus node cell (or a genetically modified myocyte), phase 4 depolarization results in initiation of an action potential. Based on cell – cell current flow via gap junctions, this action potential then is propagated to other cells of the conducting system and myocardium. If a stem cell could be genetically modified to express an HCN isoform and it formed adequate gap junctional linkages to adjacent cells, then an adjacent cell having a high level of resting potential would initiate inward (IHCN) current in the stem cell. Via current flow through gap junctions, this would depolarize the adjacent myocyte to threshold potential initiating an action potential that could propagate further in the conducting system. Moreover, as the adjacent myocyte reached the plateau range of potentials, this would serve to turn off the pacemaker current. The latter would not be reinitiated until the cell had repolarized to its maximum diastolic potential. Thus, unlike the gene therapy approach, the HCN-expressing adult human mes-

19

enchymal stem cell would not—by itself—fire spontaneously. Rather, it would provide a source of depolarizing current that is triggered by and synchronized to the diastolic potential of the electrotonically connected myocyte and turned off when the myocyte attains its plateau potential. In effect, the two coupled cells would function as a single pacemaker unit. As such, the stem cell is a ‘‘platform’’ or ‘‘vehicle’’ that delivers gene therapy. This approach does not require that the stem cell differentiate into a myocardial or pacemaker-like cell. In employing these stem cells to build a biological pacemaker, we must recall that the architecture of our template, the sinus node, is ideally suited to permit a small node to drive a large chamber (see Ref. [47]). Critical coupling along a gradient reaching from the dominant pacemaking area towards the periphery is an important aspect of this structure. Coupling also plays a role, as in specific settings it may be arrhythmogenic (see Ref. [48]). Certainly, for sinus node, the absence of IK1 and the presence of a pacing voltage range depolarized relative to that of the atrium requires sufficient isolation (i.e., uncoupling) for nodal cells to remain depolarized, and sufficient coupling to provide local circuits to activate the atrium. Downloaded from by guest on January 20, 2017

Fig. 5. Initiation of spontaneous rhythms by wild-type or genetically engineered pacemaker cells as well as by genetically engineered stem cell pacemakers. Top panel: in a native pacemaker cell or in a myocyte engineered to incorporate pacemaker current via gene transfer, action potentials (inset) are initiated via inward current flowing through transmembrane HCN channels. These open when the membrane repolarizes to its maximum diastolic potential and close when the membrane has depolarized during the action potential. Current flowing via gap junctions to adjacent myocytes results in their excitation and the propagation of impulses through the conducting system. Bottom panel: a stem cell has been engineered to incorporate HCN channels in its membrane. These channels can only open, and current can only flow through them (inset) when the membrane is hyperpolarized; such hyperpolarization can only be delivered if an adjacent myocyte is tightly coupled to the stem cell via gap junctions. In the presence of such coupling and the opening of the HCN channels to induce local current flow, the adjacent myocyte will be excited and initiate an action potential that then propagates through the conducting system. The depolarization of the action potential will result in the closing of the HCN channels until the next repolarization restores a high negative membrane potential. In summary, wild-type and genetically engineered pacemaker cells incorporate in each cell all the machinery needed to initiate and propagate action potentials. In contrast, in the stem cell – myocyte pairing, two cells together work as a single functional unit whose operation is critically dependent on the gap junctions that form between the two disparate cell types.

20

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

Fig. 7. Idioventricular rhythm resulting from implantation of adult human mesenchymal stem cells containing GFP and HCN2 5 days after implant into left ventricular epicardium of adult dog. From left to right: (A) sinus rhythm before and after vagal stimulation commences; (B) idioventricular rhythm during the course of vagal stimulation; (C) resumption of sinus rhythm on termination of vagal stimulation, and (D) ventricular pacing from the site of the injection of stem cells. Note the similarity between the paced complexes in panel D and those occurring during the idioventricular rhythm in Panel B (records provided by Drs. Alexei Plotnikov and Iryna Shlapakova).

Downloaded from by guest on January 20, 2017

Fig. 6. Pacemaker function in an in vitro model in which adult human mesenchymal stem cells loaded with either GFP (panel A) or GFP + murine HCN2 (panel B) are plated on a cover slip that is then overlaid with neonatal rat ventricular myocytes. In this figure, 4 – 5 days later, the spontaneous rate in the presence of HCN2 is markedly faster than in the presence of GFP alone (reprinted from Ref. [30] by permission).

The delivery of HCN genes via stem cells confers different issues. The pacemaker unit includes the myocyte and its IK1; hence, the inward current provided by the stem cells must be sufficient to overcome the outward IK1. Moreover, the gap junctional coupling must be sufficient to hyperpolarize the stem cells sufficiently to activate the HCN channels. Therefore, we can conceive of no optimal gap junctional coupling here; the better the coupling the more effective the unit will be. In exploring the potential of adult human mesenchymal stem cells, we learned they have little endogenous If , and that electroporation can be used to incorporate the cells with an HCN gene with an efficiency of 40% or greater [30]. Transfecting cells with both HCN and a fluorescent marker (i.e., GFP) provides an easy means of selecting for further use cells and populations of cells having sufficient HCN expression. These techniques permit genetic modification of cells via nonviral means (thereby removing all concerns associated with viral vectors) and facilitate sorting cell populations to decide which are/are not adequate for further study and administration. We have measured the pacemaker current in these cells and found it to be hyperpolarizationactivated, catecholamine-responsive and acetylcholine-responsive, and to be blocked by cesium [30]—all properties of the pacemaker current, If . We have used a cell culture system wherein a small node of stem cells expressing GFP plus the HCN variant to be studied, or GFP alone is plated on a cover slip and the node

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

6. Conclusions We have reviewed a field that is very much in its infancy. Moreover, the biological pacemaker represents a subdivision of a field that in all its aspects promises to revolutionize the way in which medicine is practiced. Certainly, the promise of biological replacements for diseased conducting systems is as revolutionary to the field as electronic pacemakers were a half century ago. With regard to the use of stem cells, particularly, we see within the realm of probability technologies completely disruptive to existing modalities for many therapies, going well beyond biological pacemakers. Yet, much must be done if probability is to become reality. Whether we are dealing with viral gene transfer, embryonic stem cells, or mesenchymal stem cells, much of the necessary toxicology is awaited. Will constructs or cells delivered stay where they are implanted, or will they migrate elsewhere? If stem cells are delivered, will they remain as stem cells or mature into other cell types? Will these function as well in the heart as in the model cell systems in which they are studied or will they evolve different functional characteristics? Can stem cells or viral constructs be guaranteed not to incorporate the potential for malignancy? Clearly, the challenges with regard to teratogenicity, infection, and rejection are imposing ones. We also need to know the longevity of the constructs used. Moreover, they must be studied in competition with electronic pacemakers to ensure which of the therapies is the more reliable. Certainly, if the biological pacemaker is not as or more reliable and longer lasting than the electronic, it

Fig. 8. Immunostaining for gap junctions in a section that interfaces myocardium and stem cells from the heart of a dog that received adult human mesenchymal stem cells transfected with GFP + HCN2. DAPI staining was used to identify nuclei. Purple arrows indicate intercalated discs; white arrows, gap junctions between stem cells; and red arrows, between stem cells and myocytes (modified after Ref. [30] by permission).

Downloaded from by guest on January 20, 2017

is then overlaid with a monolayer of neonatal rat ventricular myocytes [30]. The contiguous sheet of myocytes in contact with adult human mesenchymal stem cells housing the pacemaker construct beats significantly faster than myocytes in contact with adult mesenchymal stem cells expressing GFP only (Fig. 6). This system permits the rapid study of HCN isoforms and mutations of these isoforms to inform us of rate characteristics and autonomic responsiveness. Proof of concept for adult human mesenchymal stem cells administered as cellular nodes to the heart was obtained in canine experiments in which adult mesenchymal stem cells expressing HCN2 were injected into a small region of left ventricular epicardium [30]. A week later, vagal stimulation was performed to induce atrioventricular block. Dogs receiving HCN2 + GFP developed idioventricular rhythms having rates approximating 60 bpm that were pace-mapped to the injection site (Fig. 7). In contrast, animals that received adult mesenchymal stem cells expressing GFP alone manifested rates of about 45 bpm only ( p < 0.05), originating at various sites. Immunohistochemical studies revealed nests of adult human mesenchymal stem cells at the site of injection, as well as evidence for gap junctional coupling between adult mesenchymal stem cells and myocytes [30] (Fig. 8). Long-term experiments and comparisons with electronic pacemakers are awaited as is the demonstration of autonomic responsiveness in situ. Concerns regarding adult human mesenchymal stem cells are major and include duration of function of the pacemaker, comparison with electronic pacemakers, and issues such as rejection, neoplasia, migration to other sites, and differentiation into other cell types.

21

22

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

will remain a research tool rather than become a clinical reality. Given these concerns, is the endeavor worth the effort? If the goal of therapy in needy patients is to cure, rather than to palliate, we believe the effort is warranted. All the problems identified are amenable to solution using techniques we already have; it is simply a matter of expending the time and the energy until the answers are in our grasp.

Acknowledgements Our thanks to Eileen Franey for her careful attention to the preparation of the manuscript. Supported by USPHSNHLBI grants HL-28958, HL-67101, HL-20558, and GM55263.

References

Downloaded from by guest on January 20, 2017

[1] Ratko TA, Cummings JP, Blebea J, Matuszewski KA. Clinical gene therapy for nonmalignant disease. Am J Med 2003;115(7):560 – 9. [2] Simons M, Ware JA. Therapeutic angiogenesis in cardiovascular disease. Nature reviews. Drug Discov 2003;2(11):863 – 71. [3] Strauer BE, Brehm M, Zeus T, et al. Repair of infracted myocardium by autologous intracoronary monomuclear bone marrow cell transplantation in humans. Circulation 2002;106:1913 – 8. [4] Di Francesco D. A study of the ionic nature of the pacemaker current in calf Purkinje fibres. J Physiol 1981;314:377 – 93. [5] Di Francesco D. Block and activation of the pacemaker channel in calf Purkinje fibres: effects of potassium, caesium and rubidium. J Physiol 1982;329:485 – 507. [6] Brown HF, Kimura J, Noble D, Noble SJ, Taupignon A. The ionic currents underlying pacemaker activity in rabbit sinoatrial node: experimental results and computer simulations. Proc R Soc Lond B Biol Sci 1984;222:329 – 47. [7] Hagiwara N, Irisawa H, Kameyama M. Contribution of two types of calcium currents to the pacemaker potentials of rabbit sinoatrial node cells. J Physiol (Lond) 1988;395:233 – 53. [8] Hagiwara N, Irisawa H, Kasanuki H, Hosoda S. Background current in sinoatrial node cells of the rabbit heart. J Physiol (Lond) 1992;448: 53 – 72. [9] Noma A, Irisawa H. Effects of Na+ and K+ on the resting membrane potential of the rabbit sinoatrial node cell. Jpn J Physiol 1975;25: 287 – 302. [10] Ono K, Ito H. Role of rapidly activating delayed rectifier K+ current in sinoatrial node pacemaker activity. Am J Physiol 1995;269:H453 – 62. [11] Li J, Qu J, Nathan RD. Ionic basis of ryanodine’s negative chronotropic effect on pacemaker cells isolated from the sinoatrial node. Am J Physiol 1997;273:H2481 – 9. [12] Vassalle M, Yu H, Cohen IS. Pacemaker channels and cardiac automaticity. In: Zipes DP, Jalife J, editors. Cardiac Electrophysiology: From Cell to Bedside. 3rd ed. Philadelphia: WB Saunders; 2000. p. 94 – 103. [13] Noma A, Nakayama T, Kurachi Y, Irisawa H. Resting K conductances in pacemaker and non-pacemaker heart cells of the rabbit. Jpn J Physiol 1984;34:245 – 54. [14] Yu H, Chang F, Cohen IS. Pacemaker current exists in ventricular myocytes. Circ Res 1993;72:232 – 6. [15] Miake J, Marba´n E, Nuss HB. Gene therapy: biological pacemaker created by gene transfer. Nature 2002;419:132 – 3. [16] Michaelsson M, Jonzon A, Riesenfeld T. Isolated congenital complete heart block in adult life: a prospective study. Circulation 1995;92: 442 – 9.

[17] Zivin A, Bardy GH, Mehra R. Cardiac pacemakers. In: Spooner PM, Rosen MR, editors. Foundations of cardiac arrhythmias. New York: Marcel Dekker; 2001. p. 571 – 98. [18] Zivin A, Bardy GH, Mehra R. Implantable cardioverter defibrillators. In: Spooner PM, Rosen MR, editors. Foundations of cardiac arrhythmias. New York: Marcel Dekker; 2001. p. 599 – 619. [19] Furrer M, Naegeli B, Bertel O. Hazards of an alternative medicine device in a patient with a pacemaker. N Engl J Med 2004;350(16): 1688 – 90. [20] Martin ET, Coman JA, Shellock FG, Pulling CC, Fair R, Jenkins K. Magnetic resonance imaging and cardiac pacemaker safety at 1.5Tesla. J Am Coll 2004;43(7):1315 – 24. [21] Edelberg JM, Aird WC, Rosenberg RD. Enhancement of dmurine cardiac chronotropy by the molecular transfer of the human h2-adrenergic receptor cDNA. J Clin Invest 1998;101:337 – 43. [22] Edelberg JM, Huang DT, Josephson ME, Rosenberg RD. Molecular enhancement of porcine cardiac chronotropy. Heart 2001;86: 559 – 62. [23] Qu J, Plotnikov AN, Danilo Jr. P, et al. Expression and function of a biological pacemaker in canine heart. Circulation 2003;107:1106 – 9. [24] DiFrancesco D, Ferroni A, Mazzanti M, Tromba C. Properties of the hyperpolarizing-activated current (If) in cells isolated from the rabbit sinoatrial node. J Physiol (Lond) 1986;377:61 – 88. [25] Campbell GD, Edwards FR, Hirst GDS, O’Shea JE. Effects of vagal stimulation and applied acetylcholine on pacemaker potentials in the guinea pig heart. J Physiol (Lond) 1989;415:57 – 68. [26] Di Francesco D, Mangoni D, Maccaferri M. The pacemaker current in cardiac cells. In: Zipes DP, Jalife J, editors. Cardiac electrophysiology: from cell to bedside. 2nd ed. Philadelphia: WB Saunders; 1995. p. 96 – 103. [27] Lasker SM, Han D, Kline RP. Zatebradine slows ectopic ventricular rhythms in canine heart 2 hours after coronary artery ligation. J Cardiovasc Pharmacol 1997;29(5):662 – 9. [28] Robinson RB, DiFrancesco D. Sinoatrial node and impulse initiation. In: Spooner PM, Rosen MR, editors. Foundations of cardiac arrhythmias. New York: Marcel Dekker; 2001. p. 151 – 70. [29] Qu J, Barbuti A, Protas L, Santoro B, Cohen IS, Robinson RB. HCN2 overexpression in newborn and adult ventricular myocytes: distinct effects on gating and excitability. Circ Res 2001;89:E8 – E14. [30] Potapova I, Plotnikov A, Lu Z, et al. Human mesenchymal stem cells as a gene delivery system to create cardiac pacemakers. Circ Res 2004;94:952 – 9. [31] Plotnikov AN, Sosunov EA, Qu J, et al. A biological pacemaker implanted in the canine left bundle branch provides ventricular escape rhythms having physiologically acceptable rates. Circulation 2004; 109:506 – 12. [32] Miake J, Marba´n E, Nuss HB. Functional role of inward rectifier current in heart probed by Kir2.1 overexpression and dominant-negative suppression. J Clin Invest 2003;111(10):1529 – 36. [33] Miake J, Nuss HB. Mechanism of automaticity induced by dominantnegative suppression of IK1 in guinea pig ventricular myocytes. Biophys J 2003;84(2):1529 (Part 2 of 2). [34] Miake J, Nuss B. Multiple ionic conductances sustain IK1-suppressed biopacemaking. Circulation 2003;108:IV-35. [35] Silva J, Rudy Y. Mechanism of pacemaking in IK1-downregulated myocytes. Circ Res 2003;92:261 – 3. [36] Santoro B, Liu DT, Yao H, et al. Identification of a gene encoding a hyperpolarization-activated pacemaker channel of brain. Cell 1998; 93:1 – 20. [37] Ludwig A, Zong X, Jeglitsch M, Hoffman F, Biel M. A family of hyperpolarization-activated mammalian cation channels. Nature 1998;393:587 – 91. [38] Kaupp UB, Seifert R. Molecular diversity of pacemaker ion channels. Annu Rev Physiol 2001;63:235 – 57. [39] Strauer BE, Kornowski R. Stem cell therapy in perspective. Circulation 2003;107:929 – 34. [40] Kehat I, Kenyagin-Karsenti D, Snir M, et al. Human embryonic stem

M.R. Rosen et al. / Cardiovascular Research 64 (2004) 12–23

[41] [42]

[43]

[44]

cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J Clin Invest 2001;108:407 – 14. Gepstein L. Derivation potential applications of human embryonic stem cells. Circ Res 2002;91:866 – 76. Zhang YM, Hartzell C, Narlow M, Dudley Jr SC. Stem cell-derived cardiomyocytes demonstrate arrhythmic potential. Circulation 2002; 106(10):1294 – 9. Jongsma HJ, Masson-Pevet M, Tsjernina L. The development of beatrate synchronization of rat myocyte pairs in cell culture. Basic Res Cardiol 1987;82(5):454 – 64. Liechty KW, MacKenzie TC, Shaaban AF, et al. Human mesenchymal stem cells engraft and demonstrate site-specific differentiation after in utero transplantation in sheep. Nat Med 2000;6(11):1282 – 6.

23

[45] Huebach JF, Graf EM, Leutheuser J, et al. Electrophysiological properties of human mesenchymal stem cells. J Physiol 2003; 554.3:659 – 72. [46] Valiunas V, Doronin S, Valiuniene L, et al. Human mesenchymal stem cells make cardiac connexins and form functional gap junctions. J Physiol 2004;555.3:617 – 26. [47] van Capelle RW, Joyner FJ. Propagation through electrically coupled cells. How a small SA node drives a large atrium. Biophys J 1986;50: 1157 – 64. [48] de Groot JR, Coronel B. Acute ischemia-induced gap junctional uncoupling and arrhythmogenesis. Cardiovasc Res 2004;62:323 – 34.

Downloaded from by guest on January 20, 2017