Development, Proliferation, and Growth of the Mammalian Heart

Abstract

During development, the embryonic heart grows by addition of cells from a highly proliferative progenitor pool and by subsequent precisely controlled waves of cardiomyocyte proliferation. In this period, the heart can compensate for cardiomyocyte loss by an increased proliferation rate of the remaining cardiomyocytes. This proliferative capacity is lost soon after birth, with heart growth continuing by an increase in cardiomyocyte volume. The failure of the injured adult heart to regenerate often leads to the development of heart failure, a major cause of death. With the recent observation of a small fraction of cardiomyocytes that appear to have retained the proliferative capacity within the adult heart, as well as the identification of developmental pathways such as the Hippo-signaling pathway that can invoke mature cardiomyocyte proliferation, more studies are taking a knowledge-based mechanistic approach to heart regeneration. A key question being asked is if this knowledge can be used therapeutically to reinitiate cardiomyocyte proliferation after injury such as myocardial infarction. In this respect, uncovering and understanding the mechanisms and conditions that give rise to a fully functional and adaptive heart in the developing embryo could provide us with the answers to many of the questions that are now being asked.

The developing mammalian heart can regenerate, a capacity that is lost shortly after birth, when cardiomyocytes stop proliferating. Mechanisms and pathways that are required for the high rates of cardiomyocyte proliferation of the developing heart reveal targets to reinitiate cardiomyocyte division and enhance regeneration in the adult heart.

Main Text

Cardiomyocytes of the mature mammalian heart were considered to be essentially post-mitotic and not able to undergo postnatal cell division. However, during the last decade, studies in human have revealed a limited but significant rate of annual cardiomyocyte renewal.1, 2 Nevertheless, one can infer that this renewal rate would never be sufficient to replace the large numbers of cardiomyocytes that are lost after cardiac injuries such as myocardial infarction. Instead, the heart, upon injury, undergoes a quick patch-up repair. During a rapid inflammatory response, large numbers of fibroblasts infiltrate the damaged myocardium leading to the deposition of fibrotic scar tissue. As a consequence, cardiac function becomes compromised and heart failure, a leading cause of death, may soon ensue.³ In light of this, regeneration of human heart tissue has been and remains a major goal in cardiovascular research. Over the last two decades, research has mainly focused on tissue replacement approaches. Historically, the existence of cardiac endogenous stem cells and their ability to differentiate has been a topic of much discussion.⁴ Alternative approaches focus on deriving cardiomyocytes from various stem cell sources, including embryonic stem cells and induced pluripotent stem cells.5, 6 Stem-cell-derived cardiomyocytes mature insufficiently under in vitro conditions and hence must adopt their fate after being transplanted into the host tissue.⁶ Whereas studies administering stem-derived cardiomyocytes into rodents that underwent myocardial infarction have demonstrated the feasibility of this technique,7, 8, 9 with remuscularization of the injured area, application in non-human primates has been fraught with difficulties, including insufficient maturation of cardiomyocytes and sustained arrhythmias.5, 6

New insights and inroads have come from highly regenerative species, such as the zebrafish, as well as the regenerative capacity demonstrated in the heart of fetal and neonatal rodents, all of which can regenerate cardiac tissue by proliferation of resident cardiomyocytes, a capacity that is then lost when mammals mature.10, 11, 12, 13, 14 Current research has shifted to the idea that proliferation can be re-induced in adult cardiomyocytes. To achieve this, it is crucial to gain a deeper understanding of the switches and mechanisms of cardiomyocyte proliferation and differentiation from early embryonic development into adulthood.

Progenitor Differentiation and Cardiomyocyte Proliferation Underlie Embryonic Heart Formation

During development, the heart grows by differentiation of proliferating, multipotent cardiac progenitor cells into cardiomyocytes and controlled local proliferation of already-differentiated cardiomyocytes.¹⁵ This process requires the tight regulation of levels of extracellular growth factors as well the intracellular expression of transcription factors.

Most knowledge regarding heart development has been gained from studies of transgenic mouse embryos and histologic 3D reconstructions.16, 17 Other studies, however, have made use of MRI and episcopic fluorescence image capture to compare cardiac development in mice and humans. From these studies, it could be demonstrated that the sequence of events in cardiac development are comparable in both species, thus establishing the mouse as a useful model for the extrapolation to the study of human heart development.¹⁸

The formation of the heart starts at week 3 of human embryonic development,¹⁶ corresponding to embryonic day 7.5 (E7.5) of mouse development.17, 19 At this stage, the embryo consists of the three germ layers: the endoderm, mesoderm, and ectoderm. Most of the cardiac tissue will originate from the mesoderm. During gastrulation, early cardiac progenitor cells reside at bilateral regions of the embryonic midline. When the embryo folds, these heart-forming regions fuse across the midline, forming what is commonly referred to as the cardiac crescent. The cells of the cardiac crescent display a cardiogenic gene-expression profile, initiate calcium transients, and not long thereafter start to beat.¹⁹

In all vertebrates examined so far, two sets of precursor cells, based on temporal expression of various marker genes, can be distinguished during the early phases of cardiac development, the cells of the first heart field (FHF) and the second heart field (SHF; Figure 1).¹⁵ The division in two populations is somewhat artificial, because different HF markers indicate different populations and borders between populations and because the differentiation of the precursor cells and their addition to the heart tube is a continuous process.²⁰ Nevertheless, the FHF-derived differentiating cell populations from the cardiac crescent fuse at the midline and form the linear heart tube, enclosing the endocardial layer that is separated from the myocardial layer by the cardiac jelly, a layer of acellular matrix.¹⁷ Upon forming the heart tube, the primitive cardiomyocytes stop proliferating (Figure 1). The heart tube grows by further addition of mesodermal SHF precursor cells, which migrate from the pericardial cavity dorsal and caudal of the heart tube.²¹ SHF cells proliferate at high rates, mediated by canonical WNT/β-catenin signaling²² and are added to the heart tube at the arterial and venous poles while they are undergoing differentiation.²¹ During their addition to the heart tube, the SHF-derived cardiomyocytes also temporally cease proliferation.²¹ In mammals, cells from the FHF contribute mainly, if not wholly, to the definitive left ventricle and atrioventricular canal.²³ Cells from the SHF will contribute to the formation of the right ventricle, outflow tract, and atria.¹⁵

Figure 1.

Formation and Looping of the Human Heart Tube

Proliferation of the mesodermal precursor cells of the first heart field stop after differentiation into the cardiomyocytes of the heart tube, whereas the mesodermal precursor cells of the second heart field (SHF) remain highly proliferative. Proliferation is re-initiated after looping of the heart tube and the beginning of ventricular chamber formation at stage 10. The proliferation index indicates the percentage of Ki67-positive cells as a marker for proliferation. Courtesy of Antoon Moorman.

Various cardiac transcription factors such as GATA4, NKX2-5, ISL1, TBX5, TBX1, and MEF2C determine the cardiogenic fate of the heart field precursors.24, 25 T-box transcription factor family members that are regionally expressed in the heart play a central role in defining the future characteristics of the cardiomyocytes, such as their automaticity, conduction velocity, and contractility.26, 27 Several families of microRNAs (miRNAs), single-stranded, non-coding RNAs, have been shown to target these transcription factors and thereby regulate the cardiac fate of progenitor cells.²⁸ In this respect, the actions of the families miR-1 and miR-133a have been well studied.29, 30

Between Carnegie stage (CS)10 and CS11, corresponding to E8.25¹⁷ in the mouse, the heart tube loops.¹⁶ During this process, cardiomyocytes at the outer curve of the tube display a distinct transcriptional program and start to re-enter the cell cycle and proliferate leading to the so-called “ballooning” of the ventricles (Figure 2).16, 31, 32 Slightly later, the cardiomyocytes of the forming atria do likewise. The non-chamber regions, the atrioventricular canal, outflow tract, and sinus venosus, however, express TBX2, TBX3, MSX2, and/or ID2, all of which function to suppress proliferation and chamber differentiation.17, 21, 23 The chambers grow considerably, and the embryonic outflow tract gets taken up into the right ventricle. The atrioventricular canal will remain small in comparison to the chambers and give rise to the atrioventricular node,³³ and the sinus node will form between from the sinus venosus.³⁴

Figure 2.

Ventral View of the Proliferating Myocardium during Heart Development in Mouse E8–E17.5

The upper graph shows proliferation rates of cardiomyocytes within the developing heart. Initial formation of the heart tube takes place at E8(a), after which proliferation of the myocardium stops. Re-initiation of the proliferation starts at the left side of the heart tube at E8(b)–(c) and during looping of the heart tube at E8.25. The high proliferation at the outer curvature leads to ballooning of the future left ventricle at E8.5. At E9.5, the four chambers of the heart become visible. The non-chamber regions such as the outflow tract and the atrioventricular canal proliferate slowly (E11–E17.5). The highest proliferation rates are located at the base of the ventricular chambers and decrease from E11 onward. The lower graph demonstrates that during heart development the heart rapidly grows by an increase in cell number. The proliferation index indicates the percentage of cells that had incorporated bromodeoxy-uridine (BrdU) during DNA synthesis. Courtesy of Antoon Moorman.

Further expansion of the embryonic heart occurs primarily through cell divisions in the ventricles and, to a lesser extent, the atrial chambers (Figure 2).¹⁷ In the mouse embryo, the formation of the four chambers is visible from E9.5 onward.¹⁷ In human development, the four chambers become visible at CS11.¹⁶ At this stage, cardiomyocyte proliferation rates differ locally within the developing heart.16, 17 In distal parts of the outflow tract, proliferation rates decrease, whereas inflow tract and atria are formed by highly proliferating cardiomyocytes. During these stages, the highest proliferation rates can be observed in the ballooning ventricles (Figure 2). From E11 onward, the proliferation rate and thereby the increase in cardiomyocyte number steadily declines during the further course of development (Figure 2). Within the ventricles, the trabecules develop at the luminal site of the myocardium and become clearly visible by E9.5,³⁵ corresponding to CS12.¹⁶ Crucial signaling cues for the development of trabecules will come from the endocardial lining, located in the lumen of the chambers. Important molecular signals include the Notch signaling pathway, activating BMP10 and Neuregulin1 (NRG1).³⁶ NRG1 binds to the receptors ErbB3 and ErbB4, which will then dimerize with ErbB3, ErbB4, or ErbB2.³⁷ They induce a high rate of proliferation in the less differentiated cardiomyocytes located at the base of the trabecules, where Notch signaling is strongest,17, 36, 38 leading to trabecular elongation. Following this, the outer lining of the myocardium starts to compact, and proliferation in the trabecules decreases.16, 17 Further myocardial cell divisions are now controlled by the epicardium that covers the outer surface of the myocardium and is derived from the pro-epicardium. It produces secreted factors that control the proliferative expansion of the compact sub-epicardial ventricular myocardium.³⁹ These factors include fibroblast growth factor (Fgf)-signaling ligands (e.g., FGF9), insulin-like growth factor (Igf)-signaling ligands (IGF2), Wnt signaling ligands (e.g., WNT9B), and retinoic acid.39, 40, 41 Epicardial cells are multipotent and soon after covering the embryonic ventricular wall undergo epithelial-to-mesenchymal transformation and migrate into the myocardial wall. Here, they differentiate into fibroblasts, coronary vascular smooth muscle cells, and possibly endothelial cells.⁴² Although there is a common consensus that cardiomyocytes comprise about 30% of the ventricular wall,⁴³ there have been varying reports about the content of fibroblasts and endothelial cells in the heart, probably due to a lack of suitable detection markers, varying isolation techniques, and sampling bias.2, 43, 44 A recent study reported that the murine and human heart contains 60% endothelial cells and less than 20% of fibroblasts.⁴³

During cardiac development, the myocardium increases in volume from about 0.003 mm³ at CS9 (20 days of development) to around 2 mm³ at CS16 (6 weeks of development) to around 15 mm³ at CS23 (>8 weeks of development).⁴⁵ In mice, myocardium growths in volume from 0.0014 mm³ at E8 to almost 2 mm³ at E17.5.¹⁷ During this time, the volume of the cardiomyocytes stays constant, whereas the number of cardiomyocytes within the heart increases exponentially and reaches almost 1 million (Figure 2).¹⁷ A major regulator of the proliferation of differentiated cardiomyocytes during cardiac development is the Hippo pathway, an evolutionary conserved signaling pathway that controls organ size by regulating proliferation, apoptosis, and stem cell fate (Figure 3).⁴⁶ Suppression of this pathway underlies the high rates of cardiomyocyte proliferation during heart development.⁴⁷ When the Hippo pathway is active, the kinases mammalian sterile 20-like kinase 1 and 2 (MST1/2), which are homologous of the Drosophila Hippo kinase, form a complex with the scaffold protein Salvador (SAV), and in turn activate the large tumor suppressor homolog 1 and 2 (LATS1/2) kinases.48, 49 LATS1/2 and the Mps one binder kinase activator 1a/b (MOB1A/B) phosphorylate the downstream effector of the Hippo cascade, the Yes-associated protein (YAP).50, 51 Phosphorylation of YAP orchestrates its degradation or cytoplasmic retention by 14-3-3 binding.⁵² Under the condition of an inactive Hippo pathway, YAP and its transcriptional co-activator with PDZ-binding motif (TAZ) translocate to the nucleus where they interact with various transcription factors. Interacting transcription factors include p73,53, 54, 55 the runt-related transcription factor (RUNX)⁵⁶ and TBX5.57, 58 Interaction of YAP/TAZ with the TEA domain family members 1–4 (TEADs) has been shown to regulate the expression of target genes involved in the regulation of cell-cycle progression.59, 60, 61 Nuclear exclusion of YAP is also controlled via Hippo-independent ways, such as alpha-catenin that binds YAP under high-cell-density conditions.62, 63 The Hippo pathway’s role during cardiac development was demonstrated by various studies in mouse models. Different conditional YAP knockout mice frequently showed reduced embryonic cardiomyocyte proliferation and embryonic or perinatal lethality.64, 65, 66 The conditional deletion of SAV, upstream of YAP, also leads to a higher expression of various Wnt-target genes and accumulation of nuclear β-catenin, which suggested an interaction between YAP and β-catenin.⁴⁷ Accordingly, overexpression of a constitutively active form of YAP during cardiac development was sufficient to cause elevated cardiomyocyte proliferation and increased heart size.47, 65, 67

Figure 3.

The Hippo Pathway Regulates Cardiomyocyte Proliferation

When the Hippo pathways is off, the Yes-associated protein (YAP) and its transcriptional co-activator with PDZ-binding motif (TAZ) translocate to the nucleus where they bind various transcription factors such as the TEA domain-containing sequence-specific transcription factors (TEADs). Expressed target genes promote cardiomyocyte proliferation. Postnatal activation of the Hippo pathway, for example by formed reactive oxygen species (ROS), leads to the formation of a complex between the mammalian sterile 20-like kinase 1/2 (MST1/2) and Salvador (SAV). In turn, they activate the large tumor suppressor homolog 1/2 (LATS1/2) and the Mps one binder kinase activator 1a/b (MOB1A/B)48, 49 that phosphorylate YAP/TAZ, leading to their cytoplasmic retention by 14-3-3 binding or their degradation.50, 51 Next to the Hippo pathway, cytoplasmic retention of YAP is also caused by the dystroglycan (DGC) complex, which binds YAP with its intracellular domain, impeding its translocation to the nucleus.⁹⁴ High neonatal levels of Agrin in the extracellular matrix bind DGC leading to conformational changes, impeding the binding of YAP.⁹³

The Mammalian Neonatal Heart Grows by Hypertrophy and Cardiomyocyte Proliferation

Whereas cardiomyocytes actively proliferate before birth, the mammalian perinatal period is marked by a drop in cardiomyocyte proliferation. The subsequent increase in cardiac volume occurs almost exclusively by hypertrophy of cardiomyocytes.⁶⁸ In humans, cardiomyocytes increase about 8.6-fold in their volume within the first 20 years of life.⁶⁹ The postnatal decrease of cardiomyocyte proliferation is accompanied by the occurrence of incomplete cell cycles (Figure 4).⁷⁰ Within the first week after birth, most cardiomyocytes complete their final cell division. In mice, a wave of karyokinesis without cytokinesis leading to binucleation as well as a wave of endoreplication, leading to polyploidization, account for most of the postnatal DNA synthesis up to 14 days of age.70, 71 This will eventually result in about 80%–90% of all cardiomyocytes being binucleated.70, 71, 72 It is relevant to note that mice are born 1 week after heart septation has completed at E14.5, whereas the human fetus will continue to grow for another several months after completion of septation.¹⁸ This has led to the assumption that the heart of neonatal mice is less developed as compared to humans and some of the processes that occur during human fetal development still need to occur in mice during the postnatal period.⁷³ For example, whereas binucleation occurs during the first 2 weeks of murine postnatal development, in humans, about 25% of cardiomyocytes become binucleated already during a brief period before birth, a proportion that remains constant throughout life.2, 69, 74 An increase in DNA content leading to polyploidization of cardiomyocytes takes place mainly during the second decade after birth.² Both binucleated and polyploid cardiomyocytes are generally believed to be terminally differentiated and unable to contribute to cardiomyocyte renewal during cardiac homeostasis and injury. In contrast, the small remaining population of diploid, mononucleated cardiomyocytes still has the capacity to proliferate.75, 76, 77 Interestingly, a recent study showed that the frequency of these cells is highly variable within different mouse strains.⁷⁷ Mouse strains with a higher amount of mononucleated, diploid cardiomyocytes displayed an increased proliferation rate and a greater recovery after myocardial infarction as compared to mice carrying less of these fraction of cells. This shows that the genetic background may influence the regenerative process after injury and should be taken into account when comparing different studies. In contrast to mammals, zebrafish retain their regenerative capacity throughout adulthood.⁷⁸ Zebrafish cardiomyocytes stay mononucleated and diploid throughout their whole life, and upon injury, lost cardiomyocytes are replaced by proliferation of resident cardiomyocytes.¹⁰ When polyploidization was induced in zebrafish cardiomyocytes, the regenerative response was limited to the remaining fraction of dividing diploid cardiomyocytes, and regenerative capacity was fully lost when the mononucleaeted fraction was reduced to 50% of all cardiomyocytes.⁷⁹ This indicates that indeed the proliferative capacity is restricted by the small number of the remaining mononucleated and diploid cardiomyocytes.

Figure 4.

Schematic Overview of the Cardiomyocyte Cell Cycle Activity during Cardiac Development and Postnatal Growth in Mice

During cardiac development, the heart grows by proliferation of cardiomyocytes (green line). Already before birth, cardiomyocyte proliferation decreases, and incomplete cell cycles account for most of the neonatal cell-cycle activity. Cardiomyocytes that undergo endorepliction exit the cell cycle after the S phase and become polyploid (blue line). Karyokinesis without cytokinesis leads to binucleation of most of the cardiomyocytes (red line). Both polyploid and binucleated cardiomyocytes are unable to proliferate. After the neonatal period, annual cardiomyocyte proliferation is limited to less than 1%.

The time window and rate at which cardiomyocyte proliferation decreases after birth have been highly debated. In mice, a recent study reported a proliferative burst of cardiomyocytes at 14 days of postnatal age, already after the initial decline of cardiomyocyte proliferation.⁸⁰ However, this observation may have been the misinterpretation of endoreplication events and could be ruled out by studies showing that the rate of cardiomyocyte proliferation continuously decreases within the first 2 weeks after birth.70, 81 High cardiomyocyte-proliferation rates in young humans were reported by a study combining stereological quantification and the analysis of mitotic events.⁶⁹ This group’s findings suggested a strong increase in cardiomyocyte number within the first 20 years of human life. Conversely, the application of ¹⁴C birth-dating in combination with stereology demonstrated no increase in cardiomyocyte cell number later than 1 month after birth with a final number of about 3.2 × 10⁹ cardiomyocytes.² It could be shown that all events of subsequent cardiomyocyte proliferation account for cardiomyocyte replacement rather than number expansion, taking place majorly within the first 10 years of life.²

Numerous intrinsic and extrinsic factors that cause postnatal cardiomyocyte cell-cycle exit leading to binucleation and polyploidization of cardiomyocytes have been identified. First of all, cardiomyocytes mature, leading to structural changes: during fetal cardiomyocyte proliferation, cardiomyocytes disassemble their sarcomeric apparatus before mitosis and re-assemble it during cell division, a process known as dedifferentiation. Likewise, the postnatal increase in sarcomeric content may interfere with the correct disassembly and re-assembly of these structures.⁸² Cardiomyocytes withdraw from the cell cycle by a downregulation of positive cell-cycle regulators such as cyclins and cyclin-dependent kinases that maintain cardiomyocyte proliferation during heart development. At the same time, levels of cell-cycle inhibitors increase.⁸³ A change in the expression of transcription factors and miRNAs has been implicated in this process.84, 85 For example, increased postnatal expression of the transcription factor Meis homeobox 1 (Meis1) was shown to activate the cell-cycle-dependent kinase (CDK) inhibitors p15, p16, and p21.⁸⁵ CDKs bind to their corresponding cyclins and thereby control progression through the different phases of the cell cycle. Accordingly, the inactivation of MEIS1 led to increased postnatal cardiomyocyte proliferation.⁸⁵ In adult mice, re-entry of cardiomyocyte proliferation could already be stimulated by overexpression cyclin D2, which decreased scar formation after myocardial infarction.⁸⁶

Several miRNAs are involved in the drop of cardiomyocyte proliferation. Members of the miR-15 family are upregulated after birth and impede neonatal cardiomyocyte proliferation by the repression of cell-cycle activators such as the checkpoint kinase 1, which regulates the progression of the cell cycle through mitosis.84, 87 An increased expression of miR-128 after birth was also implicated in the drop of postnatal cardiomyocyte proliferation. Overexpression of miR-128 reduced the time course of early neonatal cardiomyocyte proliferation and consequently increased cardiac hypertrophy.⁸⁸ Likewise, early neonatal heart regeneration was impeded in these mice. MiR-128 was shown to suppress the action of the positive cell-cycle regulators cyclin E and CDK2. Levels of miR-128 stay high during adulthood, and it could be demonstrated that deletion of miR-128 enhanced cardiac regeneration after myocardial infarction in adult mice. Also, the delivery of exogenous miRNAs with viral vectors and intracardiac injections such as the human miRNAs hsa-miR-199a-3p and hsa-miR-590-3p stimulated proliferation of resident cardiomyocytes in neonatal rats and could induce cardiac regeneration in adult mice after myocardial infarction.89, 90 A safe delivery of miRNAs might be a promising way to stimulate cardiac regeneration in humans. Recently, miRNAs have also been suggested to be behind some of the beneficial effects observed after application of stem cells into the damaged cardiac tissue.⁹¹ Extracellular vesicles released from these cells were shown to discharge paracrine factors that support cardiac repair processes when injected into infarcted myocardial tissue. Among other bioactive molecules, miRNAs released by the extracellular vesicles were proposed to stimulate proliferation and angiogenesis and may represent an attractive avenue of research to explore in detail if this can further be validated.

Whereas YAP is highly active before birth, the postnatal activation of the Hippo pathway causes restrained cardiomyocyte proliferation. Mechanical cues are an important upstream regulator of the Hippo pathway.⁹² The extracellular matrix molecule Agrin and the dystrophin glycoprotein complex (DGC) were found to play an important role during postnatal cardiomyocyte proliferation (Figure 3). The transmembrane-complex DGC directly binds to YAP and thereby inhibits its translocation to the nucleus. High neonatal levels of Agrin bind to DGC and impede its interaction with YAP. Agrin levels are already markedly reduced after postnatal day 7, which may lead to increased YAP-DGC interaction and thus prevent YAP nuclear translocation.93, 94 When Agrin was administered after myocardial infarction, consequent scar formation was reduced,⁹³ suggesting that this may be a promising therapeutic approach. In addition, adult mice deficient for the Hippo pathway due to the conditional knockout of Sav in cardiomyocytes displayed a higher proliferation rate of cardiomyocytes after myocardial infarction and were enriched for nuclear localized YAP and sarcomere breakdown.61, 95

Reactive oxygen species (ROS) that are formed during the transition from the fetal hypoxic to the normoxic environment after birth have been shown to activate the Hippo pathway (Figure 3).⁹⁶ ROS develop when cardiomyocytes switch from anaerobic glycosylation to oxidative phosphorylation.⁹⁷ The DNA damage response that also becomes activated in response to ROS was shown to inhibit the cell-cycle regulators cyclin/CDKs.⁹⁸ In accordance with that, it was demonstrated that scavenging of the postnatal ROS extended the time window of cardiomyocyte proliferation after birth, confirmed by the smaller amount of hypertrophic cardiomyocytes and an increase in cardiomyocyte number. However, ROS scavenging only appeared to postpone but not impede the cardiomyocyte cell-cycle arrest, suggesting that other intrinsic or extrinsic mechanisms play a role during the postnatal cell-cycle exit.⁹⁸ In line with this, a small population of potentially proliferative mononucleated cardiomyocytes residing in a hypoxic niche within the adult heart, which supposedly is protected from oxidative damage, has been identified.⁹⁹ The oxygen-dependent degradation domain of the Hypoxia-inducible factor 1-alpha, which is stable in hypoxic cells but gets degraded in a normoxic environment, was fused to Cre, which was subsequently used to label the hypoxic cardiomyocytes followed by analysis of their fate.⁹⁹ The labeled cardiomyocytes increased in cell number throughout the whole heart with a calculated annual proliferation rate of 0.3%–1%, similar to results reported by other studies.⁷⁵

Fetal and Neonatal Hearts Can Regenerate

Interestingly, whereas adult hearts are incapable of cardiomyocyte regeneration, fetal and neonatal hearts appear to retain this capacity for a short time window after birth.11, 12, 13, 14, 87, 100 In a study introducing mosaicism for depletion of the holocytochrome-c synthase, an enzyme necessary for myocardial energy generation at E12.5 of mouse development, it was observed that this depletion was met by an increased proliferation of the remaining cardiomyocytes, compensating for the impaired proliferation of about 50% of cardiomyocytes. Heart formation in these mice was partially impaired, leading to cardiac abnormalities in almost half of the adult mice, and postnatal mortality was mostly caused by dilated cardiomyopathy.¹³ The response to direct cell loss earlier during development was studied in a genetic ablation model with the removal of up to 60% of cardiac progenitor cells and embryonic cardiomyocytes at E9.0. The depleted cardiomyocytes could be replaced by cardiac progenitor cells and, at least partly, by proliferating embryonic cardiomyocytes. In contrast to the previous study, cardiomyocyte depletion at E9.0 led to normal heart formation and fully functional hearts in adult mice.¹⁴

In neonatal mice, the regenerative response to cardiac injury is strongly limited by the extent of injury and age.11, 12 For example, it was demonstrated that surgical resection of hearts of neonatal mice at postnatal day 1 led to complete cardiac regeneration (compensatory growth), whereas the same procedure at postnatal day 7 resulted in fibrosis and consequently scar formation. New cardiomyocytes were shown to originate from proliferation of existing cardiomyocytes and regeneration was accompanied by extensive angiogenisis.¹¹ Similar to what has been observed in fetal cardiomyocytes, neonatal cardiomyocytes displayed disorganized sarcomeres, indicating their proliferative response.¹¹ Comparable results were obtained in studies with myocardial infarction leading to complete recovery in mice of 1 day of age, whereas scarring occurred in 7-day-old mice.¹² Regeneration in neonatal mice seems to largely depend upon the formation of new vessels and the re-induction of perfusion after myocardial injury.87, 101 Major cues for neoangiogenisis might originate from the inflammatory response mediated by neonatal macrophages.¹⁰¹ Whereas the adult inflammatory response after a myocardial infarction leads to fibrosis, the neonatal inflammatory response initiates angiogenesis and thereby facilitates regeneration of the injured cardiac tissue. Neonatal macrophages were demonstrated to be different in their transcriptional pattern and polarization as compared to adult macrophages.¹⁰¹

The brief period of time in which the neonatal heart can still recover from injury is termed the regenerative window, which in mice may be as short as 3 days, coinciding with the time frame in which at least a fraction of cardiomyocytes is still proliferating.¹¹ Interestingly, numerous studies have demonstrated that prolongation of the postnatal cardiomyocyte proliferative window can be achieved. For example, mice carrying a cardiomyocyte-specific knockout of the Hippo component Sav displayed enhanced regeneration by proliferating cardiomyocytes after receiving cardiac apex resection at postnatal day 8.¹⁰² Proliferation of neonatal cardiomyocytes was also increased when mice were compound deficient for SAV and DGC.⁹⁴ Sustained expression of the cardiac transcription factor GATA4, which is crucial for cardiac lineage specification during development but decreases within the first neonatal week, enhanced myocardial regeneration after injury.¹⁰³ Also, overexpression of Tbx20 increased cardiomyocyte proliferation and the regenerative response after myocardial infarction by repressing cell-cycle inhibitor genes such as Meis1 and p21.104, 105 Further, the expression of ErbB2 that stimulates cardiomyocyte proliferation during fetal heart development is markedly downregulated within the first postnatal week. Accordingly, expression of constitutively active ErbB2 during the postnatal period as well as administration of NRG1 facilitated prolonged cardiomyocyte proliferation and enhanced regeneration after myocardial infarction.76, 106

Perhaps a key question is if and for how long the regenerative window exists in humans. The higher rates of cardiomyocyte cell division detected in neonates² and young adults⁶⁹ could suggest the existence of a similar regenerative window. There is also some indication that hearts of children suffering from congenital heart defects may respond better to cardiac surgery than adults due to a greater regenerative capacity.¹⁰⁷ In line with this is a reported case of myocardial infarction in a newborn child with coronary artery occlusion. Within weeks after the infarction, the heart showed recovery, and long-term heart function and morphology was normal.¹⁰⁸ More studies are needed to assess the time span and extent of the regenerative window in humans, which might offer an important time window for therapeutic applications and corrective surgery in young patients born with a congenital heart defect.

Adult Hearts Fail to Regenerate

After the neonatal period, humans and mice show a limited rate of annual cardiomyocyte turnover of about less than 1%.1, 75 Over the last 2 decades, a contribution of resident cardiac stem cells to cardiac homeostasis and regenerative response has also been suggested.¹⁰⁹ Perhaps the most striking of these stem cell populations comprised the c-kit⁺ resident stem cells, suggested to provide a significant contribution to the formation of new cardiomyocytes after cardiac injury.¹¹⁰ In addition, in neonates, it was proposed that c-kit-expressing cardiac progenitor cells contribute to the generation of new myocardium after injury.¹⁰⁰ However, the contribution of c-kit⁺ cells to cardiac regeneration was lost in adults, c-kit⁺ cells adopting only a vascular fate instead.¹⁰⁰ Using lineage tracing, another study demonstrated that c-kit⁺ cells mainly differentiate into endothelial cells and rarely to the cardiomyocyte population.¹¹¹ Misinterpretations could arise from the fact that postnatal dedifferentiating cardiomyocytes112, 113, 114 and endocardial and epithelial cells¹¹⁵ have also been shown to express the surface marker c-kit, demonstrating the importance of accurate lineage tracing of progenitor cell populations before conclusions can be drawn. A recent study even suggested that c-kit⁺ cells might not be able to transdifferentiate in cardiomyocytes at all, as they demonstrated that c-kit⁺ cardiomyocytes originated from fusion with leukocytes.¹¹⁶ The present consensus is that in adult mammals, resident cardiac progenitor cells contribute only a very minor fraction (<0.01% per annum) to new cardiomyocytes during homeostasis and after injury.111, 117, 118 Similarly, extracardiac progenitor cell sources such as bone-marrow-derived stem cells contribute only marginally to homeostasis via fusion and transdifferentiation into cardiomyocytes.118, 119 These rates are therefore too low to replace the large numbers of cardiomyocytes that are lost after myocardial infarction, and it can be concluded that they do not significantly play a role in regeneration.

Future Perspectives: Turning on Cardiomyocyte Proliferation?

Currently, the consensus view is that cardiomyocyte proliferation occurs at extremely low rates in the homeostatic adult mammalian heart and that stem cell contribution to the heart muscle is rare. After cardiomyocyte loss during myocardial infarction, the remaining cardiomyocytes fail to proliferate, leading to heart failure and death. Causes of the postnatal cell-cycle withdrawal are complex and involve intrinsic as well as extrinsic signaling pathways. Fully unraveling these pathways may be necessary to achieve cardiac regeneration in adult hearts. The lessons learned from proliferation control during mammalian heart development and in fish heart regeneration may help us find ways to therapeutically induce cardiac regeneration in adults. Manipulating the activity of pathways that are active during cardiac development, such as the ErbB signaling pathway and the Hippo pathway, are currently among the most promising that may be harnessed for this purpose. However, YAP signaling is also a key component in the development and progression of cancer, and oncogenic risks could hinder translation to the clinic. The role of miRNAs in cardiomyocyte proliferation might be of translational value. However, as for other therapeutic molecules, specificity and side effects need to be assessed, and safe and efficient delivery methods established before these can safely be applied in the clinic. Moreover, it will be necessary to further identify and characterize the target population of potentially proliferative cardiomyocytes within the human heart that may differ between patients. Alternative approaches, such as direct reprogramming of cardiac fibroblasts into cardiomyocytes and other cell types to increase functional muscle mass, are also being explored.120, 121 Precise understanding of the developmental differentiation and epigenetic mechanisms will help to improve reprogramming efficiency and increase its clinical potential. New technologies, such as precise genome editing using CRISPR and single-cell RNA sequencing, provide promising tools to identify the molecular cues that would illicit a controlled and sufficient rate of cell-division and to investigate whether the newly formed cardiomyocytes are and remain fully functional. Furthermore, it remains to be established whether and to what extent we need to control the cell-type composition, including the endothelial cells and fibroblasts and coronary vasculature after successfully restoring the muscle cell mass.