Introduction

Heart failure, most often subsequent to ischemic heart disease, is a major cause of morbidity and mortality worldwide . After a large myocardial infarction, more than a billion cardiomyocytes are lost , and although substantial advances have been made in medical treatment improving the prognosis of these patients , none of the current therapeutic approaches directly targets the loss of cardiomyocytes, the underlying cause of heart failure.

For decades, the heart was viewed as a terminally differentiated organ lacking the regenerative capacity sufficient to replace the loss of myocytes following injury. The discovery of endogenous cardiac progenitor cells and reports demonstrating a low turnover of existing cardiomyocytes have changed this view to some extent . Whether these new cardiomyocytes are derived from the proliferation of existing cardiomyocytes or from cardiac progenitor cells is still under discussion.

First-generation adult stem cell therapies, consisting mainly of bone marrow mononuclear cells, and lately even stem cell growth factor receptor (c-kit)-positive cardiac cells expanded from heart biopsies, have been studied extensively for the improvement of cardiac function after myocardial injury . To date, however, only minor effects, if any, have been achieved through this approach , possibly attributed to paracrine mechanisms without the evidence of replacement for lost cardiomyocytes .

Second-generation therapies use alternative strategies, based on molecular pathways and transcription factors guiding cardiac differentiation, to regenerate the myocardium by delivering lineage-specified cardiopoietic mesenchymal stromal cells, transplanting cardiomyocytes derived from embryonic stem (ES) cells, or directly converting heart fibroblasts into cardiomyocytes by inducing core cardiac transcription factors .

The third generation of cardiac stem cell therapies is needed to increase the pool of cardiomyocytes to the level that prevents or reverses the negative remodeling of the myocardium leading to heart failure. This future approach should be based on a deeper understanding of cardiogenesis, and it should aim to enhance the heart’s intrinsic regenerative capacity through the activation of endogenous progenitor cells or the stimulation of cell cycle reentry and expansion of adult cardiomyocytes.

On the other side of the cardiovascular disease spectrum, there are the congenital heart defects characterized by structural abnormalities of the heart or the great vessels or functional abnormalities in electrical conduction. Although rare diseases, these are often life threatening for the fetus or the young patient, and it is therefore imperative to try to develop new therapies for these defects. Congenital heart diseases are viewed mainly to be consequences of mutations in the transcriptional pathways that regulate heart development . Thus, recent progress in uncovering the intrinsic signaling pathways important for directing the fate of the multipotent cardiac progenitors (canonical Wnt signaling, insulin growth factor (IGF), Notch, and Hippo signaling) might yield clinically relevant information that could be used to prevent and correct fetal heart malformations and for regenerating the adult human heart.

This review will highlight the contribution of different cardiac progenitors to the embryonic heart development, the endogenous regenerative capacity of the adult heart including the plasticity of cardiomyocytes, as well as the linkage between cardiac progenitors and congenital heart disease, and how cardiac progenitors might be able to contribute to heart regeneration after myocardial injury.

Cardiac development

The heart is one of the first organs to be developed during embryogenesis and the first organ that displays function. The heart is derived from the mesodermal germ layer, which is specified into the cardiac mesoderm through the interaction of inductive and inhibitory signals from the adjacent endoderm and ectoderm. These signals include wingles integrated (Wnt), fibroblast growth factor (FGF), and transforming growth factor-β (TGF-β) pathways leading to the activation of early cardiomyocyte transcriptional programs . The next step in cardiogenesis is the specification and differentiation of these cells through the development of specific heart fields . The first, or primary, heart field (FHF) forms the cardiac crescent in the anterior splanchnic mesoderm at approximately week 2 of human gestation . In the third week of human gestation, the cardiac crescent fuses at the midline, and it forms the linear heart tube, which starts to beat around embryonic day 22 . Subsequently, during week 4, the linear heart tube undergoes rightward looping , and it grows rapidly through cell proliferation and recruitment of subpharyngeal cells to the arterial and venous poles . The cells originating from the pharyngeal mesoderm migrate and enter the FHF-derived heart tube, differentiating into progenitors of the second heart field (SHF) . SHF progenitors are defined by their expression of the LIM-homeodomain transcription factor Islet-1 (Isl1), and they contribute to the atria, outflow tract (OFT), and the right ventricle, whereas the left ventricle mainly seems to be derived from the FHF cells, which are marked by their expression of T-box transcription factor 5 (Tbx5), NK2 homeobox 5 (Nkx2-5), and the ion channel hyperpolarization-activated cyclic nucleotide-gated channel 4 (HCN4) .

The early embryonic heart tube consists of two cell layers, the endocardium and the myocardium. The third layer, epicardium, is derived from the proepicardium, which in turn arises from the coelomic mesenchyme of the septum transversum roughly at embryonic day 21 . The embryonic proepicardial progenitor cells differentiate into cardiac fibroblasts, coronary vasculature, and a small number of cardiomyocytes .

The fourth progenitor population involved in cardiogenesis consists of the cranial neural crest cells, and it arises from the dorsal neural tube . The neural crest cells are crucial for the septation of the OFT, the formation of heart valves, and the full parasympathetic innervation of the heart .

Through complex interactions of FHF and SHF progenitors with proepicardial and cranial neural crest cells, the fetal heart is septated into four defined chambers, and it connects to the aorta and the pulmonary trunk approximately during gestational week 7 .

Cardiac development

The heart is one of the first organs to be developed during embryogenesis and the first organ that displays function. The heart is derived from the mesodermal germ layer, which is specified into the cardiac mesoderm through the interaction of inductive and inhibitory signals from the adjacent endoderm and ectoderm. These signals include wingles integrated (Wnt), fibroblast growth factor (FGF), and transforming growth factor-β (TGF-β) pathways leading to the activation of early cardiomyocyte transcriptional programs . The next step in cardiogenesis is the specification and differentiation of these cells through the development of specific heart fields . The first, or primary, heart field (FHF) forms the cardiac crescent in the anterior splanchnic mesoderm at approximately week 2 of human gestation . In the third week of human gestation, the cardiac crescent fuses at the midline, and it forms the linear heart tube, which starts to beat around embryonic day 22 . Subsequently, during week 4, the linear heart tube undergoes rightward looping , and it grows rapidly through cell proliferation and recruitment of subpharyngeal cells to the arterial and venous poles . The cells originating from the pharyngeal mesoderm migrate and enter the FHF-derived heart tube, differentiating into progenitors of the second heart field (SHF) . SHF progenitors are defined by their expression of the LIM-homeodomain transcription factor Islet-1 (Isl1), and they contribute to the atria, outflow tract (OFT), and the right ventricle, whereas the left ventricle mainly seems to be derived from the FHF cells, which are marked by their expression of T-box transcription factor 5 (Tbx5), NK2 homeobox 5 (Nkx2-5), and the ion channel hyperpolarization-activated cyclic nucleotide-gated channel 4 (HCN4) .

The early embryonic heart tube consists of two cell layers, the endocardium and the myocardium. The third layer, epicardium, is derived from the proepicardium, which in turn arises from the coelomic mesenchyme of the septum transversum roughly at embryonic day 21 . The embryonic proepicardial progenitor cells differentiate into cardiac fibroblasts, coronary vasculature, and a small number of cardiomyocytes .

The fourth progenitor population involved in cardiogenesis consists of the cranial neural crest cells, and it arises from the dorsal neural tube . The neural crest cells are crucial for the septation of the OFT, the formation of heart valves, and the full parasympathetic innervation of the heart .

Through complex interactions of FHF and SHF progenitors with proepicardial and cranial neural crest cells, the fetal heart is septated into four defined chambers, and it connects to the aorta and the pulmonary trunk approximately during gestational week 7 .

Defining cardiovascular progenitors

In order to define cardiovascular progenitors, it is necessary to identify their origin in an embryonic heart and to organize them within the hierarchy of cardiogenesis. Only when we have a reliable map of fetal cardiovascular cell lineages can we look for the presence of endogenous stem cells and progenitors in the adult heart.

A common primordial cardiovascular progenitor that gives rise to progenitors of both the FHF and SHF has been characterized by the expression of transcription factor brachyury (Bry), a member of the T-box family of genes, shown to be critical for mesoderm formation . Kattman and colleagues used Bry and kinase insert domain receptor (KDR), also known as fetal liver kinase 1 (Flk-1), which encodes vascular endothelial growth factor receptor 2, to identify a population of cells with cardiovascular potential during ES cell differentiation. Individual Bry+KDR+ cells generated colonies capable of differentiating into cardiomyocytes, endothelial cells, and vascular smooth muscle cells . Some colonies expressed SHF marker Isl1, whereas others were positive for Tbx5, a marker linked to FHF . The next marker in the embryonic cardiovascular hierarchy, which may identify an early progenitor population that precedes the separation of the first and second heart lineages, is the transcription factor mesoderm posterior 1 (Mesp1). This factor seems to be important for the activation of the cardiogenic transcription factors: Nkx2-5, GATA-binding protein 4 (GATA4), myocyte enhancer factor 2c (Mef2c), and Isl1 .

Downstream, the expression of Nkx2-5, indispensable for the development of ventricular cardiomyocytes, characterizes FHF- and SHF-derived cells committed to the cardiomyogenic fate . Another early marker, which may be used for sorting early cardiac progenitors, is stage-specific embryonic antigen-1 (SSEA-1) . Blin and colleagues have derived these cells from human ES cells and induced pluripotent stem cells, differentiating them into cardiomyocytes, smooth muscle cells, and the endothelium . Recently, we were able to sort out a pure population of cells expressing SSEA-1 from human embryonic hearts obtained from abortion material . Isolated cells expressed the pluripotent stem cell marker Oct4 as well as the key cardiac transcription factors Nkx2-5, GATA4, Isl-1, and Tbx5. Besides expressing the markers of both FHF and SHF, these cells co-expressed the mesenchymal stromal cell markers. The expression of the cardiomyocyte-specific proteins troponin T and actin was upregulated in in vitro human fetal heart-derived SSEA-1+ cells, and they displayed contractile filaments and Ca ion channels, which were activated by IGF-1 .

A specific FHF marker was recently identified by Spater and colleagues; HCN4, the voltage-gated ion channel, which is a marker of the conduction system, appears to specify cells that predominantly differentiate into atrial and ventricular cardiomyocytes .

The SHF develops from multipotent cardiovascular progenitors characterized by the expression of the transcription factor Isl1 . SHF progenitors migrate from the pharyngeal mesoderm into the heart tube as it grows and undergoes looping . However, Isl1 protein has been detected already at the cardiac crescent stage, suggesting the possibility that Isl1 is expressed in FHF at a very early stage . In the OFT of the human fetal hearts at gestation week 9, Lui and colleagues identified an Isl1-positive population of endothelial progenitors expressing VEGF-A receptors , whereas Vedantham et al. demonstrated that Isl1 is a regulator of sinoatrial node development . Thus, in contrast to HCN4, Isl1 expression may identify a transient progenitor state important for heart lineage diversification that rapidly converges into the vascular endothelial cell, myocardial, and conduction system cell lineages, followed by downregulation of Isl1 upon cardiomyocyte differentiation . Recently, our laboratory has focused on characterizing, isolating, and expanding cardiovascular progenitors from the human embryonic hearts. First, in the work by Genead and coworkers, we characterized the distribution and electrophysiological properties of Isl1-positive progenitor cells in the early first-trimester human embryonic hearts . In fetal hearts obtained from gestational weeks 5–10, Isl1 was mainly detected in the OFT, but also in both the atria and the right ventricle, with certain clusters co-expressing the cardiomyocyte marker troponin T. Interestingly, most of the proliferating cells were Isl1 negative but troponin T positive, showing robust proliferative potential of fetal ventricular cardiomyocytes. When the cells were cultured as spontaneously beating cardiospheres, they expressed Nkx2-5 and troponin T, whereas Isl1 was detected only in a minority of cells .

Subsequently, we were able to derive fetal human Isl1+ cells, expressing markers defining multipotent cardiac progenitors such as Tbx5, Nkx2-5, and Tbx18, from the gestational weeks 6–10 old hearts, and to expand these cells to large numbers by culturing them on biologically relevant laminins in a medium containing molecules stimulating the Wnt/β-catenin pathway (submitted manuscript by Mansson-Broberg et al.).

The embryonic proepicardial progenitor cells originate from the mesenchyme, and they express Tbx18 and Wilms tumor 1 (Wt1), or semaphorin 3D (Sema3d) and scleraxis (Scx) . They primarily differentiate into coronary smooth muscle and endothelial cells, but early in development, a subset of proepicardial progenitors express Isl1 and Nkx2.5 and may also differentiate into cardiomyocytes .

Finally, the cranial neural crest cells, which are involved in the development of heart valves and the parasympathetic innervation of the heart, are defined by the expression of Wnt1 and Pax3 markers .

Even in the adult heart, a few resident Isl1-positive cells can be found in the OFT, expressing troponin T as a sign of cardiac commitment . Other proposed cardiac progenitors in the adult heart are c-kit+ cells, Sca-1+ cells, platelet-derived growth factor receptor (PDGFR)-α+cells, SSEA-1+ cells, and side population cells . Characterization of these cells is based upon their surface markers and functional properties. During heart development, the c-kit receptor may be expressed in some cells originating from the FHF that contribute to cardiomyogenesis as well as to noncardiogenic cells. However, the pool of c-kit+ cells with the potential to differentiate into cardiomyocytes is depleted before adulthood .

The definitive proof of the existence of endogenous stem and progenitor cells in the adult heart would be of great value in the cardiac regenerative field. Nevertheless, establishing the relationship of these cells to the embryonic progenitors and arranging them within the heart development hierarchy is challenging, and it requires rigorous lineage-tracing strategies.

Pathways controlling specification and differentiation of embryonic cardiac stem cells

The process of cardiac lineage specification and differentiation is primarily controlled by activating or inhibitory signals of activin A, bone morphogenetic protein (BMP), Wnt, Notch, and FGF signaling networks, regulating key cardiac transcriptional factors such as Nkx2-5, GATA4, Tbx5, and chromatin remodeling protein SMARCD3 , which in turn are indispensable for the transcription of the cardiomyocyte contractile apparatus. The commitment of the Bry+ mesodermal progenitors toward a cardiogenic fate requires the inhibition of canonical and activation of noncanonical Wnt signaling . The expansion and differentiation of Nkx2-5 and Isl1-positive cardiac progenitors is also dependent on canonical Wnt signaling, and a sequential induction and inhibition of this pathway led to the development of a robust cardiomyocyte differentiation protocol from embryonic and induced pluripotent stem cells . The Notch, IGF, Sonic Hedgehog, and Hippo signaling pathways have recently emerged as important regulators in cardiovasculogenesis, and disturbances in these pathways cause a broad spectrum of cardiac anomalies .

Besides the transcriptional signals, which are critical for the activation of cardiac genes, a number of recent reports highlight posttranscriptional regulation by small, noncoding micro-RNAs (miRNAs) as the emerging novel regulator of cardiogenesis and the promising tool in enhancing the endogenous regenerative potential of the heart .

The relationship between mutations affecting the signaling pathways of cardiogenesis and congenital heart disease

The complexity of the molecular pathways involved in cardiogenesis may explain why congenital heart defects are the most common birth anomalies affecting around 1% of newborns , whereas in grown-up patients, the congenital conduction system defects represent a significant risk of sudden cardiac death .

During embryonic heart development, Nkx2-5 acts in synergy with basic helix–loop–helix protein Hand2 and the transcription factor Mef2c to promote cardiomyocyte differentiation and chamber identity ; together with Tbx5 and GATA4, Nkx2-5 forms a complex network needed for proper cardiac septation and conduction system development . In mice, the loss of Nkx2-5 is lethal for the embryo , whereas human mutations in Nkx2-5 locus cause atrial septal defects and disturbances of the electrical conduction system . Even mutations of GATA4 result in atrial and ventricular septal defects , whereas the mutations in Tbx5 cause atrial and ventricular septal defects as well as defects in the cardiac conduction system and limb abnormalities in patients with the Holt–Oram syndrome .

The OFT undergoes extensive morphogenetic changes during embryogenesis involving both the SHF progenitors and the neural crest cells, which makes this region particularly sensitive to developmental disturbances. In mice, the loss of Hand2 and Mef2c, transcription factors highly expressed in the SHF, leads to right ventricular hypoplasia, and Isl1 loss-of-function embryos exhibit a selective loss of SHF-derived structures . The deletion of the transcription factor Tbx1 on chromosome 22q11, which also regulates OFT development, causes Di George syndrome . This is the most common human deletion syndrome, characterized by multiple cardiac and craniofacial disorders. Mutations of FGF8 result in ventricular septal defects and persistent truncus arteriosus, whereas Notch1 mutations interfere with aortic valve development, causing a wide range of malformations that may lead to severe defects in embryonic valve opening and left ventricular hypoplasia .

Another signaling system involved in cardiogenesis and final development toward a functioning organ is the ephrin/Eph family of tyrosine kinase receptors and ligands, responding to mesodermal induction signals , influencing cell migration, neural cell migration, and cell-to- cell interaction during embryogenesis . Furthermore, EphB2 and B4 are distinguishing features of arterial and venous endothelial cells , and the lack of or defect expression of these receptors causes severe cardiovascular malformations and defect trabeculation of the ventricles due to poor cell-to-cell interaction between the endothelium and the adjacent tissue during embryogenesis. In addition, the ErbB–neuregulin signaling system plays an essential role in cardiogenesis. Defect expression of the ErbB family of tyrosine kinase receptors (ErbB2/B4) during embryogenesis causes cardiac malformations and similar defects of the left ventricular trabeculation . Neuregulin-1 activation of the ErbB4 receptor induces a cascade of intracellular transduction upon stimulation with the ErbB4 ligand neuregulin-1β (endogenously present on the surface of endothelial cells and subject to proteolytic cleavage and release upon oxidative stress). Furthermore, neuregulin-1β activation of the ErbB4 receptor is shown to enhance cardiomyocyte differentiation in ES cells .

Comparison of the postnatal cardiomyogenesis between ma mmalian and nonmammalian vertebrate species

Our present understanding of cardiac regeneration has been based mainly on studies in the newt and the zebra fish, which are able to fully regenerate hearts after amputation of up to 20% of the ventricles . Upon injury, newt and zebra fish cardiomyocytes dedifferentiate into progenitor-like cells characterized by the disassembly of their sarcomeric structure that reenter the cell cycle and subsequently redifferentiate into mature cardiomyocytes . However, the earlier work of Lepilina and colleagues suggests that a pool of undifferentiated progenitors is the basis of zebra fish cardiac regeneration . Using a genetic fate-mapping strategy, Zhang and coworkers could demonstrate that atrial cardiomyocytes contribute to ventricular regeneration through a step of dedifferentiation . Following targeted destruction of the ventricle of the zebra fish, atrial cardiomyocytes adjacent to the atrioventricular canal underwent a dedifferentiation process and started to reexpress the cardiac progenitor markers GATA4, Hand2, Nkx2-5, Tbx5, and Tbx20, while proliferating and migrating into the injured ventricle. Moreover, SHF progenitors, resident near the OFT, also seemed to be contributing to the ventricular cardiomyocytes. These findings indicate that nonmammalian vertebrate species regenerate ventricular cardiomyocytes through both dedifferentiation and proliferation of cardiomyocytes and most probably also by the activation of resident cardiac progenitors.

The regenerative capacity of the mammalian heart is very limited, as cardiomyocytes are withdrawn from the cell cycle, and they become binucleated early on in the neonatal period . In mice, especially in the neonatal period, a certain turnover of cardiomyocytes has been shown through the proliferation of already-existing cardiomyocytes . Porrello and colleagues have demonstrated that mouse heart can fully regenerate if amputation of the ventricular apex occurs on the first neonatal day. This remarkable repair seems to be achieved through a mechanism comparable to that of zebra fish regeneration, namely the dedifferentiation and proliferation of existing cardiomyocytes.

However, in contrast to the newt and the zebra fish that retain regenerative capacity throughout life, in mice, this initial robust regenerative potential is rapidly lost already within the first postnatal week .

During adolescence, the proliferation of cardiomyocytes is reduced, except for the thyroid hormone surge at postnatal day 15, which activates the IGF-1/IGF1-R/Akt pathway initiating a proliferative burst with a concomitant increase in the number of cardiomyocytes with about 40% .

In humans, Bergmann and colleagues performed carbon-14 birth-dating studies suggesting that fewer than 50% of cardiomyocytes are replaced over an entire life span, and the rate of replacement is declining with age . Several recent reports confirm this finding showing a low (<1% per year) turnover rate in the adult mammalian heart . Even if there is no evidence for thyroid hormone-induced proliferation of cardiomyocytes in early adolescence of humans, karyokinesis, division of the nucleus, was observed throughout life, whereas division of the cell cytoplasm (cytokinesis) could not be detected in human hearts after 20 years of life . Importantly, as both Bergmann and Mollova performed studies on presumably normal hearts from individuals without a known history of cardiac disease, the possible increase in the cardiomyocyte turnover rate upon injury is not excluded, as shown in mice subjected to myocardial damage .

From these human studies, we can draw the conclusion that new cardiomyocytes are generated in the postnatal heart but at a slow rate, and the cellular origin of these cells is still unknown.