Chapter 70. Stem Cell–Induced Regeneration of Myocardium

Statistics recently published by the American Heart Association clearly outline the magnitude of the medical and economic problems associated with heart failure.1 It is estimated that 6 million people in the United States are affected by this disease, with a yearly incidence of new episodes (per 1000 people) of 15.2 between 65 and 74 years of age, a figure that is two times higher between 75 and 84 years of age. Still in the United States, the number of hospitalizations for heart failure has thus increased from 877,000 to 1,106,000 in a 10-year time frame (1996–2006), but the problem is now worldwide. The mortality remains high as 80% of men and 70% of women under the age of 65 years in whom heart failure has been diagnosed are expected to die within the next 8 years with a risk of sudden death which is six to nine times that of the general population. As expected, these figures translate into a heavy economic burden with costs estimated to 37 billion dollars in 2008.

Although heart transplantation remains the only radical treatment for end-stage heart failure, its use is limited by organ shortage and the complications associated with major immunosuppression. Surgical remodeling procedures require well-defined anatomical indications, and the failure of the STITCH2 trial may lead to revisit them. Mechanical assist devices are still primarily used as bridges to transplant (or recovery) and, despite its ability to provide symptomatic relief, biventricular stimulation fails in 20 to 30% of patients.3 Finally, none of the large trials implemented over the last decade to investigate new drugs has yielded a positive outcome allowing to improve the survival of heart failure patients. Put together, these observations have provided a rationale for exploring new therapeutic options, among which “regeneration” of the chronically failing heart by stem cells has raised a tremendous interest. From the onset, however, it is fair to state that regeneration bona fide, ie, generation of a new myocardial tissue with the same properties as the native one, is a concept, a target, and a hope, but not yet a reality. This chapter first summarizes the current status of stem cell trials, then describes the strategies that can be considered for effecting a true myocardial regeneration, and finally emphasizes the importance of addressing the low cell engraftment issue to make these strategies successful.

So far, two types of stem cells, both of adult origin, have been tested clinically: skeletal myoblasts and bone marrow–derived cells.

Skeletal Myoblasts

In June, 2000, we performed the first human transplantation of autologous myoblasts in a patient with severe ischemic heart failure4 who underwent a concomitant coronary artery bypass grafting (CABG). This case initiated a 10-patient pilot trial that was soon followed by three other adjunct-to-CABG myoblast transplantation studies,5–7 which only differed from ours in that they systematically entailed a concomitant revascularization of the myoblast-injected areas.

Altogether, these studies have demonstrated the feasibility of the procedure as well as its safety with the caveat of an increased risk of ventricular arrhythmias. This has been attributed to the failure of engrafted myoblasts to express gap junction proteins and their subsequent electrical isolation from the surrounding cardiomyocytes,8 which could slow conduction velocity and predispose to reentry circuits.9 This hypothesis is supported by the in vitro finding that myoblast transfection with connexin 43 decreases arrhythmogenicity,9 but clinically other myoblast-unrelated factors are likely in play, particularly the needle-induced inflammatory tissue damage10 and the intrinsically arrhythmogenic nature of the underlying heart failure substrate.

Although these initial studies had suggested improved functional outcomes following myoblast transplantation, they were neither designed nor powered to provide efficacy data, which led us to implement a randomized, double-blind, placebo-controlled trial (MAGIC, an acronym for Myoblast Autologous Grafting in Ischemic Cardiomyopathy), which was conducted in 21 centers across Europe.11 Ninety-seven patients with severe left ventricular (LV) dysfunction, a postinfarction nonviable scar and an indication for bypass surgery were assigned to receive multiple in-scar injections of autologous myoblasts (at two different doses: 400 and 800 million) grown from a muscular biopsy taken 3 weeks earlier or similar injections of a placebo solution. An internal cardioverter-defibrillator (ICD) was implanted in each of these patients to provide objective readouts of arrhythmic events. At the 6-month study point, the proportion of patients who had experienced arrhythmias was not statistically different between the myoblast-treated and the placebo-injected groups despite a trend for a higher incidence of arrhythmic events early after myoblast implantations and there was not a single death that could be attributed to arrhythmias. This rather reassuring safety finding contrasted with the disappointing observation that regardless of the dose, myoblast injections had failed to improve LV function beyond that seen in the placebo group, even though the high-dose group experienced a significant reduction in remodeling (which was a prespecified secondary end point). Although this signal suggested that cells can exert some beneficial effects, these effects were clearly too limited to translate into meaningful improvements in contractile function, thereby ruling out the occurrence of a procedure-induced myocardial regeneration.

In parallel to these surgical trials, phase I catheter-based studies have assessed myoblast delivery through an ultrasound-guided coronary sinus catheter12 or through an endoventricular catheter under electromechanical guidance.13 Like their phase I surgical counterparts, these trials have primarily demonstrated the technical feasibility of these approaches. Efficacy data are more limited and are still controversial. At the 2008 Society of Cardiovascular Angiography and Interventions meeting, Serruys reported on 31 patients who were allocated to myoblast injections while 16 received “optimal medical therapy.” In keeping with the MAGIC data, there was no added benefit of cell therapy on LV function measured at 6 months after the procedure. In contrast, Dib and coworkers14 have reported improved outcomes in patients receiving transendocardial injections of skeletal myoblasts compared with a placebo group but with the caveat of a 1-year only follow-up.

Bone Marrow–Derived Cells

Most studies of bone marrow–derived cell therapy have been in the context of acute myocardial infarction and refractory angina and, comparatively, few of them have tackled heart failure. In this setting, they can be categorized in relation with the specific cell population that has been tested.

Mononuclear Cells

Beyond the anecdotal cases, five surgical studies have tested intramyocardial injections of bone marrow–derived unfractionated mononuclear cells (MNC) in conjunction with bypass surgery and, overall, their results are mixed. In a 36-patient trial, Mocini and coworkers15 reported a 3-month significant improvement in regional and global LV function compared with baseline, but these outcome measures did not differ from those obtained in control patients who, surprisingly, failed to improve functionally following bypass surgery. The 20-patient study of Hendrikx and coworkers16 showed that after 4 months, there was a significantly greater wall thickening in cell-transplanted patients (as assessed by magnetic resonance imaging [MRI]), but no improvement in ejection fraction or perfusion defects beyond values recorded in control patients. The third trial in which cells were injected both in the myocardium and in the bypass grafts also failed to meet its primary end point.17 Conversely, two studies,18,19 of which one was randomized,18 have reported improved function and perfusion up to 6 months18 and 5 years19 after MNC transplantation combined with CABG. Of note, whereas the average numbers of injected cells ranged from 60 to 292 million in the mentioned negative trials, those with a positive outcome entailed transplantation of up to 6.59 × 108 and 1.23 × 109 MNC,18,19 which calls attention to the importance of careful dose-effect relationship studies before definitive conclusions can be drawn.

Catheter-based studies of bone marrow cells in patients with heart failure have been initiated by an open-label nonrandomized 11-patient trial who were reported to have experienced a 1-year improvement in exercise capacity following endoventricular injections of MNC.20 A limited number of studies entailing intracoronary MNC infusions have then claimed positive functional outcomes in both ischemic21,22 and nonischemic23 cardiomyopathy settings. However, their results should be interpreted cautiously in view of the methodologic limitations inherent in the usual lack of randomization, the functional impairment of cells taken from patients with chronic ischemic heart disease24 and the small engraftment rate of intracoronary injected cells (see the following), particularly at the chronic stage, when the endothelial alterations associated with acute ischemic events and that may promote transvascular trafficking waned. Of note, the recent randomized controlled trial conducted by van Ramshorst and coworkers25 has reported a 3% improvement in ejection fraction without changes in LV volumes after endomyocardial MNC injections, but the target population was chronic ischemia not amenable to conventional revascularization and these outcomes thus likely illustrate the expected angiogenic effect of bone marrow cells rather than de novo myogenesis.

Hematopoietic Progenitors

The well-documented angiogenic properties of bone marrow–derived cells (see the following) have led other investigators to rather focus on hematopoietic progenitors featuring a high angiogenic capacity. In this setting, the most convincing results have been reported by Stamm and coworkers26 who have injected CD133+ cells epicardially during CABG. Although this trial was not strictly randomised, it has been rigorously conducted and provides encouraging hints in favor of the capacity of the CD133+ population to safely improve LV function and perfusion at 6 months postoperatively, particularly in patients with the poorest preoperative LV function. It is likely, however, that these effects, which await further confirmation by the upcoming randomized PERFECT trial, proceed from a cell-induced increase in angiogenesis and not from the conversion of the CD133+ progenitors into new cardiomyocyte. As such, they do not match the strict definition of myocardial regeneration.

Mesenchymal Stem Cells

The third category of BM-derived cells, which is raising a growing interest, even though clinical experience with them is still limited in the field of heart regeneration, comprises mesenchymal stem cells (MSC). These cells represent a rare population (0.01 to 0.001%) of the bone marrow and are usually defined by their adherence to plastic surfaces, their negativity for hematopoietic markers (CD45, CD34), the expression, among others, of the CD29, CD44, CD73, CD90, and CD105 antigens, and the capacity to differentiate along the osteogenic, chondrogenic, and adipogenic lineages in response to the appropriate cues. Most (but not all)27 experimental studies conducted in small28,29 and large30–32 animal models of chronic myocardial infarction,28,29,31,33 or dilated cardiomyopathy34 have now documented that transplantation of autologous30 or allogeneic28,29,31,33 transplantation of MSC may improve LV function and attenuate adverse remodeling in a dose-dependent fashion,32 at least at midterm.28 These patterns of improvement are usually associated with increased angiogenesis and reduced fibrosis. As all these studies have failed to document the differentiation of MSC into cardiomyocytes, it is likely that improved outcomes after MSC treatment proceed from their trophic effects35 reported to be more potent than those of MNC.34

Put together, these data have paved the way for upcoming or already ongoing trials of MSC in patients with chronic LV dysfunction in whom cells are delivered transendocardially as a stand-alone procedure or transepicardially as an adjunct to CABG or to placement of an assist device (which should yield useful histologic data when the heart will be removed for transplantation). In one of these studies, MSC are pretreated with a combination of growth factors targeted at increasing their cardiomyogenic differentiation potential36 as exposure of cells to the demethylation agent 5-azacytidine, which had been initially proposed to drive MSC toward a cardiomyogenic lineage, is unsuitable for clinical applications. Interpretation of outcomes, however, might be complicated by the fact that some companies involved in these trials have developed proprietary processes allowing isolation of specific subsets of MSC. Because there are no cell surface markers that specifically and uniquely identify MSC, it remains difficult to determine whether these different MSC subsets really reflect different cell fractions with distinct properties or simply represent the same MSC population assessed at different developmental stages and with the use of different surface markers.

Although MSC have initially been retrieved from bone marrow, alternate tissue sources have then been proposed like the placenta,37 menstrual blood,38 or the periodontal ligament.39 However, the greatest focus is currently on adipose tissue as a source of MSC derived from its stroma vascular fraction. Although there are differences in the global gene expression pattern between adipose tissue–derived MSC and their BM counterparts,40 the two types of cells share in common lineage-specific differentiation potentials, immunomodulatory proper-ties,41 and cardioprotective effects.42,43 These effects have been attributed to cell-derived antiapoptotic and angiogenic factors,44 and the latter might be even greater in the case of adipose-derived MSC,43 which also express the CD34 marker at the time of isolation.40,41 Conversely, there is no evidence for differentiation of these cells into cardiomyocytes.45 An additional attractive feature of adipose tissue–derived MSC is that their procurement by liposuction is patient friendly and their expansion capacity seems greater than that of their BM counterparts.46 To make their availability still more expeditious, a device has been developed to extemporaneous process lipoaspirates so as to yield a mixed cell population that can be immediately transplanted at the point of care and a clinical trial (PRECISE) testing the endoventricular delivery of this cell product has just been completed in patients with coronary artery disease not amenable to revascularization in the target area. How this device-generated heterogeneous cell population compares with that comprising the stroma vascular fraction specifically isolated from liposuction products remains to be established. It is also fair to temper the current enthusiasm about these cells by mentioning the negative study of van der Bogt and coworkers,47 who have used bioluminescence imaging for the in vivo tracking of injected cells and have reported an equally high rate of MSC death regardless of the tissue source (bone marrow or adipose tissue) and a corresponding lack of cardioprotection. However, it cannot be excluded that this negative outcome was strongly biased by the prolonged period of pretransplantation MSC cultures.

Regardless of the tissue source, what has largely driven the clinical interest in MSC is their alleged “immune privilege” and the related hope that they could be used as an allogeneic off-the-shelf readily available product with the attendant logistical, regulatory, and economic benefits. This paradigm is based on the in vitro mixed lymphocyte reactions-based findings that MSC do not evoke a response of allogeneic lymphocytes and inhibit lymphocyte proliferation induced by allogeneic peripheral blood MNC.41 These immunomodulatory effects, which require direct contact between MSC and the responder T cells and release of soluble mediators have been attributed to three major mechanisms: (1) MSC are hypoimmunogenic because of their lack of major histocompatibility complex (MHC) class II and costimulatory molecules. (2) They prevent T-cell responses both directly by directing CD4 T cells to a suppressive phenotype and disrupting CD8 T cell and natural killer cell function, and indirectly by interfering with dendritic cell maturation. (3) They secrete soluble mediators, which induce a suppressive local microenvironment48,49 through downmodulation of inflammatory cytokines (particularly TNF-α and IFN-γ) produced by activated T lymphocytes and inhibition of B-cell proliferation.

The current enthusiasm about MSC, however, needs to be tempered in light of several considerations. First, the plasticity of MSC renders them extremely sensitive to the physical nature of the underlying substrate; the stiffer the substrate, the greater the likelihood of a differentiation along the osteogenic lineage.50 Given the heterogeneous composition of postinfarction scars, there is a potential for some of the in-scar injected MSC to give rise to intramyocardial calcification, a complication that, so far, has only been observed experimentally.51 Second, the demonstration of the immunosuppressive effects of MSC have been largely based on in vitro experiments and despite the successful results of MSC treatment of severe graft-versus-host disease by LeBlanc and associates,52 the immune privilege of these cells is still debated. Indeed, it has been reported that MSC could function as antigen-presenting cells and intramyocardial injection of human MSC in rat myocardium have shown that although MSC could engraft in immunocompromised hosts, they triggered a robust immune response in those with an intact immune system.53 Likewise, Poncelet and coworkers54 have reported that intracardiac transplantation of allogeneic porcine MSC elicited a complete cellular and humoral immune response whereas the same cells had failed to trigger an in vitro proliferative response. A clinical industry–sponsored study of intravenous allogeneic MSC treatment has already been completed in patients with acute myocardial infarction and has yielded a quite reassuring safety record. Efficacy data are less conclusive and the study does not allow us to know whether the injected cells have been rejected or not. Fourth, the optimal route for MSC delivery is not straightforward. Their intravenous infusion is obviously the simplest and the safest approach and has been advocated on the premise that MSC would sense signals emitted by the infarcted tissue and preferentially home to the injured area. However, homing has been primarily reported in rodent experiments55 and studies in large animals56 and humans57 have failed to show MSC engraftment in the heart. On the other hand, a direct intramyocardial injection, surgically or through an endoventricular catheter, does not raise safety issues, but such is no longer the case if one considers an intracoronary infusion because the relatively large size of MSC (twice that of MNC, ie, approximately 20 μm) may result in capillary plugging.58,59 Although some data suggest that this hurdle can be overcome by appropriate cell dosing, concomitant implementation of a robust antithrombotic therapy,43 and careful control of the rate of cell delivery,59 it is clear that the intracoronary delivery of MSC in heart failure patients should still be viewed with caution. Finally, multiple passages of MSC may induce spontaneous transformation60 leading to uncontrolled growth kinetics and this potential safety issue thus requires that accurate cytogenetic studies be performed if expanded MSC products are to be used clinically.

Mechanisms of Action of Transplanted Cells: The Paracrine Paradigm

The previously mentioned discrepancy between the improved functional outcome yielded by MSC treatment and the lack of direct tissue replacement by the grafted cells is indeed a phenomenon that encompasses the other cell types tested so far, and this has led some to hypothesize that the benefits of cardiac cell therapy reported so far were not mediated by a new onset graft–derived myogenesis, but rather by an increase in cell-associated wall thickness leading to a reduction in chamber volume and consequently in wall stress and/or the release of trophic factors promoting survival and repair of injured tissue. Support to the latter hypothesis has come from the finding that skeletal myoblasts61,62 and bone marrow–derived cells63–66 stimulate endogenous pathways leading to increased angiogenesis, decreased apoptosis, and a favorable shift of extracellular matrix remodeling toward improved scar elasticity, even though the two cell types differ by the nature of the paracrine mediators involved, the downstream signal transduction pathways, and the ultimate specific effects.67 In line with this hypothesis, it has been reported that the beneficial effects of intravenously injected MSC, despite their low intramyocardial engraftment rate, were caused by trapping of the cells within the lungs and a subsequent upregulation of the anti-inflammatory factor TSG-6.68 Importantly, the cell-triggered upregulation of host-associated cytoprotective mediators seems to be prolonged, which explains the functional recovery of the heart at a time in which the trigger cells have faded away.69 By analogy with the ability of transplanted neural stem cells from fetal origin into the brain ventricles of aged rats to stimulate endogenous neurogenesis,70 the recruitment of dormant cardiac progenitor cells residing in the myocardium has been proposed as another target for cell-released factors.71 However, this hypothesis is more tenuous in view of the persisting uncertainties surrounding the existence of this stem cell niche in the adult heart (see the following). Finally, MSC have also been shown to release mediators that increase sympathetic nerve sprouting,72 which could contribute to increase cardiac performance, but also carries the risk of promoting arrhythmias related to heterogeneous cardiac innervation.

The recognition that cells primarily act as delivery vehicles not by physically substituting for dead host cardiomyocytes has led to speculation that a more straightforward means of inducing protection could be the sole injection of the conditioned medium derived from these cells. This hypothesis has been pioneered by the group of Dzau and coworkers73 who documented, in acutely infarcted rat hearts, infarct size limitation, and functional recovery as early as 72 hours after intramyocardial injection of the conditioned medium from hypoxic MSC genetically modified to overexpress the prosurvival gene Akt1. Likewise, delivery of the conditioned medium derived from MSC in a pig model of acute myocardial infarction has been shown cardioprotective, with the caveat of a very short (4 hours) period of observation.74 Although these studies provide compelling evidence for paracrine protection, their relevance to regeneration of the chronically failing myocardium remains more questionable because of the specificities of their protocol design (acute model of infarction and short follow-up). Indeed, in a study in which MSC were compared with injection of angiogenic growth factor genes, the former were found to better improve myocardial performance,75 which suggests that the physical presence of the cells, even transient, may be critical because the expected fast clearance of injectants from the target tissue likely decreases their efficacy (unless they are supplied with slow-release delivery systems). In addition, it may be difficult for conditioned media retrieved at a single time point to fully replicate the complex blend of trophic factors released by the engrafted cells over time.

Although one cannot dispute that the mechanical (reduction in wall stress) and paracrine effects of engrafted cells may effect some cardioprotection, it is still quite uncertain that these effects may be robust enough to translate into clinically meaningful improvements in patient outcomes, which stimulates the search for strategies that could achieve such an improvement through a true myocardial regeneration.

Conceptually, there are three major approaches that can be considered for inducing myocardial regeneration.

Stimulation of Endogenous Cardiomyocyte Proliferation

Whereas it has long been considered that the heart was a terminally differentiated organ, recent data have challenged this paradigm and elegant studies using 14C dating have provided evidence for some renewal of human cardiomyocytes spanning the lifetime but with a low turnover, as percentages of dividing cells decreased from 1% per year at the age of 25 to 0.45% at the age of 75 years.76 This has led some investigators to hypothesize that differentiated quiescent cardiomyocytes, which did not proliferate under resting conditions, could be induced to re-enter the cell cycle in response to extracellular mitogens. It has thus been demonstrated in mouse models of myocardial infarction that factors such as periostin77 or neuregulin78 triggered cardiomyocyte cell-cycle activity and that the resulting regeneration was associated with improved function, reduced fibrosis and infarct size, and increased angiogenesis. However, whether and how these basic science data can be translated into a safe and efficacious clinical therapy remains to be elucidated.

Transplantation of Putative Extracardiac Cell Niches Endowed with a Cardiomyogenic Differentiation Potential

In 2006, Witnisky and coworkers79 described in the mouse skeletal muscle a population of cells committed to a cardiomyogenic lineage. Unfortunately, using similar surface markers, we failed to show that such a cardiac-directed cell fraction existed in the human muscle.80 These findings are consistent with the increasing recognition that stem cells harbored in adult tissues are unlikely to cross lineages, except possibly under clinically irrelevant experimental conditions. It remains to be established whether this hurdle can be overcome by small molecules such as those of the sulfonyl hydrazone family, which have recently been reported as able to activate cardiac differentiation of human mobilized peripheral blood mononuclear cells,81 thereby effecting myocardial regeneration in a nude model of myocardial infarction.

Transplantation of Cardiac Stem/Progenitor Cells

Endogenous Sources: The Heart Itself

The 14C experiments mentioned in the preceding have actually been preceded in the early 2000s by studies from the group of Anversa who first claimed that despite being traditionally considered as a postmitotic organ, the heart actually harbored dividing cardiomyocytes evidenced by a positive expression for the cell proliferation marker Ki67, and that the proliferation rate of these cells was dramatically increased in patients with heart failure.82 The phenotypic characterization of these cells, found in multiple species, has yielded puzzling results in that different surface markers occasionally expressed in a mutually exclusive fashion have been identified (reviewed in Barile)83 and it is still uncertain whether this variability in marker distribution actually reflects distinct cell populations with specific functional properties or the ability of a unique pool of stem cells to express different markers at different stages of its maturation. C-kit + cells have been the most extensively investigated; because they lacked expression of hematopoietic markers, a bone marrow origin, such as that reported in sex-mismatched heart transplant recipients in whom 0.04% cardiomyocytes were found to be repopulated from extracardiac cells,84 has been excluded. Rather these c-kit + cells have been considered to represent a myocardial pool of progenitor cells committed to a cardiomyogenic lineage. Proof-of-concept was putatively provided by demonstration of donor-derived cardiomyocytes (along with smooth muscle cells and endothelial cells) when human cardiac progenitors were injected intramyocardially85 in a rat model of myocardial infarction and resulted in an improvement in contractile function, an attenuation of remodeling and histologic patterns of regeneration. This paradigm found a clinical application in July 2009 when a first patient was injected in Louisville, KY with his own cardiac stem cells derived from a right atrial biopsy that had been taken 4 months earlier during a CABG and subsequently expanded in Anversa's lab in Boston. Other investigators have adopted a slightly different approach in that instead of trying to isolate these putative cardiac stem cells, they grew fragments of myocardial tissue as cell clusters termed cardiospheres. These aggregates are composed of c-kit + progenitor cells primarily in their core and of a mix of cells expressing cardiac, endothelial, and mesenchymal markers at their periphery. Following preclinical studies showing that intramyocardial86 and intracoronary87 delivery of human and porcine cardiospheres could regenerate acutely and chronically infarcted myocardium, respectively, the group of Marban in Los Angeles moved to the clinics in June 2009 and performed a first intracoronary transfer of these cardiospheres 1 month after removal of a right ventricular specimen by an endomyocardial biopsy. A trial exploring the potential of this approach (CADUCEUS for CArdiosphere-Derived aUtologous Stem Cells to Reverse ventricular dysfunction) is currently under way.

Whereas the use of cardiac stem cells is appealing because of the cardiac lineage commitment of these cells and the avoidance of immunological issues owing to their autologous origin, several caveats must be mentioned. The major concern is that it is still uncertain whether these cardiac stem cells really exist in the adult diseased heart. In our experience, we have identified a few c-kit + cells in right atrial tissue taken during CABG or through endomyocardial biopsies,88 but all of them expressed hematopoietic markers and none expressed specific cardiac markers. Likewise, a high proportion of the c-kit + cells isolated from failing human hearts by Kubo and coworkers89 were positive for markers of the hematopoietic lineage, and although their number was fourfold higher that in nonfailing hearts, their reparative capacity was clearly inadequate because all failing hearts required transplantation. In keeping with these data, cells expressing both stem and differentiated cardiac markers have been identified in the human neonatal myocardium, but with a progressive decline in cell density over the first postnatal month,90 thereby raising serious doubts about the persistence of cardiac “stem” cells in adult patients with coronary artery disease, ie, those who would them need most. The fact that 4 months were required for growing these cells in the recently reported clinical experience also highlights the technical challenges associated with their isolation and scale-up and raises some doubts about a realistic widespread clinical use. Finally, a recent study has further challenged these findings by showing that spontaneously beating neonatal rat cardiospheres resulted from contamination by myocardial tissue fragments and that these cardiospheres actually represented aggregated fibroblasts rather than clonally expanded cardiac stem cells.91 Another more recently described variety of stem cells of cardiac origin is represented by those derived from the epicardium, but these cells also seem to improve function of infarcted hearts paracrinally as they fail to acquire a cardiac phenotype.92

Embryonic Stem Cells

Human embryonic stem cells (hESC) are derived from the inner cell mass of blastocyst-stage leftover embryos, which have been generated in the setting of assisted fertilization. Their most appealing feature is an intrinsic pluripotentiality that allows them to be committed in vitro toward the cardiac lineage in response to appropriate differentiation cues recapitulating those in play during the embryonic life and among which bone morphogenetic proteins (members of the TGF-β superfamily), Wnts, and fibroblast growth factors play a predominant role.93 It has thus been demonstrated that hESC could generate cells displaying the three major attributes of cardiomyocytes: (1) excitation-contraction coupling, demonstrated by the expression of ionic currents;94 intracellular calcium transients triggered by electrical activation and the resulting synchronous contractions; (2) responsiveness to the chronotropic effects of cardioactive drugs,95 and (3) the capacity to couple with neighboring cells through connexin 43-supported gap junctions.96 Following transplantation in rodent models of myocardial infarction, these cells have been shown to improve LV function, an effect usually associated with a beneficial effect on remodeling.97–99 Of note, mouse ESC have also been shown functionally effective in a nonischemic genetic cardiomyopathy model.100 One could legitimately argue that these cardioprotective effects have been reported with most cell types. However, an observation more specific for ESC is that, in contrast to their adult counterparts, they really differentiate into cardiomyocytes in transplanted areas, an event likely driven by cues originating from the infarcted tissue.101 However, injected ESC derivatives share with other cell types a limited amount of long-term engraftment,102 and this finding, along with the fact that increasing graft size has been reported not to positively correlate with a gain of function, has led to hypothesize that ESC-induced stimulation of vascularization was actually a major contributor to the cardiac functional improvement.103 Indeed, whereas ESC derivatives may well act paracrinally by activating endogenous cytoprotective pathways,104 one cannot fully exclude that they also directly contribute to cardiac contractility by generating a force transmitted to the surrounding tissues.105 Furthermore, it is important to recognize that the assessment of hESC transplantation is hampered by the lack of appropriate preclinical models as all of them are fraught with the confounding effect of xenotransplantation. For this reason, we have conducted experiments in nonhuman infarcted primates transplanted with primate ESC–derived cardiac progenitors, thereby mimicking an allogeneic situation closer to clinical practice. The data have confirmed the ability of these progenitors to achieve their differentiation into cardiomyocytes following in-scar engraftment. Furthermore, the finding made in a pig model of iatrogenic atrioventricular block, that intramyocardial transplantation of hESC-derived cardiac derivatives could restore electrical activity,106 provides compelling evidence for the ability of these cells to achieve an effective electromechanical integration within the recipient heart.

Asides from ethical considerations, the clinical use of hESC still raises several translational issues107 that fall into three main categories: (1) derivation and propagation of a line endowed with a cardiogenic differentiation potential, under good manufacturing practice (GMP) conditions; this first step of the process should allow to scale-up the number of cells without the use of xeno-components while maintaining their pluripotentiality and ensuring their genetic and epigenetic stability over time; (2) directed differentiation of the propagated pluripotent cells toward a cardiomyogenic lineage; (3) sorting of the cells, so as to yield a “pure” population of cardiac progenitors for clinical use. (Progenitors are likely preferable to a terminally differentiated cell type because of their ability to still proliferate in the host tissue to a certain extent, thereby effecting myocardial regeneration.) Several purification methods are available and rely on negative or positive selection, depending on whether they target the elimination of still undifferentiated cells or the enrichment of the lineage-committed progenitors from other unwanted cell types, respectively. Introduction of a reporter-suicide gene into stem cells that could serve as a safety net against cellular misbehavior is another option that warrants further investigation. In our experience, a positive selection approach based on identification of the surface marker SSEA-1 (or CD15) has turned out to be efficacious in that this antigen reliably labels cells that have entered a differentiation pathway,108 thereby allowing immunomagnetic sorting using an anti-CD15 antibody to selectively yield the clinically required cardiac progenitor cell population. The elimination of the contaminating cells is actually critical not only for efficacy, but also for safety reasons as a major concern with ESC is the development of a teratoma. Teratomas are complex tumors consisting of cell lines originating from the three initial germ layers. Although several factors (cell number, transplant site, degree of immunosuppression) may affect whether these tumors form after transplantation of ESC,109 a key role is played by the state of differentiation of the transplanted cells. Indeed, teratomas occur when cells are used in a still undifferentiated or poorly differentiated state and have thus retained an intrinsic capacity of uncontrolled proliferation, which can overcome the cardioinstructive signaling of the host heart. Conversely, animal models have failed to show evidence for teratomas following transplantation of hESC-derived cells committed to a given cell lineage,101,105 as demonstrated by their loss of pluripotent markers and the concomitant upregulation of the lineage-specific ones. These considerations highlight the importance of the purification step during preparation of a clinically usable ESC-derived cell therapy product.

Another major clinical issue raised by the use of this allogeneic cell source is immunogenicity. Despite initial hopes, it is now widely recognized that although undifferentiated ESC lack expression of MHC type II antigens and costimulatory molecules, they trigger a cellular and humoral immune response,110 which does not spare their differentiated progeny.111 Even though the lack of antigen-presenting cells within ESC-derived grafts may contribute to a reduced immunogenicity compared with adult tissues,111 it is still necessary to implement some form of immunosuppressive strategy to avoid cell loss by this immune process. Different approaches have thus been considered. Banking of hESC lines with a range of MHC profiles allowing an appropriate donor-recipient matching is likely not unrealistic, because about 150 randomly obtained cell lines could provide a worthwhile human leukocyte antigen (HLA) match for most potential recipients.112 This figure, however, was based on matching only selected MHC loci and thus does not dismiss that additional immune suppression would be required to overcome residual immunogenicity. Also, it may underestimate the ethnic diversity of most Western populations. Furthermore, such a banking approach raises complicated logistical and economic issues and, as such, first requires the convincing demonstration that the use of hESC-derived cells is therapeutically effective. Given the risks associated with a protracted use of immunosuppressive drugs, induction of tolerance in the recipient looks a particularly appealing approach should it be possible at the cost of minimal host conditioning.113 In practice, it might be achieved by the administration of hematopoietic stem cells derived from a given “donor” ESC line114 and the resulting induction of a microchimerism ensuring specific tolerance to cardiac (or other) progenitors derived from that same donor ESC line. The use of costimulation blockade for inducing peripheral tolerance mediated by regulatory T cells could be another strategy.115 In contrast, the creation of a “universal” cell line, for example by knockout of the β-2 microglobulin gene, which is essential for MHC-I presentation, remains a more uncertain perspective because MHC molecules also represent a surveillance mechanism whose deletion could create a favorable environment for viral infections and development of malignancies. Thus, from a practical standpoint, early clinical trials of hESC will likely have to rely on more conventional therapies based on immunosuppressive drugs. The challenge is then going to find an acceptable trade-off between the efficacy of these drugs and their well-known side-effects so as to optimize the risk-to-benefit ratio of the procedure. The additional use of autologous or allogeneic mesenchymal stem cells for further mitigating the immunorejection process is supported by a study in kidney transplant recipients,116 although we have experimentally failed to show such a benefit in a rat model of hESC transplantation.99

The recent presidential lift on the US federal funding ban on ESC research and the attendant increase in resources allocated to this area make likely that many of the remaining hurdles will be overcome in a not too distant future.

Induced Pluripotent Cells

In 2006, Yamanaka made a real breakthrough in the stem cell field when he reported that mouse embryonic and adult fibroblasts could acquire properties similar to those of ESC after retrovirus-mediated transfection of genes encoding the four transcription factors Oct3/4, Sox2, Klf4, and c-Myc.117 Since then, several groups have embarked in this research on induced pluripotent stem (iPS) cells and have successfully reprogrammed human adult somatic cells into an embryonic-like state, including from patients with a variety of neurodegenerative and genetic diseases.117 Reprogrammed cells can then be redirected toward a given cell lineage, including cardiomyocytes, and this cardiomyogenesis process actually involves a temporal sequence of gene expression (specific for cardiac mesoderm, cardiac transcription factors, and cardiac-specific structural proteins, successively), which parallels that observed during differentiation of hESC.118

The use of iPS could potentially allow to address two limitations of ESC, ie, immune rejection and ethical concerns raised by the use of human embryos, with still the caveat that one does not know whether the reprogramming-redifferentiation process could not cause expression of otherwise hidden-self and lead to autoimmunity events. However, their clinical application also faces important obstacles. Some of them are shared with ESC, such as teratoma formation following unwanted transplantation of still undifferentiated cells; other hurdles are rather unique to iPS cells and specifically include the risk of incomplete/aberrant reprogramming that may impair their ability to be redifferentiated toward the required cell type. It also remains to determine which is the best somatic cell source to use (eg, skin fibroblasts, keratinocytes, blood cells) and how to increase the still low efficiency of the reprogramming process while improving its safety. This is indeed a key question as a major risk of the initially used retroviruses and lentiviruses is transgene reactivation leading to tumorigenesis, hence the more recent interest in plasmids, proteins,119 and even chemicals that represent a mandatory step on the pathway to production of iPS for clinical use. However, even these virus-free approaches may induce genetic alterations and therefore would call for a sequencing of the whole genome of iPS cell clones to detect these alterations before any clinical application.

These considerations suggest that although iPS cell technology may be soon ready for prime time toxicology screening and patient-specific disease modeling, its application to “regenerative” medicine still requires many challenges to be surmounted. Nevertheless, encouraging data have already been reported. Thus, the potential therapeutic value of iPS cells has yet been successfully tested in three noncardiac disease models (sickle cell anemia, Parkinson's disease, and hemophilia A). These data have more recently been extended to the heart by Nelson and coworkers,120 who have pioneered the transplantation of fibroblast-derived iPS cells in a murine model of myocardial infarction and shown that it resulted in an improvement in LV function compared with the parental fibroblasts, associated with a differentiation of the iPS cell progeny into cardiomyocytes. However, because this study did not include “true” ESC-derived cardiomyocytes as another control, additional experiments remain necessary to thoroughly assess the comparative efficacy of these two types of pluripotent stem cells. This is particularly relevant in view of recent data that indicate that regardless of their origin or the method by which they have been generated, iPS retain a gene expression pattern distinct from that of hESC and that extends to the expression of noncoding (mi)RNAs.121 Therefore, it is critical to assess the functional significance, if any, of these differences. Finally, although iPS have the major theoretical advantage over hESC of being patient-specific, one should not underscore the limitations of autologous cell therapy products when large patient populations are targeted; these limitations include intrinsic interindividual variability in cell functionality, which complicates the production of consistent cell products, logistical constraints inherent in cell shipments, and the huge cost of customized batch controls.

Regardless of the cell type used, a consistent finding is that only a tiny number (usually <1%) of the cells that have been initially delivered can still be identified after a few days or, at best, weeks. This massive attrition rate proceeds from two sequential events: an initially low retention of the grafted cells into the target tissue and a subsequent low survival rate of the cells that have been initially retained. It is thus clear that the optimization of regenerative cardiac cell therapy requires to address these two issues.

Low Retention of Grafted Cells

This low retention is caused by the interplay of several factors: mechanical leakage of the cells along the needle tracks, washout through the venous and lymphatic drainage systems of the heart122 with a subsequent large extracardiac cell redistribution (primarily in the lungs and the spleen), and limited myocardial homing in cases in which cells are delivered intravascularly.

Factors of Cell Retention

Three factors of cell retention must be considered: the route of cell delivery, the phenotype of the cells, and the nature of the target tissue.

Cells can be delivered by intravenous infusion, intracoronary or endomyocardial catheterization (the latter usually under electromechanical mapping), and open-chest transpericardial injections. Each of these routes has advantages and drawbacks.

The intravenous route is clearly the simplest and the safest. Unfortunately, it is associated with a minimal, if any, myocardial uptake of the injected cells,123 suggesting that homing signals are unlikely to be powerful enough to overcome trapping (particularly in the lungs) of the circulating cells and drive them toward the injured areas. Intracoronary injections with the stop-flow technique result in minimal myocardial retention and dynamic tracking of 18F-fluorodeoxyglucose–labeled progenitor cells has clearly documented a loss of approximately 80% of the myocardiac activity of the tracer after balloon deflation.124 Furthermore, direct intracoronary infusion of cells may be harmful if their size is such that it can result in microvascular obstruction, as reported with MSC.58 Indeed, direct intramyocardial injections seem to be the most efficient route for delivering cells,125 and at least in the case of skeletal myoblasts, it does not seem that engraftment is then different whether these intramyocardial injections are made from the endocardium or transepicardially.126 In the latter case, the role of the heart status remains uncertain with one study reporting no difference between beating-heart and arrested-heart injections,127 whereas another, using more sophisticated assessment tools (MRI and quantitative polymerase chain reaction) documented the higher intramyocardial cell numbers when injections were made in the cardioplegic heart.128 The major limitation of this epicardial approach is clearly its invasiveness (unless an associated surgical procedure is indicated); hence, the potential interest of a recently developed miniature robotic device that can be introduced through a subxiphoid approach, navigates over the epicardial surface and, owing to a fine-positioning control system, allows multiple injections in predefined target spots.129 Finally, retrograde infusion through the coronary sinus with a dedicated ultrasound-guided catheter130 has gained limited clinical acceptance and would likely be technically challenging in heart failure patients whose coronary sinus is now often already filled with leads connected to a synchronization device.

The phenotype of delivered cells is also likely to affect the magnitude of their retention as demonstrated by the finding that intracoronary delivery of CD34+ cells results in greater numbers of these progenitors in the myocardium compared with unfractionated MNC.131

The nature of the target tissue also plays a role in cell retention. This is illustrated by the finding that intramyocardial retention of MNC is dramatically reduced in chronic heart failure compared with acute myocardial infarction,132 likely because chemoattractive forces that may promote cell trafficking at the time of acute injury are no longer present (or less operative) in the setting of a remodeled old infarction.

Improvements in Cell Retention

Improvements in cell retention need to be considered differently depending on whether cells are delivered intramyocardially or intravascularly.

Intramyocardial Delivery

A first strategy consists of acting on cells and to increase the viscosity of the injectate to minimize leakage at the time of injections. This can be achieved by incorporation of cells into biomaterials or by encapsulating them in microparticles.

Incorporation into Biomaterials

Cells sense matrix nano topography and stiffness cues133 as well as mechanical signals such as strain and stretch and these factors can thus modulate important cell functions such as proliferation, alignment, and differentiation. This implies that the choice of a scaffold has to be customized to a given stem cell type. Basically, polymers can be categorized as naturally derived (collagen, fibrin, decellularized matrices) or synthetic. The interest of embedding cells into one of these biomaterials is not only to improve their retention for purely physical reasons (increased viscosity) but also to facilitate some key functions such as survival, proliferation, and differentiation owing to cell patterning within a tridimensional microenvironment. These speculations are supported by the finding that incorporation of different types of cells into supports such as collagen,134 fibrin glue,135 or self-assembling nanopeptides136 have been successful in increasing angiogenesis and preserving cardiac function. In addition, these polymers can serve as delivery vehicles for growth factors, thereby contributing to further enhance the efficacy of the cell transplantation procedure.137

In line with this reasoning, some investigators have successfully tested injection of cell-free dermal fillers138 or alginate-derived hydrogels139 for limiting postinfarction remodeling. These results are consistent with the finding that injecting passive materials alone into a postinfarct scar and consequently replacing a stiff tissue by a more compliant material may improve function and reduce wall stress in the ventricle, thereby contributing to normalize mechanotransduction signaling.140 However, exclusive reliance on the passive function of acellular biomaterials is unlikely to affect myocardial regeneration, which requires incorporation of contractile cells able to line up with the appropriate geometry and to be excited in synchrony with the native cardiomyocytes.

Encapsulation

This approach entails incorporation of cells into porous microbeads the membrane of which is engineered in such a way that cell-secreted therapeutic factors can freely diffuse into the surrounding myocardial environment, whereas influx of oxygen and nutrients, but not of immune cells, is made possible by the size of the pores. In addition of increasing cell retention because of the size and viscosity of the particles, encapsulation may protect nonautologous cells from the host immune system without drug-based immunosuppression.141 In fact, this technology has been successfully used for encapsulating xeno Langerhans' islets in animal models of diabetes.142 Some clinical trials are currently under way in this indication. However, in the context of myocardial regeneration, this approach is questionable because it exclusively relies on the paracrine effects of cells and inherently prevent any form of functional integration of the graft within the host myocardium.

A second strategy consists of acting on the delivery systems. If the system is based on conventional injections, the design of the device could be conceivably improved to maximize cell retention, for example through changes in needle configuration allowing better spatial dispersion of cells and/or computerized driving to allow for a more accurate control over flow rate, duration, and injection pressure.

However, needle-based injections are fraught with several limitations: leakage along the needle tracks; random and poorly reproducible distribution of cells; creation of multiple intramyocardial clusters that can physically impede the propagation of electrical impulses and result in conduction blocks facilitating arrhythmias;10 induction of volume-dependent tissue damage; and dissociation of the cells promoting a particular form of apoptotic cell death called anoikis</i>,143 which occurs when anchorage-dependent cells lose their attachment to the extracellular matrix and the related survival signals.

Consequently, one can take advantage of open-chest interventions to step away from injections and switch to the epicardial delivery of cell constructs overlying the infarct area and simply glued or sutured to its edges. The group of Eschenhagen has shown that if the selected cells featured a contractile phenotype as occurs with rat cardiomyocytes, it was possible, by mixing these cells with collagen, to generate force-generating beating heart constructs that, following implantation onto the scar tissue, were found to improve function, prevent dilation, and provide electrical coupling with native myocardium.144 In the future, it is possible that cell-scaffold interactions might benefit from the bioprinting technology that enables transfer of cells onto the matrix (just like an inkjet is printed by a desktop printer) according to a precisely defined pattern.145

One step further, Okano and his group in Japan have developed an original technology based on the manufacture of cell sheets cultured on temperature-sensitive dishes.146 Upon cooling, a layer of contiguous viable cells can be collected without disrupting the cellular microenvironment caused by enzymatic digestion otherwise required for cell detachment. The multilayered construct is then applied onto the surface of the heart where it adheres spontaneously. This technique has been used successfully to a variety of cells (eg, skeletal myoblasts, fibroblasts, epithelial cells, rat cardiomyocytes, mesenchymal stem cells) and has the advantage of avoiding the presence of any foreign material and allowing cells to be anchored through the extracellular matrix adhesive proteins that they have secreted themselves. In cardiac applications, the cell sheet technology has been shown more effective than injections in both ischemic147 and nonischemic148 animal models of heart diseases and a myoblast cell sheet has already been used clinically as an adjunct to a LV assist device. In a head-to-head comparison of conventional injections of myoblasts versus the epicardial delivery of either a myoblast-seeded collagen patch or a myoblast cell sheet,149 we have been able to confirm the superiority of the two latter techniques with regard to improvement of postinfarction function, reduction of fibrosis, and increase in angiogenesis.

Conceptually, the use of epicardial cell constructs raises the major concern of the coupling of superficially delivered cells with the underlying host cardiomyocytes. However, migration of the patch-bound cells in the myocardium has been reported150 as well as their coupling with host cardiomyocytes through gap junction proteins.150,151

Intravascular Delivery

If cells are to be injected into the coronary arteries, a first strategy consists of acting on the delivery systems. This approach is illustrated by the recent development of a perforated intracoronary catheter that allows cell delivery in the perivascular space.

A second strategy consists of acting on the homing signals responsible for cell trafficking. This can be achieved by interventions targeting cells directly or the recipient tissue.

Cell Targeting

Homing of cells to sites of injury is orchestrated by a tightly coordinated sequence of events involving “sensing” of the damaged tissue–emitted signals, selectin-mediated rolling, integrin-mediated firm adhesion to the vascular wall and transendothelial migration in the extracellular matrix.152 Among the multiple factors that participate in the recruitment and trafficking of circulating cells, a crucial role is played by the chemokine CXC receptor (CXCR) 4, which is expressed on early hematopoietic stem cells and endothelial progenitors and its receptor, the stromal cell–derived factor (SDF)-1α.

Interventions designed at enhancing homing have thus logically relied on cell engineering to make them overexpressing one component of the receptor-ligand pair, ie, CXCR4153 or SDF-1.154 Transfection of MSC with insulin growth factor-1155 has been another means of reactivating SDF-1 signaling. Aside from these gene-based approaches, drugs that activate endothelial nitric oxide synthase (NOS) such as statins or NOS enhancers have also been successful in enhancing cell homing.152

Tissue Targeting

In this setting, direct myocardial SDF-1 gene transfer156 has been proposed for re-establishing tissue levels of SDF-1 known to decrease shortly after infarction. However, the anticipated regulatory hurdles, resulting from the combination of cell and gene therapy make appealing potentially safer and simpler techniques based on physical interventions. So far, three of them have been successfully tested: (1) Low-energy shock wave aims to induce expression of SDF-1,157 and the CELLWAVE trial is already testing the effects of extracorporeal shock wave before bone marrow progenitor cell therapy in an effort to improve homing of these cells and subsequent angiogenesis in patients with chronic ischemic heart disease following anterior myocardial infarction. (2) Focused ultrasound-mediated destruction of microbubbles has been shown to create a series of biochemical responses (eg, release of proinflammatory cytokines, induction of metalloproteinase activity, reduction of laminin content), which result in enhanced transendothelial migration and interstitial invasion of MSC without causing endothelial damage.158 (3) Finally, magnetic targeting could also be useful for driving cells toward the site of injury based on the seminal proof-of-principle finding that an externally applied magnetic device was effective in enhancing iron-loaded endothelial progenitor cell engraftment at a site of vascular injury.159

It is finally important to note that while enhancement of homing is particularly important in the context of acute myocardial infarction where the intravascular route is the only one available for stem cell delivery, this issue is somewhat less critical in the context of regeneration of the chronically failing heart, which allows a greater versatility in the stem cell transfer approaches and, in particular, a direct catheter-based or surgical intramyocardial delivery.

Low Survival of Retained Cells

Although cell death results from the complex interplay of multiple factors, ischemia inherent in the poor vascularization of the target areas, apoptosis subsequent to the loss of cell anchorage to their matrix and immune destruction of allogeneic cells have been identified as key players. Consequently, adjunctive interventions targeted at optimizing graft survival have focused on these events.

Induction of Angiogenesis

When the grafted area cannot be revascularized directly by CABG or angioplasty, two different types of strategies can be considered for increasing blood supply to the areas of engraftment and therefore promoting survival of the injected cells.

The first is based on the direct intramyocardial injection of exogenous angiogenic inducers which fall into four main categories: (1) genes-encoding angiogenic factors;160 (2) growth factors themselves with the caveat that their fast clearance requires appropriate delivery systems; (3) genetically modified cells overexpressing an angiogenic factor;161,162 and (4) cotransplanted cells featuring an angiogenic potential such as CD133+ hematopoietic progenitors,163 or the more recently described epicardium-derived cells.164 Although there are no conclusive data about the most efficacious of these interventions, the importance of a cellular carrier is suggested by the finding that myoblast transfection with VEGF provides better cardiac protection than direct adenoviral VEGF delivery.165

The second strategy relies on the induction on an endogenous angiogenesis through an enhanced homing of circulating bone marrow cells in the areas of cell engraftment. As mentioned, this approach has largely relied on manipulation of the CXCR4-SDF1 axis to successfully increase the influx of angiogenic cells into the cell-transplanted regions.

Although one is missing a comprehensive head-to-head comparison of these various interventions, overall they all have met the objective of increasing angiogenesis, which has usually translated into a better LV function and an enhanced survival of the cellular graft. An advantage of targeting SDF-1 could be that in addition of its stem cell-recruiting effects through its interaction with CXCR4, it may exert direct cardioprotective effects by activating downstream signaling pathways involved in cardiomyocyte survival.166 However, a potential clinical use of SDF-1 should be cautiously considered as the proteolytic cleavage of SDF-1α by matrix metalloproteinase-2 could yield a highly neurotoxic molecule, SDF.5–67,167 Along the same line, malignant transformation has been reported in a mouse model of myocardial infarction following intramyocardial transplantation of MSC that had been engineered to overexpress stem cell factor.168

Limitation of Apoptosis

As pointed out, injection of dispersed cells promotes their apoptotic death because of the loss of survival signals originating from physiologic cell-cell and cell-matrix interactions. A first means of addressing this issue is to deliver cells that have been transfected with antiapoptotic factors such as Akt169 or Bcl2.170 However, given the safety issues associated with gene therapy, a more attractive option could be to upregulate endogenous survival pathways though preconditioning of cells either physically by heat shock171 or pharmacologically by potassium channel agonists.172 Finally, where direct access to the heart is feasible, the use of three-dimensional cell-seeded patches or scaffold-free cell-sheets, particularly if they are multilayered,173 is probably the best approach with regard to the risk-to-benefit ratio. The advantage of these constructs is not only that they do not cause disruption of myocardial tissue and preserve cell cohesiveness, but also that they allow incorporation of several different cell populations such as cardiomyocytes (regardless of their origin), endothelial cells, and fibroblasts.174–176 Cross-talks between these cell populations may synergize trophic and particularly angiogenic effects which may enhance survival of the contractile cells “of interest.”150,176 Of note, studies of composite sheets made of cardiomyocytes and endothelial cells have shown that blood vessels originated from the construct and bridged to connect with host capillaries,175,176 thereby providing additional evidence for the coupling between these epicardial cell constructs and the myocardium underneath.

Limitation of the Immune Response to Allogeneic Cells

This issue can be addressed by immunomodulatory strategies (such as those described about ESC) when the objective is a sustained cell engraftment or by encapsulation (also described in the preceding) if one admits that cells will not integrate within the recipient myocardium and exclusively relies on their paracrine effects.

To close this section, it is fair to acknowledge that the importance of keeping cells alive has recently been challenged by a study showing in a mouse model of myocardial infarction that extracts from unfractionated bone marrow cells, ie, their intracellular soluble contents, were actually as cardioprotective as factors released by living cells.71 This conclusion, however, implies that protection exclusively relies on paracrine effects and is therefore not relevant to the generation of new myocardial tissue by donor-derived cells.

Tracking of Stem Cells

Clearly, the objective assessment of the above mentioned approaches for optimizing cell engraftment requires the use of safe, reliable, and noninvasive techniques for tracking cell fate over time, both in the target myocardium and remote organs.177 Specific constraints on contrast agents include the lack of genetic modification of the labeled cells, of dilution with cell division and of transfer to non-stem cells. Experimentally, radionuclides, MRI, and reporter genes are the most extensively used molecular imaging techniques. Reporter genes are particularly attractive. Once transfected into the cell of interest, they yield a product that, following exposure to the corresponding radioactive or optical reporter probe, turns on a signal that can be imaged by positron emission tomography, single-photon emission computed tomography, or MRI. A first advantage of this approach is that it should only detect viable cells because expression of the imaging signal requires intracellular expression of the reporter gene product. This contrasts with MRI-based tracking of iron-loaded cells, which may generate false-positives when iron released by dead stem cells is engulfed by host macrophages and results in MRI signals mistakenly interpreted as originating from the transplanted cells.178 Second, the reporter gene can be genetically integrated and transmitted to the transfected cell's progeny, thereby allowing to monitor cell proliferation. Finally, combination of several reporters can allow to set up a multimodal imaging platform such as that achieved with a triple-fusion reporter gene, which allowed a successful noninvasive monitoring of the kinetics of cell survival, proliferation, and migration by bioluminescence and PET imaging.179 Validation of these data is supported by the critical finding that reporter genes do not seem to adversely affect cell integrity and function. Multimodality imaging is likely important because there is currently no single technique that can optimally combine resolution, sensitivity, safety, accuracy of cell quantification, and clinical applicability. Bioluminescence, for example, which uses light generated by the enzyme luciferase is limited to small animals because of the poor tissue penetration of external illumination sources and a similar limitation applies to near-infrared fluorophores. In the future, however, this issue might be addressed by nanometer-sized light-emitting quantum dots conjugated to bioluminescent agents. Improved imaging techniques are thus eagerly awaited for experimental stem cell tracking and subsequent clinical translation without adding significant cost or regulatory roadblocks.

Because there is no animal model that can fully reproduce the complex situation of patients with advanced heart failure in need for myocardial regeneration, it seems legitimate to continue clinical trials, provided that their experimental grounds are robust and consistent enough to suggest that the selected cell type has a reasonable potential of effecting some myocardial regeneration without causing adverse events. Beyond the stage of pilot small-sized feasibility studies, these trials should comply with the standard methodologic guidelines (randomization, control groups, blind assessment) and their protocol should carefully take into account some key variables such as cell dosing, delivery modalities, and realistic approvability by the regulatory authorities. Until large mortality trials can be launched, surrogate end points and imaging modalities should also be selected in light of the expected objective of regeneration and, as such, the commonly used measurements of global LV function may have to be shifted toward the assessment of cell-related morphologic changes such as infarct size, LV remodeling, or regional wall thickness areas. At the end, a continuous cross-talk between different disciplines including clinical cardiology, developmental biology, cell culture technology, tissue engineering, imaging and high-throughput production of scalable, and reproducible biologics appears mandatory for moving the field forward and successfully achieving the goal of myocardial regeneration under conditions of an optimal risk-to-benefit ratio.