Stimulating Cardiogenesis as a Treatment for Heart Failure

Abstract

Following myocardial injury, cardiomyocyte loss cannot be corrected by using currently available clinical treatments. In recent years, considerable effort has been made to develop cell-based cardiac repair therapies aimed at correcting for this loss. An exciting crop of recent studies reveal that inducing endogenous repair and proliferation of cardiomyocytes may be a viable option for regenerating injured myocardium. Here, we review current heart failure treatments, the state of cardiomyocyte renewal in mammals, and the molecular signals that stimulate cardiomyocyte proliferation. These signals include growth factors, intrinsic signaling pathways, microRNAs, and cell cycle regulators. Animal model cardiac regeneration studies reveal that modulation of exogenous and cell intrinsic signaling pathways can induce re-entry of adult cardiomyocytes into the cell cycle. Using direct myocardial injection, epicardial patch delivery or systemic administration of growth molecules, these studies show that inducing endogenous cardiomyocytes to self-renew is an exciting and promising therapeutic strategy to treat cardiac injury in humans.

Current heart failure treatments

Heart failure (HF), in which the heart fails to pump sufficient blood to meet the body’s need, is the leading global cause of death¹. There are currently estimated 38 million patients suffering from HF worldwide, and this number continues to rise². HF mortality rates are high, and currently about half of HF patients die within 5 years of diagnosis. Heart failure is a complicated pathological process initiated by heart injury, which arises most commonly from ischemic heart disease associated with coronary vascular disease^{3, 4}. In addition, HF can result from chemotherapy exposure and other forms of cardiomyopathy, such as congenital and viral cardiomyopathies. Due to modern medical advancements, survival rates for pediatric congenital heart disease (CHD) patients have increased in recent years, resulting in more patients surviving into adulthood. Unfortunately however, a significant number of adult CHD patients have shown a dramatic increase in HF risk^{5, 6}. Although current treatment options are useful in the management of HF, there remains an overwhelming need to develop a definitive modality to reverse HF and restore function of damaged tissue.

Pharmacological treatment

Depending on the symptoms that HF patients present, they may require only one or a combination of drugs. Currently available drugs target a symptom of the cardiovascular system, or a contributing factor to HF. Major classes of available drugs include: 1) drugs that help to reduce blood pressure by reducing afterload, such as angiotensin-converting enzyme (ACE) inhibitors and their alternative Angiotensin II receptor blockers, isosorbide dinitrate, beta-blockers, and hydralazine hydrochloride, 2) drugs that reduce blood volume and thus decrease cardiac workload, such as aldosterone antagonists and diuretics, and 3) drugs that increase cardiac contractility, improve cardiac pumping function and slow heart rate, such as digoxin⁷. Use of these drugs is helpful to prolong the life of HF patients improve cardiac function and reduce symptoms. However, none of these can efficiently reverse heart failure⁷.

Heart Transplant and Left Ventricular Assist Devices (LVAD)

Patients with severe HF require implantation of a left ventricular assist device (LVAD) or transplantation of an entire donor heart. The only definitive treatment for HF is heart transplantation—an option that is severely limited by the scarcity of donor hearts. From 1987 to 2012, 40,253 people were on a waiting list for a heart transplant, yet only 26,943 received a transplant⁸. The LVAD, a mechanical pump that is surgically implanted into the left ventricle of the failing heart, has been commonly used as a bridge to heart transplantation. Due to the lack of available donor hearts, the number of LVAD implantations has largely increased, which has improved survival rates of HF patients⁸. However, LVAD implantation has also resulted in increased costs, with a $234,808 mean cost of LVAD-related hospitalizations in 2011⁸. Although current HF treatments such as drug therapies, LVAD implantation, and heart transplantation have value in improving heart function, hospitalized HF patients still experience very high post-discharge mortality and readmission rates, a number which has not decreased in the last twenty years¹. Thus, there is an urgent need to develop new HF therapies.

Cell transplant treatment

In addition to clinical trials, considerable basic and translational research has been conducted with the aim of developing a cell transplantation-based treatment of heart failure. Several sources have been proposed for generating new cardiomyocyte cell sources, including embryonic stem cells (ESC), induced pluripotent stem cells (iPSC), mesenchymal stem cells, cardiac progenitor cells, and transdifferentiation of non-cardiomyocytes^{7, 9–12}. Considerable effort has been made to overcome the roadblocks of cell transplant treatment. Tissue and genetic engineering techniques are complementary to cell-based therapy and are being developed with the goal of improving survival and integration of transplanted cells into native myocardium. Studies indicate that application of tissue-engineered cell sheet-based patches can increase the survival and engraftment of transplanted cells, help to improve cardiac function and help to prevent cardiac dilation of the infarcted heart¹³. However, after cell transplantation, engraftment and differentiation of administered cells into the native myocardium is very inefficient¹³. Recent innovative work indicate that cell-free delivery of extracellular vesicles, or exosomes, via patches to the injured heart improves cardiac recovery^{14, 15}. Thus, patches offer a promising strategy for improving HF treatment. Direct reprogramming of non-cardiomyocytes such as cardiac fibroblasts, is another strategy to generate new cardiomyocytes to restore cardiac function during HF. In vitro and in vivo animal studies indicate that a variety of signaling cocktails can convert cardiac fibroblasts into induced cardiomyocyte-like cells, which helps to reduce the cardiac scarring and improve cardiac function recovery after infarction, yet there are still numerous barriers to overcome before any clinical application should be considered¹².

Activation of endogenous cardiac repair

Recent reports have revealed that endogenous adult cardiomyocyte renewal is measurable, yet inefficient and incapable of adequately responding to extensive acute heart damage. Because of these limitations, therapies based on endogenous cardiomyocyte self-renewal are currently unavailable. It was previously thought that cardiomyocyte self-renewal is rare and is unlikely to be a useful method to treat HF. However, recent studies in animal models, such as zebrafish and mouse, have yielded mechanistic insights into the endogenous cardiac regeneration process, including findings concerning cardiomyocyte proliferation, inflammation, fibrosis, neovascularization, and the extracellular matrix¹⁶. These findings suggest that it is therapeutically possible to stimulate endogenous heart repair to treat HF following myocardial infarction. Importantly, recent discoveries indicate that inhibiting a critical regulatory genetic pathway called Hippo unleashes a powerful self-reparative capacity in the mammalian heart, and results in reversal of established HF^17–27. In addition, other signals and pathways such as IGF, Cyclin proteins, and microRNAs, have been reported to improve endogenous cardiac repair by stimulating cardiomyocyte proliferation in rodents following cardiac injury. We will specifically discuss these signals in more depth below. Although we have not reached the point that we can translate these animal model findings into clinical treatments, these exciting advances have uncovered new feasible therapeutic avenues for treating HF by inducing endogenous heart muscle to self-repair. A comparison of potential advantages of creating therapeutics to target endogenous cardiac repair is outlined in Table 1.

Table 1.

Comparison of Inducing Endogenous Repair to Other Potential Therapeutics

Treatment Type	Advantages of Endogenous Repair Over Treatment
Current Pharmacological Therapeutics (ACE inhibitors, beta-blockers, diuretics, digoxin, etc.)	Endogenous repair can lead to functional improvement and reversal of course of the disease.
	Potential for fewer or shorter treatment period, reducing responsibility of patient compliance to medications
Left Ventricular Assist Device (LVAD) Placement	LVAD placement requires major surgical intervention, which creates high cost due to hospitalization
	LVADs are only rarely curative and often are only act as a bridge to transplant
Heart Transplantation	Major shortage of implantable hearts compared to total HF patient population
	Patients require immunomodulatory therapies to help slow possibility of organ rejection
Cell Transplantation Therapy	Cell transplantation requires incorporation of new cardiomyocytes, both electrically and structurally, to prevent arrhythmia formation

The State of Cardiomyocyte Renewal in Mammals

During neonatal life, the mammalian heart undergoes a period of hyperplastic expansion followed by a stage of hypertrophic growth to accommodate the demand for cardiac output²⁸. In rodents, studies show that neonatal cardiomyocytes undergo a 10% rate of DNA synthesis within the first 4 days of life, but by day 7, this rate quickly reduces down to 1% concomitant with terminal differentiation^29–33. To investigate the regenerative capacity of neonatal mice, multiple groups have employed various injury models: 1) surgical resection of the ventricular apex, 2) occlusion of the left anterior descending artery (LAD-O), and 3) cryo-infarction. Although informative and useful, these models present limitations. Determining regeneration in the apex resection model is dependent on resection size as neomyogenesis is limited in hearts that receive large resections and have thicker scarring³⁴. Moreover, apex resection does not model physiological injury that the mammalian heart typically experiences, such as an MI. Importantly however, use of the apex resection model led researchers to the novel discovery that cardiac regeneration requires reinnervation of the heart^{35, 36}. Although LAD-O is better suited to model MI, it is more technically demanding than apex resection and requires experience to have consistent infarcts. Cryo-infarction represents an intriguing model as this technique generates an injury and regenerative response similar to LAD-O³⁷, yet can be more easily examined. Importantly however, the LAD-O injury models MI most accurately. Importantly, seminal work in mice revealed that hearts of neonatal mice can regenerate following apex resection or MI injury^{38, 39}. Using these models, groups have reported various degrees of scar formation and regeneration, and the consensus agreement is that the regenerative window closes after day 7, whereby cardiomyocytes terminally differentiate^{33, 34, 37, 40, 41}.

Adult cardiomyocytes are highly specialized, differentiated cells with very limited regenerative capacity, and until recent years, it was believed that cardiomyocytes were not capable of proliferation. However, in an influential study, researchers retroactively analyzed ¹⁴C isotope integration into cardiomyocytes of humans who lived during atmospheric nuclear bomb testing conducted from the early 1950s through 1963⁴². Using carbon dating techniques and mathematical modeling, they calculated that human postnatal cardiomyocytes renew at a ~1% rate per year that decreases with age to ~0.3% (Fig. 1)⁴². Based on these calculations, they found that 40% of cardiomyocytes in the heart have been produced throughout a full life span, whereas the additional 60% are originally formed during prenatal development^{42, 43}. These studies further reveal that hearts of humans 10 years or younger show ¹⁴C incorporation rates of ~5–6%, and this rate sharply declines to 1.9% between the ages of ten to twenty (Fig. 1)⁴³.

Figure 1: Renewal rates of human cardiomyocytes.

A) Timeline and rate calculated retrospectively using mathematical modeling and carbon dating measurements. B) Current theories describing the source of new ¹⁴C during aging: Left-labeled non-cardiomyoctyes fuse with cardiomyoctyes, leading to false-positive labeling. Middle-increased ploidy of cardiomyoctyes without cell division, Right-bona fide cardiomyoctye renewal. (Illustration Credit: Ben Smith).

Mounting evidence suggests that postnatally born cardiomyocytes are generated from pre-existing cardiomyocytes^44–48. Using thymidine-based DNA labeling and mitotic markers in mice, these studies indicate that a small percentage of mammalian adult cardiomyocytes are cycling^{29, 30, 49, 50}. Both of these measuring techniques simply measure cell cycle re-entry, and do not measure completion of cell division. A group in 2014 devised an approach to directly measure cell replication and completion of cytokinesis in mice using a Cre-dependent GFP and RFP labeling system “mosaic analysis with double markers”⁵¹. Consistent with other data, they concluded that postnatal cardiomyocytes are born at very low rates, are cardiomyocyte-derived, and do not undergo a second round of division⁵¹. Immunofluorescence-based snap-shot profiling of M-phase and cytokinesis in cardiomyocytes from human donor hearts revealed a cardiomyocyte M-phase index of 0.04% from birth to 10 years of age which decreases to 0.009% during the next decade⁵². However, cardiomyocyte cell division indexes were less than 0.01% and 0.005% at these same respective stages⁵². It is clear that while adult cardiomyocytes are capable of cell division, these events naturally happen at very low rates that are insufficient to renew myocardium following injury.

Role of Ploidy in Cardiomyocyte Renewal

While the adult mammalian heart has little ability to regenerate following injury, this is not the case of all animal species. It is well established that both zebrafish and newt heart tissue can be repaired following injury through true cardiomyocyte division^{53, 54}. Furthermore, the neonatal mouse heart also has a seven-day regenerative window following birth³⁸. Within this timeframe, cardiomyocytes are able to successfully undergo cellular division and contribute to tissue renewal⁵⁵. Both of these observations have led to a critical examination of the differences in cardiomyocyte cell biology across species. One interesting difference is the ploidy, or relative DNA content, of the cardiomyocytes. Zebrafish, newt, and neonatal mouse cardiomyocytes are all diploid (2n) cells, containing two sets of chromosomes, while adult mammalian cardiomyocytes exist largely as polyploid cells, containing four or more sets of chromosomes.

During mammalian development, cardiomyocytes are able to freely divide. However, following birth, the renewal capacity of the mammalian cardiomyocytes quickly dissipates^{56, 57}. Instead, the cardiomyocytes exit the cell cycle prematurely. This can either happen following DNA synthesis, a process known as polyploidization, that results in a cardiomyocyte with a single tetraploid (4n) nucleus. The cardiomyocyte can also exit the cell cycle following mitosis but without undergoing cytokinesis, known as binucleation, resulting in a cardiomyocyte with two diploid (2n) nuclei. The relative abundance of polyploidization and binucleation differs between mammalian species, with a majority of adult human cardiomyocytes resulting from polyploidization and a majority of mouse cardiomyocytes resulting from binucleation^56–58. While the molecular mechanisms explaining these differences in cardiomyocyte cell cycle exit are not well understood, these differences demonstrate that there are potentially different mechanisms regulating cell cycle exit in different mammalian species. There is growing circumstantial evidence of an inverse correlation between cardiomyocyte ploidy and renewal potential. Strikingly, the mouse neonatal regenerative window highly correlates with increased cardiomyocyte ploidy. At birth, all mouse cardiomyocytes are diploid. However, by the end of the first week of life, around half of the cardiomyocytes have undergone binucleation, and cardiomyocyte division following injury also has greatly decreased⁵⁰. There is also correlative evidence that the pool of diploid cardiomyocytes in the adult mouse are important contributors to the small amount of cardiomyocyte division seen after cardiac injury. For instance, it has been shown that cycling adult cardiomyocytes in the mouse marked by hypoxia fate mapping are more likely to be mononucleated⁵⁹. Furthermore, mouse strains with greater numbers of diploid cardiomyocytes demonstrate greater functional recovery following myocardial infarction⁶⁰.

While the evidence that diploid cardiomyocytes have greater regenerative potential in mouse is only correlative, recent work in zebrafish has been able to more directly address this hypothesis. Using a genetic model capable of increasing ploidy in zebrafish cardiomyocytes, it was found that diploid cardiomyocytes were significantly more likely to contribute myocardial renewal following injury⁶¹. Interestingly, converting 50% of zebrafish cardiomyocytes to a polyploidy state was sufficient to inhibit cardiac regeneration completely – this same percentage of binucleated cardiomyocytes is found at the end of the regenerative window in neonatal mice. Future work in mammalian systems investigating the role of polyploidy in the mammalian cardiomyocyte will not only improve our understanding of cardiomyocyte renewal but may also lead to new approaches to stimulate cardiogenesis therapeutically as the molecular underpinnings of this phenomenon are revealed.

Measuring True Cardiomyocyte Division

Awareness of polyploidization and binucleation cell cycle variants in cardiomyocytes is also critical when studying cardiac regeneration. While the ultimate goal of regeneration studies is to stimulate cardiomyocyte proliferation to contribute to damaged myocardium, measuring true cardiomyocyte division remains a technical challenge. This is due, in large part, to the fact that cardiomyocyte cell cycle re-entry does not necessarily indicate increased proliferation. In fact, cardiomyocyte ploidy frequently increases in heart failure, suggesting that cell cycle re-entry is an adaptive response to cardiac injury⁶². Therefore, it is critical that cell cycle assays are properly interpreted when reporting results in order to ensure reproducible conclusions. There are many cell cycle assays commonly used that only indicate cell cycle re-entry, with no indication whether cytokinesis will occur. Nucleotide analogs, such as EdU and BrdU, are incorporated during DNA synthesis and only indicates that the cell has progressed through S phase following analog administration. Histone H3 phosphorylation (pHH3) is a histone modification that occurs in G2 phase and continues to be present throughout mitosis and therefore relays no information regarding cytokinesis⁶³. Aurora B kinase, which is the kinase responsible for Histone H3 phosphorylation, has the added benefit of being expressed throughout cytokinesis. However, Aurora B kinase is expressed in cardiomyocytes both undergoing binucleation and true cellular division, so its expression alone is not sufficient to identify definitive cardiomyocyte division events⁶⁴.

While each of the previous markers are helpful for building evidence that cardiomyocytes are re-entering the cell cycle, they should be accompanied with more direct proof of cardiomyocyte proliferation. Stereology is a powerful technique for calculating total cardiomyocyte numbers in the heart by measuring cardiomyocyte number and size in successive tissue sections⁶⁵. While other techniques exist for counting, such as performing counts after isolating cardiomyocytes through tissue dissociation, these often result in inconsistent results due to variation in digestion efficiency. Another useful technique is clonal analysis. There are multiple genetic models that have been developed in mice to perform these experiments, including the Brainbow and Mosaic Analysis with Double Markers (MADM) systems, both of which have been successfully used to demonstrate cardiomyocyte proliferation^{19, 51, 66, 67}. Finally, recent work has indicated that midbody, a critical structural component of the contractile ring, location and distance between daughter nuclei may also be indicative of true cardiomyocyte division⁶⁸. Time-lapse imaging experiments of acute heart slices demonstrated that the midbody, which contains both Aurora B kinase and Anillin, is expressed symmetrically between daughter nuclei if the cardiomyocyte will undergo cellular division and asymmetrically if binucleation will follow. Furthermore, daughter nuclei of cardiomyocytes that underwent division were much farther apart than daughter nuclei of cardiomyocytes undergoing binucleation. It is possible that adoption of these methods will be helpful in cardiac regeneration studies to more accurately measure cardiomyocyte cell division.

Molecular Signals that Stimulate Cardiomyocyte Proliferation

Given the sharp decline in cell cycle activity and loss of regenerative potential in adult mammalian hearts, considerable effort has been made to understand the cellular mechanism(s) that underlie cardiomyocyte cell division. A number of groups have determined that within the cardiomyocyte, signaling from growth factors, intrinsic pathways, microRNAs and cell cycle regulators stimulate cardiogenesis following injury (Fig. 2).

Figure 2: Molecular signals that stimulate cardiac repair.

Exogenous signaling by growth factors, such as NRG/ERBB and IGF family members, activates intracellular pathways that promote cardiomyocyte (CM) proliferation and regeneration mechanisms. The Hippo kinase pathway is an intrinsic signaling pathway that triggers inhibitory phosphorylation of the CM proliferation protein Yap. This leads to either Yap sequestration at the plasma membrane with the dystrophin glycoprotein complex (DGC) or in the cytosol with 14–3-3 (dashed arrows). Treatment of CMs with agrin or anti-Hippo (shSalv) molecules promotes localization of dephosphorylated Yap to the nucleus to drive expression of CM renewal genes. MicroRNAs (miRs) can positively or negatively regulate CM proliferation, and development of anti-miR antagomiR molecules have innovated cardiac repair strategies. The CM cell cycle can be reactivated via modulation of regulators such as Cyclin A2 and member of the Cyclin B and Cyclin D families. Delivery of Wee1 and TGFb inihibitor molecules enhances cyclin/CDK-induced cardiac repair. CM proliferation: positive regulators (green), negative regulators (red). . (Illustration Credit: Ben Smith).

Cell Cycle Regulators

As neonatal life ends, cardiac expression of cell cycle regulators such as cyclins and cyclin-dependent kinases is largely silenced³². During the cell cycle, Cyclin A2 positively regulates G1/S and G2/M transitions, whereas Cyclin D2 drives G1/S progression. Overexpression studies in mice suggest that these genes promote proliferation of mature cardiomyocytes in adult hearts. Constitutive cardiac expression of Cyclin A2 in transgenic mice produces cardiomegaly due to the upregulation of cardiomyocyte proliferation⁶⁹. Following ischemic injury, Cyclin A2 transgenic mice show less ventricular dilation and enhanced cardiac function, concomitant with cell cycle reentry in infarct and border zone myocardium⁷⁰. Studies in rats and pigs revealed that viral delivery of Cyclin A2 to infarcted hearts was protective to ischemic injury^{71, 72}. Likewise, transgenic overexpression of Cyclin D1 and D2 promotes cardiomyocyte cell cycle activity in infarcted hearts as evidenced by increased multinucleation, DNA synthesis, and proliferation^{73, 74}. Recent investigation of the cardiomyocyte cell cycle reveals that combinatorial overexpression of cell-cycle regulators CDK1/CCNB/CDK4/CCND (referred to as 4F) activates proliferation of post-mitotic cardiomyocytes⁷⁵. In this study, the authors used rigorous lineage tracing to determine that viral delivery of 4F induced cardiomyocyte proliferation, cell survival and functional recovery to infarcted hearts⁷⁵. The replacement of CDK1/CCNB with Wee1 and TGFβ inhibitor molecules further enhanced cell-cycle re-entry of cardiomyocytes⁷⁵, supporting the notion that cell cycle reactivation, if used in a safe manner, may be useful as a therapeutic strategy for cardiac repair.

Growth Factors

In addition to cell-cycle regulators, a growing body of signal transduction studies indicates that several classes of ligand/cell surface receptor complexes can exogenously induce cell cycle re-entry of cardiomyoctes and facilitate regeneration. The Neuregulin1 (NRG1) ligand is an epidermal growth factor that signals through ERBB2–4 tyrosine kinase receptors in the myocardium to regulate cardiac development (reviewed in⁷⁶). Key regeneration studies reveal that NRG1/ERBB2/ERBB4 signaling promotes cardiomyoctye proliferation in the contexts of cardiac injury and homeostasis^{77, 78}. Exogenous administration of NRG1 into adult mice was shown to stimulate cardiomyocyte proliferation and regeneration⁷⁷. However, pediatric clinical trials and mice studies reveal that NRG1 delivery stimulates cardiomyocyte proliferation more effectively during a restricted period of early life⁷⁹. In support of this, cardiac regeneration potential declines significantly after neonatal life due, at least in part, to reduced expression of ERBB2⁸⁰. Deletion of ERBB2 in embryonic or neonatal cardiomyocytes results in myocardial thinning marked by reduced cardiomyocyte numbers, whereas ERBB2 overexpression in neonate and adult mice produces cardiomegaly due to increased cardiomyocyte proliferation, de-differentiation and hypertrophy⁸⁰. Moreover, transient induction of ERBB2 expression after LAD-O ischemic injury improves functional recovery and regeneration significantly⁸⁰. These data indicate that increasing NRG1/ERBB activity is a potential avenue for enhancing cardiac regeneration. However, further investigation of NRG1/ERBB2/ERBB4 signaling is warranted, as little is known regarding NRG1 upstream regulators or downstream targets. Importantly however, the ERK, Akt and GSK3β/β-catenin pathways were found to be indispensable downstream mediators of ERBB2-induced cardiomyocyte de-differentiation, proliferation and hypertrophy⁸⁰. Dual treatment of FGF1 and a p38 MAPK inhibitor (p38i) following myocardial infarction in rats led to increased cardiomyocyte mitoses⁸¹. Although cardiomyocyte proliferation is similarly increased, heart function is not rescued by p38i treatment alone⁸¹. Consistent with this finding, a 2016 clinical trial revealed that the p38 MAPK inhibitor losmapimod failed to improve the outcome of patients hospitalized after acute myocardial infarction⁸². Follow-up preclinical studies revealed that catheter delivery of FGF/NRG1-loaded microparticles to ischemia-reperfusion model swine hearts lead to increased myocardial vascularization and remodeling, accompanied by improved cardiac function⁸³.

Epicardial growth factors, including follistatin-lilke 1 (FST1) and insulin growth factor 2 (IGF2) have also been shown to promote cardiomyocyte cell cycle re-entry. Epicardial patch delivery of FST1 to pig and mouse model infarcts promotes proliferation of pre-existing cardiomyocytes, leading to improved cardiac function and survival⁸⁴. Notably, this effect was only observed when FST1 was epicardially-derived⁸⁴. This specificity suggests that the origin of a signaling factor influences its capacity to trigger a response in cardiomyocytes. Similarly, IGF2 signaling in the epicardium during mouse development is required for ventricular cardiomyocyte proliferation⁸⁴. Sustained co-administration of the growth factors insulin growth factor 1 (IGF1) and hepatocyte growth factor (HGF) improved cardiac function and stimulated cardiomyocyte production and capillarization to chronic MI pig hearts⁸⁵. In addition, injection of growth hormone-releasing hormone (an activator of IGF1 signaling) into pig hearts two weeks after LAD-O injury led to significantly reduced scarring, yet cardiac function was not rescued⁸⁶. Periostin, an extracellular matrix component, can also promote cardiomyocyte proliferation. Delivery of recombinant periostin via biodegradable Gelfoam extracellular scaffolds to rat hearts after myocardial infarction stimulates cell cycle re-entry of differentiated cardiomyocytes, reduces scaring and fibrosis, and improves cardiac function⁸⁷. Delivery of recombinant periostin into the pericardial space of post-MI swine stimulated myocardial regeneration, improved cardiac function, and increased capillary density⁸⁸. However, myocardial fibrosis was significantly increased in these hearts⁸⁸, indicating that periostin treatment to direct cardiac repair in large mammals likely presents challenges (e.g. defective remodeling and arrhythmias).

Intrinsic Signaling

In recent years, investigators have identified a number of intrinsic cardiomyocyte signaling pathways that regulate cardiac regeneration. Among these pathways are those that execute crucial functions during heart development, cardiomyocyte maturation and postnatal life. Several groups have led the way to determine the transcriptional regulatory network(s) that regulate cardiomyocyte proliferation and regeneration. Recent cardiac regeneration studies identified the transcription factor Meis1 as a crucial regulator of postnatal cardiomyocyte proliferation⁸⁹. In this work, the authors observed that deletion of Meis1 in cardiomyocytes after the regenerative phase of neonatal life extended the timeframe that cardiomyocytes normally cycle and regenerate⁸⁹. In adult hearts, Meis1 deletion activated cardiomyocyte mitosis, without compromising cardiomyocyte function, whereas Meis1 overexpression was inhibitory to neonatal cardiac regeneration⁸⁹. Given that the CDK inhibitors p15, p16, and p21 are Meis1 transcriptional activation targets⁸⁹, these data collectively indicate that Meis1 is a negative regulator of cardiomyocyte proliferation. Using cryoinjury and apex resection models, investigators discovered that the cardiac transcription factor GATA4 is essential for cardiac regeneration by regulating the paracrine factor FGF16⁹⁰. Viral overexpression of FGF16 partially rescued the injury-induced defects in GATA4-ablated hearts⁹⁰, supporting the notion that paracrine factors can be targeted for cardiac repair. Overexpression of the transcription factor Tbx20 activates the YAP, BMP and AKT proliferative pathways, whereas the anti-proliferation genes p21 and Meis1 are repressed by Tbx20⁹¹. Importantly, Tbx20 overexpression in adult cardiomyocytes after MI reduced infarct size and markedly improved cardiac function and survival⁹¹.

The Hippo signaling pathway is an evolutionarily conserved tissue and organ size regulatory pathway first described in flies⁹². Core kinase components Mst/Hpo and Lats/Wts relay inhibitory phosphorylation of the downstream effector Yap/Yki, a transcriptional coactivator that promotes expression of survival and pro-growth genes⁹². In recent years, researchers have identified key Hippo roles in cardiac development, disease, cardiomyocyte homeostasis and regeneration. During development, Hippo signaling restrains cardiomyocyte proliferation to maintain proper heart size, while preserving overall tissue patterning⁹³. Cardiac-specific deletion of the core Hippo component Salvador (Salv) during embryogenesis produced cardiomegaly due to hyper-proliferation of cardiomyocytes, and knockout of Mst and Lats kinases produced similar phenotypes⁹³. Conversely, Yap deletion during cardiogenesis leads to embryonic lethality with hearts exhibiting pronounced myocardial thinning^{94, 95}. More recent studies reveal that Hippo signaling represses a number of cellular mechanisms that are critical for endogenous heart repair. Several reports indicate that Hippo inhibits adult cardiac regeneration^17–27. In rodents, deletion of Hippo signaling components results in improved cardiac repair and reversal of established HF after myocardial infarction²⁰. In this study, the authors also delivered a small molecule Hippo pathway inhibitor (shSalv) via an adeno-associated virus 9 (AAV9)²⁰, a small virus shown to be safe in humans. AAV9 preferentially infects the heart muscle in mammalian species including mice, pigs, and humans and is a feasible choice for clinical use. Direct myocardial delivery of AAV9-shSalv during ischemic injury, or even weeks after injury, improves heart function²⁰. These data support the possibility of using gene therapy to induce cardiomyocyte regeneration with minimal toxic side effects. Yap overexpression in mouse hearts after MI promoted cardiomyocyte proliferation and reduced infarct size, while preserving heart function^{19, 96}. In addition, recent concurrent reports reveal that the Hippo effector Yap interacts with the dystrophin glycoprotein complex (DGC)^{21, 26} and the extracellular protein Agrin at the plasma membrane to regulate cardiomyocyte proliferation. One study revealed that Hippo phosphorylation of Yap triggers DGC sequestration of Yap at the plasma membrane in a mechanism to inhibit cardiomyocyte proliferation²¹. In another study, administration of Agrin to mice after MI promoted cardiac regeneration²⁶. Biochemical assays revealed that Agrin promotes DGC disassembly, disruption of the Yap-DGC interaction and Yap translocation to promote cardiomyocyte proliferation²⁶. Other studies revealed that injury-response genes are activated in regenerating Hippo-deficient hearts^{20, 23}. Following MI, levels of the transcription factor Paired-like homeodomain-2 (Pitx2) are upregulated in Hippo-deficient myocardium²³. In this context, Pitx2 and Yap co-regulate expression of antioxidant genes to protect the heart from injury-induced oxidative stress and promote regeneration²³. Like Pitx2, the mitochondrial quality control stress response gene Park2 is upregulated in Hippo-deficient hearts and is required for regeneration²⁰. These studies investigating Hippo signaling reveal that besides cardiomyocyte renewal, failing hearts have previously unappreciated cardiac repair genetic programs. Recent profiling of human patient hearts reveal that Hippo signaling is maladaptively upregulated in HF samples²⁰. Altogether, these findings raise the possibility that endogenous cardiomyocytes can be mobilized by inhibiting the Hippo pathway, and that inhibiting Hippo can be used as a strategy to directly treat loss of pumping function in human HF patients.

MicroRNAs

MicroRNAs (miRs) are small non-coding RNAs, which repress gene expression, are required to maintain proper cardiomyocyte homeostasis. Disruption of the miR biogenesis protein Dicer in adult mouse myocardium leads to dilated cardiomyopathy and heart failure^{97, 98}. Moreover, Dicer expression is reduced in end-stage heart failure patients, suggesting that DICER and miRs have important role(s) in cardiomyocyte progression and heart failure⁹⁷. Indeed, human studies have shown that miR levels are altered in cardiac disease states^99–101. High-throughput screening of human miRs revealed that miR590 and miR199a promote cardiomyocyte cell cycle re-entry and promote cardiomyocyte proliferation in neonatal and adult mice¹⁰². Additional analyses showed that following myocardial infarction, co-delivery of these miRs stimulates cardiac regeneration and rescue of cardiac function¹⁰². Individual knockdown of these miRs did not upregulate cardiomyocyte proliferation¹⁰², suggesting functional redundancy between miR590 and miR199a targets. Mouse neonatal LAD-O studies revealed that miRs of the miR15 family inhibit cardiomyocyte proliferation³⁹. Furthermore, inhibition of miR15 in adult hearts promotes cardiomyocyte proliferation and improves functional recovery after MI³⁹. Therapeutic targeting of miR15 in mice and pigs using anti-miR oligonucleotides reduced infarct size following ischemia-reperfusion¹⁰³. miR-17–92, an oncogene that promotes cell proliferation, is required for myocyte proliferation in the heart¹⁰⁴. Transgenic miR-17–92 overexpression in mice hearts was protective to ischemic injury¹⁰⁴. Cardiac function was improved, scarring was reduced, and border-zone cardiomyocyte proliferation was increased in these hearts¹⁰⁴. Expression of miR-31a-5p is upregulated in neonatal rat cardiomyocytes during early regenerative phases¹⁰⁵, suggesting a positive regulation of cardiomyocyte proliferation by miR-31a-5p. In support of this notion, cardiomyocyte proliferation marker levels were reduced following injection of neonatal rats with miR-31a-5p antagomiRs¹⁰⁵. miR302–367 is essential for cardiomyocyte proliferation during development and overexpression of miR302–367 in the adult heart reduces fibrotic scarring after MI¹⁰⁶. However, these hearts exhibited cardiomyocyte de-differentiation, ventricular dilation and reduced cardiac function¹⁰⁶. Interestingly, miR302–367 targets the Mst and Lats kinases of the Hippo pathway¹⁰⁶, suggesting that miR302–367 regulation of cardiomyocyte renewal and regeneration is mediated, at least in part, by Hippo inhibition. Taken together, these studies reveal that targeting miRs remains a promising therapeutic strategy to induce cardiac reparative processes.

Barriers to current treatments

To varying degrees, pharmacological intervention, LVAD implantation or heart transplantation, and cell transplantation treatments have significant limitations and new approaches are urgently required. Current pharmacological treatments mainly target symptoms or HF contributing factors to improve heart performance but cannot reverse the process of heart failure. Besides the severe scarcity of donor hearts, immune responses and organ rejection are common issues following heart transplantation. Efficacy of LVAD implantation varies in individual patients. Whereas some patients respond favorably to LVAD support and recover cardiac function, other patients do not respond to LVAD support and require heart transplantation. In addition, there are heavy health and financial burdens that come with both LVAD implantation and heart transplantation.

Cell transplantation has been wildly shown to improve cardiac function in cardiac injury animal models, but clinical trials have not revealed consistent benefits¹⁰⁷. For example, in the REPAIR-AMI trial, autologous bone marrow-derived mononuclear cell (BMMNC) transplantation led to strong improvement trends of 5-year survival in patients with acute myocardial infarction¹⁰⁷. However, there is no evidence of improved left ventricular function in any BMMNC clinical trial¹⁰. Observations and outcomes can vary in commonly used lab animal models like rodents, large mammals like swine and dog, and preclinical models like nonhuman primates. Implantation of ESC-derived cardiomyocytes do not cause arrhythmia in rodents but cause ventricular arrhythmia in primates with engrafted cells¹⁰⁸. Therefore, variation of interspecies response to cell therapy is a critical issue to keep in mind and varying genetic backgrounds can affect interpreting experimental results. Hence, it is necessary to compare and validate data collected from different species before translating preclinical research to clinical trials. Moreover, these studies rely almost exclusively on ejection fraction as a prognostic measure¹⁰⁷. In future studies, an obvious improvement over measuring ejection fraction would be to perform global longitudinal strain to assess newly described forms of HF¹⁰⁹. One obvious common biological barrier to cell treatment is that the adult heart is continuously beating. As such, injected cells fail to colonize the diseased heart and are expelled into the circulatory system. In general, there are several challenges for cell transplantation treatments, such as poor engraftment and survival of implanted cells, limited contribution to cardiac contraction activity, immune responses, increased arrhythmogenic risks, risk of teratoma formation, and ethical considerations¹³.

What needs to be improved

As stated previously, the current therapies of pharmalogical intervention, LVAD placement, and heart transplantation all have drawbacks. Pharmalogical intervention treats HF symptoms and slows disease progression but is not capable of reversing the natural progression of the disease. LVAD implantation prolongs survival, but not all patients respond to treatment, and a majority still eventually require heart transplantation. While heart transplantation is the only curative treatment available, there are not enough donor hearts available for the large number of HF patients and it is a costly procedure. Clearly there is an urgent need for improved HF therapies. In recent years, the heart regeneration field has grown considerably, and has focused almost exclusively on stem cell-based treatments. Nonetheless, researchers have made steady progress towards inducing endogenous cardiomyocyte self-renewal. As a result, numerous animal model studies have shown that inducing endogenous cardiomyocytes to self-renew is a viable and undoubtedly exciting approach to generate new, healthy cardiomyocytes^17–27. Resident cardiomyocytes can be induced to re-enter cell cycle and proliferate to generate new cardiomyocytes, and a number of cellular signals have been identified as regulatory nodes for this newly unearthed phenomenon^17–27. Stimulating cardiomyocyte self-renewal through endogenous signaling is an appealing approach, as it would increase the volume of working myocardium after injury and would have the potential to reverse HF progression. Even small increases in heart function may have a large impact for HF patients, potentially prolonging survival and increasing quality of life, meaning combinatorial treatment plans with current standard of care plus novel therapies stimulating cardiomyocyte proliferation have the ability to make a large impact on the growing HF patient population.

Identification of modulators that target cardiomyocyte proliferation signaling pathways is imperative for the development of efficient cardiac repair therapies. Ideally, small molecule targeting of these pathways can be used to promote cardiomyocyte proliferation while minimizing cardiotoxicity. The Hippo-DGC-Agrin studies mentioned above suggest that viral delivery or direct injection of cardiomyocyte-proliferation agonists may be a viable cardiac repair strategy in humans. However, translation of technologies such as heart-specific viral vectors into human therapies requires thoughtful examination especially considering the oncogenic potential of activating growth pathways. In addition to intrinsic activation of cardiomyocyte proliferation pathways, administration of defined extracellular factors via epicardial patch delivery represents a complementary strategy to stimulate endogenous repair. Advantages of this approach include: 1) a cell-free system; reduces the likelihood of recipient cells rejecting delivered factors 2) delivery of peptides would cause transient effects, thereby minimizing the possibility of tumor formation and 3) newly produced cardiomyocytes would more likely incorporate into myocardium and function properly. Importantly however, path delivery can stimulate proliferation of non-cardiomyocytes. Similarly, current anti-MiR cardiac repair strategies lack cardiomyocyte specificity. Hence, innovations to the available patch and miR technologies are needed to overcome these barriers. Moreover, miR studies in non-human primates are warranted to demonstrate whether miR treatment might sufficiently repair cardiac damage in humans.

In addition to these improvements, there remain a number of basic science and clinical challenges to overcome to make stimulating cardiogenesis as a HF treatment strategy a reality (Table 2). Refined and rigorous cardiomyocyte proliferation assays should be developed to ascertain true cardiomyocyte proliferation during regeneration. In addition, the increased use of large animal models in conjunction with careful interpretation of the injury model used will facilitate the translation of HF treatment in human studies. As the clinical application of these findings are being developed, issues such as scalability, tissue-specificity, and integration of newly produced cardiomyocytes into post-injury myocardium need to be thoroughly addressed (Table 2).

Table 2.

Current Challenges for Stimulating Cardiogenesis as a Treatment for Heart Failure

Challenge	Potential Solutions
Basic Science
Measuring True Cardiomyocyte Proliferation	Practice of rigorous proliferation assays (stereology, lineage tracing, etc.) when measuring and reporting cardiomyocyte proliferation
Variation Among Animal Models	Increase use of both large mammal models such as dog and swine at early stages of therapeutic development
Variation Among Cardiac Injury Models	Interpretation of results should be specific to the cardiac insult performed and any limitations of injury technique explained
Translational
Scalability of Therapeutic to the Human Heart	Increased testing in animal models with similar organ size, such as swine, to determine appropriate delivery techniques
Cell- and Tissue-Type Specificity	Further research into epicardial patches, viral delivery modes (AAV), and direct myocardial injection techniques to ensure minimal off-target cellular proliferation / tumor formation
Integration of New Cardiomyocytes into Existing Myocardium (Electrical and Morphological)	Extensive characterization of both electrical and structural incorporation of novel cardiomyocytes must be performed

Non-standard Abbreviations and Acronyms:

HF

heart failure

CHD

congenital heart disease

ACE

angiotensin-converting enzyme

LVAD

left ventricular assist device

ESC

embryonic stem cell

iPSC

induced pluripotent stem cell

IGF

insulin growth factor

LAD-O

left anterior descending coronary artery occlusion

MI

myocardial infarction

GFP

green fluorescent protein

RFP

red fluorescent protein

EdU

5-ethynyl-2´-deoxyuridine

BrdU

bromodeoxyuridine

pHH3

phosphohistone H3

MADM

mosaic analysis with double markers

4F

four factors

shSalv

short hairpin Salvador RNA

DGC

dystrophin glycoprotein complex

miR

microRNA

REPAIR-AMI

reinfusion of enriched progenitor cells and infarct remodeling in acute myocardial infarction

BMMNC

bone marrow-derived mononuclear cell