Heart Regeneration: 20 Years of Progress and Renewed Optimism

Summary

Cardiovascular disease is a leading cause of death worldwide, and thus there remains great interest in regenerative approaches to treat heart failure. In the past 20 years, the field of heart regeneration has entered a renaissance period with remarkable progress in the understanding of endogenous heart regeneration, stem cell differentiation for exogenous cell therapy, and cell delivery methods. In this review, we highlight how this new understanding can lead to viable strategies for human therapy. For the near term, drugs, electrical and mechanical devices and heart transplantation will remain mainstays of cardiac therapies, but eventually regenerative therapies based on fundamental regenerative biology may offer more permanent solutions for patients with heart failure.

Lee ETOC

Cardiovascular disease is a leading cause of death, and there is therefore great interest in regenerative approaches to treat heart failure. Gargern and Lee review the remarkable recent progress in the field of heart regeneration and discuss how this improved understanding can lead to viable strategies for human therapy.

Introduction

Cardiovascular disease remains the leading cause of death worldwide due in part to the limited capacity of the adult mammalian heart to regenerate (Roth et al., 2020). The heart is the first organ formed during embryonic development (Sakabe et al., 2005) and is composed of not only the contractile cardiomyocytes, but other cell types that support cardiac function and homeostasis, including endothelial cells, fibroblasts, vascular smooth muscle cells, immune cells, adipocytes, and neuronal cells (Litvinukova et al., 2020). Although the heart retains the ability to regenerate in some lower vertebrates (e.g. zebrafish, newts) (Becker et al., 1974; Oberpriller and Oberpriller, 1971; Poss et al., 2002) and during the early neonatal period in mammals (Haubner et al., 2016; Porrello et al., 2011), adult mammals do not retain the capacity to replenish the heart with sufficient cardiomyocytes to restore function following injury, despite a limited ability of adult mammalian cardiomyocytes to re-enter the cell cycle (Bergmann et al., 2009).

Current standard of care for patients with heart failure is suboptimal, with medications primarily aimed at managing symptoms and neurohormonal activation but not targeting the underlying myocardial pathophysiology (Committee et al., 2021). Mechanical circulatory support with left ventricular assist devices can delay the need for heart transplantation but are associated with significant adverse events including bleeding or thrombosis (Petricevic et al., 2015). Availability of organ donors for heart transplantation remains inadequate to meet demand (Bakhtiyar et al., 2020), thus methods to regenerate the heart are highly attractive to treat the growing population of patients with cardiac disease.

In the past 20 years, we have seen tremendous progress in the science of heart regeneration. Although the field may still be in its infancy, with successful clinical translation of regenerative therapies in the heart yet to occur, we have learned critical lessons about molecular mechanisms for heart regeneration (Eschenhagen et al., 2017). These lessons represent important steps on the road to realizing the full potential of regenerative medicine for heart disease. Heart regeneration is a broad term that includes the concepts of native cardiomyocyte renewal potential, cell therapy, direct reprogramming, and tissue engineering approaches to rebuild damaged hearts (Figure 1). Here, we focus on five major lessons learned in the field of heart regeneration over the past 20 years and then speculate on what the next 20 years might bring.

Figure 1. Overview of heart regeneration approaches.

Approaches to regenerate myocardium include renewal of pre-existing cardiomyocytes by stimulating de-differentiation and proliferation of existing, mature cardiomyocytes, trans-differentiation of non-cardiomyocytes into cardiomyocytes such as with gene therapy methods, and delivery of stem cell-derived cardiomyocytes either as an injectable system or as a tissue engineered patch.

Lesson 1: Mammalian adult cardiomyocytes can re-enter the cell cycle.

Mammalian post-natal cardiomyocytes were once thought to be incapable of cell cycle re-entry due to evidence of neonatal cell cycle arrest in rodents (Soonpaa et al., 1996). However, there is definitive evidence for ongoing, low levels of cardiomyocyte cycling in adult mammals. Atmospheric levels of carbon-14 (¹⁴C) increased during the 1950s-1960s due to nuclear bomb testing during the Cold War, leading to incorporation of ¹⁴C in cycling cardiomyocytes for individuals born shortly before and during that time (Bergmann et al., 2009). Using this approach, Bergmann et al. calculated that adult human cardiomyocytes show cell cycle activity at a rate of ~0.5–1% per year (Bergmann et al., 2009). Similar rates of adult cardiomyocyte cycling have been shown in mice using labeling with ³H-thymidine (<1% (Soonpaa and Field, 1997)) or ¹⁵N-thymidine (0.76% per year (Senyo et al., 2013)) and in humans using the mitotic marker, phosphorylated histone H3 (pH3) (0.04–1.6% per year (Mollova et al., 2013)). Taken together, these reports provide strong evidence for a low level of annual cardiomyocyte cell cycle re-entry, diminishing further into late adulthood, and the rates of cardiomyocyte cycling in different mammalian species have been remarkably consistent. We note that it is possible that many of these cycling cardiomyocytes become polyploid without producing a daughter cell, as these methods do not all clearly distinguish between cell cycle re-entry and true proliferation (i.e. completed cell division). In addition, if a daughter cell is produced, it remains unclear whether this contributes to increasing cardiomyocyte numbers or whether the total number of cardiomyocytes remain stable over time. One report suggests that there is a burst of cardiomyocyte proliferation in pre-adolescent mice between postnatal day 15 (P15) and ~P18 in response to a thyroid hormone surge, increasing cardiomyocyte numbers by approximately 40% during that week (Naqvi et al., 2014). However, another report suggests that the total number of cardiomyocytes remains constant over time, although cardiomyocyte exchange is higher during young hearts compared to aged hearts (Bergmann et al., 2015). New techniques that better distinguish between cell cycle re-entry and true cell division will be needed to resolve this discrepancy (Hesse et al., 2018; Kadow and Martin, 2018), but there is consensus that cell cycle re-entry is present at low levels in adult cardiomyocytes.

Animal studies have demonstrated that the native renewal capacity of adult mammalian cardiomyocytes can be augmented through various interventions, such as through regulation of transcription factors (Padula et al., 2021), stimulation of selected microRNA pathways (Aguirre et al., 2014; Tian et al., 2015), exposure to hypoxic conditions (Nakada et al., 2017), or activation of selected growth factor signaling pathways (D’Uva et al., 2015; Wei et al., 2015) (Figure 2). Transcription factors involved in early cardiomyocyte differentiation also promote cardiomyocyte proliferation, including GATA4 (Rojas et al., 2008), NF-κβ (Karra et al., 2015), TBX20 (Chakraborty et al., 2013), Notch-activated CSL (Collesi et al., 2008), and Wnt/β-catenin activated TCF/LEF (Heallen et al., 2011). Meis1 is a homeodomain transcription factor that promotes cell cycle arrest through upregulation of cell cycle inhibitors, p15, p16, and p21, and deletion of Meis1 in adult mouse cardiomyocytes leads to cell cycle re-entry (Mahmoud et al., 2013). Cardiac-specific deletion of p53 and Mdm2 in mice leads to increased expression of Cdk2 and cyclin E, decreased expression of p21 and p27, and enhanced proliferation of cardiomyocytes (Stanley-Hasnain et al., 2017). Conversely, overexpression of YAP1, a target of Hippo kinase, promotes cell cycle re-entry via interaction with the TEAD transcription factor family in adult mouse hearts (von Gise et al., 2012). Overexpression of cell cycle genes, including cyclin-dependent kinases 1 and 4 (CDK1 and CDK4, respectively), cyclins B1 and D1, and members of the E2F family of transcription factors, induces cell cycle re-entry in mammalian cardiomyocytes (Eghbali et al., 2019; Mohamed et al., 2018; van Amerongen et al., 2010). In addition, overexpression of reprogramming factors, Oct4, Sox2, Klf4, and c-Myc, the same factors used to generate the first reported induced pluripotent stem cells from fibroblasts (Takahashi and Yamanaka, 2006), can promote de-differentiation and proliferation of adult cardiomyocytes (Chen et al., 2021).

Figure 2. Mechanisms regulating the cardiomyocyte cell cycle.

Proliferation of existing cardiomyocytes requires de-differentiation and sarcomere disassembly followed by re-entry into the cell cycle. Growth factors such as bone morphogenic protein (BMP), transforming growth factor (TGF), or follistatin like-1 protein (FSTL1), or proliferative miRNAs can lead to stimulation of pro-proliferative developmental pathways such as YAP, NOTCH, and Wnt/B-catenin and/or affect expression of transcription factors promoting cardiomyocyte proliferation. In contrast, transcription factors p53 or MEIS1 lead to upregulation of cell cycle inhibitors, and Hippo pathway activation leads to inhibition of the pro-proliferative YAP pathway. miRNAs can also inhibit cardiomyocyte proliferation by inhibiting sarcomere disassembly and/or promoting cell cycle arrest. Pro-proliferative factors shown in blue; inhibitory factors shown in orange.

Delivery of small interfering RNAs (siRNAs) to p21, p27, and p57 can stimulate proliferation of adult murine cardiomyocytes (Di Stefano et al., 2011) while siRNA-mediated inhibition of Rb1 and Meis2 promotes cardiomyocyte proliferation in rats following myocardial infarction (Alam et al., 2019). Rb1 also acts with p130 and p107 to promote cell cycle exit into a quiescent (G₀) state (Park et al., 2013; Sdek et al., 2011). Noncoding RNAs (ncRNAs) are important regulators of cardiomyocyte development (Braga et al., 2021; Kay and Soltani, 2021). Expression of specific microRNAs (miRNAs), including miR-199a-3p (Torrini et al., 2019), miR-302 (Tian et al., 2015), or miR-17–92 family members (Chen et al., 2013) (such as miR-19 (Gao et al., 2019)) can promote cardiomyocyte proliferation, while expression of other miRNAs, such as miR-1 (Gan et al., 2019), miR-15 family members including miR-195 (Porrello et al., 2013), or miR-133a-1/2 (Liu et al., 2008) promotes cardiomyocyte cell cycle exit. Exercise upregulates a number of circulating miRNAs, including miR-222 (Baggish et al., 2011), which protects cardiomyocytes from ischemic injury and promotes cardiomyocyte growth and proliferation, through downregulation of homeodomain interacting protein kinase 1 (HIPK1) (Liu et al., 2015; Vujic et al., 2018).

De-differentiation of cardiomyocytes with sarcomere disassembly is needed to re-enter the cell cycle (Zhu et al., 2021), and protein factors such as neuregulin (acting via its receptor, ERBB) (Bersell et al., 2009; D’Uva et al., 2015) or oncostatin M (Kubin et al., 2011), have been shown to promote sarcomere disassembly. Conversely, miRNAs miR-99/100 or let-7a/c can inhibit sarcomere disassembly by downregulating farnesyltransferase subunit β (FNTβ) and SMARCA5 (Aguirre et al., 2014). In addition to stimulating cardiomyocyte de-differentiation, extracellular growth factors such as the transforming growth factor-β (TGF-β) superfamily of proteins (which includes the TGF-β proteins and bone morphogenic proteins (BMPs)) can promote cardiomyocyte cell cycle progression (Peng et al., 2021; Sorensen and van Berlo, 2020). Epicardial, but not myocardial, follistatin-like 1 (Fstl1) protein can stimulate cell cycle re-entry and proliferation of cardiomyocytes in mice and pigs (Wei et al., 2015). Many pathways have been shown to promote or inhibit cardiomyocyte cell cycle activity. However, the extent to which cardiomyocytes complete cell division versus become polyploid remains unclear in most studies. In addition, whether a level of ~0.5–1% cycling cardiomyocytes is sufficient to support heart functional improvement after different manipulations is uncertain.

Lesson 2: Multiple models reveal mechanisms for successful heart regeneration.

Model organisms including zebrafish, urodele amphibians (newts, axolotl), and small mammals (mice, rats, rabbits, guinea pigs) have all been used to define molecular mechanisms for postnatal heart regeneration. Zebrafish and urodele amphibians have been particularly helpful due to their intrinsic ability to undergo complete heart regeneration after a significant portion (20–50%) of the apical myocardium is transected (Becker et al., 1974; Oberpriller and Oberpriller, 1971; Poss et al., 2002). In these animals, new cardiomyocytes arise from de-differentiation of pre-existing mature cardiomyocytes that proliferate and then re-differentiate into cardiomyocytes (Jopling et al., 2010; Laube et al., 2006). Partial ventricular resection in zebrafish and newts triggers sarcomere disassembly (Jopling et al., 2010) and reactivation of transcription factor pathways involved during heart development, such as Notch, Islet1 and Gata4, to induce cardiomyocyte proliferation after myocardial injury (Kikuchi et al., 2010; Raya et al., 2003; Witman et al., 2011). The epicardium may also contribute cells that undergo an epithelial-to-mesenchymal transition to revascularize the regenerated myocardium (Lepilina et al., 2006).

While irreversible scar formation appears to be detrimental to cardiac regeneration (as discussed further in medaka and mammals below), transient fibrosis appears necessary for promoting regenerative capacity, suggesting that regulation of non-cardiomyocyte activity in the heart can dictate the ability of cardiomyocytes to re-enter the cell cycle (Sanchez-Iranzo et al., 2018). Activated fibroblasts deposit extracellular matrix proteins that are resorbed after fibroblast inactivation following myocardial cryoinjury in zebrafish, and deletion of col1a2 both inhibits the myocardial fibrotic response and prevents cardiomyocyte proliferation in zebrafish (Sanchez-Iranzo et al., 2018). While robust cardiac regeneration following injury occurs in zebrafish well into adulthood (Poss et al., 2002), it should be noted that this is not a universal trait across all teleost species. The medaka is another fish model that, unlike zebrafish, exhibits a fibrotic response in the heart leading to permanent scar formation following injury with a significantly impaired cardiac regenerative capacity compared to zebrafish (Ito et al., 2014). This is attributed to differences in immune responses to injury in medaka compared to zebrafish, with medaka having delayed macrophage infiltration and neutrophil clearance and reduced neovascularization after myocardial cryoinjury (Lai et al., 2017). Conversely, stimulation of Toll-like receptor signaling can accelerate immune responses and promote cardiomyocyte proliferation in medaka (Lai et al., 2017), suggesting a powerful role of the immune system and inflammatory response in controlling regeneration capacity.

While adult mammals have a poor natural history following myocardial injury leading to cardiomyocyte necrosis, myocardial fibrosis, and heart failure (Prabhu and Frangogiannis, 2016), the ability of postnatal mammalian cardiomyocytes to proliferate is not lost until after the first few days of life (Porrello et al., 2011). Neonatal mice retain substantial ability to regenerate myocardium following apical resection (Porrello et al., 2011) or coronary artery ligation (Porrello et al., 2013) in the first week of life. Similar findings have been reported in neonatal rats, with the optimal window for regeneration likely being within the first 2–3 days after birth (Lam and Sadek, 2018). There is also clinical evidence to suggest that human neonates may have a more robust capacity to regenerate myocardium after infarction compared to adults (Haubner et al., 2016). As in lower vertebrates, new cardiomyocytes are thought to arise from mature cardiomyocytes that de-differentiate into a progenitor-like state before proliferation postnatally in neonatal mammals (Lam and Sadek, 2018; Porrello et al., 2011). In contrast, myocardial injury occurring after the first week of life in mammals leads to fibrosis and scarring with minimal to no evidence of cardiomyocyte proliferation (Porrello et al., 2011; Zogbi et al., 2014).

Differences in the immune response between lower vertebrates that retain cardiac regenerative potential and mammals may help to explain the limited cardiac regenerative potential in post-natal mammals (reviewed in (Farache Trajano and Smart, 2021)). Following a myocardial infarction, the innate immune system initiates an acute inflammatory response characterized by recruitment of neutrophils and monocytes (Farache Trajano and Smart, 2021). Neutrophil activation leads to increased reactive oxygen species production and cytokine release that further exacerbates inflammation in the heart (Ma, 2021). Monocytes differentiate into macrophages of different subtypes – the M1 subtype secretes additional pro-inflammatory cytokines while the M2 subtype activates fibroblasts leading to secretion of extracellular matrix proteins and subsequent scar formation (Peet et al., 2020). Lower vertebrates that retain the capacity for heart regeneration have robust recruitment of eosinophils (Bevan et al., 2020) which secrete anti-inflammatory cytokines such as IL-4, and this may tilt the immune system balance toward a regenerative rather than fibrotic state in the heart (Toor et al., 2020). Understanding the molecular differences between lower vertebrates and neonatal mammals versus adult mammals is critical to enabling new therapies that regenerate the heart.

Several mechanisms leading to cardiac regenerative potential loss in postnatal mammals have been identified (Figure 3). Prior to birth, the mammalian fetus derives oxygen, nutrition, temperature stability, and immune protection from the mother; however, after birth, the neonate must breathe spontaneously, initiate enteral nutrition, maintain body temperature, and establish immunity independently, and the processes that establish neonatal independence also contribute to loss of cardiomyocyte proliferative potential. Initiation of spontaneous breathing in an oxygen-rich environment and transition to enteral nutrition, which is high in fatty acids (Girard et al., 1992), prompting mitochondrial maturation, leads to increased production of reactive oxygen species (ROS) (Puente et al., 2014; Zhao et al., 2019). Induction of a DNA damage response by ROS appears to contribute to the loss of regenerative capacity, as it is possible to prolong the regenerative window in a hypoxic environment or through use of ROS scavengers (Puente et al., 2014). DNA damage can lead to post-natal changes in DNA methylation that may downregulate bone morphogenic protein, transforming growth factor-β, Hedgehog, and Wnt/β-catenin signaling, likely contributing to cell cycle arrest in cardiomyocytes (Sim et al., 2015). Acute inflammation following myocardial injury in neonatal mice activates cardiomyocyte proliferation and angiogenesis through interleukin-6 (IL-6)/signal transducer and activator of transcription 3 (STAT3) signaling (Han et al., 2015; Ni et al., 2021); however, chronic inflammation in adult mice promotes fibrosis and scar formation while inhibiting angiogenesis and cardiomyocyte proliferation (Suthahar et al., 2017). Neonates have underdeveloped immune systems, with immature neutrophils, monocytes, macrophages, and B and T cells exhibiting a blunted response to stimuli (reviewed in (Simon et al., 2015)), that gradually develops over the first few months of life in response to environmental pathogen exposures and changes in the gut microbiome (Olin et al., 2018).

Figure 3. Mechanisms of cardiac regenerative potential loss in postnatal mammals.

The switch from anaerobic to aerobic respiration after birth leads to increased production of reactive oxygen species and induction of a DNA damage response pathway, which leads to cell cycle arrest and cardiomyocyte maturation. Similarly, transition from placental to enteral nutrition, with breastmilk containing higher levels of fatty acids, leads to mitochondrial maturation which also releases reactive oxygen species during oxidative phosphorylation. The DNA damage response pathway leads to histone modifications and subsequent shifts in gene expression from early to late cardiomyocyte genes, which promote cardiomyocyte cell cycle arrest and maturation. Maturation of the immune system, due to changes in the gut microbiome and pathogen exposure after birth, shifts the post-injury inflammatory response toward fibrosis and reduced angiogenic potential. Increased metabolic rates are required to maintain temperature outside of the womb, which is regulated by thyroid hormone and beta-adrenergic signaling and simultaneously leads to cell cycle arrest by downregulation of Ect2. Establishment of the circadian cycle after birth also acts via the sympathetic nervous system to upregulate Per1/2, which promote cell cycle arrest in cardiomyocytes.

The immature immune system of neonates allows for a more balanced regenerative response in the mammalian heart, while the pro-inflammatory and pro-fibrotic response in adult mammals leads to scar formation and lack of a regenerative response (Sattler and Rosenthal, 2016). The ability to quickly revascularize the heart after injury also appears to be an important component of retaining cardiomyocyte regenerative capacity. Loss of revascularization potential following deletion of C-X-C motif chemokine ligand 12 (CXCL12) or its receptor CXCR4 in neonatal mice contributes to loss of cardiomyocyte proliferation, while exogenous delivery of CXCL12 in adult mice promoted formation of collateral vessels by arterial endothelial cells (Das et al., 2019). Endocrine adaptations following birth also affect the regenerative potential of the heart – a sharp increase in thyroid hormone levels after birth drives increases in metabolic rate that facilitate maintenance of endothermy via interaction with adrenergic receptors (Payumo et al., 2021), and these changes work in concert to inhibit cardiomyocyte proliferation (Hirose et al., 2019). β-adrenergic receptor signaling suppresses cardiomyocyte expression of Ect2, leading to cell cycle arrest, while inhibition of β-adrenergic receptor signaling with propranolol can promote cytokinesis in cardiomyocytes of human infants with congenital heart disease (Liu et al., 2019). Sympathetic nervous system activity also suppresses cardiomyocyte cell cycle activity by activation of Period1 (Per1) and Period2 (Per2), two genes that regulate the circadian clock (Tampakakis et al., 2021); despite the loss in regenerative potential, establishment of a circadian rhythm is important for promoting postnatal maturation of cardiomyocytes (Bray et al., 2008; Durgan et al., 2006).

Genetic lineage tracing experiments have defined the origin of new cardiomyocytes as arising from pre-existing cardiomyocytes in adult zebrafish and neonatal mouse heart regeneration. Whether endothelial cells, fibroblasts, epicardial cells and inflammatory cells are major determinants or functional modulators of the cardiac regenerative window and capacity awaits future investigation.

Lesson 3: Adult stem cells do not participate in cardiomyocyte regeneration.

The past two decades have seen a dramatic shift in the concept of adult stem cells and their potential to contribute to new cardiomyocytes. Early evidence suggested that extracardiac progenitor stem cells might exist that can differentiate into cardiomyocytes (Hsieh et al., 2007; Laflamme et al., 2002), with early indications that these might arise from the bone marrow (Deb et al., 2003; Jackson et al., 2001). Multiple surface markers had been proposed to identify potential adult cardiac stem cells, including stem cell antigen-1 (Sca-1) and c-kit (Jackson et al., 2001) that are now known to primarily mark endothelial cell precursors. Interest in identification of adult stem cells that could differentiate into cardiomyocytes may have been driven in part by the political landscape in the United States ~20 years ago, with restrictions on the use and study of embryonic stem cells due to ethical concerns prior to the discovery of induced pluripotent stem cells derived from adult cells.

Some studies suggested that Sca-1+ cells were a source of new cardiomyocytes in adult mouse hearts (Noseda et al., 2015; Oh et al., 2003; Uchida et al., 2013); however, these studies used techniques that may have led to an overestimation of the percent of Sca-1+ cells. Development of robust lineage tracing techniques over the past 20 years have largely eliminated a major role for an adult cardiac stem cell population in cardiomyocyte regeneration (Lee, 2018). Multiple investigators have now demonstrated that Sca-1+ cells do not contribute significantly to cardiomyocytes, but rather are endothelial cell precursors (Neidig et al., 2018; Soonpaa et al., 2018; Tang et al., 2018; Vagnozzi et al., 2018; Zhang et al., 2018a). Similarly, c-kit is a receptor tyrosine kinase that is found on a variety of cell types but is now thought in the heart to be predominantly expressed by a cardiac endothelial cell population rather than cells with cardiomyogenic potential (Sultana et al., 2015; van Berlo et al., 2014). While there is some evidence that a small number of neonatal c-kit positive cardiac cells differentiate into cardiomyocytes, adult c-kit positive cells do not appear to retain the ability to differentiate into cardiomyocytes (Jesty et al., 2012; Zaruba et al., 2010).

Initial studies of bone marrow cells transdifferentiating to cardiomyocytes in mice two decades ago (Orlic et al., 2001) were rigorously challenged soon thereafter (Murry et al., 2004; Wagers et al., 2002). However, enthusiasm for bone marrow therapies for the heart persisted, partly on the basis of potential angiogenic benefits (Yoon et al., 2010). To date, ~80 randomized clinical trials involving several thousand patients have been performed using bone marrow-derived cells in the heart (Michler, 2018; Pompilio et al., 2015). Meta-analyses aggregating data from multiple clinical trials demonstrate that while bone marrow-derived cells are largely regarded as safe, with few adverse events and no significant risk of arrhythmias, the clinical benefit is likely modest with only small improvements in global left ventricular systolic function, if any (Abdel-Latif et al., 2007; Afzal et al., 2015). Benefits of bone marrow therapies may be attributed to paracrine effects on the heart, rather than generation of new cardiomyocytes (Michler, 2018) (Figure 4).

Figure 4. Adult stem cells can have paracrine effects but do not become new cardiomyocytes.

Bone marrow derived cells have been studied in multiple clinical trials to test whether they can promote cardiomyocyte maturation. While there is some evidence of clinical improvement, these findings do not appear to be due to replacement with new cardiomyocytes but rather through paracrine effects leading to enhanced angiogenesis.

Cardiosphere-derived cells (CDCs) are derived from a heterogeneous population of cells (including c-kit+, CD31/CD34+, and CD90+ subpopulations of cells) isolated via endomyocardial biopsy from patients that are then cultured and expanded in vitro (Davis et al., 2009). These cells were also previously thought to have cardiomyogenic potential but are now also thought to act primarily through paracrine mechanisms (Lopez et al., 2020). CDCs have shown mixed results in animal studies (Kasai-Brunswick et al., 2017; Malliaras et al., 2012), but given some evidence of functional improvement and overall reassuring safety profile, CDCs have proceeded to clinical trials, with possible scar size reduction seen in the CADUCEUS trial but not the larger ALLSTAR trial. However, the ALLSTAR trial did demonstrate improvements in left ventricular end diastolic volume and N-terminal pro B-type natriuretic peptide (NT-proBNP) with CDC delivery (Makkar et al., 2020; Malliaras et al., 2012). When the CDC subpopulation of c-kit+ cardiac derived cells were delivered in combination with mesenchymal stem cells (MSCs) to patients with ischemic cardiomyopathy, there was a significant improvement in quality of life scores but not left ventricular ejection fraction or left ventricular volumes at 12 months in the CONCERT-HF trial (Bolli et al., 2021).

A major contribution of c-Kit+ and Sca1+ progenitor cells to newly regenerated cardiomyocytes has been ruled out by rigorous genetic fate mapping studies. While early studies suggested that unidentified progenitor cells could give rise to new myocytes after injury (Hsieh et al., 2007; Laflamme et al., 2002), this no longer appears to be the case. Benefits observed after stem cell transplantation are now largely attributed to either paracrine or immunomodulatory effects, and further investigation will be needed to identify exactly what factors improve outcomes (Vagnozzi et al., 2020). Furthermore, c-Kit+ and Sca1+ progenitor cells, as well as other progenitor cells, could provide an important role by generating non-myocyte cells including vasculature cells (Marino et al., 2019).

Lesson 4: Multiple approaches can lead to new cardiomyocytes for failing hearts

Because non-myocytes do not appear to contribute importantly to new cardiomyocytes in the post-natal mammalian heart, many avenues are under investigation for new cardiomyocytes after injury. Early studies in rats, guinea pigs, and rabbits demonstrated that cardiac transplantation of skeletal myoblasts results in successful engraftment in the heart, and the skeletal myoblasts showed evidence of partial transdifferentiation into cardiomyocytes (Murry et al., 1996; Taylor et al., 1998). In 2001, the first use of cells as a potential therapy in the heart was reported in a human with severe heart failure following myocardial infarction, with autologous skeletal myoblasts injected into the myocardium at the time of coronary artery bypass grafting (Menasche et al., 2001). This was followed by the MAGIC trial, where surgical myoblast delivery did not significantly improve left ventricular systolic function at 6 months and an increased risk of ventricular arrhythmias following myoblast transplantation was observed (Menasche et al., 2008). Given lack of evidence of functional improvement, concern for the risk of post-operative arrhythmias, and need for a cell phenotype that can specifically meet the demands of the myocardium, recent approaches have focused on generating bona fide cardiomyocytes for failing hearts.

Primary cardiomyocytes isolated from adult mammals rarely proliferate in vitro (Alam et al., 2020), and cell lines mimicking cardiomyocytes that are capable of proliferation do not completely resemble mature cardiomyocytes (Claycomb et al., 1998), making in vitro studies and expansion for clinical therapies extremely challenging prior to the development of pluripotent stem cell-derived cardiomyocytes (PSC-CMs). Directed differentiation of mouse (Maltsev et al., 1993) and human (Xu et al., 2002) embryonic stem cells (ESCs) into cardiomyocytes enabled what has become a highly promising approach for heart regeneration by allowing production and expansion of cardiomyocytes in vitro. More recently, availability of induced pluripotent stem cells (iPSCs) derived from adult cells, first described in 2006 (Takahashi and Yamanaka, 2006), has permitted a world-wide effort of research into lab-grown cardiomyocytes, bypassing the ethical concerns of using embryonic stem cells.

PSC-CMs are appealing because these cells can clearly remuscularize areas with damaged myocardium. Initial studies testing the use of PSC-CMs in small animals showed promise in restoring cardiac function following delivery PSC-CMs (Lalit et al., 2014). However, large animal studies revealed a risk of potentially life-threatening ventricular arrhythmias following injection of PSC-CMs into non-human primate or pig models (Chong et al., 2014; Liu et al., 2018; Romagnuolo et al., 2019; Shiba et al., 2016). The mechanisms for this increased arrhythmia risk, termed engraftment arrhythmia, are thought to be due to automaticity of immature PSC-CMs in the first few weeks after injection (Liu et al., 2018; Romagnuolo et al., 2019). It may be possible to partially overcome this risk through pharmacologic suppression (Nakamura et al., 2021), and research is also ongoing to test in vitro maturation methods prior to delivery to further reduce this risk.

The Embryonic Stem Cell-derived Progenitors in Severe Heart Failure trial (ESCORT) is the first (and to date, only completed) clinical trial involving PSC-CMs, delivered in an epicardial patch to six patients with severe heart failure following myocardial infarction at the time of coronary artery bypass surgery (Menasche et al., 2018). All patients had improved systolic function of the regions treated with a cell patch, although the contribution of patch versus coronary artery bypass grafting could not be easily distinguished (Menasche et al., 2018). Only one of these patients developed arrhythmias following epicardial patch placement (Menasche et al., 2018), in contrast to the majority of macaques and pigs experiencing arrhythmias in preclinical trial following intramyocardial delivery (Chong et al., 2014; Liu et al., 2018; Romagnuolo et al., 2019), suggesting that method of delivery may impact the risk of arrhythmia. A number of early phase industry-sponsored clinical trials involving PSC-CMs are now recruiting, with even more clinical studies in planning stages (reviewed in (Eschenhagen et al., 2021)).

Direct reprogramming of non-cardiomyocytes to cardiomyocytes in the heart offers another approach to potentially replenish cardiomyocytes without requiring in vitro culture and expansion while avoiding the need for immunosuppression. Transduction of three transcription factors, Gata4, Mef2c, and Tbx5 (GMT), into neonatal mouse cardiac fibroblasts directly reprogrammed these cells into cardiomyocytes (Ieda et al., 2010) and retroviral delivery of these same GMT factors, with (Song et al., 2012) or without (Qian et al., 2012) the addition of the Hand2 transcription factor, can direct fibroblast-to-cardiomyocyte transdifferentiation in mice in vivo. Human fibroblasts have proven to be more challenging to reprogram into cardiomyocytes than neonatal mouse cardiac fibroblasts, requiring additional factors (GMT + ESSRG, MESP1, MYOCD, and ZEPM2 in (Fu et al., 2013), GMT + Hand2, T-box5, myocardin, microRNA-1 (miR-1) and miR-133 in (Nam et al., 2013), or GMT + miR-133 in (Zhou et al., 2019)) for successful transdifferentiation. Further work will need to be performed to ensure safety and efficacy of this approach.

Studies to date suggest that cardiomyocyte replacement therapy can remuscularize the heart and improve systolic function, but this still needs to be confirmed in larger studies in both animals and humans. The optimal method for delivery of allogeneic stem cell-derived cardiomyocytes remains under debate (Li et al., 2021), and successful strategies will need to minimize arrhythmogenic risk and bypass rejection by the immune system. Direct reprogramming methods will need to be optimized to provide precise spatiotemporal control of cells undergoing transdifferentiation.

Lesson 5: Important barriers to human therapy are being addressed by fundamental research

Technological advances over the past 20 years have fueled the advances in the field of heart regeneration. For instance, application of genetic lineage tracing methods permitted more robust identification of sources of proliferating cardiomyocytes (Porrello et al., 2011). In addition, the ability to generate induced pluripotent stem cells from adult cells overcame the ethical constraints of using embryonic stem cell-derived cardiomyocytes for clinical use or disease modeling (Takahashi and Yamanaka, 2006). This also prompted discovery of factors that enable transdifferentiation of non-cardiomyocytes to cardiomyocytes (Ieda et al., 2010). Completion of the Human Genome Project in 2003 (International Human Genome Sequencing, 2004) and the invention of CRISPR/Cas9 gene editing (Jinek et al., 2012) have accelerated the ability to precisely target and modify genes to either model or cure disease throughout the body. However, barriers to clinical translation of regenerative approaches in the human heart remain (Figure 5). Gene therapy methods to stimulate cardiomyocyte proliferation or induce transdifferentiation of fibroblasts into cardiomyocytes have been proposed, but further work will need to enhance delivery efficiency and ensure safety (Cannata et al., 2020; Yamakawa and Ieda, 2021). As described, cardiomyocyte replacement methods are hampered by inadequate cardiomyocyte maturation likely increasing the risk of ventricular arrhythmias after delivery and rejection of allogeneic cells by the immune system.

Figure 5. Challenges to clinical translation of myocyte therapy approaches.

Cell therapy with pluripotent stem cell-derived cardiomyocytes can lead to ventricular arrhythmias, thought to be due to delivery of immature cardiomyocytes that continue to exhibit automaticity, or spontaneous beating. Combination of non-cardiomyocytes into a tissue engineered structure can promote maturation of cardiomyocytes. Allogeneic stem cells have risk of immune rejection, thus strategies to eliminate human leukocyte antigens may permit cell survival without immunosuppression. Methods to promote direct reprogramming of non-cardiomyocytes into cardiomyocytes within the heart are inefficient and long-term safety concerns such as uncontrolled transdifferentiation or tumor formation are unclear. Use of transient delivery systems with improved efficiency may make this approach more clinically feasible.

Early attempts at gene therapy in the heart involved delivery of DNA encoding vascular endothelial growth factor (VEGF) to promote therapeutic angiogenesis in patients with coronary artery disease, but results have been overall unsuccessful at improving cardiac perfusion (reviewed in (Giacca and Zacchigna, 2012)). Adenovirus and adeno-associated virus (AAV) vectors have been designed that preferentially target the heart, although route of administration remains a challenge to minimize systemic effects and maximize delivery efficiency (Cannata et al., 2020). Newer generation AAV vectors with capsid variants selected through directed evolution can reach target organs with higher efficiency (Tabebordbar et al., 2021). Gene therapy methods to stimulate cardiomyocyte proliferation have been proposed (Cannata et al., 2020) but safety mechanisms will be needed to prevent indefinite reactivation of the cell cycle. New methods such as CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9-mediated gene editing approaches have significant promise for correcting gene mutations causing heritable diseases. Current attempts have led to mosaic patterns of gene editing in the heart in mice (Johansen et al., 2017), thus improved methods of delivery and targeting will need to be developed before this technology will be ready for clinical trials. Lipid nanoparticles offer an additional method to enhance delivery of nucleic acids that have recently been clinically used in RNA vaccines for COVID19 and show promise for delivery to other organs (Hou et al., 2021). Modified RNA (modRNA) has also become increasingly attractive as a method of gene delivery to the heart owing to its more immediate gene expression kinetics, improved gene transfer efficiency, and reduced safety concerns compared to viral vectors used to delivery DNA (reviewed in (Chanda et al., 2021; Kaur and Zangi, 2020)). RNA interference with small interfering RNAs (siRNAs) can be used to silence expression of targeted genes, with much promise at treating a variety of cardiovascular diseases (Laina et al., 2018) and/or stimulate cardiomyocyte proliferation (selected references described in Lesson 1). In addition to miRNAs to promote cardiomyocyte proliferation described in Lesson 1, miRNA strategies may also modify cardiac hypertrophy, as overexpression of miR-133 can inhibit cardiac hypertrophy in mice (Care et al., 2007), while inhibition of miR-212/132 can prevent cardiac hypertrophy induced by pressure-overload in mice (Ucar et al., 2012). Long non-coding RNAs (lncRNAs) can act as a counter balance for miRNAs; for example, the lncRNA, CRRL, binds to miR-199a-3p, leading to suppression of cardiomyocyte proliferation, while silencing of CRRL preserves cardiac function in adult rats after myocardial infarction (Chen et al., 2018). Further work will be needed to better understand the mechanisms of action of ncRNAs, optimize delivery methods, and test safety and efficacy in animal and human studies.

Partial maturation of PSC-CMs can be achieved via prolonged time in culture (Lundy et al., 2013), electrical stimulation (Nunes et al., 2013), or through affecting metabolic pathways (Feyen et al., 2020; Hu et al., 2018; Yang et al., 2014). Advanced maturation of PSC-CMs with adult CM-like contractile phenotype can be achieved when PSC-CMs are cultured in combination with dermal fibroblasts and with electrical stimulation in a custom bioreactor (Ronaldson-Bouchard et al., 2018). Similarly, engineered human myocardium (EHM) structures formed from a combination of PSC-CMs with fibroblasts exhibit structural, contractile, and electrophysiological phenotypes similar to post-natal myocardium (Tiburcy et al., 2017). EHMs are currently being studied in the BioVAT-HF phase I/II clinical trial in Germany to test whether such structures can be used as a cardiac patch (ClinicalTrials.gov, 2020). Preclinical strategies to perform three-dimensional printing of cardiovascular tissue (Kozaniti et al., 2021), develop electrically conductive biomaterials (Esmaeili et al., 2021), or optimize hydrogel properties for cardiac tissue engineering (Sharma et al., 2021) have been reviewed recently in detail.

Use of allogeneic stem cells for cardiac cell therapy will require chronic immunosuppression to prevent rejection of the donor cells. While it may be possible to produce autologous pluripotent stem cells, autologous cell therapy is less feasible in most situations due to the cost of producing individual cell lines and the need for off-the-shelf therapies in most heart diseases. Strategies to develop hypoimmunogenic autologous cell lines have shown promise in cardiac applications in animal models. Early studies demonstrated that depletion of human leukocyte antigen class I (HLA-I) molecules by suppression or deletion of beta-2-microglobulin (B2M) with or without deletion of HLA class II (HLA-II) molecules by deletion of class II major histocompatibility class transactivator (CIITA) did not affect the ability of PSCs to differentiate into cardiomyocytes in vitro (Karabekian et al., 2015; Mattapally et al., 2018). More recently, cardiomyocytes differentiated from genetically edited PSCs with deletion of HLA-I and HLA-II and overexpression of CD47 (to suppress phagocytosis) can evade the immune system in allogeneic recipients and successfully engraft without use of immunosuppression in mice (Deuse et al., 2019; Deuse et al., 2021). Such strategies that minimize the need for immunosuppression may make allogeneic cell therapy approaches more clinically feasible.

Application of fundamental principles to improve human health requires ongoing discovery and innovation, and this process has accelerated over the past 20 years due to numerous scientific breakthroughs. Many strategies to regenerate the heart in mammals have been tested in animal models but have yet to move into the clinic. We anticipate that the next 20 years will bring successful clinical translation of new therapies in the heart as the field overcomes existing barriers to ensure safe and efficacious methods to treat heart disease.

Future Directions

Future work will require optimization of current technologies to refine cell or nucleic acid delivery methods, dosing, and patient selection, while also expanding the repertoire of potentially clinically viable approaches to promote heart regeneration. Our understanding of genetic causes of disease has deepened tremendously over the past 20 years, and we are realizing the importance of epigenetic mechanisms such as DNA methylation, histone acetylation, or non-coding RNAs during development and disease (van Weerd et al., 2011; Zhang et al., 2018b). Further work will be needed to better understand how epigenetic mechanisms can be harnessed to regenerate the heart (Cui et al., 2018; Soler-Botija et al., 2020). Metabolic signaling via nutrient availability (Bae et al., 2021) or reactive oxygen species (Liang et al., 2020) can also be manipulated to promote cardiac regeneration. Extracellular vesicles are membrane-enclosed vesicles secreted by most cell types that can carry proteins and/or nucleic acids and can facilitate communication between cells (Saludas et al., 2021), and they can be engineered to enhance delivery of bioactive materials to the heart (de Abreu et al., 2020). Delivery of extracellular vesicles secreted by cardiac MSCs after myocardial infarction can increase angiogenesis, cardiomyocyte proliferation, and cardiac function in mice (Ju et al., 2018). Studies are needed to understand the mechanisms by which extracellular vesicles may be able to promote cardiac regeneration as a cell-free approach is attractive for potential clinical translation. Finally, delivery of autologous mitochondria via intracoronary delivery can support cardiac function in a pig model of myocardial ischemia-reperfusion injury (Blitzer et al., 2019). Additional studies are needed to understand how mitochondrial transplantation might promote heart regeneration (Mietsch and Hinkel, 2021).

Current animal models used to represent human disease, including coronary artery ligation to mimic ischemic cardiomyopathy or transverse aortic constriction to mimic pressure overload as seen in aortic stenosis, do not fully recapitulate the pathophysiology of chronic heart failure and do not adequately capture the diversity of molecular mechanisms that lead to heart failure (reviewed in (Houser et al., 2012)). For example, chronic atherosclerosis typically precedes acute myocardial infarction in adults, which leads to vascular remodeling not represented in acute coronary ligation in a mouse model (Riehle and Bauersachs, 2019). More accurate small animal models to represent human heart failure with are needed to perform early efficacy testing of potential novel therapies, whether it be stimulation of native cardiomyocyte proliferation potential, cell delivery, or direct reprogramming, to facilitate identification of promising approaches and accelerate clinical translation.

Conclusions

There is enormous demand for regenerative therapies to treat heart failure. The past 20 years provided a strong foundation of rigorous studies that have redefined our understanding of cardiomyocyte biology. We are now appropriately focusing on the cardiomyocyte as a key cell in heart regeneration. We anticipate that the next 20 years will lead to acceleration of academic-industry partnerships, as promising approaches to promote heart regeneration move towards clinical translation. Improved understanding of molecular mechanisms that guide cardiomyocyte proliferation, differentiation, and maturation and the role of non-cardiomyocytes in supporting cardiac function will be necessary to target desired pathways to achieve robust cardiac regeneration. Although devices, drugs and heart transplantation will remain mainstays of heart failure treatment, there may be a time when heart regeneration biology provides long-term biological solutions for many patients.