The Hippo–YAP (Yes-associated protein) pathway is an evolutionarily and functionally conserved regulator of organ size and growth with crucial roles in cell proliferation, apoptosis, and differentiation. This pathway has great potential for therapeutic manipulation in different disease states and to promote organ regeneration. In this Review, we summarize findings from the past decade revealing the function and regulation of the Hippo–YAP pathway in cardiac development, growth, homeostasis, disease, and regeneration. In particular, we highlight the roles of the Hippo–YAP pathway in endogenous heart muscle renewal, including the pivotal role of the Hippo–YAP pathway in regulating cardiomyocyte proliferation and differentiation, stress response, and mechanical signalling. The human heart lacks the capacity to self-repair; therefore, the loss of cardiomyocytes after injury such as myocardial infarction can result in heart failure and death. Despite substantial advances in the treatment of heart failure, an enormous unmet clinical need exists for alternative treatment options. Targeting the Hippo–YAP pathway has tremendous potential for developing therapeutic strategies for cardiac repair and regeneration for currently intractable cardiovascular diseases such as heart failure. The lessons learned from cardiac repair and regeneration studies will also bring new insights into the regeneration of other tissues with limited regenerative capacity.
The Hippo–YAP (Yes-associated protein) pathway is an evolutionarily conserved pathway that controls organ size.
Hippo signalling restrains cardiomyocyte proliferation during development to control cardiac size.
The Hippo–YAP pathway regulates the activity of growth pathways during prenatal and postnatal life and is important for cardiac homeostasis.
Hippo signalling inhibits adult cardiac regeneration.
The Hippo–YAP pathway regulates various events during cardiac regeneration, including cardiomyocyte proliferation and differentiation, injury resistance, stress response, and mechanical signals.
Manipulating the Hippo–YAP pathway is a potential therapeutic tool for treating cardiac injury.
How is organ size reproducibly regulated during development? Can tissues and organs achieve self-renewal during homeostasis and after injury, and how is self-renewal achieved? These fundamental questions are of great interest to biologists, and tremendous progress has been made in answering them. In the past decade, attention has been intensely focused on the Hippo–YAP (Yes-associated protein) pathway, which is an organ size and growth regulator that is evolutionarily and functionally conserved among diverse species. Since the initial discovery of the Hippo–YAP pathway in genetic screens in Drosophila melanogaster, this pathway has emerged as an important regulator of tissue renewal. Considerable insights into the Hippo–YAP pathway have been generated with the use of genetic models and biochemical studies1,2,3,4,5,6.
After the first report showing that repression of Hippo signalling increases heart size during mouse development7, numerous studies have revealed a pivotal role of the Hippo–YAP pathway in cardiac development, growth, homeostasis, disease, and regeneration. Because the human heart lacks the capacity to self-repair after cardiac injury such as myocardial infarction (MI), cardiac pumping function is progressively lost and eventually leads to heart failure — the leading cause of death in the Western world8. An enormous unmet clinical need exists for alternative methods of treating heart failure. Currently, the only definitive treatment for heart failure is cardiac transplantation, which is limited by a scarcity of donor hearts. Thus, developing new approaches to induce endogenous repair after cardiac injury is critical. Notably, studies in the past 5 years have indicated that inhibition of Hippo signalling results in cardiac regeneration9,10,11,12. These important studies suggest that endogenous cardiac regeneration can be induced to repair and regenerate an injured heart and that the Hippo–YAP pathway is a promising therapeutic target to trigger endogenous heart regeneration. Moreover, this work has shed light on adult tissue renewal and regenerative medicine.
In this Review, we provide a summary of the Hippo–YAP pathway and discuss the current understanding of the roles of this pathway in heart development, growth, homeostasis, disease, and regeneration. We focus particularly on the role of the Hippo–YAP pathway in regulating cardiomyocyte proliferation and cardiac repair and regeneration.
Overview of the Hippo–YAP pathway
Although much remains to be learned about the Hippo–YAP pathway, elegant genetic and biochemical studies have provided important insights into this pivotal pathway (summarized in Fig. 1). The core mammalian Hippo signalling components include the tumour suppressors mammalian STE20-like protein kinase 1 (MST1; also known as STK4) and MST2 (also known as STK3), which are orthologous to Hpo in Drosophila; the scaffold protein salvador homologue 1 (SAV1; orthologous to Salvador in Drosophila); large tumour suppressor homologue 1 (LATS1) and LATS2 (which are orthologous to nuclear Dbf2-related family protein kinase Warts in Drosophila); and the scaffolding proteins MOB domain kinase activator 1 A (MOB1A) and MOB1B (which are orthologous to Mats in Drosophila). As shown in Fig. 1, MST1, MST2, and SAV1 form a complex that phosphorylates and activates LATS1 and LATS2 kinases that interact with the cofactor MOB1, which further phosphorylate the transcriptional co-activators YAP and TAZ (transcriptional co-activator with PDZ-binding motif; also known as WWTR1), which are the downstream effectors of the Hippo signalling pathway. Activation of Hippo signalling inhibits the transcriptional activity of YAP and TAZ by preventing their translocation into the nucleus and promoting their degradation in the cytoplasm. In the absence of repression by Hippo signalling, YAP and TAZ can partner in the nucleus with different transcription factors such as TEA domain transcription factor family members (TEADs) to regulate the transcription of target genes, such as genes encoding proteins regulating cell proliferation and survival4.
Although YAP and TAZ most commonly interact with TEADs, these co-activators have also been shown to interact with other transcription factors including CCAAT/enhancer-binding protein-α (C/EBPα)13, cAMP-responsive element-binding protein (CREB)14, early growth response protein 1 (EGR1)15,16, receptor tyrosine-protein kinase erbB4 (ERBB4)17, proto-oncogene Fos (FOS)18, forkhead box protein O1 (FOXO1)19, FOXM1 (ref.20), hypoxia-inducible factor 1α (HIF1α) and HIF1β21,22, zinc-finger GLI proteins23, Krüppel-like factor 5 (KLF5)24,25, myoblast determination protein 1 (MYOD1)26, octamer-binding transcription factor 4 (OCT4; also known as POU5F1)27, paired box protein Pax3 (PAX3)28,29, pituitary homeobox 2 (PITX2)30, peroxisome proliferator-activated receptor-γ (PPARγ)31, tumour proteins 63 (p63)32,33 and 73 (p73)34,35,36, runt-related transcription factor 1 (RUNX1) and RUNX2 (refs31,37,38,39), SMAD family members40,41,42,43,44,45,46, T-box transcription factor TBX5 (refs47,48), and homeobox protein Nkx2.1 (NKX2.1)49. In addition, some factors inhibit the activity of YAP and/or TAZ as nuclear transcription cofactors. The scaffold protein angiomotin-like protein 1 (AMOTL1) can physically interact with YAP and protocadherin Fat4 (FAT4) to form a complex that sequesters YAP in the cytoplasm50. Transcription cofactor vestigial-like protein 4 (VGLL4) competes directly with YAP for the binding of TEADs in the nucleus51,52,53. The tumour suppressor NF2 (neurofibromin 2; also known as Merlin) also functions as an inhibitor of YAP, primarily by activating Hippo kinase activity54,55.
Hippo–YAP pathway in heart development
Overview of heart development
During early mammalian development, morphogen gradient signalling in the epiblast gives rise to primitive cardiac mesodermal cells marked by expression of the transcription factor mesoderm posterior protein 1 (MESP1)56. Subsequent canonical WNT–β-catenin signalling directs cardiac cell commitment and the migration of MESP1+ cells to form the anterior lateral plate mesoderm, which later gives rise to the cardiac crescent and the pharyngeal mesoderm, situated dorsomedial to the cardiac crescent57,58,59. In MESP1+ cells, bone morphogenetic protein (BMP) signalling then triggers cardiomyogenesis by inducing the expression of the cardiac transcription factors genes Gata4, Isl1, Mef2c, Nkx2-5, and Tbx5 (ref.60). Importantly, inhibitory WNT signalling immediately after this stage is critical for the commitment of cardiac precursors to maintain the cardiac fate61. This cardiac precursor commitment marks the establishment of two cardiac progenitor fields: the first heart field (FHF) in the cardiac crescent and the second heart field (SHF) in the pharyngeal mesoderm. At this stage, FHF cells differentiate via BMP, fibroblast growth factor (FGF), and noncanonical WNT signalling pathways62. By contrast, the proliferation of SHF cells is maintained by canonical WNT, FGF, and sonic hedgehog signalling pathways62. Importantly, WNT signalling in the SHF is crucial for the development of the right ventricle and interventricular myocardium63. Genetic studies in mice revealed that the deletion of the gene encoding the WNT effector β-catenin in the SHF produces hearts with grossly undersized right ventricles63. In addition, several studies have indicated that the transcription factor TBX5 is required for proper formation of the vertebrate heart64. Whereas expression of mutant tbx5 in Xenopus inhibits the formation of the primitive heart altogether, Tbx5-mutant mice have defects in heart tube, septation, and cardiac conduction64. The final major stage of prenatal cardiac development is morphogenesis, which progresses as the linear heart tube forms, followed by cardiac looping and septation. Cardiac development is reviewed in further detail in another article in this Focus Issue65.
Hippo–YAP pathway in cardiac development
As highlighted above, much is known about how cell-extrinsic growth factor signalling from BMP, WNT, insulin-like growth factor (IGF), and neuregulin pathways regulates cardiomyocyte proliferation during cardiac development66. The cell-intrinsic mechanisms that regulate cardiac growth and size control largely remained a mystery until 2011, when an exciting flurry of genetic studies in mice revealed that Hippo signalling has major cell-intrinsic roles in maintaining proper heart size throughout prenatal development and adult life7,67,68.
As mentioned previously, studies in Drosophila indicated that components of the Hippo pathway are required for maintaining overall organ size. In vitro studies revealed that the Hippo effector TAZ is a co-activator of TBX5 transcriptional activity, suggesting that the Hippo pathway has roles in cardiac development47. To investigate this possibility further, Heallen and colleagues deleted Sav1 specifically in the embryonic mouse heart by crossing mice with a Sav1 conditional-knockout allele with mice with a Nkx2-5–Cre driver7. In this Sav1 conditional-knockout strain, Sav1 is deleted in cardiac progenitors by embryonic day 7.5 (E7.5)69. By E12.5, Sav1-knockout hearts showed expanded ventricular myocardial layers without a change in cardiomyocyte size7. At birth, these mice had cardiomegaly owing to increased cardiomyocyte proliferation while maintaining preserved patterning in the heart7. Conditional inactivation of other core Hippo pathway genes, Mst1, Mst2, or Lats2, confers similar cardiac phenotypes7. Moreover, embryonic deletion of the gene encoding the Hippo effector YAP by using cardiac-specific Tnnt2–Cre and Nkx2-5–Cre results in prenatal lethality characterized by hearts with severely thinned myocardial layers67,68. In these hearts, cardiomyocyte proliferation is substantially reduced67,68. Conversely, activation of YAP in embryonic mouse hearts increases cardiomyocyte proliferation67,68. Consistent with these data, studies using gene trapping in embryonic mouse hearts revealed that disruption of TEAD1 results in thinned ventricular walls, causing embryonic lethality70, indicating that TEAD1 function is essential for heart development. Steadily increasing the expression of Tead1 in postnatal mouse hearts leads to contractile dysfunction and interstitial fibrosis, which are hallmarks of remodelling and pathological hypertrophy in failing hearts71. Collectively, these observations established that the regulation of cardiomyocyte proliferation by the Hippo–YAP pathway is an organ-intrinsic pathway that controls heart size during cardiogenesis4.
In 2018, Xiao and colleagues showed that Hippo signalling also has a critical role in epicardium-related heart development72. The epicardium contains essential noncardiomyocyte progenitors that undergo epithelial-to-mesenchymal transition (EMT) and give rise to epicardial-derived cells (EPDCs). EPDCs develop into cardiac fibroblasts and vascular smooth muscle cells73,74, which are the primary supporting cells of the heart and are essential for myocardial and coronary vascular development75. Xiao and colleagues found that deletion of Lats1 and Lats2 specifically in the epicardium led to embryonic lethality by E15.5, with coronary vasculature defects72. In addition, single-cell RNA sequencing showed that epicardial cells with deletion of both Lats1 and Lats2 did not differentiate into fibroblasts and remained in an intermediate subepicardial-like state in which the cells displayed both epicardial and fibroblast characteristics72. Sequential genetic and pharmacological studies further revealed that fibroblast differentiation arrest was mediated by YAP, which concurrently controlled extracellular matrix (ECM) composition and vascular remodelling during heart development72. Notably, Lats1−/−Lats2−/− mutant embryos with heterozygous deletion of Yap and Taz were viable at E15.5 without major coronary vasculature defects72. Another study published in 2016 reported EMT defects in mouse embryos with epicardial deletion of Yap and Taz76. Although the role of the Hippo–YAP pathway in noncardiomyocytes remains less well understood, these findings indicate that the Hippo–YAP pathway is required for cardiac fibroblast differentiation and coronary vessel development.
Crosstalk with other signalling pathways
In many contexts, the Hippo–YAP pathway crosstalks with other major signalling pathways, including WNT–β-catenin, BMP, transforming growth factor-β (TGFβ), Notch, G protein-coupled receptor (GPCR), and phosphoinositide 3-kinase (PI3K)–RACα serine/threonine-protein kinase (AKT) signalling pathways1,2,3,4,5,6. In addition, a broad range of factors regulate activity of the Hippo–YAP pathway, including epigenetic factors, mechanical and hormonal signals, cell–cell contact, cell polarity, and the cytoskeleton1,2,3,4,5,6. Thus, the Hippo–YAP pathway is a ‘signal integrator’, in turn regulating a wide variety of genes such as Aurkb, Dhrs3, Dpp4, Fgd4, Ldha, Lin9, Myh7, Ndufb3, Nppa, Nppb, Oxnad1, Park2 (also known as Prkn), Pkp4, Sntb1, and Sgcd, which encode proteins that are critical for various cellular events such as cell proliferation, survival, differentiation, migration, and metabolism1,2,3,4,5,6,10,30,72.
Expression profiling studies of embryonic hearts from Sav1 conditional-knockout mice revealed the upregulation of transcript levels of known canonical WNT–β-catenin target genes7, including Sox2 (encoding a protein that has important roles in proliferation, cellular reprogramming, and cardiac repair), Snai2 (encoding a tumorigenesis factor that regulates EMT), and the cell survival-related genes Birc2 and Birc5. Immunofluorescence studies of Sav1-knockout mouse hearts revealed that the levels of nuclear β-catenin (a readout of WNT signalling) were significantly upregulated in Sav1-mutant cardiomyocytes7. Furthermore, β-catenin levels were stabilized in mouse hearts expressing a constitutively active form of YAP (YAP-S112A)68. Together, these data support the notion that Hippo signalling inhibits WNT signalling in developing hearts. Interestingly, studies in noncardiac tissues indicated that the Hippo effector TAZ binds and inhibits the WNT pathway component dishevelled in the cytosol, thereby promoting β-catenin degradation77. This observation raises the intriguing possibility that, in the heart, Hippo signalling regulates WNT in a biphasic fashion. Given the previously reported role for WNT signalling in promotion of cardiac size, investigators generated mice mutant for both Sav1 and Wnt to examine genetically whether Hippo signalling antagonizes WNT-dependent cardiac growth7. Indeed, compared with Sav1 conditional-knockout mouse embryos, Sav1 conditional-knockout mouse embryos with a lower dosage of β-catenin showed significantly reduced myocardial thickness and ventricular cardiomyocyte proliferation rates7. Furthermore, biochemical studies revealed that YAP–TEAD forms a complex with β-catenin–T cell factor/lymphoid enhancer factor (TCF/LEF) on the promoter elements of several proliferation-related genes7. These data collectively indicate that Hippo signalling inhibits the interaction between YAP and β-catenin in the nucleus, resulting in the suppressed expression of growth-related genes during cardiac development (Fig. 2).
Crucial developmental studies have revealed that the IGF–PI3K–AKT signalling pathway is an essential downstream node of Hippo–WNT signalling during the regulation of cardiac growth. Transcriptome profiling revealed that the expression of core genes in the IGF signalling pathway is significantly upregulated in mouse embryonic hearts expressing the constitutively active YAP-S112A68. Among these genes are Ctnnb1, Igf1, Igfbp2, Igfbp3, and several β-catenin transcriptional target genes. In addition, levels of PI3K and phosphorylated AKT are increased in mouse cardiomyocytes expressing YAP-S112A, indicating that YAP stimulates IGF pathway activity68. Importantly, IGF–AKT signalling triggers the inhibitory phosphorylation of the β-catenin destruction complex protein glycogen synthase kinase 3β (GSK3β)68. Accordingly, levels of active β-catenin were elevated in YAP-S112A-expressing cardiomyocytes, as reported previously78. Furthermore, inhibition of IGF reduces YAP activity and β-catenin levels in cardiomyocytes68. Consistent with these data, Tumaneng and colleagues have reported that YAP activates the PI3K–mechanistic target of rapamycin (mTOR) growth pathway via miR-29-mediated downregulation of the tumour suppressor phosphatase and tensin homologue (PTEN)79. Whether Hippo regulates PI3K–mTOR-dependent growth in the heart remains to be determined. However, these findings further define the Hippo–WNT interaction during cardiac development and reveal that YAP transcriptional activity is crucial for activating the IGF pathway to promote β-catenin-dependent cardiomyocyte proliferation (Fig. 2).
Hippo–YAP pathway in postnatal heart
During fetal development, cardiac growth is primarily mediated by the proliferation of pre-existing cardiomyocytes, whereas postnatal heart growth is predominantly governed by physiological hypertrophy of cardiomyocytes rather than by proliferation80. Pathological conditions such as ischaemia–reperfusion injury, MI, and valvular heart disease induce adaptive hypertrophy, which increases cardiac wall thickness to reduce stress, which in turn leads to reduced myocardial oxygen consumption as a compensation for insufficient contractile mass81. Common injuries to the human heart such as ischaemia–reperfusion injury can lead to cardiomyocyte death, mainly because of the production of reactive oxygen species (ROS)82, and this cardiomyocyte loss usually leads to the development of heart failure.
During embryonic heart development in mice, YAP is required for cardiomyocyte proliferation and does not affect fetal cardiomyocyte size, suggesting that YAP does not promote physiological hypertrophic growth of cardiomyocytes67. Interestingly, expression of the constitutively active YAP form, YAP-S112A, driven by the promoter of Myh7 (encoding β-myosin heavy chain) — the transgene activity in the heart begins at E9 — increases cardiomyocyte proliferation; however, when these transgenic mice reach adulthood, the size of the heart is normal, suggesting that a decrease in cardiomyocyte size occurs to compensate the cardiac mass68. Postnatal, cardiomyocyte-specific deletion of Yap with the use of an Myh6 (encoding α-myosin heavy chain)–Cre driver resulted in elevated cardiomyocyte apoptosis and dilated cardiomyopathy in mice83. Mice with heterozygous deletion of Yap had no obvious cardiac phenotype at baseline but after MI showed increased cardiomyocyte apoptosis, decreased compensatory cardiomyocyte hypertrophy, and cardiac dysfunction83. Thus, in response to stress, compensatory hypertrophy requires a certain level of YAP, and YAP loss of function leads to insufficient cardiac hypertrophy and cardiac dysfunction. YAP has been reported to activate mir-206 expression in cardiomyocytes84. Increased mir-206 expression induces cardiac hypertrophy. Conversely, decreased mir-206 expression attenuates YAP-induced cardiac hypertrophy84. In addition, both postnatal overexpression of Yap and inhibition of Hippo signalling (for example, through inactivation of SAV1) have protective effects on the injured heart, improving cardiac function and survival after MI9,11.
In addition to the primary role of the Hippo–YAP pathway in regulating cell proliferation during embryonic cardiac development, the Hippo–YAP pathway has important functions in cardiac homeostasis and cell survival at postnatal stages. However, the role of the Hippo–YAP pathway in the regulation of cardiomyocyte hypertrophy remains unclear. YAP might act through a dose-dependent mechanism to regulate either cardiomyocyte proliferation or hypertrophy, suggesting that the role of YAP in cardiac hypertrophy is more complex than previously believed.
Hippo–YAP pathway in cardiac disease
In heart samples from patients with hypertrophic cardiomyopathy and in mice with transverse aortic constriction (TAC), YAP levels are elevated and the inhibitory phosphorylation at Ser127 of YAP is decreased85, suggesting a role for nuclear YAP in hypertrophic heart disease. Another study showed that the levels of phosphorylated YAP and phosphorylated LATS were elevated in heart samples from patients with ischaemic or nonischaemic heart failure compared with nonfailing heart samples, whereas levels of SAV1 were unchanged, indicating that Hippo activity is increased in patients with heart failure10.
Hippo kinases induce the inhibitory phosphorylation at Ser127 of YAP and thereby, in many contexts, promote YAP degradation in the cytoplasm. Repression of upstream Hippo kinases induces the nuclear accumulation of YAP. Studies have shown that injury stress, such as that from ischaemia–reperfusion injury, MI, or pressure overload, activates MST1, which increases caspase activation and cardiomyocyte apoptosis86,87,88 and presumably increases levels of phosphorylated YAP. Cardiac-specific overexpression of Mst1 in mice results in the progressive deterioration of cardiac function, perhaps as a result of elevated cardiac wall stress that leads to increased myocardial oxygen consumption and apoptosis86. Unlike Mst1 overexpression that leads to dilated cardiomyopathy and premature death in mice, the overexpression of a dominant-negative form of MST1, which suppresses endogenous MST1, does not cause significant changes in cardiac function or cardiac hypertrophy in mice at baseline conditions. By contrast, overexpression of a dominant-negative form of MST1 improves cardiac function and protects the heart against cardiomyocyte death caused by ischaemia–reperfusion injury and MI86,87.
Ras association domain-containing protein 1 isoform A (RASSF1A) is an endogenous activator of MST1 in the heart, and RASSF1A protein and mRNA levels are significantly increased in mouse hearts 1 week after TAC, whereas RASSF1A protein levels are significantly decreased in long-term hypertrophic mouse hearts (12 weeks after TAC) and in human failing hearts88,89. Accordingly, both cardiomyocyte-specific deletion of Rassf1a with the use of Myh6–Cre and expression of an MST1-binding-deficient form of RASSF1A under the Myh6 promoter (which blocks the activation of MST1 specifically in cardiomyocytes) attenuates the hypertrophy and fibrosis caused by TAC-induced pressure overload in mice88. Unlike the overexpression of a dominant-negative form of MST1, the specific overexpression of a dominant-negative form of LATS2 by using an Myh6–Cre driver increases cardiomyocyte size, with modest cardiac hypertrophy90.
A 2016 study showed that oxidative stress by hydrogen peroxide treatment activates NF2 expression in isolated cardiomyocytes in vitro and that ischaemia–reperfusion injury in mice activates NF2 expression in the myocardium in vivo55. In the adult mouse heart, elevated NF2 levels increased cardiomyocyte apoptosis through the regulation of Hippo signalling mediated by the activation of MST1 and the inhibition of YAP. Cardiomyocyte-specific knockout of Nf2 protects mice against ischaemia–reperfusion injury, but mice with deficiency in both NF2 and YAP lose the protection against ischaemia–reperfusion injury, indicating that NF2 functions primarily through YAP activation in this context55. In addition, the Hippo pathway has been reported to have a role in the pathogenesis of arrhythmogenic cardiomyopathy, in which cardiomyocytes are replaced by fibro-adipocytes. In both a mouse model of arrhythmogenic cardiomyopathy and in patients with arrhythmogenic cardiomyopathy, the protein constituents of intercalated discs are altered, leading to the pathogenic activation of NF2 in the heart91. Consequently, activated NF2 triggers the phosphorylation of Hippo kinases MST1, MST2, LATS1, and LATS2 and of the Hippo effector YAP91.
Hippo–YAP in cardiac regeneration
Unlike the hearts of lower vertebrates, such as amphibians and fish, which have a robust capacity for regeneration throughout life92,93, the mammalian heart retains its regenerative capacity only during early life and loses this potential postnatally78,94. Because the human adult heart lacks regenerative capacity, the massive loss of cardiomyocytes due to cardiac injury such as MI cannot be reversed. Instead, fibrosis and scar formation ensue, compromising heart function95,96. As compensation for the reduced pumping action and to maintain cardiac function, the body activates the sympathetic adrenergic system, which acutely increases cardiac output while also increasing strain on residual cardiomyocytes through elevated oxygen consumption and increased heart rates. These events produce a vicious cycle that leads to deleterious cardiac remodelling and additional cardiomyocyte loss, eventually ending in death95,96. The primary reason for poor heart regeneration is that the mammalian heart cannot efficiently generate new cardiomyocytes in response to injury. However, seminal studies in zebrafish and neonatal mice have revealed that their heart can regenerate after cardiac damage78,97,98. In these models, differentiated cardiomyocytes re-enter the cell cycle and replace scar tissue at the injury site78,97,98. Until postnatal day 7, the newborn mouse heart maintains its regenerative capacity in response to injury, but this capacity is lost in the adult mouse heart78. These discoveries have sparked interest in the mechanisms that control the cardiac cell cycle, and the identification of such mechanisms has laid the groundwork for cardiac regenerative therapies. Several lines of evidence have revealed that Hippo signalling prevents cell cycle re-entry in postnatal cardiomyocytes. Remarkably, genetic knockout of Hippo pathway components in adult cardiomyocytes stimulates cardiomyocyte progression through the S phase, mitosis, and cytokinesis9. Likewise, the overexpression of activated YAP is sufficient to promote the proliferation of postmitotic cardiomyocytes67. These studies have also revealed that, as the heart ages and cardiomyocytes exit the cell cycle, Hippo activity markedly increases whereas YAP activity sharply declines9,67. Thus, the strategy of exploiting the young heart’s innate regenerative capacity to enhance adult heart regeneration has become a particularly attractive approach for cardiac repair.
Hippo–YAP and cardiomyocyte renewal
The Hippo–YAP signalling pathway has a role in regulating cell proliferation and organ size99,100. Heallen and colleagues first showed that blocking the Hippo pathway by deleting Lats2, Mst1, Mst2, or Sav1 promoted cardiomyocyte proliferation during development7. Several groups have shown that YAP is essential for cardiomyocyte proliferation during mouse embryonic cardiogenesis67,68. Activated YAP promotes neonatal cardiomyocyte proliferation and cytokinesis in vitro67,68, and YAP and TAZ are required for cardiomyocyte proliferation during postnatal life. Postnatal deletion of Yap and/or Taz in mice by using an Myh6–Cre driver101 results in hypoplasia of the myocardium and lethality owing to decreased cardiomyocyte proliferation12. By contrast, postnatal overexpression of Yap increases cardiomyocyte proliferation12,67. Studies from the Radice group have indicated that α-catenins negatively regulate the activity of YAP in mouse heart. Perinatal deletion of αE-catenin and αT-catenin results in increased cardiomyocyte proliferation102,103. Concurrently deleting one Yap allele can rescue this phenotype, suggesting that YAP is the essential downstream effector of α-catenins in the regulation of proliferation102,103. Consistent with the findings described above, a study published in 2016 showed that the cardiac protein VGLL4 antagonizes YAP-dependent cardiac growth in adult mice by disrupting the interaction between YAP and its transcriptional partner TEAD1 (ref.104). In postnatal mouse hearts, VGLL4 is acetylated, binds TEAD1, and targets TEAD1 for degradation. As a result, cardiomyocyte proliferation is restricted104. Taken together, these studies suggest that as the heart develops, increased Hippo signalling and expression of YAP antagonists prevent cardiomyocyte cell cycle re-entry, thereby inhibiting cardiac repair.
Several other studies have examined the role of the Hippo–YAP pathway in models of cardiac injury. Inactivation of Hippo signalling by deleting Sav1 or Lats1 and Lats2 in adult cardiomyocytes promotes efficient heart regeneration in mouse models of postnatal cardiac apex resection and of adult MI9. In addition, YAP has been shown to be required for neonatal mouse heart regeneration12; indeed, forced expression of a constitutively active form of YAP in the adult heart stimulates cardiac regeneration and improves contractility after MI. Similarly, another study reported that YAP deficiency attenuated heart regeneration and reduced cardiomyocyte proliferation83. Of note, in all these studies, the Hippo–YAP pathway was manipulated before injury. To determine the role of the Hippo–YAP pathway in heart repair after MI, Lin and colleagues induced Yap overexpression in adult mouse myocardium 1 week after MI by using the recombinant adeno-associated virus subtype 9 (AAV9) and found that Yap overexpression improved cardiac function and survival after MI11. Importantly, in mice with established ischaemic heart failure, cardiomyocyte-specific deletion of Sav1, 3 weeks after MI, resulted in increased cardiomyocyte renewal, scar size reduction, and the reversal of heart failure10. AAV9-mediated knockdown of Sav1 in cardiomyocytes also resulted in an effective stress response and promoted the recovery of heart function in these mice10. These studies indicate that both the repression of Hippo signalling and the activation of YAP increase cardiomyocyte proliferation by threefold to fivefold compared with controls10,11, suggesting that targeting the Hippo–YAP pathway has potential therapeutic benefits for patients with heart failure (Fig. 3).
Hippo–YAP and the stress response
YAP expression is high in neonatal and juvenile mouse hearts and declines with ageing67. This phenomenon is correlated with the proliferative capacity of cardiomyocytes because adult cardiomyocytes rarely proliferate under homeostatic conditions9,105. Why YAP activity decreases with age remains unclear, but studies suggest that YAP activity is likely regulated through epigenetic or metabolic shifting during cardiomyocyte maturation106,107,108. Energy generation in the neonatal heart occurs mainly through glycolysis, which switches to fatty acid oxidation in the adult heart106,107,108. Changes in metabolic substrate provision during early postnatal life might account for this metabolic switching106. Consistent with this idea, a study reported in a 3D, ex vivo, cardiac organoid culture system that the switching in metabolic substrates from carbohydrates to fatty acids is a central driver of cardiac maturation109. Interestingly, YAP activity was repressed during this maturation process109. An interesting area of future investigation will be studying the underlying mechanisms of this metabolic shift and how these metabolic cues regulate the Hippo pathway.
During myocardial ischaemia, millions of cells die, and the immune system response is triggered, rapidly deploying inflammatory cells to the damaged site. Although the underlying mechanisms regulating inflammation during heart regeneration remain elusive, studies have indicated that inflammation probably has both positive and negative roles in heart regeneration110. Studies have revealed that neonatal heart regeneration requires the secretion of signalling factors from macrophages111 and that the loss of macrophages during cardiac injury in adults increases scar formation112,113. Moreover, cardiac overexpression of Ccl2, encoding a chemokine that recruits and activates monocytes, was shown to improve heart function by inducing macrophage infiltration, angiogenesis, myocardial IL-6 secretion, and the accumulation of cardiac myofibroblasts114. Whether one or several of these inflammatory factors is associated with Hippo signalling in heart regeneration remains uncertain. However, in the gut, IL-6 can increase YAP activity through proto-oncogene tyrosine-protein kinase SRC family kinases and promote intestinal regeneration115 (Fig. 4).
ROS have an essential role in cardiac diseases, as discussed above, as well as in heart regeneration. The level of ROS in the mouse neonatal heart is markedly increased during the first week after birth116 and induces the DNA damage response, which has an important role in the transition from the regenerative to the nonregenerative stage of the heart116. Inhibition of ROS generation and the DNA damage response prolongs the regenerative window of the mouse neonatal heart116.
MST1 is activated by oxidative stress86,87,88,117. In mice, transgenic overexpression of Mst1 is associated with progressive deterioration of cardiac function86. Conversely, inhibition of endogenous MST1 improves cardiac function and reduced fibrosis and cell death after MI87. Ischaemia–reperfusion injury is associated with increased production of ROS82, and RASSF1A has been shown to activate MST1 in the myocardium under ischaemia–reperfusion injury88. Whether and how RASSF1A responds to oxidative stresses such as ROS needs further study. Interestingly, NF2 is also activated by oxidative stress and promotes apoptosis through the activation of MST155 (Fig. 4). MST1 also inhibits the antioxidant response through the suppression of the YAP and FOXO1 interaction, given that the YAP–FOXO1 complex induces the expression of genes related to the antioxidant response19. However, YAP activity was not directly examined in most of these studies. Unlike mice deficient in SAV1 or in both LATS1 and LATS2, which have increased proliferation of cardiomyocytes7,9,10,118, cardiomyocyte proliferation is surprisingly unaffected in Mst1-knockout mice subjected to ischaemia–reperfusion injury54,85, MI86, or TAC87. Possible explanations for these findings are that another protein has redundant functions with MST1, such as MST2, or that MST1 exerts its function via alternative pathways, such as by inhibiting autophagy through the phosphorylation of beclin 1 (ref.119). Alternatively, a threshold of YAP activity might be required for the regulation of apoptosis and proliferation.
Interestingly, ROS activate nuclear factor erythroid 2-related factor 2 (NRF2), which is a transcriptional regulator of the antioxidant response and induces the expression of the Pitx2 homeobox gene after cardiac injury30. NRF2 also facilitates the translocation of PITX2 from the cytoplasm to the nucleus, where PITX2 interacts with YAP and activates the expression of genes encoding electron transport chain components and ROS scavengers. Pitx2 conditional knockout in mouse cardiomyocytes is detrimental for the regeneration of the neonatal myocardium30.
Epicardial-mediated heart regeneration
The epicardium is activated after myocardial injury120,121,122,123. The activated epicardium provides mechanical support to the heart and secretes cytokines to modulate revascularization and repair of the damaged heart. Deletion of Yap and Taz in the epicardium in mice results in cardiomyopathy, death, and profound pericardial inflammation and myocardial fibrosis after MI124. Unlike the role of RASSF1A and MST1 in the regulation of proliferation, apoptosis, and the inhibition of nuclear factor-κB (NF-κB) and tumour necrosis factor (TNF) in fibroblasts88, the loss of Yap and Taz in the epicardium results in decreased levels of the anti-inflammatory cytokine IFNγ and increased inflammation and inflammation-associated fibrosis124 (Fig. 4). More information on the role of the epicardium in cardiac regeneration is described in another article in this Focus Issue125.
Hippo–YAP and mechanical signalling
The Hippo–YAP pathway has been shown to be closely related to mechanical stress responses126,127,128,129,130,131. The mechanical stress response directly associates with heart physiology; therefore, changes in physical pressure by the ECM can alter Hippo signalling during heart development and regeneration. Indeed, nuclear YAP levels in normal adult cardiomyocytes are low, but nuclear YAP is present in infarcted cardiac tissue with stiffer ECM132. Cytoskeleton rearrangement in response to local mechanical cues is also essential for mediating YAP activation103. In agreement with this idea, treatment with a RHO kinase inhibitor blocks the YAP response to mechanical cues103,133.
Interestingly, studies have provided direct evidence connecting the Hippo–YAP pathway with the ECM. Bassat and colleagues reported that ECM composition can regulate cardiomyocyte growth and differentiation during postnatal mouse heart maturation134. The investigators initially identified agrin in an ECM-mediated cardiomyocyte proliferation screen and found that agrin is required for neonatal mouse heart regeneration and promotes cardiomyocyte proliferation after MI134. Mechanistically, the study showed that both YAP and agrin interact with the dystrophin–glycoprotein complex (DGC)134, a multicomponent transmembrane complex linking the actin cytoskeleton to the ECM. Agrin can bind the DGC, resulting in YAP disassociating from the DGC with release of YAP into the cardiomyocyte nucleus, which in turn promotes cardiomyocyte proliferation134. In an independent study, Morikawa and colleagues135 reported that phosphorylated YAP directly interacts with dystroglycan (DAG1), an essential component of the DGC, indicating that Hippo signalling is required for the interaction between YAP and the DGC135. The investigators further studied the function of this interaction in a strain of mice with a null allele of Dmd, which is an experimental model for the debilitating muscle-wasting disease Duchenne muscular dystrophy that is often accompanied by dilated cardiomyopathy. In these mice, Hippo pathway inactivation suppressed cardiomyopathy and resulted in reduced apoptosis and increased cardiomyocyte proliferation135. YAP has a central role in this context, given that the disrupted DGC cannot sequester YAP, resulting in YAP nuclear localization135. Of note, the DGC assembles at the plasma membrane of mature cardiomyocytes136,137, which might be a reason for the decreased YAP activity in the adult heart. Collectively, these studies reveal a pivotal role for the ECM via its interaction with the DGC in regulating cardiomyocyte proliferation and apoptosis through the Hippo–YAP pathway (Fig. 4). Given that ECM components are essential for cardiac physiology, the cell-to-cell interactions that occur during heart regeneration require further study. Studying the expression of ECM components such as agrin in different cardiac cells might provide clues about how the Hippo–YAP pathway can be manipulated to promote heart regeneration.
Whether mechanical stress directly regulates YAP activity through Hippo signalling remains debatable. Studies in the past 7 years have reported that mechanical cues regulate YAP activity through a noncanonical Hippo signalling mechanism (that is, Hippo-independent)50,102,103,128,129,138. In particular, in the Fat4-knockout mouse heart, which has increased nuclear YAP activity, the phosphorylation levels of the Hippo kinases LATS and MST are unchanged50. However, as mentioned above, YAP phosphorylation by Hippo signalling is required for its interaction with the DGC135. Furthermore, in vitro studies have indicated that cytoskeleton reorganization activates YAP through the inhibition of LATS kinases129,139,140. In agreement with this idea, studies in Drosophila showed that the Hippo pathway is suppressed by F-actin130,141. Interestingly, several components of the Hippo pathway, including AMOTL1 (refs50,142), MST1 (ref.140), MST2 (ref.140), and NF2 (ref.143), bind to F-actin. Understanding how Hippo or YAP and TAZ signalling integrates with mechanical cues during heart regeneration requires further study.
Hippo–YAP regulators and targets
MicroRNAs are essential players in the regulation of cardiac repair144. Interestingly, in 2015, studies uncovered a connection between microRNAs and the Hippo pathway. The microRNA cluster miR-302–367 regulates heart development and regeneration through the regulation of cardiomyocyte proliferation145. Mechanistically, miR302–367 suppresses the expression of Mst1, Mob1b, and Lats2, which in turn leads to the activation of YAP145. MicroRNAs have also been identified as downstream targets of YAP. YAP induces the expression of mir-206, which has an essential role in YAP-mediated cardiac hypertrophy and survival84.
As transcriptional co-activators, YAP and TAZ function by interacting with transcriptional factors such as TEAD1–TEAD4 (refs146,147,148). The YAP–TEAD interaction is required for YAP to stimulate cardiomyocyte proliferation67. Other studies have also indicated that YAP promotes adult cardiomyocyte proliferation by stimulating IGFI and AKT signalling and fetal gene expression (for example, Myh7, Nppa, and Nppb)68,83,102,145. In agreement with this idea, Pik3cb, which encodes an isoform of the catalytic subunit of PI3K that associates with the PI3K–AKT signalling pathway, has been identified as a direct target of YAP and is associated with promotion of cardiomyocyte proliferation149. In addition, chromatin immunoprecipitation sequencing analysis and mRNA expression profiling of Hippo-deficient mouse hearts showed that YAP directly regulates the expression of genes related to cell cycle progression (for example, Aurkb and Lin9), F-actin polymerization, and actin cytoskeleton linkage to the ECM (such as Fgd4, Pkp4, Sgcd, and Sntb1)118. Expression of the gene encoding the transcription factor PITX2 is induced in the injured ventricles of Hippo-deficient mice, and Pitx2 expression is required for neonatal heart regeneration30. In addition, as mentioned above, PITX2 directly interacts with YAP in the nucleus and activates the expression of a subset of genes encoding components of the electron transport chain and ROS scavengers (such as Ldha, Ndufb3, and Oxnad1)30. A study using a technique to capture mRNA specifically from cardiomyocytes showed that the expression of genes related to cell proliferation and the stress response was increased in Hippo-deficient mouse cardiomyocytes10. The study also revealed that E3 ubiquitin-protein ligase parkin (PARK2; also known as PRKN), which is associated with mitochondrial quality control, is essential for heart repair10 (Fig. 3).
In adult zebrafish, approximately 3% of cardiomyocytes proliferate at basal conditions; after injury, nearly ten times more cardiomyocytes re-enter the cell cycle93. However, in the adult mouse heart, even with Hippo deficiency and YAP activation, < 5% of cardiomyocytes proliferate, as determined by 5-ethynyl-2′-deoxyuridine (EdU)-incorporation studies9,10,11,118. This limited capacity of cardiomyocytes to self-renew might not be sufficient to support cardiac regeneration, and other additional mechanisms might be involved. Indeed, Yap overexpression in cardiomyocytes is associated with reduced levels of apoptosis12. Furthermore, blocking Hippo signalling increases scar border vascularity with reduced fibrosis during heart regeneration10. Notably, in this study, the Hippo pathway was manipulated specifically in cardiomyocytes. How Hippo-deficient cardiomyocytes affect noncardiomyocytes such as fibroblasts and endothelial cells remains unclear, but intercellular signal transmission is a possibility.
One explanation for the limited proliferation rate of cardiomyocytes in the adult mammalian heart might be the negative feedback loop of the Hippo pathway. The DGC has been reported to be a direct target of YAP118, and the DGC in turn inhibits YAP activity134,135. Another possibility is that during maturation, the morphology of cardiomyocytes changes, and structures such as the intercalated discs sequester YAP in the cytoplasm102,103,150. Moreover, Lats2 has been shown to be a downstream target of YAP and TAZ, and together they form a Hippo–YAP/TAZ–LATS2 negative regulatory loop151. Whether and how this negative feedback loop regulates heart regeneration requires further study.
As mentioned above, the Hippo pathway interacts with WNT–β-catenin signalling. For example, YAP and TAZ have been reported to suppress the WNT–β-catenin pathway through direct binding to dishevelled77,152 or β-catenin153. Nuclear YAP cooperates with β-catenin to promote the expression of genes related to heart development7. The observation that YAP and TAZ have both positive and negative roles in the WNT pathway seems contradictory. Reconciliation of these roles has been partially achieved by the finding that YAP and TAZ act as components of the β-catenin destruction complex and as β-catenin inhibitors in the ‘WNT off’ state; however, in the ‘WNT on’ state, YAP and TAZ are dislodged from the β-catenin destruction complex, resulting in β-catenin stabilization and activation154,155. The role of WNT signalling in heart regeneration is not fully understood. WNT–β-catenin signalling was reported to be essential for cardiomyocyte differentiation and proliferation during regeneration156,157. However, knockout of the gene encoding β-catenin in cardiomyocytes is beneficial for heart regeneration158,159, possibly because the disrupted destruction complex leads to YAP and TAZ nuclear localization and activation154,155. Solid evidence has shown that Hippo and WNT signalling are tightly associated during heart development7,160, although the association between the Hippo and WNT pathways during heart regeneration remains unclear. Of note, studies in the gut showed that YAP inhibits WNT signalling and induces cell reprogramming into a fetal stage, which promotes intestinal regeneration133,161. How Hippo and WNT pathways interact during cardiac regeneration is an important area of future exploration.
Extensive effort has been devoted to the field of Hippo–YAP pathway research. Many studies have uncovered the central regulatory role of the Hippo–YAP pathway in various physiological and pathological contexts. The Hippo–YAP pathway has a critical role in regulating cardiac development, growth, and homeostasis. Importantly, many interesting findings suggest that the Hippo–YAP pathway is a promising therapeutic target for cardiac regenerative medicine. Although essential, developing novel regenerative strategies to treat heart failure is challenging, especially given that mammalian hearts, including the human heart, lack self-reparative capacity. Both inhibition of Hippo signalling and expression of Yap in adult mouse cardiomyocytes (with the use of an engineered AAV9 vector) markedly improved both cardiac function and survival after MI10,11. Of note, the use of AAV9 in humans is safe and elicits a minimal immune response, making this approach feasible for manipulating the Hippo–YAP pathway for clinical therapeutic use. However, many unsolved questions remain regarding the function of the Hippo–YAP pathway in the heart. Further investigation is needed to clarify the answers to important questions, such as what is the exact role of the Hippo pathway in cardiac hypertrophy and in maintaining cardiac homeostasis and how does Hippo function in cardiac protection. In addition, although the inhibition of Hippo kinases and YAP gain of function can promote myocardial regeneration after cardiac injury10,11, what the potential long-term consequences are of chronic Hippo pathway reduction remain unclear. Hippo kinases have been shown to be tumour suppressors, whereas elevated YAP levels have previously been shown in patients with hypertrophic cardiomyopathy, suggesting that long-term YAP activation is deleterious to the heart85. Another challenge for researchers is understanding the complex regulation of the Hippo–YAP pathway and the crosstalk between several important signalling pathways and environmental cues, which are outlined in this Review. Studies published in 2017 have indicated that the Hippo–YAP pathway interacts with the ECM in regenerating mouse hearts134,135 — findings that require further investigation. The Hippo–YAP pathway is also regulated by microRNAs, which have been shown to have important roles in heart development, homeostasis, and regeneration145. Together, these and other studies have pointed to new potential therapeutic targets that might be manipulated to improve cardiac regeneration in the adult heart. Future work must bring forth the data required to further understand the mechanisms involved and address the current uncertainties about the role of the Hippo–YAP pathway in the heart.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors apologize to researchers whose work is not cited here because of space constraints. The authors thank the following funding sources: the AHA National Center Scientist Development Grant (14SDG19840000 to J.W.; 16SDG26460001 to T.H.) and Postdoctoral Fellowship (18POST34060186 to S.L.), the NIH (DE026561 and DE025873 to J.W.; DE 023177, HL 127717, HL 130804, and HL 118761 to J.F.M.), the DOD (W81XWH-17-1-0418 to J.F.M.) and the Vivian L. Smith Foundation (to J.F.M.). J.F.M. received support from the LeDucq Foundation’s Transatlantic Networks of Excellence in Cardiovascular Research (14CVD01) and the MacDonald Research Fund Award 16RDM001. N. Stancel (Texas Heart Institute, USA) provided editorial support.
Signalling molecule that forms a concentration gradient to guide and determine tissue pattern formation during morphogenesis. Cell fates and responses depend on morphogen concentration, thus morphogens are required for differentiation and position determination of the various cell types in a tissue and have a crucial role in development.
Group of cells (also known as primitive ectoderm) that form the outermost layer above the hypoblast of the embryo. The epiblast is derived from the inner cell mass and gives rise to the three primary germ layers (ectoderm, mesoderm, and endoderm), the extra-embryonic mesoderm, the amniotic ectoderm, and the allantois.
- Cardiac crescent
During mammal heart formation, before the heart tube forms, heart progenitor cells in the cranio-lateral mesoderm fuse at the midline to form a bilateral crescent-shape structure known as the cardiac crescent.
- Pharyngeal mesoderm
The mesodermal cell population located in the head region of the embryo; contributes to the pharyngeal arch cores and the second heart field during embryonic development.
- Gene trapping
A high-throughput mutagenesis approach for introducing insertional mutations thoughout the genome.
- Intercalated discs
Unique junctions that connect cardiomyocytes together and define their borders, which is a special feature of cardiac muscle and is required for cardiac cell–cell communication and coordination of muscle contraction.
Metabolic pathway that breaks down glucose to pyruvate and releases energy to form ATP and NADH for cellular metabolism.
- Fatty acid oxidation
A multistep catabolic process (also known as β-oxidation) in which fatty acids are broken down to generate acetyl-CoA, which then enters the citric acid cycle that produces energy for cellular metabolism.
Cells that have a phenotype between fibroblasts and smooth muscle cells, which is usually defined by expression of α-smooth muscle actin (also known as ACTA2); myofibroblasts are crucial in wound repair.