The Utility of Mesenchymal Stem Cells as Biological Pacemakers

The origins of biological pacing are multicentric, beginning in the 1960s as different investigators generally unsuccessfully attempted sinoatrial node transplants.1-3 A more mechanistic approach to biological pacing was afforded by the discovery of If, the inward current that initiates phase 4 depolarization4-6 (Figure 1). Boosting the translation of these considerations into action were advances in identification of the genetic determinants of ion channels and the ability to clone, mutate, and overexpress using viral and other vectors. The telling question of why bother to build a biological pacemaker has 2 responses: one is the George Mallory/Mt Everest answer, transliterated to “Because we can.” This is not as trite as it sounds: by attempting to (re)create the method of pacing used by the normal heart, we learn more about what makes the heart function and sharpen our abilities to modify that function to a constructive end. The upper panel illustrates representative sinoatrial node action potential (control: solid lines) and some ion channels and exchangers that contribute to it. If is activated on hyperpolarization, providing inward current to initiate phase 4. T- and L-type inward Ca currents commence near the end of phase 4. Na/Ca exchange current also influences the phase 4 slope. Outward potassium current IK is responsible for repolarization. Norepinephrine (NE) effect to increase phase 4 and impulse initiation is represented as broken lines. The lower panel illustrates an HCN pacemaker channel, having 6 transmembrane spanning domains. When the channel is open, Na is admitted. Cyclic adenosine monophosphate (cAMP) binding sites are present near the carboxy terminus.β1-adrenergic and M2-muscarinic receptors provide NE and acetylcholine (Ach) binding sites, respectively, and regulate adenylyl cyclase (AC) activity via G-protein coupling. AC, in turn, regulates intracellular cAMP levels, determining the availability of the second messenger for binding and channel modulation. Adapted with permission from Biel et al.6 The second answer focuses on advancing or replacing current state-of-the-art therapy, the electronic pacemaker. This lifesaving device had its origins in the 19th century,7 was developed largely in the latter half of the 20th century,8-13 and provides a prime example of how knowledge derived from basic research can translate to a clinically effective outcome. Only because this highly successful therapy is not perfect is there room to consider biological alternatives. By “not perfect” we mean that the requirement for hardware implantation carries with it the possibility of infection and the need for hardware removal/ replacement (due to aging of batteries, imperfections in leads, and—in pediatric patients—the changing anatomy that accompanies growth and development).14, 15 Other issues such as lack of autonomic responsiveness of electronic pacemakers and inability in individual patients to personalize the implantation of electrodes at sites that optimize stroke volume are being met in part by rate-responsive pacemakers and “leadless” electrodes.16, 17 Whether these will achieve optimal function (optimal being defined by normal cardiac pacemaker function) is yet to be determined. In confronting electronic pacing, the biological approach is a disruptive technology. Whereas this may discomfit some in the “business” of pacing, this disruption is actually welcome. That is, if we think of electronic and biological pacing as competing technologies, then the beneficiaries of advances made in either or both areas are scientific knowledge and patient care. Regardless of whether the future ultimately rests with electronics, biologics, or a tandem approach, we couldn't ask for anything better. Although our focus in this article is on human mesenchymal stem cells (hMSCs) in biological pacing, it is worthwhile to briefly consider the evolution of the field. We and others have worked with tissue culture and small animal models of biological pacing for some time,18, 19 but the first breakthrough in a large animal model was Edelberg and colleagues'20 demonstration that overexpressing β2-adrenergic receptors in pig hearts could increase pacemaker rate. The problem in further developing this approach was the inherent arrhythmogenicity of catecholamine actions on the heart. The first use of ion channel modulation in biological pacing was Miake and colleagues'21 creation of a dominant negative construct for the repolarizing potassium current, IK1, in the guinea pig heart. This increased phase 4 depolarization and generated pacemaker function but also prolonged repolarization and was arrhythmogenic.22 Our group reported the use of HCN genes (encoding the hyperpolarization-activated, cyclic nucleotide-gated channel that is responsible for If).23-25 We used this approach because If not only initiates the pacemaker potential but operates solely during diastole and, as a result, does not prolong repolarization. Others have since used this same approach, and all have focused either on the HCN1, HCN2, or HCN4 channel isoforms or on mutant or chimeric constructs.25-27 The above-mentioned approaches have used either naked plasmids or viral vectors to deliver the pacemaker constructs. What all share is the addition to the normal ion channel “machinery” of the cardiac myocyte of a means for increasing the rate of impulse initiation. In other words, the intent is to convert nonpacemaker cells (or nondominant pacemaker cells such as those in the specialized conducting system) into a dominant pacemaker. While success has been achieved in proof-of-concept experiments, the naked plasmids have the disadvantage of poor efficiency, and the viral vectors bring concerns about both the long-term impact of introducing a viral particle on an organism and the duration of effect. For these reasons, 2 types of stem cell approaches have been used. One is the human embryonic stem cell, which is pluripotent and can be forced along a pacemaker lineage in cell culture. The resultant cells incorporate the entire compliment of channels needed to provide pacemaker function. They have been delivered to porcine hearts, have been effectively coupled with native myocytes, and have provided an extrinsically administered pacemaker cell population to the heart.28 Their use, however, has required immunosuppression, an intrusion that is not favorable for human pacemaker therapy. Moreover, questions remain about persistence of cells at the site administered as well as their differentiation into another cell type (nonpacemaker and/or neoplastic). As stated above, both embryonic stem cells and viral vectors provide viable options for fabricating biological pacemakers. However, concerns about both are sufficient that alternative methods should be explored. Another such method is platform delivery using hMSCs. While hMSCs are multipotent, they have little to no compliment of the channels necessary to generate a cardiac action potential. Hence, they generate neither depolarizing nor repolarizing currents.14, 15 For this reason, we considered engineering them into a platform for adding pacemaker genes to the cardiac syncytium. Moreover, hMSCs have 2 unique characteristics that favor their use as platforms for biological pacing: First, they have an abundant complement of connexins 43 and 40 (2 of the building blocks of cardiac gap junctions).29 This suggests that hMSCs loaded with pacemaker genes should effectively couple with and transfer current to cardiac myocytes. Second, detailed literature suggests that hMSCs are immunoprivileged, producing autocrine factors that mask their surface antigens such that they are not recognized by a host's immune system.30, 31 These attributes led us to propose (Figure 2) that we could load hMSCs with HCN genes via electroporation, thereby avoiding use of viral vectors; that hMSCs, thus loaded and injected into the heart would couple via gap junctions to adjacent myocytes; and that the myocytes would hyperpolarize the hMSCs, causing pacemaker current to flow and excite the myocytes.14, 15 In other words, the 2 cells would operate together in a push-pull fashion to initiate an action potential: the hMSC would provide the machinery to bring the cell pair to threshold potential and the myocyte would provide the machinery for generating the action potential itself. Finally, we proposed that the nature of the immunoprivilege would be such that the hMSCs could be given as a xenotransplant into a canine heart without the need for immunosuppression.32, 33 In other words, we viewed the hMSCs as a biological platform for delivering pacemaker genes: the hMSCs would not become pacemakers but they would reside in the heart, providing an essential part of the machinery, along with myocytes or specialized conducting cells to express dominant pacemaker function. Panel A: Rationale for biological pacing Top: Initiation of spontaneous rhythms by sinoatrial node cells. Action potentials (inset) are initiated via inward current flowing through transmembrane HCN channels. Current flowing via gap junctions to adjacent myocytes results in excitation and impulse propagation through the conducting system. Bottom: A stem cell electroporated to incorporate HCN channels in its membrane. When the membrane is hyperpolarized via current flow through gap junctions, the HCN channels open to induce inward current, which excites the coupled myocyte to initiate an action potential, propagating through the conducting system. (Adapted with permission from Rosen et al.14) Panels B and C: Biological pacemaker function in a canine heart in situ. Top to bottom, electrocardiographic (ECG) leads I, II, III, AVR, AVL, and AVF. Panel A: saline control. Panel B received human mesenchymal stem cells (hMSCs) loaded with HCN2 and GFP. Left panels: Pacing from hMSC injection site showing ECG configuration. Center: Spontaneous rhythms about 7 days later. In saline, rate is low and escape time after overdrive pacing at 80 beats per minute (bpm) is long. In hMSCs+HCN2, the rhythm pace-maps to site of injection, and the last 2 beats of overdrive pacing (shaded) at 80 bpm are followed by a 1-second escape time. (Adapted with permission from Potapova et al.32) Panel D: hMSCs 6 weeks post-left ventricular anterior wall intramyocardial injection. Hematoxylin and eosin stain: basophilic cells that are CD44 positive (human) and GFP positive (peroxidase stain). These findings (arrows) suggest that these are the cells we injected 6 weeks earlier. There is no labeling/binding of canine-IgG to their surface, providing evidence against humoral rejection. CD3-positive T lymphocytes are rarely seen, providing evidence against cellular rejection. Staining is negative for activated CASPASE 3, providing evidence against apoptosis (original magnification: ×400). (Adapted with permission from Plotnikov et al.33) In all of these expectations, we were right—in part. We were correct in the assumption that we could load hMSCs with the HCN2 gene and express a robust pacemaker current, that the hMSCs would couple to myocytes and permit the passage of sufficient current to generate biological pacemaker function, that we could implant these constructs in dogs that were not immunosuppressed and in complete heart block, and see the hMSC-based pacemakers function for periods of 6 weeks (the longest period attempted by us to date) with no evidence of rejection, loss of function, or apoptosis (Figure 2).29, 32, 33 We also found that 700,000 hMSCs, approximately 50% of which are loaded with the HCN2 gene, appear to be the threshold cell number for initiating robust pacemaker function. Below this number, function is sporadic or nonexistent; above this number (through 1,200,000), excellent function occurs, with no apparent dose-response relationship.33 But we were incorrect in a very important area: that is, whereas some dogs maintained robust pacemaker function over the 6-week period, others not only showed loss of function but had histological evidence of rejection.33 In retrospect, it is not surprising that rejection occurred in a xenotransplant (indeed, we might have reasonably expected it in all cases). The critical question becomes, why did some animals reject and others did not? While this question is of interest with regard to xenotransplantation, it is not readily applicable to the use of hMSCs in humans. Indeed, the literature in humans strongly suggests that allogeneic administration of hMSCs is safe, with regard to concerns about rejection.34, 35 However, the literature in humans for a variety of stem cell types, including allogeneically administered hMSCs, raises another concern: that is, of hMSCs “wandering” to other sites.14, 15 In biological pacemaker studies, this would translate to their loss from the region at which one wants to generate pacemaker function and also provides the possibility of initiating unwanted impulses at other sites in the body. This latter concern encouraged us to develop a means to label hMSCs so that their location can be tracked over time. A variety of methods for tracking cells has been tested, and we and others have commented on these.36, 37 All have shortcomings that have led us to use nanotechnology approaches for labeling. Here, we have employed quantum dots, which provide a strong signal well outside the range of tissue autofluorescence, do not cross gap junctions, and are removed by the reticuloendothelial system when cells die.37 Using quantum dots, we have easily sited cells within sections of myocardium and have performed 3-dimensional reconstructions of their distribution. The challenges that remain are being able to track the particles in the body in situ and understanding what minimum number of cells will be detectable should they wander to other sites. As stated in above, there are 2 goals in this field: one is the advancement of knowledge, the other is the advancement of pacemaker therapy for patients. We have no doubt that both goals will be met; although, whether the patient in need of a pacemaker 50 years from now will be treated electronically, biologically, or via the tandem type of approach we recently described,25 is uncertain. With respect to issues of heart failure, per se, any approach that permits individual patients to be electrophysiologically tested to identify the optimal locus for impulse initiation resulting in pacemaker placement at that site will be a major advance. This would minimize the concern over long-term evolution of heart failure in patients whose pacemaker electrodes currently are affixed in the right ventricular endocardial apex.38 As for more complex issues such as biventricular pacing for the treatment of heart failure,39 while dual siting of biological pacemakers and signal coordination might be a possibility, this is far closer to the drawing board than an experimental reality. The authors express their gratitude to Ms Eileen Franey for her careful attention to the preparation of this manuscript. This work was supported in part by United States Public Health Service/National Heart, Lung and Blood Institute (USPHS-NHLBI) grant HL-28958 and by Boston Scientific. The authors receive research support from Boston Scientific.