Author contributions: J.W.B.: conception and design and manuscript writing; P.v.d.M.: conception and design, financial support, and manuscript writing; J.P.G.S.: financial support and manuscript writing; I.J.D.: conception and design, financial support, manuscript writing, and final approval of manuscript.
Advanced heart failure represents a leading public health problem in the developed world. The clinical syndrome results from the loss of viable and/or fully functional myocardial tissue. Designing new approaches to augment the number of functioning human cardiac muscle cells in the failing heart serve as the foundation of modern regenerative cardiovascular medicine. A number of clinical trials have been performed in an attempt to increase the number of functional myocardial cells by the transplantation of a diverse group of stem or progenitor cells. Although there are some encouraging suggestions of a small early therapeutic benefit, to date, no evidence for robust cell or tissue engraftment has been shown, emphasizing the need for new approaches. Clinically meaningful cardiac regeneration requires the identification of the optimum cardiogenic cell types and their assembly into mature myocardial tissue that is functionally and electrically coupled to the native myocardium. We here review recent advances in stem cell biology and tissue engineering and describe how the convergence of these two fields may yield novel approaches for cardiac regeneration. Stem Cells2013;31:2587–2598
Heart failure is a leading cause of death and hospitalization in the developed world [1-3]. Most commonly this supply/demand mismatch results from a loss of fully functional myocardial tissue and an inability of the heart to meet the metabolic demands . Current therapies of heart failure focus on symptomatic treatment of volume overload, prevention of ventricular remodeling, modulation of maladaptive neurohumoral responses, or device-based mechanical and electrical support . Of great significance, however, these therapies are not directly aimed at correcting the underlying pathophysiology of an inadequate number of normally organized functional myocardial cells. Cell-based therapy aimed at replacing or augmenting the number of functional myocardial cells therefore represents an attractive therapeutic approach for heart failure. For such a cell-based approach to be successful, several major hurdles will have to be overcome. The optimum cell type(s) will have to be purified and expanded to result in a sufficient number of mature cardiomyocytes for robust myocardial regeneration. These cells will have to be assembled into an effective three-dimensional pumping machinery. This grafted tissue will then have to be electrically and functionally integrated with native myocardium in order to be capable of significantly augmenting the cardiac output of the failing heart, without resulting in arrhythmias or rejection. In this review, we will explore the various stem cells populations thus far used in cardiac regeneration, the different tissue engineering approaches that have been used to assemble functional myocardial tissue, and the future work that lies ahead.
The Human Experience: Clinical Trials of Cell Therapy
After initial promising results of bone marrow stem cells therapy in animal studies, clinical trials in patients with acute myocardial infarction (MI) were initiated (Table 1). The first study, transplantation of progenitor cells and regeneration enhancement in acute myocardial infarction (AMI), was performed more than a decade ago. This phase-1 study allocated 20 patients with acute MI to receive either bone marrow-derived stem cells or circulating blood-derived progenitor cells into the infarct related artery . In this open-label, uncontrolled trial, left ventricular ejection fraction (LVEF) and myocardial viability in the infarct zone improved significantly in both groups. After these promising initial results, several mid-sized randomized studies demonstrated a modest but statistically significant improvement in LVEF in post-MI patients, including the BOOST and REPAIR-AMI trial [7, 8]. A post hoc subgroup analysis of the REPAIR-AMI trial showed that bone marrow stem cell therapy was most effective in patients with a clearly depressed left ventricular (LV) function, which might prevent adverse ventricular remodeling to some extend and improve quality of life. Unfortunately, 5-year follow-up of the BOOST trial revealed that the improvement in LVEF was transient . These early results were subsequently confirmed by several international trials that did not find a beneficial long-term effect of bone marrow-derived stem cell therapy, including the REGENT trial, ASTAMI, and the trial by Janssens et al. [10-12]. More recently yet, similar negative results were observed in the HEBE trial . In this multicenter trial, 200 patients with large first MI were randomized to mononuclear bone marrow cells, mononuclear peripheral blood cells, or standard medical therapy. After 4 months of follow-up, there was no difference in regional myocardial function as assessed by magnetic resonance imaging between the three different groups. In addition, two randomized phase-2 multicenter studies, performed by the cardiovascular cell therapy research network (CCTRN), did not find any beneficial effect of cell therapy in different patient groups and at various time points. These two double-blind placebo-controlled studies were geared at determining the optimum timing of BMMC coronary infusion after acute MI. The TIME trial compared intracoronary cell infusion at 3 and 7 days after acute MI to placebo, and the LateTIME trial compared bone marrow mononuclear cell (BMMNC) coronary infusion to cell-free carrier control infusion 2–3 weeks after acute MI. Neither of these two phase-2 CCTRN studies showed any beneficial effect of BMMC infusion on either global or regional LV-function [14, 15].
Table 1. Selected large scale (patients [mt] 50) randomized controlled trials of intracoronary bone marrow-derived cell infusion after myocardial infarction
No. of patients
Effect (BM vs. control)
Abbreviations: BMC, bone marrow cells; f/u, follow-up; LVEF, left ventricular ejection fraction; MI, myocardial infarction.
BMC injection vs. no therapy 5 days after acute MI
LVEF + 6% (p < .05) in BMC group at 6 months, no significant difference after 5 years f/u
Whereas the above-mentioned trials were performed in acute MI, Strauer et al. performed a study in patients with chronic heart failure . Three hundred and ninety-one patients with an ischemic etiology of heart failure were included, and 191 of these patients received an intracoronary bone marrow cell infusion. Patients treated with bone marrow stem cells had improved LV-function and reduced mortality compared to untreated controls. Although the study is the largest of its kind, it needs to be emphasized that the study was open-label and was not randomized: Patients who refused to undergo bone marrow cell infusion were included in the control group. This could well have lead to a significant patient-selection bias and therefore severely limits the value of this trial. Furthermore, the FOCUS-CCTRN study explored transendocardial delivery of BMMNC in patients with chronic ischemic heart disease and LV-dysfunction who had no revascularization options. In this double-blinded placebo-controlled study, transendocardial BMMNC injections were compared to injections of a cell-free substrate. BMMNCs did not improve myocardial perfusion, maximal oxygen consumption, or LV end-systolic diameter compared to controls .
A recent meta-analysis included 2,625 patients with acute MI or ischemic heart disease from 50 trials . This study estimated the mean weighted differences for changes in LV ejection fraction, infarct size, LV end-systolic volume, and LV end-diastolic volume in all individuals. Taking together, the selected studies in this meta-analysis showed a modest improvement in LV ejection fraction (3.96%), smaller infarct size (4.03%), and decreased end-systolic and end-diastolic volumes, in patients treated with BMC therapy compared to untreated controls.
Even this modest improvement, however, is undermined by the fact that the largest published trails to date, the REGENT, HEBE, and TIME did not find a beneficial effect of bone marrow-derived stem cell therapy and were not included in this meta-analysis. Thus, studies bone marrow-derived stem cells have shown at best ambiguous results with no evidence of engraftment of transplanted cells.
As a result of these findings, it has been proposed that transplanted bone marrow-derived stem cells engraft transiently while exerting paracrine effects and releasing growth factors promoting angiogenesis, modulating the immune-response, and stimulate endogenous cardiac repair [19, 20]. This hypothesis is supported by a study in a large animal model of acute MI that showed that the injection of bone marrow derived stem cells increased capillary density and improved collateral perfusion and regional cardiac function . Furthermore, fate-mapping studies in transgenic mouse models showed that after MI endogenous stem cells refresh adult cardiomyocytes . This endogenous repair mechanism of the mammalian heart could be enhanced through injection of bone marrow-derived c-kit+ cells  or cardiac-derived cells . Despite these disappointing initial results, bone marrow cells remain the safest and source of human cells with the largest clinical experience . As a result, several trials including BAMI and RENEW are now entering phase-3.
The finding that paracrine factors secreted by transiently engrafted cells mediate most (if not all) of the clinical benefit has generated new interest in the use of allogenic cell sources for cardiac regeneration. Despite the promise of this approach, the specific cytokines and growth factors secreted by transplanted cells and their mechanism of action remain unknown. Collectively, these studies highlight the fact that future research should be aimed at developing a mechanistic understanding of any putative beneficial effects of bone marrow-derived stem cell therapy for the treatment of heart failure. Future advances in this field could cover the delivery of genetically or pharmacologically modified bone marrow cells or the direct application of a cell-derived cocktail of paracrine and/or growth factors or exosomes , and therefore facilitate the use of these cell types in cardiac regenerative medicine.
Mesenchymal Stem Cell Therapy
Mesenchymal stem cells (MSCs) are multipotent stromal cells with a predominantly mesodermal origin. MSCs possess the capacity for self-renewal and the potential to phenotypic adopt a spectrum of different somatic cell types. Most of the reported in vitro data on MSC differentiation into cardiac-like cell types relied on the exposure to the DNA demethylating chemical 5-azacytidine. The endogenous roles described for MSC are maintenance of the hematopoietic stem cell niche, organ homeostasis, wound healing, and aging. Therefore, bone marrow-derived MSCs appeared as an appealing cell source for cardiac regeneration [27-29]. In the past decade, a number of phase-1/2 trials were introduced. In an open-label randomized trial, 96 patients were included who underwent primary percutaneous coronary intervention within 12 hours after acute MI. Patients received either autologous bone marrow-derived MSCs or saline intracoronary infusion. LV ejection fraction was significantly improved by 8% in patients treated with MSCs compared to saline infusion . To investigate the safety and efficacy of intravenous allogeneic MSCs after acute MI, a double-blinded placebo-controlled trial was set up. Up to 6 months no difference in adverse cardiac events was observed, and renal, hepatic, and hematologic laboratory indexes were similar between groups. Additionally, MSC-treated patients showed a 6% increase in ejection fraction at 3 months . The percutaneous stem cell injection delivery effects on neomyogenesis (POSEIDON) trial aimed to compare the difference between transendocardial injection of autologous and allogeneic MSCs, in patients with ischemic cardiomyopathy. The first analyses revealed that both autologous and allogenic MSC treatment have low rates of serious adverse events . Furthermore, the mesenchymal stromal cells in chronic ischemic heart failure (MSC-HF trial) and transendocardial injection of autologous human cells (bone marrow or mesenchymal) in chronic ischemic left ventricular dysfunction and heart failure secondary to MI (TAC-HFT) trials are underway to examine the efficacy of MSC therapy in patients with chronic heart failure.
Cardiac Progenitor Cell Therapy
Work from multiple laboratories has shown that resident cardiac progenitor cells (CPCs) (c-kit+, Sca-1+) isolated from adult heart auricles or heart biopsies can become functional cardiomyocytes through coculture or epigenetic manipulation (with 5-Azacitidine) followed by TGF-β stimulation [33-35]. This opened up the strategy of autologous CPC-based therapy for the injured heart. The CADUCEUS (CArdiosphere-Derived aUtologous stem CElls to reverse ventricUlar dysfunction) study was designed to evaluate cardiac stem cell infusions in patients after acute MI. In the treatment group, infarct-size-reduction and an improved regional function were observed up to 1-year after treatment . Similar beneficial effects were found in the initial analysis of the still running (only the 1-year results have been published up-to-date) SCIPIO (stem cell infusion in patients with ischemic cardiomyopathy) trial. In this randomized open study design, 20 patients received coronary infusion of one million CPCs 4 months after CABG. In the treatment group, infarct-size-reduction and an increased LVEF were observed up to 1-year after treatment [37, 38].
Currently, the AutoLogous Human CArdiac-Derived stem cell to treat Ischemic cArdiomyopathy (ALCADIA) study is aiming to evaluate the safety and efficacy on the transplantation of autologous human cardiac-derived stem cells with the controlled release of basic fibroblast growth factor (bFGF) in patients with severe refractory heart failure. The first results of the ALCALDIA trial are expected later this year. Furthermore, the Allogeneic Heart Stem Cells to Achieve Myocardial Regeneration (ALLSTAR) trial will be the first to examine the safety of cardiac-derived cell therapy in a phase II trial. Thus, although the early results show the promise of autologous CPC infusion therapy, the long-term benefits remain uncertain. Furthermore, it remains unclear if the observed improvements in cardiac function are due to true regeneration or due to cardio-protective or paracrine effects as a result of transient engraftment (as earlier discussed for bone marrow cells).
Potential of Skeletal Myoblast in Cell Therapy
Skeletal muscle has the intrinsic capacity of repair and regeneration via the proliferation and differentiation of skeletal myoblasts. Since this tissue can be acquired in an autologous manner, it lacks the risk of immunological rejection [39, 40]. Although, early trials demonstrated an improvement in LV-function as well as an improvement of symptoms [41-44], this effect was not observed in the follow-up double-blinded placebo-controlled trial (MAGIC trial). In contrast, there was evidence of early postoperative arrhythmic events and poor engraftment and/or survival of injected cells . Thus, despite the theoretical appeal of skeletal myoblasts as a cell source for regenerative myocardial therapy, the clinical outcomes severely limit their applicability [39, 40].
From Embryos to Embryonic Stem Cells: Lessons From Development
Understanding the normal cellular differentiation of cardiac progenitors and the assembly of their differentiated progeny into the highly specialized three-dimensional structure of the four-chambered mammalian heart will likely represent a key milestone in the quest for genuine cardiac regeneration. In this regard, recent studies have uncovered a series of early heart progenitors that bring about the generation of the diverse cell types that constitute the mature mammalian heart (Fig. 1).
The early heart arises from primordial cardiogenic mesoderm progenitors that express the related transcription factors Mesp1/2. These progenitors give rise to multipotent progenitors in the first-heart-field (FHF) and the second-heart-field (SHF), a closely related set of cells in the developing mesoderm [46-48]. The FHF arises from the anterior lateral mesoderm and forms the cardiac crescent in early embryogenesis. Later in development, these cells coalesce into the linear heart tube and ultimately give rise to the left ventricle of the mature four-chambered mammalian heart. Progenitors of the SHF arise in the pharyngeal and splanchnic mesoderm and subsequently migrate into the developing heart to give rise to the right ventricle, the outflow tract (OFT), and portions of the inflow tract. The LIM-homeodomain transcription factor Islet1 (Isl1) appears to mark these early SHF cardiac progenitors and has shown to be functionally important for their specification and differentiation [49-51]. The early progenitors of the FHF, however, appear to be marked by the T-box transcription factor Tbx5 [52-54]. In contrast to murine cardiogenesis, human cardiac development occurs over a considerably longer period of time and requires several orders of magnitude more cells. Accordingly, the period of cardiac progenitor expansion and lineage diversification must also be extended. Nonetheless, recent work has identified human multipotent fetal Isl1-positive cardiovascular progenitor populations that give rise to the cardiomyocyte, smooth muscle, and endothelial lineages in the developing heart. A subset of these early cardiac progenitor populations, expressing Isl1, are localized to the fetal atria and proximal/medial OFT and give rise to lineage-restricted intermediates marked by Isl1 in combination with Nkx2–5 or other cardiac differentiation markers .
Non-cell Autonomous Cues in Embryonic Development
The early cardiac progenitors of the developing heart are subject to a continuously changing microenvironment within the developing embryo. Changes in the activation of cell signaling pathways, cell migration as well as changes of the three-dimensional architecture during the morphogenesis of the primitive heart form the dynamics. The developing cardiac progenitor populations are therefore exposed to significant temporally and spatially regulated non-cell autonomous signaling molecules. During the later cardiac development, the robust expansion of cardiac progenitors is critical for the ventricular chamber to achieve sufficient muscle mass necessary for the mechanical work to maintain adequate blood flow in the exponentially growing embryo. This increase in muscle mass has to be coordinated between the different chambers of the mammalian heart in order to avoid congenital malformations. Recent advances have suggested that canonical Wnt signaling appears to play a critical role in the expansion of Isl1+ SHF cardiac progenitors and their differentiated progeny [56, 57]. Gain of function mutations of β-catenin that result in the constitutive activation of canonical Wnt signaling in E8.5–9.5 SHF progenitors results in robust expansion and inhibition of differentiation of these cell populations both in vitro and in vivo [57-60]. Interestingly, recent work now also suggests that Hippo/Yap signaling, which is well recognized to control organ size in Drosophila , may control mammalian heart size by suppressing Wnt signaling and thereby restricting cardiomyocyte proliferation [62-64]. Similarly, stage-specific expression of the Activin A and BMP4 signaling proteins appears to be critical in mesodermal and cardiac differentiation during normal in vivo cardiac development as well as in vitro cellular differentiation [65-68]. Thus, defining the cell autonomous and non-cell autonomous cues that control cardiac progenitor expansion and differentiation will be critical to understanding normal development and generating the cell populations necessary for regenerative cardiovascular medicine [60, 62, 65, 66, 68].
Proliferation of Postnatal Cardiac Myocytes
Annual cardiomyocyte turnover in the mammalian heart was estimated between 1% and 4% and predominantly occurs through renewal of pre-existing cardiac myocytes . Unlike the adult mammalian heart, certain fish, and amphibians retain the capacity to regenerate from cardiac damage throughout life . And while the adult mammalian myocardium has almost no capacity to refresh the cardiac myocytes lost after injury, it was shown that the early neonatal myocardium has an intrinsic capacity to regenerate. When hearts of 1-day-old mice were cryo-injured at the apex, a regenerative response was characterized by proliferation of pre-existing cardiomyocytes resulting in minimal formation of fibrotic tissue. This capacity of cardiomyocytes (as opposed to progenitor cells) to proliferate is lost in the first week after birth. This neonatal intrinsic cardiomyocyte regenerative response is similar to the sustained regenerative capacity of zebrafish hearts [71, 72].
Pluripotent Embryonic Stem Cells as Model Systems for Cardiac Development
The capacity of embryonic stem cell (ESC) lines to differentiate in vitro and to recapitulate many of the in vivo developmental programs provides an attractive model system for studying lineage commitment (Fig. 1). Significantly, ESC in vitro differentiation can be scaled up to generate large numbers of cardiac progenitors or myocytes.
A major challenge for the use of pluripotent stem cells in cardiac tissue regeneration has been the isolation of large numbers of cardiac progenitors from the heterogeneous products of ESC in vitro differentiation, robustly controlling cardiac lineage differentiation and purifying of differentiated cardiomyocytes. Recently, purified populations of cells that express Isl1 and the VEGF receptor (KDR) have been shown to give rise to endothelial, smooth muscle, and cardiac muscle cells of the SHF [50, 66, 73]. Similarly, the cardiac-specific Nkx2.5 enhancer has been used to isolate bipotential progenitors from murine embryos as well as murine ESCs , and the mesoderm marker Brachyury T in combination with a VEGF receptor cell surface marker (Flk1/KDR) was used to successfully isolate multipotent cardiac progenitors from human and murine ESCs [65, 66].
Unlike Isl1+ progenitors, both Nkx2.5+ and Flk1+ progenitors likely represent a heterogeneous mix of FHF and SHF progenitors. In order to distinguish between these two progenitor populations, the cardiac-specific enhancer of the Nkx2.5 gene driving the expression of green fluorescent protein along with a SHF-specific enhancer of the Mef2C gene driving the expression of red fluorescent protein (dsRed) was used to generate a double color murine transgenic system . This allowed the isolation of purified populations of FHF and SHF cardiac progenitors. A subset of these SHF progenitors appeared to be completely committed to the ventricular myogenic cell fate and at the same time be capable of limited expansion prior to differentiation.
Current understanding of the molecular pathways that control in vivo cardiogenesis has facilitated the development of several different approaches for the in vitro-directed differentiation of pluripotent stem cells toward the cardiac cell fate. These approaches have largely relied on varying the timing of signaling proteins and small molecules to yield higher numbers of cardiac lineage committed cells. The Wnt/β-catenin signaling pathway, for example, has a biphasic role during differentiation. In the early developmental stage, activation of Wnt signaling, via Wnt3a and GSK-3-inhibition, and in the later stage downregulation, via Dkk1, are associated with enhanced cardiac differentiation . Similarly, a recent report demonstrates that stage-specific Activin A/Nodal and BMP4 signaling directs early cardiac mesoderm specification and late cardiac myocyte differentiation . Related directed differentiation approaches have relied on small noncoding regulatory micro-RNAs (miRNAs). In mouse and human, miR-1 and −133 have been shown to promote mesoderm formation from ESC, but have opposing functions during further differentiation into cardiac muscle progenitors . Further cardiomyocyte differentiation of ESCs or cardiac-derived stem cells can be improved through overexpression of miR-1 and miR-499 [77, 78]. Despite these advances in promoting cardiac differentiation from ESCs, cell line-specific optimization of concentrations and timing of signaling molecules remain essential for successful directed differentiation. In addition to ESC-derivates, the isolation of parthenogenetic (uniparental parthernodes) stem cells (PSCs) derived from murine oocytes forms a potential source of autologous pluripotent stem cells [79, 80]. In a recent report, oocyte-derived PSCs were efficiently differentiated into beating cardiomyocytes, and subsequently applied in preclinical cardiac regenerative approaches . Future work should aim to validate ESC findings in human PSCs.
Induced Pluripotent Stem Cells and Direct Reprogramming
Until recently, the derivation of a pluripotent stem cell line was thought to require the use of discarded embryos from fertility treatment. Remarkably, introduction of only four factors (Oct3/4, Sox2, Klf4, and c-Myc) into mouse and human embryonic fibroblasts successfully generated induced pluripotent stem cell (iPSC) lines that showed features similar to normal ESC lines (Fig. 2) [82, 83]. In a similar approaches, it has now also been possible to generate iPSCs by the expression of specific noncoding regulatory miRNAs , synthetic modified mRNAs , purified recombinant proteins , whole-cell extracts from ESCs , or from genetically engineered HEK293 cells . While these types of approaches represent an attractive option for the generation of transgene-free iPSCs, their efficiency remains low. Taken together, these studies point to a novel and evolving approach of generating patient-specific iPSC lines.
The postnatal human heart contains a large pool of fibroblasts. The reprogramming of these cells into functional cardiac myocytes that are electrically and functionally coupled to the normal heart represents an attractive approach for cardiac regeneration. Recent reports have in fact demonstrated the direct reprogramming of mouse somatic cells into ventricular myocytes in vitro  and in vivo [90, 91], with the overexpression of a group of cardiac transcription factors (Gata4, Mef2c, Tbx5, and/or Hand2) (Fig. 2). Although this early work clearly shows the induction of functional ventricular myocytes from fibroblasts, it remains an extremely rare event . Remarkable, this efficiency is markedly enhanced when direct reprogramming occurs in vivo, raising the possibility for clinically useful cardiac regeneration.
ESC in Cardiac Regeneration: the Early Animal Experience
The ability of ESC lines to differentiate in vitro into cardiac lineages raised the question of whether these cells and their derivates may be suitable for repairing the injured heart. To assess this possibility, undifferentiated mouse ESCs were transplanted into the hearts of adult mice. ESC transplants did not lead to the generation or engraftment of ESC-derived cardiomyocytes but did lead to teratoma formation within 4 weeks, highlighting an important potential pitfall for the use of pluripotent stem cells in regenerative cardiovascular medicine [93-96]. Moreover, the syngenic but not the allogenic mouse ESC transplants were detectable at 8 weeks post-transplantation suggesting that innate host immunity resulted in the rejection of allogenic mouse ESC and their differentiated progeny .
This early work underscored several requirements for successful cardiac regeneration: transplanted cells must be able to survive and differentiate into clinically relevant cell types in vivo, they must be able to undergo electromechanical coupling with the host myocardium, they must lead to measurable functional improvement in myocardial function, and they must not result in teratoma formation, arrhythmias, or other adverse clinical outcomes. Given that undifferentiated ESC satisfied none of these requirements, follow-up experiments explored the use of cardiac-enriched differentiated human ESC. Work from several laboratories demonstrated that the direct intramyocardial injection of such cardiac-enriched human ESC derivatives into rodent hearts resulted in the generation of human ESC-derived myocardial grafts [97-100]. However, injection of ESC derivates only transiently improved LV-function .
More recent progress has allowed for the isolation of a highly purified population of human cardiomyocytes from human ESC via directed-differentiation and an improvement of the survival of these cells by a prosurvival cocktail that limits cardiomyocyte death after transplantation . These improvements allowed for the consistent formation of human ESC-derived myocardial grafts in infarcted mouse and rat hearts. The engrafted human myocardium attenuated ventricular dilation and preserved regional and global contractile function after MI compared with controls. Thus, human ESC-derived cardiomyocytes appear to partially remuscularize myocardial infarcts and augment heart muscle in rodent model systems [100-102]. It remains unclear, however, if engrafted ESC-derived cardiomyocytes cause adverse arrhythmic side effects, since both antiarrhythmic  and proarrhythmic  effects were described for mouse cardiac myocyte grafts in injured mouse hearts. Recent work has demonstrated that human ESC-derived cardiomyocytes engraft in a guinea pig MI model, electromechanically integrate with host myocardium, contract synchronously with host hearts, and significantly reduce the incidence of spontaneous and induced ventricular tachycardia . This shows the potential for transplantation of human cardiomyocytes derived from a renewable cell source.
As noted above, clinically meaningful engraftment of transplanted myocardial tissue requires proper electromechanical coupling with the host heart. This in turn requires the formation of the intercalated discs to allow the electrophysiological coupling of transplanted cardiomyocytes to host cardiomyocytes. During normal development, postnatal remodeling and adaptation of the ventricular myocardium to the hemodynamic changes lead to localization of the gap-junctions to the transverse terminals of the cells. This allows for the rapid spread of action potentials necessary for normal cardiac conduction. Adherens-junctions anchor sarcomeric actin filaments between myocytes and are required for correct gap-junction formation [106, 107]. Cadherins are an important functional component of these junctions. Transplanted ESC-derived cardiomyocytes did show a pattern of diffuse N-Cadherin (adherens-junction) expression but no Connexin-43 (gap-junction) expression [97, 100].
The in vivo differentiation of human ESC-derived myocytes into engrafted tissue points to the promise of pluripotent stem cells for regenerative medicine. Nonetheless, the existing body of literature does not support the hypothesis that the direct intramyocardial injection of cardiogenic cells will result in sarcomeric aligned, fully electromechanically coupled tissue that can result in long-term improvement in myocardial function. Indeed, current evidence would seem to suggest that augmenting the advances in cell biology with tissue engineering technologies may be necessary to meet these functional requirements.
Generating Myocardium In Vitro: Engineering Cardiac Tissue
Even optimum cardiogenic cell populations are unlikely to result in mature functional myocardial tissue in the absence of appropriate cellular and sarcomeric alignment. It will therefore be important to define the three-dimensional structure to guide cell growth, differentiation, and maturation . While reconstructing the entire four-chambered mammalian heart may be difficult, recent advances raise the possibility of engineering-specific cardiac components such as ventricular myocardium, heart valves, pacemaker conduction systems, and coronary vasculature. In the future, these heart parts could be used as replacement parts for diseased hearts. Coupled with advances in delivery systems, such a piecemeal approach may provide an alternative to direct cell transplantation .
Sources of Cardiomyocytes for Cell and Tissue Engineering Applications
To date, most attempts at engineering ventricular myocardial tissue have relied on rat neonatal cardiomyocytes as a cell source (Table 2). This is largely due to the wide availability of these cells, and to their capacity to proliferate and differentiate into functional cardiac myocytes ex vivo . In the absence of external stimuli, however, the in vitro organizational capacity of neonatal rat cardiomyocytes is limited and does not result in the sarcomeric alignment nor in myofibrillar organization like in myocardium of the native heart . While useful for developing novel tissue engineering-based platforms for cardiac regeneration, rodent cardiac myocytes will not have a role in cardiac regeneration in humans. It will therefore be necessary to adapt tissue-engineering advances from rodent neonatal cardiac myocytes to cardiac myocytes derived from renewable human patient-specific cell sources.
Table 2. Recent in vitro strategies in cardiac cell and tissue engineering
In vivo application
Abundance of cytoplasmic troponin T and ß-myosin heavy chain protein.
Same technique used for the cell layers as in the cell sheet approach.
The emergence of microfabrication and micropatterning techniques in the early 1990s allowed for novel cell engineering approaches to investigate of the interaction between cell shape and function [135-137]. Photolithography was initially used to generate cell culture surfaces with alternating lines of high and low cell-adhesion potential . This approach was refined by microcontact printing techniques that allowed extracellular matrix (ECM) proteins to be imprinted on a polydimethylsiloxane (PDMS)-coated surface. This allowed for electromechanical coupling between cardiac myocytes and the generation of anisotropic cardiac tissue [111, 112].
In a powerful proof-of-principle series of experiments, it was demonstrated that it is possible to manipulate cell shape and two-dimensional myofibrillar organization of rat cardiomyocytes by culturing them on PDMS thin films micropatterned with fibronectin in specific shapes such as triangles, rectangles, and star shapes. These constructs, called muscular thin films, adopted functional, three-dimensional conformations when released from the temperature responsive polymer poly(N-isopropylacrylamide) (PIPAAm) . This work demonstrated that there is a direct correlation between sarcomeric alignment, force generation, and work performed.
As discussed above, recent advances have allowed for the isolation of so-called mouse ESC-derived ventricular progenitors . When the ventricular progenitors were cultured on micropatterned lines of fibronectin, their differentiated progeny aligned spontaneously to form anisotropic two-dimensional myocardial tissue with uniaxial sarcomeric alignment . Of great importance is the finding that the generation of force generating two-dimensional myocardial tissue was only possible from cells that were completely committed to the ventricular cell fate, underscoring the importance of isolating highly purified populations of ventricular progenitors for the generation of tissue engineering constructs .
Although micropatterning approaches have been used to promote cellular alignment and generate two-dimensional muscular constructs, they have not yet been successfully used to generate the three-dimensional constructs required for adequate force generation. In order to overcome this barrier, a number of laboratories have attempted to use biodegradable scaffolds to generate three-dimensional cardiac tissue [139-141]. Early efforts focused on the generation of disk-shaped poly(lactic-co-glycolic) acid (PLGA) scaffolds that could serve as a three-dimensional substrate for cell adhesion in vitro and, at the same time, could be gradually absorbed in vivo [115, 116]. PLGA is metabolized into glycolic acid monomers and these have been shown to stimulate collagen synthesis [142, 143]. PLGA scaffolds were coated with ECM proteins such as laminin , gelatin [118, 119], and collagen  and seeded with rat neonatal cardiac myocytes. This resulted in three-dimensional multilayered (1–5 mm thickness) cardiomyocyte tissue within 1 week of implantation of [115, 116, 120]. Electrophysiological studies revealed local functional coupling and alignment by cellularity, conduction velocity, signal amplitude, capture rate, and excitation threshold measurements of PLGA scaffold based constructs . When electrically stimulated for extended periods of time, the PLGA constructs had improved tissue morphology and contractile function compared to nonstimulated controls . Subcutaneous implanted engineered tissue survived and formed beating vascularized grafts. Engraftment upon 3-week-old cardiac cryoinjury scars showed a 5-week survival of patches .
In a similar series of experiments, an accordion-like honeycomb scaffold was designed to replicate the structural and mechanical properties of the native heart . A specific pore microarchitecture design of the scaffold generated directionally dependent structural and mechanical properties. The mechanical properties of these scaffolds were shown to closely match those of native heart tissue: the scaffold was stiffer when stretched circumferentially as compared to longitudinally [122, 139, 140]. Although the scaffold provided mechanical anisotropy, it resulted in more isolated compartments of myocytes that were not electrically coupled over the whole construct. In addition, the bioreactor culture conditions did provide improved oxygenation, but the absence of vasculature of the new tissue dictated an upper limit on the thickness of the three-dimensional tissue [120, 144].
Further steps in the field were made with the so-called “cotton-candy technique.” Sucrose was melted in a cotton candy machine, spun out to an optimum fiber length and thickness, and coated with a fibronectin-PLGA polymer. Cardiac myocytes cultured on this substrate aligned into unidirectional three-dimensional tissue with macroscopic anisotropy .
In an effort to fabricate three-dimensional pulsatile cardiac grafts without the use of scaffolds, a novel technology that layers cell sheets three-dimensionally was developed. Rat cardiomyocytes were cultured on a polystyrene surface coated with the temperature-responsive polymer PIPAAm. When the cultured cells had coalesced into a confluent layer, the temperature was reduced thereby detaching the cardiac myocytes as a thin cell sheet without the scaffold backbone. Multiple sheets were then overlaid to generate three-dimensional heart muscle constructs [125, 145]. These constructs became electrically coupled via connexin-43 and began to pulse simultaneously [126, 127, 146]. Although conceptually appealing, at this point it remains unclear as to whether these constructs can be engineered to replicate the three-dimensional organization of the native heart.
An alternate approach for the fabrication of myocardial tissue consists of culturing cardiac myocytes in hydrogels consisting of ECM proteins to provide the three-dimensional environment for cell growth. Rat neonatal cardiomyocytes were cultured in circular hydrogel molds to generate spontaneously contracting muscular rings that could be attached directly to mechanical force transducers [128, 147]. Interestingly, cyclical stretch of these engineered heart muscle rings resulted in myocyte hypertrophy and increased force generation [128, 129, 131, 148-150]. Similarly, when electrically stimulated for extended periods of time cardiac muscle constructs had improved tissue morphology and contractile function compared to nonstimulated controls . Thus, it appears that both electric stimulation and mechanical stretch may be able to improve cardiac myocyte function although at this point it is unclear whether this occurs through the same or related pathways.
Perhaps the most ambitious approach to date involves the generation of a bioartificial hearts. This approach has the tremendous appeal of capitalizing on the intrinsic three-dimensional architecture of the native heart . Rat hearts were decellularized by the infusion of the coronary vasculature with detergents resulting in an acellular perfusable natural scaffold for cell culture. These natural scaffolds were then reseeded with rat neonatal heart cells and within 1-week after reseeding, macroscopic contractions were visible, and constructs could generate force equivalent to approximately 2% of adult hearts. An important limitation of this approach was the failure of the reseeded cardiomyocytes to completely recapitulate the micro architecture of the native heart . Nonetheless, the successful decellularization of porcine hearts raises the intriguing possibility that such an approach may yet be adapted to the treatment of human disease [151, 152].
Human Cardiac Myocytes in Tissue Engineering
Until recently, the use of human cardiac myocytes to generate tissue engineered myocardial tissue has been hindered by the lack of renewable sources of human cardiomyocytes. Initial attempts focused on the generation of synchronously contracting myocardial tissue from human ESCs containing endothelial vessel networks . In a related series of experiments, human ESCs were differentiated to cardiomyocytes and cultured with endothelial cells, and fibroblasts in a rotating orbital shaker to create vascularized human cardiac tissue patches. Implantation of these patches resulted in better cell grafts compared with patches composed only of cardiomyocytes . When cardiac-derived stem cells were combined with bioprinting technology it increased expression of the cardiac genes in the three-dimensional macroscopic generated tissue . In addition, when cultured in a three-dimensional collagen matrix, uniaxial mechanical stress conditioning promoted myofibrillogenesis, sarcomeric banding, and cardiomyocyte matrix fiber alignment. Furthermore, the addition of endothelial cells markedly increased proliferation of myocytes by 20%. Thus, controlling the mechanical load as well as vascular cell contact will be necessary for engineering human myocardium .
In order to promote three-dimensional cardiomyocyte alignment microtemplating techniques were used to micropattern a poly(2-hydroxyethyl methacrylate-co-methacrylic acid)-based hydrogel to generate tissue-engineering scaffolds. The constructs contained parallel channels designed to augment the formation of aligned cardiomyocyte bundles, supported by spherical interconnected pores intended to enhance angiogenesis and reduce scar formation. Cardiomyocytes survived for approximately 2 weeks in vitro on these scaffolds and proliferated to approximate the cell densities of the adult heart. Implantation of acellular pore scaffolds resulted in angiogenesis as well as a reduced fibrotic response . The advances made in human cardiac tissue engineering illustrate the potential for novel cardiac regenerative therapies. Future work should attempt to unravel what cues are necessary to design heart tissue meeting the functional properties and electromechanical requirements for permanent engraftment in native myocardium.
From Here to There: Combining Stem Cell Biology and Tissue Engineering Technology
Thus, despite nearly a decade of intensive research and billions of dollars of public money, at this point in time there is no clinically accepted therapy aimed at replacing the loss of functional cardiac myocytes and repairing failing myocardial tissue. Many hurdles remain, including the isolation of committed ventricular progenitors from patient-specific cell sources, the differentiation, and assembly of these progenitors into three-dimensional myocardial tissue with sufficient mass to achieve hemodynamically significant force generation, the production of an adequate blood supply for this tissue, and the electrical coupling of the tissue to the native myocardium (Fig. 3). Nonetheless, the recent advances in stem cell biology when coupled to the remarkable progress in tissue engineering technology point to a powerful emerging approach for regenerative cardiovascular medicine.
We acknowledge financial support from NIH/NHLBI U01HL100408-01 to I.J.D. and Netherlands Organization of Scientific Research, VENI Grant 016106013 to P.v.d.M.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.