The notion that mammalian hearts cannot fully regenerate in response to injury has been recently challenged [41]. Using a similar approach as in the newt and zebrafish, an apical resection of the mouse heart was performed (◘ Fig. 23.1). However, instead of performing the procedure on adult mice, the apical resection was performed on neonatal mice the day following birth [41]. Surprisingly, these neonatal mice survived a 15–20 % resection of the heart, and complete regeneration was observed without scar formation. Again, as observed in the newt and zebrafish, the genesis of many of the newly formed cardiomyocytes in the neonatal mouse was from preexisting cardiomyocytes. While this robust regenerative response could be observed during the first several postnatal days, by 7 days following birth, the heart was unable to fully regenerate and was noted to have extensive scar formation following apical resection. Therefore, it appears that the extensive regenerative capacity in neonatal mice is fully dependent on the retained ability of the neonatal cardiomyocytes to proliferate. By postnatal day 7, this ability is largely extinguished due to ongoing maturation and inability of the existing cardiomyocytes to reenter the cell cycle.

Moreover, these studies also determined the relative rapid regenerative response in neonatal mice. Adult newt and zebrafish require at least 30–60 days for complete cardiac regeneration in response to apical resection, but neonatal mice have already completed the regenerative process by day 21 [41, 42]. This may be due, in part, to the accelerated stimulation of cardiomyocyte proliferation, as the 7-day post-resection period is marked by a large number of cardiomyocytes that stain positive for either BrdU, a thymidine analog, or for phospho-Histone H3, a marker of the G2/M cell cycle phase .

One important difference that was uncovered between the apical resection at postnatal day 1 vs. 7 was the associated immune response [43]. This response included the magnitude and type of immune cells that responded to the inflamed myocardium, which was distinct and included increased numbers of monocytes and macrophages infiltrating shortly after myocardial injury. Further evidence for the importance of these immune cells was provided using a clodronate-mediated macrophage depletion experiment, which resulted in myocardial scarring in response to apical resection, even in neonatal mice. The extent of innate immunity as an important mediator of regeneration after myocardial infarction is unknown. However, it is noteworthy that galectin-3-expressing cells (a marker for macrophages) are present in high numbers in failing hearts, especially those that are fibrotic [44, 45].

The heart is one of the first organs to form during embryonic development [46]. A well-defined cascade of transcription factors and progenitor cells has been shown to be essential for cardiac development [47]. The necessity of the heart is highlighted as embryos without a properly developed heart will not survive beyond mid-gestation (mice) or the first trimester (human) [48]. The heart is a derivative of the mesodermal germ layer that receives cues (i.e., growth factors) from adjacent lineages (i.e., endodermal derivatives). Before specification of the heart begins, a number of factors are activated to initiate migration of cells that emerge from the primitive streak and populate the heart fields (◘ Fig. 23.3) [47]. The factors that regulate migration and specification during this developmental stage are not completely defined, but from zebrafish it appears that retinoic acid signaling, Wnt signaling, Notch signaling, and other pathways collectively function to specify fates and organogenesis [49–53]. Additional factors regulating mesodermal specification include Nodal and BMP as they signal to activate Brachyury and Flk1 [54]. These factors modulate mesodermal progenitors to differentiate to form hemangioblast precursor cells under immediate specification by Etv2 [55, 56].

The second wave of mesodermal progenitors, under the influence of reduced Wnt/β-catenin signaling, activates the earliest marker of cardiac progenitor cells, Mesp1 [57, 58]. However, Mesp1 is a broadly expressed mesodermal marker, and genetic lineage-tracing studies using a Mesp1-Cre transgenic mouse model label all mesoderm-derived cells in the murine heart as well as a number of other noncardiac mesodermal derivatives [57]. Mesp1 is expressed during gastrulation, when progenitor cells are migrating from the primitive streak to form the splanchnic mesoderm, where they eventually fuse together in the midline to form the cardiac crescent .

At this stage of development, cell fates are still pliable as genetic deletional studies (i.e., global loss of Etv2) in mice result in increased numbers of cardiac progenitors. During formation of the cardiac crescent, two distinct populations of cardiac progenitors exist—those residing in the first heart field and those in the second heart field (◘ Fig. 23.3). The main difference that distinguishes the two populations is the level of exposure to different growth factors. The cells in the first heart field are exposed to higher levels of BMP and FGF signaling, while Wnt signaling is inhibited. The combined action of these morphogens results in cardiac differentiation of the first heart field progenitors as they begin to express cardiac markers, including Nkx2-5 and Gata4 . Given the position in relation to the aforementioned morphogens, the progenitors in the second heart field remain undifferentiated and begin to express the second heart field marker, Isl1.

Following the onset of Isl1 expression, the second heart field progenitors begin to differentiate and express Nkx2-5 and other cardiac markers. As a portion of the second heart field, a specialized group of cells in the posterior aspect of the second heart field will develop and become the pro-epicardium under the influence of BMP and FGF signaling [59]. These signaling pathways initiate Twist expression with subsequent activation of Wt1, Tbx18, scleraxis, and semaphorin 3D. These pro-epicardial cells will envelope the heart to ultimately form the epicardium. Epicardially derived cells will undergo epithelial-to-mesenchymal transformation (EMT) to invade the heart and differentiate into smooth muscle cells, fibroblasts, and cardiomyocytes [60].

An additional source of progenitor cells that contribute cells to the developing heart is derived from the cardiac neural crest [61]. These ectodermal cells are generated as neural crest, but later migrate under the influence of semaphorins to the outflow tract of the heart, where they are responsible for septation of the aorta and pulmonary artery and will give rise to vascular smooth muscle cells that line the large arteries [62, 63]. In addition, these cardiac neural crest cells provide the neurons that enable parasympathetic innervation of the heart [64]. While the cardiac neural crest was shown to contribute cardiomyocytes to the developing zebrafish heart, this may not be observed in the mammalian heart [61, 65–67]. Despite the abundant presence of these sources of progenitor cells that together form the heart during development, it is unclear whether they also contribute to newly formed cardiomyocytes in the adult mammalian heart following injury .

Most adult tissues harbor somatic stem cells that aid in ongoing maintenance and regeneration following injury [5]. Some of these somatic stem cells have a developmental origin, such as the hematopoietic stem cell, which is derived from fetal liver hematopoietic stem cells that invade the bone marrow niche during the later stages of murine development [68]. After arrival in the bone marrow niche, the majority of these hematopoietic stem cells become quiescent and are reactivated in response to distinct cues [69]. Similarly, the adult satellite cells that form the somatic stem cells for skeletal muscle are derived from the same dermomyotome as the developmental muscle progenitors that give rise to embryonic myoblasts [70]. Recent evidence suggests that the heart may also harbor a limited number of these developmentally derived progenitor cells [6]. It is unclear whether these progenitors are cells that were never fully differentiated but are already fully committed to become cardiomyocytes or whether they are uncommitted progenitor cells that continue to have the capacity to give rise to all cell types in the heart [71].

During development, Nkx2-5 is the most specific and widespread marker of progenitor cells that will ultimately generate cardiomyocytes [72, 73]. Using a Nkx2-5-GFP reporter transgenic mouse line, it was determined that these developmental progenitors represent a bipotential pool of uncommitted cells that give rise to cardiomyocytes and smooth muscle cells [74]. Surprisingly, a number of these uncommitted progenitor cells resided in the postnatal heart, with numbers decreasing from 5 % at 1–2 weeks of age to less than 0.3 % of nonmyocytes at 8 weeks [75]. Recent studies have identified a Tgf-β signaling pathway that regulates the proliferation of these progenitor cells, both in vitro and in vivo [76]. Specifically, the inhibition of Tgf-β family receptors Alk4, Alk5, and Alk7 significantly enhanced both the numbers of Nkx2-5 progenitor cells and their differentiation toward cardiomyocytes. Moreover, the inhibition of Tgf-β receptor 1 enhanced cardiomyoblast proliferation and differentiation in vivo following myocardial infarction and improved cardiac function [76]. To what extent new cell formation was responsible for the improvement in cardiac function is not clear. Nevertheless, these exciting findings provide a platform for using embryonically derived progenitor cells that remain in the adult heart for regenerative strategies.

A second population of embryonic progenitors express the transcription factor Islet1 (Isl1) [77]. During development, the Isl1-expressing progenitors are generated in the second heart field. While recent data suggest a more widespread expression pattern for Isl1, its role in the first heart field is less prominent. The phenotype of Isl1 null mice, which lack a right ventricle and outflow tract, emphasizes the importance of this factor in the second heart field [78, 79]. Since Isl1-expressing cells are essential for the formation of the right ventricle and the outflow tract, they are important developmental progenitor cells . Interestingly, not all of the Isl1-expressing progenitor cells differentiate during development [80]. The majority of progenitor cells differentiates during cardiac development and can no longer be identified by Isl1 expression at the time of birth , but a limited number of cells remain positive for Isl1 expression (500–600 per rat heart). Isl1⁺ progenitor cells have the potential to differentiate into endothelial cells and smooth muscle cells, as well as cardiomyocytes in vivo and in vitro. Importantly, both in rodents and in humans, postnatal Isl1⁺ cardioblasts have been identified and shown to have the same capacity to differentiate into different cardiac lineages as embryonic Isl1⁺ progenitor cells. Moreover, a recent study identified a small molecule that enhanced the renewal of Isl1⁺ progenitor cells [81]. Importantly, these studies defined a triphasic role for Wnt signaling during cardiomyocyte differentiation. That is, prior to specification of cardiac progenitor cells, Wnt signaling is inhibited to allow a maximal number of progenitor cells. During the progenitor cell phase, when these cells are actively proliferating, Wnt is stimulated to maximize the number of progenitors. But immediately thereafter, Wnt signaling is once again inhibited; otherwise, cardiomyocyte differentiation will be suppressed. This time-sensitive dependency on Wnt signaling has been used during embryonic stem cell and induced pluripotent stem cell differentiation to generate cardiomyocytes [82–84]. Although these findings have uncovered a critical dependency on Wnt signaling during cardiac progenitor cell proliferation, and while the Isl1⁺ progenitor cells display multi-lineage differentiation potential, the overall number of Isl1⁺ progenitors that reside in the postnatal heart is most likely too small to produce significant cardiac regeneration, even if the numbers are enhanced in response to Wnt signaling [80, 85].

A third population of embryonic progenitors that could potentially be employed for cardiac regeneration is the epicardial-derived progenitor cells [30, 86, 87]. As previously outlined, during development, epicardial progenitor cells contribute limited numbers of cardiomyocytes to the developing heart. This was clearly shown using Wt1 as an epicardial-specific marker [30]. Genetic lineage-tracing studies demonstrated that Wt1-derived cells generate cardiomyocytes during development. Moreover, it was shown that Wt1 cells have the ability to contribute cardiomyocytes during post-myocardial infarction cardiac remodeling [31]. While the levels of new cardiomyocyte formation were not quantified, the Wt1 lineage was shown to contribute cardiomyocytes. However, this contribution only occurred when the heart was pretreated (before initiation of myocardial infarction) with thymosin β4. Subsequent studies that attempted to replicate these findings could not identify newly established cardiomyocytes in the absence of thymosin β4 priming [31, 32].

A plenitude of transcription factors has been studied during cardiac development for their role in cardiac morphogenesis [88, 89]. Many of these cardiac transcription factors can cause congenital heart defects when perturbed or mutated [90]. For example, mutations in the transcription factors Gata4 and Nkx2-5 can cause septal defects (atrial, ventricular, or combined) or even tetralogy of Fallot [91]. As previously noted, Nkx2-5 is a critical transcription factor during cardiac development and is one of the first transcription factors that specify cardiac progenitor cells [73]. Gata4 becomes expressed in cardiomyocytes early after Nkx2-5 and is also a critical transcription factor for cardiomyocytes [92, 93]. Genetic deletion of Nkx2-5 or Gata4 in mouse embryos results in cardiac developmental defects and lethality early after formation of the heart [72, 92, 93]. A third transcription factor that has a crucial role during cardiac development is Tbx5 [94]. Again, deletion of this gene in mouse embryos results in embryonic lethality soon after formation of the heart [95]. Mutations in Tbx5 result in Holt-Oram syndrome and are associated with a combination of atrial septal defects and digit abnormalities [96].

Although all of these transcription factors are critically important for cardiac development, their role during cardiac regeneration is not clear [97]. The most prominent strategy that uses these transcription factors for regenerative purposes aims to transdifferentiate fibroblasts to cardiomyocytes [98]. This is performed by overexpression of a set of at least three transcription factors in combination, including Gata4, Mef2c , and Tbx5 , as discussed later in this chapter.

A gene network that has gained more attention in recent years for its potential role in regeneration is the Hippo signaling pathway [99]. This pathway was originally described in Drosophila and was shown to be important for the determination of organ size during development [100]. Although the upstream activator of this kinase-signaling cascade has not been firmly established, it is clear that the activation of Hippo (Mst1/2 in mammals) leads to phosphorylation and activation of the downstream kinase, Warts (Lats1/2 in mammals), which in turn can phosphorylate and inactivate Yorkie (Yap/Taz in mammals) [99]. During development, it was determined that perturbation of this signaling cascade leads to increased heart size due to enhanced and persistent proliferation of cardiomyocytes [101]. These findings are consistent with previous results in Drosophila, and also in other organ systems, indicating that the Hippo signaling pathway is involved in regulating organ size. While the precise mechanisms of regulation are not entirely clear, it has been shown that Hippo signaling could affect both differentiated cells, as well as progenitor cells [99]. Thereby, the overall effect of uninhibited Hippo-mediated gene expression could be cumulative, resulting in more progenitor cells and more differentiated cells.

Importantly, the developing heart regulates its size via both mechanisms. However, to the extent Hippo signaling could also play a role in the adult heart was not entirely clear.

The downstream effectors of the Hippo signaling pathway in mammals are Yap and Taz. To determine the importance of Yap in the developing and postnatal heart, conditional deletion cardiac mutants were generated, resulting in early postnatal lethality, indicating important roles for Yap in the heart [102]. In addition, transgenic strategies were used to express a constitutively active form of Yap in the adult heart [103, 104]. Interestingly, the activation of YAP stimulated cardiomyocyte proliferation and enhanced cardiac regeneration after experimental myocardial infarction. This is likely due to the direct effect on cardiomyocyte proliferation, since neonatal mice can completely recover from an experimental myocardial infarction due to the capacity of neonatal cardiomyocytes for proliferation [41, 105].

One week after birth, when neonatal cardiomyocytes no longer proliferate, experimental myocardial infarction no longer results in complete regeneration, but rather in scar formation [41, 42]. Importantly, activation of YAP was capable of inducing cardiac regeneration both 1 week and 4 weeks after birth, circumventing the proliferative block of most postnatal cardiomyocytes. These results were further confirmed using adeno-associated virus (AAV)-mediated activation of Yap in adult mice [103]. Finally, these results have been replicated and expanded to include additional regulators of the Hippo pathway, including the miR302-367 cluster as critical regulators of cardiomyocyte proliferation and cardiac regeneration [106].

The heart consists of multiple cell types, and, during homeostasis, all of these cell populations have low turnover rates [107]. The regulation of cellular turnover in the heart is unknown. Whether a stem cell population exists that orchestrates these proliferative events depending on demand or whether they are individually regulated by dedicated progenitor cells is an area of intense interest [6, 71, 108, 109]. Histological examination of mitotic figures has identified differences in the normal turnover rates, not only of certain cell types, such as cardiomyocytes vs. endothelial cells, but also involving anatomically distinct regions of the heart [11, 12]. For example, the atria appear to be more amenable to proliferative stimuli than the cardiac ventricles. For most cell types, the assessment of mitotic figures or even incorporation of DNA nucleosides such as BrdU or EdU is sufficient to assess their turnover rate. However, cardiomyocytes have a peculiar characteristic that allows them to undergo endoreduplication, which increases ploidy per nucleus by undergoing DNA synthesis without mitosis, or by completing mitosis to generate two nuclei without undergoing cytokinesis [6, 110]. The mechanisms underlying this characteristic are not clear, and distinct species differences exist regarding the abundance of binucleated vs. mononucleated cardiomyocytes [110, 111]. To assess new cardiomyocyte formation, a number of recent technological advances have enabled an assessment of the turnover rates of cardiomyocytes.

An innovative experimental approach was undertaken in 2009 to evaluate cellular proliferation and turnover in the adult heart [112] (◘ Fig. 23.4). Radiocarbon dating is a well-known technology used to determine the age of biological materials [113]. ¹⁴C is generated in the atmosphere, incorporated into plants by photosynthesis, and constantly exchanged during the life of animals, plants, or humans who ate the animals or plants. After plants or animals die, the amount of ¹⁴C slowly decays over time at a constant rate. This constant rate is used to determine age by comparing the atmospheric levels of ¹⁴C to the measured levels in the tested sample.

This technology has been used reliably to determine the age of fossils. Recently, this approach was adopted to determine the age of individual human cardiomyocytes. During the Cold War, aboveground nuclear testing released a pulse of ¹⁴C into the atmosphere. This pulse was finite as all aboveground nuclear tests were eliminated following the Nuclear Test Ban Treaty in 1963. Therefore, this ¹⁴C atmospheric pulse allowed investigators to determine the age of human cells [114]. Initially, the investigators established a protocol to distinguish cardiomyocyte nuclei from non-cardiomyocyte nuclei based on troponin expression [115]. In addition, the analyses were performed to account for increases in nuclear ploidy.

Using this strategy, the overall rates of renewal as detected by ¹⁴C dating were relatively low, but measurable [112] (◘ Fig. 23.4). About 1 % new cardiomyocytes were formed per year, which leads to more than half of the heart being replaced over the average lifetime of humans. Although yearly renewal rates may be low, the total amount of renewed myocardium over a lifetime is likely to be important for cardiac function. Therefore, these studies emphasized that any perturbation in the regenerative capacity of the heart may lead to a reduction in cardiac function. Importantly, these findings provide convincing evidence of cardiomyocyte turnover and ongoing renewal of the heart, albeit at low levels [116].

The rates established by ¹⁴C dating of individual cardiomyocyte nuclei in humans were recently reproduced in mice using a separate strategy. Here, a novel approach was used to feed mice thymidine, a nucleoside that is incorporated into DNA during DNA synthesis and cellular proliferation [117]. However, instead of feeding mice regular thymidine, the authors used ¹⁵N thymidine , an isotope that is relatively rare in nature (0.3 %). An advanced imaging modality (multi-isotope imaging mass spectrometry) was used to quantify the ratio between ¹⁵N and ¹⁴N. Using this strategy, any signal that exceeded the natural ratio was clearly visible [117]. This method was first validated using the kinetics of intestinal crypt stem cells that continuously cycle to produce new cells. Further, the method was used to falsify the immortal strand hypothesis, which suggested that new DNA strands were preferentially segregated in 1 daughter cell [117]. When the investigators applied this new imaging modality to quantify the numbers of cardiomyocytes that undergo proliferation at baseline and following injury, they noted that the total proliferation rates were very low, with calculated renewal rates of just below 1 % per year during normal aging [109].

The ability to quantify cardiomyocyte renewal is difficult due to the complex nature of cardiac tissue [6, 118, 119]. The sizes of cells are not uniform, and non-cardiomyocyte nuclei are in close proximity to cardiomyocytes. Furthermore, the non-cardiomyocyte fraction has a greater tendency to undergo proliferation compared to cardiomyocytes [107]. These characteristics make it difficult to assign a given nucleus to a cardiomyocyte in histological sections. To correctly identify cardiomyocyte nuclei, a genetic mouse model that specifically expresses a nuclear localized β-galactosidase protein in cardiomyocytes was used to precisely quantify the cardiomyocyte renewal rates at baseline and following injury [119, 120]. This genetic mouse model was combined with injection of tritiated thymidine to precisely measure DNA synthesis in cardiomyocytes [120].

The rates reported using this genetic mouse model were very low, which reflected that they were established from a single injection of thymidine. However, when these rates were extrapolated to yearly renewal rates, they were estimated to be about 1 % per year. One important caveat, however, is the observation that cardiomyocytes can undergo endoreduplication, which results in positive nuclei that are not indicative of cardiomyocyte renewal. It is further appreciated that, in response to pathological stimuli such as hypertension or following myocardial infarction, there would be a hypertrophic response of cardiomyocytes with low-level stimulation of proliferation resulting in increased ploidy [110, 111, 121].

A clear positive correlation was shown between cardiomyocyte size and DNA content. Yet any increase in DNA synthesis does not necessarily mean an increase in the number of cardiomyocytes. Quantification of the regenerative response following an injury was even more difficult due to increased numbers of immune cells invading the heart that were often positive for incorporated DNA nucleosides such as BrdU or ³H-thymidine. All these factors complicated precise quantification of new cardiomyocyte formation [6, 121].

With advances in genetic manipulations in mice, a new strategy was devised to more accurately assess whether new cardiomyocytes are actively being formed by existing cardiomyocytes or by non-cardiomyocytes. To that end, a cardiomyocyte-specific inducible Cre line was combinatorially mated to a double reporter that expressed β-galactosidase in all cells at baseline, but GFP upon recombination [108]. After induction of recombination in adult mice, a maximum of 85 % of all cardiomyocytes was recombined to express GFP.

The hypothesis that was being tested stated that if cardiac progenitor cells were important contributors of new cardiomyocytes, then there would be an increase over time of non-GFP-expressing cardiomyocytes, since the Cre recombinase was only induced in already existing cardiomyocytes. So any newly generated cardiomyocytes from progenitor cells would not express GFP . However, no statistically significant increase was noted up to a year after labeling, indicating no contribution of progenitor cells to new cardiomyocyte formation during normal aging in mice [108]. In response to injury, however, a significant increase in the number of GFP ⁻ cardiomyocytes was noted, with as much as 15 % new cardiomyocytes being generated from non-cardiomyocytes.

Although these data were initially interpreted as unequivocal evidence of new cardiomyocyte formation by progenitor cells, the investigators of this study reinterpreted their data in light of the fact that the majority of proliferative events that gave rise to new cardiomyocytes were actually not occurring in GFP⁻, but in GFP⁺ cardiomyocytes [109]. Therefore, no good explanation exists for the observed increase in GFP⁻ cardiomyocytes. If they were progeny from a progenitor cell pool, this pool was most likely depleted due to lack of ongoing renewal, given the absence of DNA duplication events in GFP⁻ cardiomyocytes. Other lines of evidence suggest that both cardiomyocytes and non-cardiomyocytes could contribute new cardiomyocytes to the adult heart. For example, recent evidence suggested that cardiac cell therapy—where cells were injected (either intravenous or intracardiac) following myocardial infarction—resulted in the activation of endogenous progenitor cells that were stimulated to generate new cardiomyocytes [122, 123].

Another line of evidence provided direct proof that cardiomyocytes were capable of generating new cardiomyocytes in the adult mouse heart [124]. In these studies, the authors used a genetic mouse model that exchanged reporter cassettes between sister chromosomes during mitosis. This strategy enabled the direct visualization of newly formed cells that had completed cytokinesis, although the labeling efficiency was difficult to assess. The results showed that new cardiomyocytes were being formed from existing cardiomyocytes. The main discrepancy with previous findings was the lack of increased new cardiomyocyte formation in response to myocardial infarction.

The data from these studies provide convincing evidence that existing cardiomyocytes can give rise to new cardiomyocytes during normal aging. The majority of data suggest that injury such as myocardial infarction also induces new cardiomyocyte formation, although there are some conflicting data. Importantly, the reported rates of new cardiomyocyte formation from existing cardiomyocytes are mostly in agreement with each other and suggest cardiomyocyte turnover occurs at a rate of about 1 % per year in both mice and humans [109, 112]. While this low level of ongoing renewal may not be physiologically relevant for cardiac function in the mouse (which has a lifespan of 2–3 years), it would have a major impact in humans, with an average lifespan of over 70 years. This ongoing renewal may very well be extremely important to maintain normal cardiac function.

In addition to cardiomyocytes serving as a source for generating new cardiomyocytes, as discussed in the previous section, there is convincing evidence that the adult heart contains a limited number of uncommitted progenitor cells [27, 28, 71, 125] (◘ Fig. 23.5). These progenitor cells can be extracted from the heart, mostly based on marker gene expression, and maintained in their undifferentiated state in cell culture [27, 28]. Upon activation, these cells undergo differentiation both in vitro and in vivo to give rise to new cardiomyocytes, endothelial cells, and smooth muscle cells. The developmental origin of these progenitor cells is not known, and a number of different markers have been used to identify progenitor cells, with limited overlap between these populations [71].

The most extensively studied marker of cardiac progenitor cells is the tyrosine kinase receptor, c-kit [126] (◘ Fig. 23.5). This receptor is a well-known marker of hematopoietic progenitor cells, both in mice and humans [127–129]. It is also expressed on Leydig cells in the testes, on melanoblasts and melanocytes in the skin, the interstitial cells of Cajal in the intestine, etc., where the exact function of c-kit-expressing cells is incompletely defined [130–132]. Based on the hypothesis that states that markers of progenitor cells in one tissue also mark progenitor cells in another tissue, c-kit cardiac progenitor cells were identified in 2003 [27]. These cardiac progenitor cells were isolated based on the absence of the panhematopoietic lineage marker, CD45, and the presence of c-kit (lin⁻c-kit⁺). In culture, these lin⁻c-kit⁺ CPCs could be maintained in a proliferative undifferentiated state, they were clonogenic, and, importantly, they could be differentiated into all three major cardiac cell lineages: cardiomyocytes, endothelial cells, and smooth muscle cells. Furthermore, upon injection of undifferentiated CPCs into the infarcted rodent heart, they readily differentiated and improved cardiac function and repaired the infarct by approximately 70 %.

The existence of uncommitted c-kit⁺ CPCs in the rat has been confirmed in other species, including mouse, dog, pig, and humans [133–135]. The isolation procedures vary somewhat and range from the immediate isolation using magnetic sorting beads to obtain lineage-negative, c-kit-positive cells, to the plating of isolated non-cardiomyocytes overnight before the selection of lineage-negative, c-kit-positive cells [27, 136, 137]. A strong argument for the progenitor status of c-kit CPCs is that they can be grown (or expanded) clonally in an undifferentiated state and then differentiated to the three main cardiac lineages: cardiomyocytes, endothelial cells, and smooth muscle cells [27, 133, 137].

A number of studies have identified factors that have a role in forming the niche for c-kit progenitor cells in the heart [138–140]. Since these progenitor cells are capable of both proliferation and differentiation, their control should be governed by intrinsic and extrinsic cues to guide cell fate decisions. In the bone marrow, supporting cells that interact with the hematopoietic stem cells provide these cues and maintain them in a quiescent state [141, 142]. While the precise requirements for the hematopoietic niche are also incompletely defined, recent evidence suggests tight interactions with a single mesenchymal stromal cell, together with supporting endothelial cells, are critical for the formation of a niche and to provide cues for cell cycle reentry and cell division, ultimately leading to either self-renewal and/or differentiation [141, 143].

The major benefit of the hematopoietic system is the availability of a depletion strategy followed by transplantation and homing of stem cells to the hematopoietic niche [144]. This, however, is not possible in the heart. Therefore, it has been more difficult to identify factors that regulate the cardiac stem cell niche [145]. Although the stem cell niche in the heart is less well defined, a number of factors are known to play important roles. First, a number of studies have characterized the cardiac progenitor cell niche and consistently find fibronectin present in the extracellular matrix surrounding the progenitor cells [27, 146]. Second, to the extent there is an interaction with supportive cells, as has been described for the hematopoietic stem cell niche, these cell-cell interactions in the cardiac progenitor cell niche are unclear. Several reports have shown connexin43 expression on CPCs and an interaction with the surrounding myocardium and other cells, but how these interactions regulate CPCs is not known [139, 147]. One potential mechanism is the exchange of genetic material, such as miRNAs through gap junctions. Active exchange of miRNAs to guide differentiation decisions has been demonstrated between cardiomyocytes and CPCs [148]. However, these experiments were performed in cell culture, with no in vivo proof of such an exchange mechanism.

In addition to niche factors that may promote CPC proliferation or differentiation, it is hypothesized that various signaling pathways are likely to have a role in CPC differentiation [53, 149–151]. One of the most prominent pathways suggested to have a role is the Notch signaling pathway. Notch signaling has been described as an important receptor for proliferation and differentiation in other organ systems [152, 153]. For example, in hematopoietic stem cells, the activation of Notch signaling initiates erythroid differentiation. In other organs, Notch signaling has been shown to be important in regulating cell fate and cellular proliferation [153]. In the heart, Notch signaling has been shown to initiate cardiomyocyte differentiation while repressing vascular differentiation [150].

The molecular mechanism underlying these events involves direct binding on the internally cleaved active Notch receptor domain to RBP-Jk, which in turn binds to the Nkx2-5 promoter to initiate cardiomyocyte differentiation. Moreover, the inhibition of Notch signaling in neonatal mice was sufficient to cause dilated cardiomyopathy due to reduced numbers of cardiomyocytes, suggesting a crucial role of Notch signaling in the maturation of the early postnatal heart.

A second marker of CPC differentiation that initiates vascular differentiation is the expression of the vascular endothelial growth factor (VEGF) receptor, Kdr [154]. In both human and mouse CPCs, the expression of Kdr coincided with CPCs that would adopt vascular endothelial fates, while Kdr-negative CPCs showed a higher tendency to differentiate toward a cardiomyocyte fate .

Although these initial studies have uncovered some potentially important signaling events, the lack of genetic models to evaluate the impact of signaling networks specifically in CPCs has limited our understanding of the pathways that govern CPC lineage specification . Recently, however, a mouse model was engineered where Cre recombinase was knocked into the murine Kit locus. This genetic mouse model was initially used for genetic lineage tracing to determine the extent that CPCs generate differentiated cells in the developing and adult heart [155]. Surprisingly, CPCs contributed minimally to the generation of cardiomyocytes, even after myocardial injury, while the main cell type generated by c-kit CPCs was endothelial cells. These genetic mouse models will prompt more mechanistic studies focused on cell-autonomous factors within CPCs that drive cell fate decisions. However, to date, the exact mechanisms regulating cardiac progenitor cell proliferation and differentiation remain largely unknown.

In addition to c-kit-expressing CPCs, other markers have been used to isolate cardiac progenitor cells. Sca1 expression has been used to identify an alternative progenitor cell population in murine hearts [28]. The expression of Sca1 and c-kit resembles expression of these two markers on hematopoietic stem cells, although the overlap in expression of these two markers on cardiac progenitor cells is limited. Sca1 is expressed in a greater number of cells than c-kit, and most of these are vascular or perivascular cells. Based on published literature, the ability of murine Sca1-expressing CPCs to differentiate into cardiomyocytes is more limited than c-kit⁺ CPCs [27, 28].

Given the widespread expression of Sca1 in the heart, not all of these cells are considered progenitor cells. The Sca1-expressing cells that do not express the endothelial marker Pecam1 (CD31 ) are typically considered CPCs. These Sca1⁺CD31⁻ CPCs can be isolated, cultured under undifferentiated conditions, and have the ability to differentiate into cardiomyocytes upon injection into the infarcted mouse heart. Moreover, the deletion of Sca1 has been shown to negatively impact cardiac function and is associated with increased Wnt signaling, suggesting an important role for endogenous Sca1 progenitor cells in normal cardiac performance [156].

More recently, an attempt to perform genetic lineage tracing of Sca1 cardiac progenitor cells indicated cardiomyocyte differentiation from Sca1 CPCs [157]. However, the genetic strategy that was used may have overestimated the abundance of cardiomyocyte differentiation due to more widespread expression of the genetic driver than endogenous Sca1 expression [6]. Whether humans express Sca1 is widely debated, but we know the Sca1 antibody has been used to isolate human cardiac cells [158, 159]. Upon isolation, these human cardiac cells appear to be progenitor cells with the ability to proliferate and differentiate into cardiomyocytes. More importantly, these human cardiac progenitor cells promote cardiac repair upon transplantation into mouse hearts post-myocardial infarction and differentiate into cardiomyocytes [159]. To date, no clinical trials have been performed using Sca1⁺ cardiac progenitor cells.

A third source of endogenous cardiac progenitor cells is not based on the expression of a single marker, but the extrusion of the DNA dye, Hoechst 33342, in a subset of cardiac cells (◘ Fig. 23.5). It allows for the isolation of these cells using flow cytometry (fluorescence-activated cell sorting, FACS) [160]. These cells sort to the side of the main population, hence their designation as side population (SP) cells. A number of ABC (ATP-binding cassette) transporter proteins (also referred to as multidrug resistance proteins) have been identified as conferring the capacity to extrude the DNA dye, most notably Abcg2 and Mdr1 [161].

The abundance of SP cells in the heart is relatively low, and, typically, multiple hearts are pooled to isolate cardiac SP cells. SP cells express other markers, such as Sca1 and CD31, supporting the notion that SP cells are a specialized subpopulation of Sca1 CPCs [162, 163]. When isolating cardiac SP cells, they are maintained in an undifferentiated state and can be made to differentiate toward cardiomyocyte fates by addition of oxytocin or the histone deacetylase inhibitor trichostatin A or by coculture with adult rat ventricular cardiomyocytes [67, 125, 162, 164, 165]. Transplanted SP cells can home to the infarcted heart and differentiate into multiple cardiac lineages, such as cardiomyocytes, endothelial cells, and smooth muscle cells [166]. Given the lack of a clear molecular marker for side population cells, there is currently no proof of the contribution of SP cells, in vivo, to promote endogenous cardiac repair [6, 161].

In recent years, the epicardium has received intense interest as an important source of progenitor cells (◘ Fig. 23.5). Especially during development, it was shown that the epicardium contributes multiple cell types to the developing heart, including cardiomyocytes, fibroblasts, and smooth muscle cells [30, 167]. The source of these various cell types can be traced by different epicardial markers, including Wt1 and Tbx18 [30, 86, 87]. During development, these epicardial progenitor cells contribute various cell types to the heart, but the extent to which this occurs in the adult injured heart is unclear.

A recent study identified Wt1-derived cells as epicardial-generated cardiomyocytes after myocardial infarction [31]. However, a different study was unable to confirm these results using the same mouse model to perform genetic lineage tracing [32]. The only difference between these two studies was the administration of thymosin β4 before the onset of myocardial infarction—in the case when Wt1-derived cardiomyocytes were detected. The exact function of thymosin β4 in this context is not clear, but, previously, thymosin β4 was shown to ameliorate cardiac remodeling through the activation of integrin-linked kinase and Akt and to support survival of cardiomyocytes in culture [168].

Although the cells responsible for cardiac repair may no longer express the epicardial marker gene in the adult, it is still conceivable that the epicardium deploys a number of progenitor cells during development that, upon injury, can be activated to initiate cardiac repair. Evidence to support this notion used strategies similar to the ones used to analyze hematopoietic cells. It was reported that the heart contains mesenchymal cells that can form colonies similar to bone marrow mesenchymal cells [169]. These colonies are formed by undifferentiated cells that can differentiate into many cardiac cell types, including cardiomyocytes, endothelial cells, and smooth muscle cells. The extent to which these pro-epicardial-derived, colony-forming mesenchymal cells contribute to cardiac repair in the adult heart remains an unsettled question .

Although cardiac progenitor cells reside in the adult heart, their regenerative capabilities are insufficient to repair the damaged heart following a severe insult such as a myocardial infarction [170]. Regeneration with cardiomyocytes and a vascular support system would be preferable, but the post-injured heart typically is marked by scar tissue and fibroblasts [171, 172]. One clever strategy takes advantage of these large areas of fibrosis in an attempt to convert the fibroblasts into cardiomyocytes (◘ Fig. 23.6).