| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;75:S20-S28
© 2003 The Society of Thoracic Surgeons
a Department of Cardiovascular Surgery, Hôpital Européen Georges Pompidou, Paris, France
* Address reprint requests to Dr Menasché, Department of Cardiovascular Surgery, Hôpital Européen Georges Pompidou, 20, rue Leblanc, 75015 Paris, France
e-mail: philippe.menasche{at}hop.egp.ap-hop-paris.fr
Presented at the Heart Failure & Circulatory Support Summit, Cleveland, OH, Aug 22–25, 2002.
Abstract
Cell transplantation is gaining a growing interest as a potential new means of improving the prognosis of patients with cardiac failure. The basic assumption is that left ventricular dysfunction is largely due to the loss of a critical number of cardiomyocytes and that it can be partly reversed by implantation of new contractile cells into the postinfarction scars. Primarily for practical reasons, autologous skeletal myoblasts have been the first to undergo clinical trials but other cell types are also considered, particularly bone marrow stem cells, which are attractive because of their autologous origin and their purported cardiomyocyte/endothelial transdifferentiation potential in response to cues provided by the target organ. However several key issues still need to be addressed including (1) the optimal type of donor cells, (2) the mechanism by which cell engraftment improves cardiac function, (3) the optimization of cell survival, and (4) the potential benefits of cell transplantation in nonischemic heart failure. In parallel to the experimental studies designed to address these issues clinical trials are under way and should hopefully allow assessing to what extent cell transplantation may improve the outcome of patients with heart failure.
The management of patients with heart failure is receiving a continuously growing interest because of the increased prevalence (approximately 5 million US citizens) and incidence (400,000 to 600,000 new patients every year) of this condition. The magnitude of the problem is expected to be even amplified in the forthcoming years because of the increased age of the population and the improved postinfarction survival rates resulting from recent pharmacologic and interventional treatments.
Contemporary medical therapy has dramatically improved the prognosis of heart failure which will possibly be further ameliorated by the new drugs currently under investigation. In many cases however medical therapy is simply palliative and only shifts the survival curve rightwards, which accounts for a persistently high mortality that can reach 60% within 1 year for patients in New York Heart Association functional class IV. These figures obviously translate into tremendous financial costs, primarily hospital driven, and are estimated to consume 1% to 2% of the total health care budget of Western countries [1].
Although cardiac transplantation remains the only radical treatment of the most advanced forms of heart failure the limitations of this approach, largely related to organ shortage, have led to a continued endeavor for designing alternate options. Most of them have focused on reshaping the dilated left ventricle primarily by endocardial patch plasty [2] and more recently by passive constraint (Acorn) and shape-change (Myosplint, Cardioclasp) devices. In parallel improvements have been made in ventricular assist devices particularly as destination therapy [3] but the use of permanently implantable blood pumps still remains investigational. In patients with wide QRS complexes, cardiac resynchronization has also emerged as a promising treatment, which does not exclude additional approaches more directly targeted at improving pump function. In this setting cell transplantation is currently generating a great deal of interest, which is further enhanced by the doubts of several investigators regarding the relevance of gene therapy to the clinical management of heart failure.
Basic assumptions
Cell transplantation is based on two major assumptions: (1) heart failure develops when a critical number of cardiomyocytes has been irreversibly lost, and (2) function can thus be improved by repopulating these areas of "dead" myocardium with a new pool of contractile cells. One of the most compelling observations that validate this concept of functional replacement therapy comes from studies of Langerhans islet transplantation in diabetic patients and showing that after a median follow-up of 10.2 months 11 of 12 patients intraportally injected with allogeneic islets and receiving appropriate immunosuppressive therapy were able to achieve insulin independence [4].
Assuming that spontaneous multiplication of adult cardiomyocytes, if it occurs [5, 6], is in any way too low for compensating for the loss of infarct-injured cells and that conversion of in-scar fibroblasts into contractile cells would require genetic manipulations of questionable clinical relevance, the most realistic approach consists of exogenously supplying a new pool of contractile cells and to engraft them into the postinfarct scars. The requirement for a discrete area to target cell injections has resulted in most of the preclinical studies using models of segmental ischemic cardiomyopathies. Preliminary data suggest however that the benefits of cell transplantation might also extend to the setting of globally dilated, idiopathic [7], or drug-induced [8] cardiomyopathies.
Cell types
The prerequisite for implanted cells to improve cardiac function is that they feature contractile properties. Consequently although some positive data have been reported with fibroblasts [9], smooth muscle cells [10], and endothelial cells [11] these results remain inferior to those yielded by contractile cells whether this contractility is naturally present (fetal cardiomyocytes and skeletal myoblasts) or has been induced by transdifferentiation (bone marrow stem cells). The results yielded by these different cell types first need to be critically analyzed in light of the various experimental models.
Experimental models
Cell transplantation has first been tested in rodent models of myocardial infarction. This infarction in turn is created by coronary artery ligation or cryoinjury. The former method results in rather heterogenous areas of fibrosis that closely mimic the patchy pattern of human infarcts but the interindividual variations in infarct size require substantial numbers of animals to draw statistically meaningful conclusions. In contrast application of a cryoprobe results in discrete scars of a consistent size but the clinical relevance of this type of injury is more limited. The use of nonischemic models of cardiac dysfunction is more challenging. While protocols of doxorubicin administration have been developed that result in heart failure, it may be difficult to find the optimal trade-off between a sufficient degree of left ventricular dysfunction allowing demonstration of treatment effects and an acceptable survival rate allowing long-term studies. From this standpoint the development of transgenic strain of mice made knock-out for genes that control key steps of the excitation-contraction machinery is of utmost importance for assessing the effects of cell transplantation in clinically relevant models of nonischemic dilated cardiomyopathy. Once the screening phase has been achieved in rodents, investigators have usually moved to large animal studies before undertaking clinical trials. Ischemic cardiomyopathy has then been modeled relatively easily in pig or sheep by open-chest coronary artery ligation or endovascular release of thrombogenic coils. The creation of nonischemic heart failure is again far more challenging, particularly because the usual model, namely rapid pacing, is not well suited for cell transplant studies (maintenance of pacing may interfere with the function of grafted cells while its interruption is usually followed by a prompt normalization of heart function). The observation that ventricular arrhythmias which may represent an adverse event after clinical cell transplantation (see below) has never been reported in any of the basic animal studies further demonstrates the limitations of all these experimental models and emphasizes the caution with which data derived from these studies should be interpreted.
The same limitations apply to the methods used for assessing the functional results. Although the Langendorff perfusion technique has been used in some studies with the purported advantage of allowing control over preload, afterload and heart rate, the clinical relevance of data derived from isolated hearts can be questioned. It therefore appears that in vivo assessment techniques are better suited for cell transplantation studies. In this setting M-mode and bidimensional echocardiography has been widely used because of its simplicity and limited invasiveness that allows sequential measurements. Its major limitation however is that volume determinations are based on an elliptical model of left ventricular geometry that may not apply to segmental dysfunction as occurs in myocardial infarction. This problem can be overcome in large animals where regional function studies are made possible by tissue Doppler imaging, which allows the measurements of transmyocardial velocity gradients and thus provides accurate information on the intrinsic contractile and relaxation properties of the target area. 99 mTechnetium sestamibi single photon emission computed tomography is another attractive method for assessing regional and global function as well as coronary perfusion although resolution of the system may not allow an accurate absolute quantification of these indices. Furthermore this technique shares with echocardiography the major disadvantage of being dependent of preload, afterload, and heart rate. For this reason some groups have rather put emphasis on techniques that generate load-independent data. In this setting generation of pressure-volume curves by a conductance catheter is particularly attractive because it allows a separate assessment of systolic and diastolic function. However, in addition of being technically demanding this approach only explores global function and is fairly invasive in small animals, which usually precludes serial studies; the lack of base line pretransplant measurements may then complicate interpretation of posttransplantation between-group differences. Sonomicrometry is another load-independent technique that differs from the preceding one in that it provides information on regional function with the caveat that measurements can be critically dependent on location of the ultrasonic transducers. Furthermore its invasiveness makes equally difficult longitudinal studies in small animals. Thus this brief critical review outlines the limitations of the various techniques used for assessing the effects of cell transplantation and a therapeutic "trend" is often best derived from a combination of several of them. In the future, some of these issues might be addressed by magnetic resonance imaging with tagging and contrast enhancement but the application of this approach to experimental studies still remains associated with significant cost, logistical and technical problems.
Fetal and neonatal cardiomyocytes
Early studies with fetal and neonatal cardiomyocytes have been pivotal to establish the "proof of concept" by showing in small animal models of myocardial infarction induced by coronary artery ligation or cryoinjury that these cells effectively engrafted into injured areas, developed communications with host cardiomyocytes through connexin 43–supported gap junctions and improved left ventricular function [12–14]. The stability of these results over time is supported by recent data showing that grafted neonatal cardiomyocytes are still detectable in infarcted areas as long as 6 months after transplantation and are associated with thickening of the left ventricular wall, increased ejection fraction and reduced dyskinesis as assessed angiographically [15]. That transplanted cells can functionally integrate within the recipient tissue is further evidenced by the findings that fetal cells harvested from the sinoatrial area trigger a pacemaker activity after their transplantation in animals whose conduction system has been irreversibly damaged [16]. However, from a clinical perspective the transplantation of fetal or neonatal cardiac cells raises significant issues related to ethics, availability, and antigenicity that question the wide-scale clinical applicability of this approach.
The challenge is even greater for embryonic cells, which generate a continued state of turmoil that extends to the lay press. There is no question that embryonic cells are conceptually attractive because their totipotency should make possible to prepare cardiomyocyte cell lines in vitro before injecting them into myocardial scars. These cells can be derived from fertilized oocytes that are no longer targeted for childbearing; alternatively they could be obtained after nuclear transfer into enucleated recipient oocytes (therapeutic cloning). The subsequent development of the blastocyst would then be stopped at an early stage and the embryonic cells, which recapitulate the whole genetic progam of the future recipient, could then be appropriately cultured so as to drive them toward a cardiomyogenic lineage. However, apart from the major ethical and regulatory issues raised by this approach, and which go far beyond the scope of this review, the clinical applicability of embryonic cell transplantation is plagued by major technical challenges largely due to the fact that differentiation of these cells into the target cell lineage should be complete before transplantation, as contamination of the injectate with residual undifferentiated cells carries the major risk of tumor development (teratoma). In the pivotal study of Kehat and associates [17] in which human undifferentiated embryonic stem cells were grown from a single-cell clone, only 8.1% of the embryoid bodies generated from these cells spontaneously contracted and stained positively for cardiac-specific markers, thereby indicating that the objective of a 100% pure and homogeneous cell yield is still far from being achieved.
The above considerations easily explain the attention that has been rapidly paid to an alternate type of intrinsically contractile cells, namely skeletal myoblasts.
Skeletal myoblasts
These skeletal muscle stem cells (also known as satellite cells) normally lie in a quiescent state under the basal membrane of skeletal muscular fibers. After tissue injury they are rapidly recruited, proliferated, and fused, thereby effecting repair and regeneration of the damaged fibers. From the perspective of clinical applications, these cells feature several attractive characteristics: (1) an autologous origin, which overcomes problems related to availability, ethics, and antigenicity; (2) a high proliferative potential under appropriate culture conditions that allows a standardized scale-up and is a key factor for wide-scale clinical applicability; (3) a commitment to a well-differentiated myogenic lineage, which virtually eliminates the risk of tumor development; and (4) a high resistance to ischemia, which is a major advantage given the hypoxic environment of postinfarct scars in which they are implanted.
Morphologically the injected myoblasts differentiate into typical multinucleated myotubes that, at least in a sheep model of myocardial infarction, tend to repopulate the areas of fibrosis [18]. It is now clearly established that engrafted myoblasts do not transdifferentiate into cardiomyocytes [19, 20] and previous reports claiming such a transdifferentiation have most likely been methodologically flawed by methodological problems, namely failure to conclusively establish either that the detected cells were derived from the donor (and not from the host) or that engrafted donor cells truly expressed markers specific for the cardiac phenotype [20]. However although the cardiac environment is not able to reverse the commitment of engrafted myoblasts to a skeletal muscle-type lineage, our observation that the proportion of fibers demonstrating a purely slow or composite (fast and slow) myosin isoform pattern increases over time suggests that stretch or repeated electromechanical stimulation may induce some phenotypic changes, as previously reported after dynamic cardiomyoplasty. Furthermore in contrast to fetal cardiomyocytes, engrafted skeletal myotubes do not physically couple with host cardiac cells. Indeed cultured skeletal myoblasts express N-cadherin and connexin-43 (the major proteins constitutive of fascia adherens and gap junctions and therefore responsible for mechanical and electrical coupling, respectively, in heart tissue) but expression of these proteins is down-regulated after intramyocardial implantation [21].
The functional translation of this engraftment is an improvement in left ventricular function, which has been demonstrated in small and large animal models of myocardial infarction created by coronary artery ligation, cryoinjury, or toxic chemicals [18, 22–25]. In the study by Taylor and colleagues [23] function of cryoinjured rabbit hearts injected with autologous myoblasts was only improved in those hearts where transplanted cells were identified, thereby suggesting the causal relationship between engraftment of cells and functional outcome. Importantly the functional benefits of myoblast transplantation seem to be sustained over time as suggested by our 1-year follow-up data that show ejection fraction values unchanged from those measured at the 2-month posttransplant study point [26]. This long-term benefit could conceivably be related to the fatigue-resistance of engrafted fibers expressing increased proportions of slow-type myosin [19]. Finally although we have found that the posttransplant improvement in function was tightly related to the number of injected myoblasts [27] another study (which used cardiomyocytes and did not include a functional assessment) has failed to document an increase in graft size with increasing donor cell number [28]. Dose-escalation studies are clearly required to better characterize the relationship between the number of grafted cells and the functional outcome.
The mechanisms by which implanted myoblasts improve function have not yet been elucidated and at least three hypotheses, which are not mutually exclusive, can be put forward.
First, the elastic properties of implanted cells could act as a scaffold reinforcing the ventricular wall and subsequently limiting postinfarct scar expansion. However while such a mechanism may be operative when cells are injected shortly after the infarction and can thus prevent ventricular dilatation, it is less likely that myoblast-derived grafts can reverse an already completed remodeling process. This assumption is supported by the experimental finding that fetal cardiomyocyte transplantation fails to reverse left ventricular dilatation at a stage where it tends to improve indicators of systolic function [29]. These data are further strenghened by our clinical observation of unchanged end-diastolic volumes in patients having undergone late myoblast implantation in old infarcts.
The second hypothesis postulates a direct contribution of grafted cells to improved systolic function. Indeed several lines of evidence support this possibility: (1) the lack of connexin-43 does not necessarily preclude normal heart function as shown by studies of conditional knock-out mice for this junction protein [30]; in keeping with this observation, we previously reported [31] that both fetal cardiomyocytes, which express connexin-43, and skeletal myoblasts, which do not, improve postinfarction function to a similar extent; (2) experimentally both pressure-volume loops [26] and tissue Doppler imaging [18] have provided more direct evidence that engrafted skeletal myoblasts increase global and regional contractile function, respectively; (3) in our clinical phase I a new-onset systolic thickening was observed in 60% of the scarred segments that were implanted with myoblasts and autopsy of 1 patient who died late from a noncardiac cause revealed a remarkable preservation of the contractile apparatus of the in-scar embedded myotubes; and (4) a just completed study from our laboratory using epifluorescence microscopy provides direct evidence that intramyocardially engrafted skeletal myotubes have retained their contractile properties in that a strong depolarizing current can elicit action potentials followed by active contractions. However since the lack of connexin-43–supported gap junctions precludes a classic electromechanical coupling, the putative contribution of contracting myoblasts to the improvement of overall pump function should necessarily involve alternate mechanisms like stretch-induced contractions of transplanted cells in response to that of surrounding recipient cardiomyocytes or direct transmembrane channeling of electrical currents generated by these cardiomyocytes (field effects). So far, however, these mechanisms remain purely hypothetical and our electrophysiological studies have failed to demonstrate a synchronous coupling between grafted and host cells, with the caveat that the limitations of animal models make it extremely difficult to extrapolate these findings to the human heart. Nevertheless the observation that engrafted myoblasts are capable of actively contracting suggests that at least some of them are alive and it then becomes conceivable that they contribute to augment inotropism of the recipient heart by mechanisms independent of their own contractile activity, particularly release of pleiotrophic factors.
The third hypothesis is thus based on paracrine effects of the implanted myoblasts. Interestingly skeletal muscle tissue has been shown to express, among others, hepatocyte growth factor (HGF) [32], a pleiotrophic compound whose receptor, the c-Met protooncogene product, is widely expressed on different cell populations including those of the ischemic myocardium [33]. The cardioprotective effects of HGF have been established both in vitro [33] and in vivo [34]. Furthermore HGF has been reported to exert antifibrotic effects through stimulation of matrix-degrading pathways and inhibition of matrix-producing pathways [35]. Actually a marked reduction of fibrosis has been observed in a rat model of myocardial infarction in which cardiomyocyte transplantation was combined with direct intramyocardial administration of HGF and these histologic findings correlated with an improvement in cardiac performance and myocardial perfusion [36]. Insulin growth factor-1 is another compound that could mediate the paracrine effects of engrafted cells. Put together these data fit a paradigm where transplanted myoblasts would behave as platforms releasing factors that might rescue reversibly damaged periinfarct native cardiomyocytes or trigger recruitment and expansion of cardiac stem cells, thereby promoting a pool of new contractile elements [37]. Even if this paracrine mechanism is operative, it does not preclude the potential interest in strategies such as transfection of myoblasts with the gene encoding connexin 43 [38] as reestablishment of near-physiologic cell-to-cell communications could be an additional means of potentiating the functional benefits of the transplantation procedure.
Bone marrow stem cells
Transplantation of bone marrow cells is raising a growing interest because these cells share with myoblasts the possibility of being used as autografts and they could have the additional advantage of a transdifferentiation potential allowing them to convert into cardiac or endothelial cells or both. The reality is more complex, particularly because bone marrow is a mix of different cell populations whose distinct properties need to be analyzed separately.
Transplantation of total, unfractionated bone marrow is clinically appealing because of its apparent simplicity in that it entails aspiration of bone marrow from the iliac crest and, after removal of red blood cells, immediate reinjection of the aspirate into the postinfarction scar. This procedure has already been used in patients but so far, only some preliminary data have been reported in abstract form [39]. Although few details are available, the results seem to demonstrate the feasibility of the technique rather than its efficacy. Indeed our experiments in a sheep model of myocardial infarction have shown that extemporaneous injection of unpurified bone marrow into the scarred area failed to improve indicators of regional and global systolic and diastolic function (unpublished data). This correlated with the lack of differentiation of the injected cells into a "cardiac-like" tissue. In another study we have used a transgenic strain of mice that express the gene encoding ß-galactosidase under the control of a desmin promoter that allows an easy identification of the grafted cells, which turn blue if they acquire a muscular phenotype. In a model of global cardiomyopathy, we only found very few of these blue cells. Furthermore although these cells were morphologically close to native host cardiomyocytes, colocalization immunohistochemical experiments failed to conclusively establish that they expressed a highly cardiospecific marker (C protein). Put together these data suggest that the small percentage of multipotent cells present in bone marrow makes unlikely that injection of the unfractionated, unexpanded aspirate could yield an improvement of function. Interestingly Hamano and colleagues [40] have recently reported in a pig model of myocardial infarction that this approach was ineffective in improving function of the scarred area, a finding similar to ours; however, in their study, the marginal area, namely the area intermediate between normal and infarcted myocardium, was found to have increased angiogenesis and improved wall thickening. These data are consistent with the previous demonstration that cardiomyocyte conversion of endothelial cells requires cell-to-cell contact [41] and more generally emphasize the importance of delivering cells not only in the fibrous core of the scar but also in the border zones that harbor still viable myocardium.
The second option is to select well-defined populations of hematopoietic progenitors with the assumption that their plasticity would allow them to transdifferentiate in response to the environmental cues present in the target organ, and more specifically to convert into cardiac or endothelial cells or both after engraftment in the myocardium. This theory has been primarily advocated by Orlic and coworkers [42, 43] who reported "regeneration" of infarcted mouse myocardium by Lin-ckitPOS cells injected intramyocardially or endogenously translocated by cytokines. This regeneration was demonstrated by the formation of new progenitor-derived myocytes and vascular structures that translated into an improvement in function. However the design of these proof-of-concept experiments entails several features which seriously limit their clinical applicability. Thus in one study [42] intramyocardial injections were done a few hours after the infarct with the use of allogeneic cells. In another [43] mobilization of endogenous progenitors by a cytokine cocktail was started before the infarction (and continued thereafter) and in addition the animals were splenectomized before starting the cytokine treatment (to limit trapping of circulating cells). Notwithstanding these clinical caveats the above mentioned studies should be credited for having drawn attention to the influence of myocardium-specific cues on the developmental pathway of adult hematopoietic progenitors. Additional support to this concept has been provided by the study of Kocher and colleagues [44] showing that intravenous injection of human CD34+ cells expressing the phenotypic characteristics of embryonic hemangioblasts into athymic nude rats previously subjected to a myocardial infarction resulted in angiogenesis, decreased apoptosis in the periinfact region, and improved function. Again however cell delivery was performed early (48 hours) after the infarct, which is consistent with the persistence of local signals that may have driven circulating progenitors to home to sites of fresh ischemic injury. It remains to determine what would be the fate of this cell population after its late engraftment into an old fibrous scar as their sensitivity to local cues might actually lead to their differentiation into fibroblasts [45]. Indeed even if an appropriate milieu-induced transdifferentiation phenomenon occurs in border zones it still remains to define the most appropriate cell population and to subsequently develop clinically usable expansion methods as the basal percentage of these multipotent progenitors is extremely small (the CD34+ lineage represents approximately 1% to 2% of the total bone marrow). The critical importance of adequately addressing this scale-up issue is exemplified by the finding that in a mouse model of coronary occlusion-reperfusion intravenously injected with hematopoietic stem cells, donor-derived cardiomyocytes and endothelial cells only averaged 0.02% and 3.3% respectively [46]. Indeed the common embryologic origin of hematopoietic and endothelial cells suggests that upstream precursors of these lineages might be better suited for increasing angiogenesis [47] in ischemic patients, rather than for improving function of scarred areas in those with heart failure.
A third option is to use bone marrow mesenchymal cells, which are easy to collect and expand. In both rat [48] and swine [49] infarction models, these cells have been shown to differentiate into cardiac and blood vessel cells, which correlated with improved regional perfusion and wall motion, greater scar thickness, and augmented global heart function. However the prerequisite for stromal cells to acquire a myogenic phenotype is their pretreatment, during cultures, by 5-azacytidine, a demethylation agent that can cause an out-of-control upregulation of a wide variety of genes and as such raises clinically relevant safety concerns. Using a proprietary process that does not involve exposure to 5-azacytidine Shake and coworkers [50] recently reported in a swine model of occlusion-reperfusion that intramyocardially injected autologous mesenchymal cells attenuated regional systolic dysfunction without affecting ultimate infarct size. Importantly engrafted cells expressed a wide variety of myogenic markers, none of which was cardiospecific. In another study [51] where human mesenchymal cells were also cultured in the absence of 5-azacytidine before their injection into normal (noninfarcted) mice, cardiomyogenic markers were reported to be expressed by the grafted cells but the relevance of this result is hampered by the fact that only 0.44% of them were found to have survived after 4 days. In this setting a great deal of interest is being paid to a select subpopulation of bone marrow stromal cells recently identified by the group of Verfaillie and coworkers [52] in mice, rat, and human and which have been called multipotent adult pluripotent cells (MAPCs). This denomination clearly outlines that these cells, which can be expanded in culture for more than 80 population doublings without losing their pluripotency differentiation potential, can give rise to most somatic cell types. A major practical issue is that they cannot be sorted from the onset on the basis of specific phenotypic characteristics and are only identified retrospectively after several weeks in culture under tightly defined conditions, by their negative staining for several surface markers (CD34, CD 44, CD45, ckit, major histocompatibility complex class I and II). Furthermore when infused intravenously in postnatal animals they engraft and differentiate into several tissue-specific cell types but without any apparent contribution to cardiac muscle [52]. Additional studies in infarcted models are clearly warranted to better delineate the place, if any, of these cells in the context of bone marrow transplantation. It is also probably important that these studies do not exclusively use animals only injected with culture medium as controls since the already validated efficacy of skeletal myoblasts also makes this lineage a benchmark against which new cell types should be compared.
Route of cell delivery
So far, cell injections have usually been accomplished under direct control through multiple epicardial punctures. To reduce the invasiveness of the procedure, percutaneous approaches are undergoing a largely industry-driven extensive development. Much emphasis is put on endoventricular injections that benefit from improvements in catheter design and navigation systems. Surprisingly however the growing number of patients undergoing these procedures (often as life cases during well-attended meetings) sharply contrasts with the paucity of robust animal data showing (as has been the case for epicardial injections) that this "blind" approach is not only technically feasible but also functionally efficacious. A recent experimental study [53] has reported a higher intramyocardial retention of microspheres after endoventricular injections compared with epicardial injections but whether these results can be extrapolated to the use of cells remains uncertain. In the setting of these percutaneous techniques, the transvenous approach using a specifically dedicated coronary sinus catheter is particularly attractive because of its greater simplicity compared with the endoventricular route. Initial studies have established the effectiveness of bone marrow stem cell transvenous transfer into the myocardium [54] and our early pig and sheep data allow to extend these observations to the use of myoblasts. It now remains to assess the functional efficacy of this technique in infarct models. Finally, two groups in Germany have recently reported on intracoronary injections of bone marrow mononuclear cells concomitant with angioplasty at the acute stage of myocardial infarction [55, 56]. Their promising results need to be validated by randomized comparative efficacy trials.
Regardless of the route of delivery, cell death remains a major limitation of cell transplantation as up to 90% of cells may die shortly after injections and it is uncertain whether multiplication of those that have survived can catch up this high attrition rate. Several factors contribute to cell death including physical strain during injections, inflammation, apoptosis, and the hypoxic environment inherent in postinfarction scars. The importance of the latter factor is illustrated by the finding that the survival of cardiomyocytes grafted into highly vascularized granulation tissue is twofold higher than that observed after grafting into acutely necrotic myocardium [28]. As long as the proof-of-concept has not been clearly established in humans, we believe that cell-transplanted segments should not be concomitantly bypassed so as to limit confounding factors in the interpretation of functional outcomes. Should however the efficacy of the procedure be demonstrated by the forthcoming phase II trials, it then looks sound to try optimizing the benefits of cell transplantation by providing them with a vascular environment, whether obtained by surgical or catheter-based restoration of flow, coadministration of angiogenic factors [57], or engineering of cell so as to make them overexpressing some of these factors [58]. Apart from these antiischemic strategies pretreatment of cells with heat shock has also been shown to markedly increase survival of both cardiomyocytes [28] and skeletal myoblasts [59] after implantation into infarcted myocardium.
The timing of injection also possibly affects cell survival and there may be an optimal time window for cell transplantation. Too-early postinfarct injections may fail because of a high rate of cell death due to the infarct-induced inflammatory reaction. Late injections may be equally ineffective because of the previously mentioned inability of cell grafts to reverse the remodeling process once it has been completed [60].
Clinical trials
After several years of laboratory work, we initiated the first phase I human trial of autologous skeletal myoblast transplantation on June 15, 2000 [61]. Eligibility for inclusion in this trial was based on the following three criteria: (1) severe left ventricular dysfunction (ejection fraction ≤ 0.35); (2) history of myocardial infarct with a residual discrete, echocardiographically akinetic (after dobutamine challenge), and metabolically nonviable scar; and (3) indication for concomitant coronary artery bypass grafting in remote (ie, different from the transplanted area) ischemic myocardium.
The protocol involves three steps. First a biopsy of the vastus lateralis is retrieved from the thigh under local anesthesia. This muscle is then minced and grown for 2 to 3 weeks in the Cell Cultures Laboratory so as to obtain a highly purified, viable, and abundant cell yield (at least 400 x 106 cells and 50% myoblasts). Cells are subsequently reimplanted across the postinfarct scar while remote ischemic areas are revascularized by bypass grafts.
This hospital-sponsored phase I, which has included 10 patients, is now completed [62]. It has allowed to establish the feasibility of the procedure, as demonstrated by the ability to reach the target numbers of cells within the preset time frame (2 to 3 weeks). The operation by itself was shown to be safe, without specific procedure-related complications. The only adverse event that might be ascribed to cell transplantation is ventricular tachycardia, which has occurred in 4 patients. The mechanisms of these arrhythmias are being investigated. One possibility is that differences in action potential duration between engrafted myotubes and native cardiomyocytes may set the stage for reentries but this would imply cell-to-cell communications and subsequent propagation of electrical impulses, which has not been established yet. Indeed the relatively early postoperative onset of these arrhythmias may be rather consistent with the release of arrhythmogenic byproducts by the inflammatory cells invading the transplanted areas. Regardless of the mechanism, our data suggest that the incidence and severity of these arrhythmias could be reduced by an appropriate prophylaxis by amiodarone. At most, implantation of a defibrillator may be required, a strategy supported by the survival benefits expected from the device in this high-risk subset of heart failure patients as reported in the multicenter automatic defibrillator implantation trial (MADIT) II investigators [63]. While the small sample size (10 patients) of our trial, the lack of a control group, and the confounding effect of the concomitant revascularization preclude any definite conclusion, the finding that approximately 60% of the initially akinetic cell-implanted scar areas demonstrated a new postoperative systolic thickening is encouraging but needs to be validated by the forthcoming multicenter randomized phase II trial whose primary end point will be efficacy.
Two other phase I safety trials have been conducted in the United States. They are both industry-sponsored and have entailed delivery of autologous myoblasts through a surgical approach. One study was designed similar to ours as an adjunct to CABG cell transplantation; in the other, cells were planned to be grafted at the time of left ventricular assist device implantation used as a bridge with the ultimate objective of assessing the fate of grafted cells after the heart had been subsequently removed for transplantation. So far data from one of these trials have only been reported in an abstract form [64] and primarily allow to support the feasibility and safety of the procedure. Another trial involving endoventricular cell injections has been initiated in Europe but is currently stopped because of early postprocedural serious adverse events. Press releases rather than scientific peer-reviewed journals have also indicated that isolated cases had been performed in various locations but these "me-too" operations are unlikely to contribute to significant advances in the area.
Indeed it is hoped that the design of future studies will adhere to rigorous methodologic guidelines commonly used in drug trials as this is the only means of accurately assessing whether and to what extent cell transplantation can really impact on the outcome of patients with advanced postinfarction left ventricular dysfunction. In parallel with the numerous experimental studies that are under way, these clinical trials should also help answering critical and yet unsettled questions such as the optimal cell type, method of delivery, and cell survival–enhancing adjunctive strategies.
References
This article has been cited by other articles:
|
|
 |
|
  G. Forte, F. Carotenuto, F. Pagliari, S. Pagliari, P. Cossa, R. Fiaccavento, A. Ahluwalia, G. Vozzi, B. Vinci, A. Serafino, et al. Criticality of the Biological and Physical Stimuli Array Inducing Resident Cardiac Stem Cell Determination Stem Cells, August 1, 2008; 26(8): 2093 - 2103. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Siminiak, E. Meliga, O. Jerzykowska, and P. W. Serruys Percutaneous transplantation of skeletal myoblast in the treatment of post-infarction injury Eur. Heart J. Suppl., December 1, 2006; 8(suppl_H): H57 - H64. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Hata, G. Matsumiya, S. Miyagawa, H. Kondoh, N. Kawaguchi, N. Matsuura, T. Shimizu, T. Okano, H. Matsuda, and Y. Sawa Grafted skeletal myoblast sheets attenuate myocardial remodeling in pacing-induced canine heart failure model J. Thorac. Cardiovasc. Surg., October 1, 2006; 132(4): 918 - 924. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Davis, P. C. H. Hsieh, T. Takahashi, Q. Song, S. Zhang, R. D. Kamm, A. J. Grodzinsky, P. Anversa, and R. T. Lee Local myocardial insulin-like growth factor 1 (IGF-1) delivery with biotinylated peptide nanofibers improves cell therapy for myocardial infarction PNAS, May 23, 2006; 103(21): 8155 - 8160. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Patel, L. Geffner, R. F. Vina, J. Saslavsky, H. C. Urschel Jr, R. Kormos, and F. Benetti Surgical treatment for congestive heart failure with autologous adult stem cell transplantation: A prospective randomized study J. Thorac. Cardiovasc. Surg., December 1, 2005; 130(6): 1631 - 1638. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Borenstein, Z. Jian, G. Fromont, P. Bruneval, M. Hekmati, L. Behr, F. Laborde, D. Montarras, and E. Le Bret Noncultured cell transplantation in an ovine model of right ventricular preparation J. Thorac. Cardiovasc. Surg., May 1, 2005; 129(5): 1119 - 1127. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Narmoneva, R. Vukmirovic, M. E. Davis, R. D. Kamm, and R. T. Lee Endothelial Cells Promote Cardiac Myocyte Survival and Spatial Reorganization: Implications for Cardiac Regeneration Circulation, August 24, 2004; 110(8): 962 - 968. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. C. Ott, S. Berjukow, R. Marksteiner, E. Margreiter, G. Bock, G. Laufer, and S. Hering On the fate of skeletal myoblasts in a cardiac environment: down-regulation of voltage-gated ion channels J. Physiol., August 1, 2004; 558(3): 793 - 805. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
