Mammals have limited ability to regenerate damaged cardiac muscle following injury. Instead, the injured area is replaced by fibrotic scar tissue, leading to irreversible cardiac remodeling and organ failure. Conversely, zebrafish possess an extraordinary capability for heart regeneration. Following injury, zebrafish cardiomyocytes proliferate to generate new myocardium. In addition, other cells in the heart, including endothelial cells, are activated in response to injury and provide regenerative cues to cardiomyocytes. However, heart injury results in the activation of developmental and regeneration-specific gene regulatory networks, the transcriptional basis of endothelial gene activation during regeneration, and the gene products responsible for promoting efficient heart regeneration in zebrafish have not been fully delineated. Here, we performed ATAC-seq on zebrafish regenerating hearts to identify active endothelial enhancers during development and regeneration. Among the many responsive enhancers, we identified a cis-acting element in the intronic region of mmp14b, a known effector of angiogenesis. This element activated expression in endothelial cells in regenerating zebrafish and neonatal mice. mmp14b enhancer activation in response to injury depends on a consensus TEAD-binding motif. Moreover, we found that deletion of mmp14b or the mmp14b regeneration enhancer identified in these studies abrogated efficient cardiac regeneration, demonstrating an essential role for mmp14b in heart regeneration. These findings also reveal shared transcriptional circuitry between fish and mammals for endothelial activation following heart injury, providing a potential therapeutic platform for enhancing cardiac regeneration

Defective formation or remodeling of the Pharyngeal Arch Arteries (PAAs) result in congenital heart disease phenotypes, such as those seen in 22q11 deletion syndrome. We have previously shown that the PAA endothelium originates in the Second Heart Field (SHF). However, whether SHF-derived endothelial cells (ECs) are required for PAA formation, and the mechanisms regulating SHF progenitor contribution to the PAAs are unclear. To address these questions, we crossed VEGFR2^+/-;Mef2c-AHF-Cre and VEGFR2^f/f mice, to conditionally ablate VEGFR2 expression in the SHF and inhibit SHF progenitor contribution to the PAA endothelium. Although the loss of VEGFR2 in the SHF abrogated the contribution of the SHF to the PAA endothelium, ECs from a compensatory source migrated into the pharyngeal arches, resulting in the eventual rescue of PAA formation and aortic arch morphogenesis.  The extent of the rescue depended on the proportion of SHF-derived ECs in the pharyngeal arches. In embryos with at least one functional copy of the VEGFR2 in the SHF, the contribution of the SHF-derived ECs to the PAAs was on average 41%. In these cases, the rescue by the alternative EC source was complete. To investigate the source of the compensatory endothelium, we examined the arterial-venous identity and fate of the ECs within the pharyngeal arches. We found that embryos with defective SHF-derived EC contribution had an increased number of venous ECs compared with controls. Altogether, our studies demonstrate that: 1) SHF-derived ECs are required for PAA formation and remodeling, 2) VEGFR2 is essential for these processes, 3) a threshold number of SHF-derived ECs is necessary for proper PAA formation, and 4) the loss of SHF-derived ECs in pharyngeal arches can be compensated by angiogenesis from a venous vascular source. Understanding the mechanisms regulating EC compensation can lead to the prevention or treatments of congenital heart disease-associated conditions such as 22q11 deletion syndrome.

Arterial occlusion deprives billions of cardiac myocytes of oxygenated blood, damaging cardiac muscle and inciting scar formation, which ultimately impairs cardiac function. We aim to discover the developmental pathways that activate coronary vessel (CV) differentiation from their progenitors and establish whether reactivating these pathways can stimulate significant coronary genesis during cardiac injury. The endocardium is a progenitor source for CV endothelial cells (EC), but the signaling pathways driving the endocardium-to-coronary cell fate are unknown. Combining temporally regulated lineage tracing and single-cell RNA sequencing (scRNAseq), we demonstrated that the endocardial-to-coronary vessel transition is strictly regulated in time and space and correlates with Bmp2 expression. Time-specific lineage labeling revealed that endocardial cells largely transition to CVs at the onset of septal development. Transitioning cells were detected in scRANseq data at e12 where they downregulated endocardial markers and upregulated BMP2 and Vegfa-dependent genes. This cell fate transition and BMP2 expression in coronary ECs was absent in scRNAseq data at e17.5. Exogenous administration of Bmp2 and VegfA in an in vitro coronary angiogenesis assay induced a large number of endocardial-derived ECs. In vivo, upon coronary artery ligation, neonatal mice treated with adeno-associated virus expressing Bmp2 (AAV-Bmp2) exhibited a dramatic increase in newly formed endocardium-derived CVs at the injury site.  Finally, analysis of Cxcl12 and Cxcr4 knockout hearts showed that this signaling pathway functions after the endocardial-to-coronary transition to mediate artery differentiation. Our data shed light on the molecular mechanism underlying the endocardial to CVs transition and how endocardial-derived ECs contribute to coronary arterial formation, opening a new potential therapeutic target that could promote revascularization of the injured heart.

The adult zebrafish heart has a high capacity for regeneration after injury. While several pro-regenerative factors have been identified, the cell types orchestrating heart regeneration remain largely elusive. To overcome this limitation, we dissected the diversity of activated cell states in the regenerating zebrafish heart based on single-cell transcriptomics and spatiotemporal analysis. We discovered a dramatic induction of several cell states with fibroblast characteristics upon injury, and we validated the pro-regenerative function of transient collagen 12 expressing fibroblasts. To understand the cascade of events leading to heart regeneration, we determined the origin of these cell states by high-throughput lineage tracing. We found that activated fibroblasts are derived from two separate sources, the epicardium, and the endocardium. Mechanistically, we determined Wnt signalling as a key regulator of the endocardial regenerative response. In summary, our results uncover specialized activated fibroblast cell states as major drivers of heart regeneration, thereby opening up new possibilities to modulate the regenerative capacity of the vertebrate heart.

Cardiac fibrosis remains a major challenge in the field of cardiac repair, as there is no effective therapy for the prevention or reversion of cardiac scarring. We asked what are the underlying molecular principles in the development and maintenance of cardiac fibrosis. Recently, we have established an in-silico model for a minimal macrophage (Mf)-myofibroblast (mF) cell-circuit that orchestrates tissue repair and fibrosis. We set out to explore this cell-circuit in the setting of acute myocardial infarction (AMI) that results in sustained fibrosis. By employing a series of unbiased, high-throughput techniques and mathematical modeling we demonstrated that in a regenerative, neonatal model of AMI, both Mf and mF numbers rise and then return to baseline levels. This resembles the theoretical cellular trend for tissue healing. In contrast, in a non-regenerative model of AMI in adult mice, while both Mf and mF rise together for a prolonged period, only Mf numbers return to pre-injury levels. This resembles “cold fibrosis”, a theoretical point in which mF are self-sustained and Mf are depleted. Intriguingly, under cold fibrosis, Mf regain their homeostatic functions, while cardiac fibroblasts acquire a de-novo pro-fibrotic phenotype which persists within the mature scar. In-silico simulation of factors controlling the cell-circuit revealed that the most prominent factor governing the regenerative window is the mF population growth rate, controlled by an autocrine loop. NicheNet analysis predicted a novel cardiac mF autocrine growth factor, Timp1, that drives cardiac fibrosis. Accordingly, In-vivo neutralization of Timp1 by polyclonal antibodies reduced cardiac fibrosis following AMI in adult mice. These findings suggest a central role for the mF autocrine signaling loop in the development of cardiac fibrosis. Such insights may be utilized in future pursuit of effective therapy for cardiac fibrosis.

The cardiac developmental network has been associated with myocardial regenerative potential. However, the embryonic signals triggered following injury have yet to be fully elucidated. Nkx2.5 is a key causative transcription factor associated with human congenital heart disease and one of the earliest markers of cardiac progenitors, thus it serves as a promising candidate. Our previous studies illustrate an early requirement for nkx2.5 in maintenance of ventricular dentity during embryogenesis, however, its molecular function in the adult zebrafish heart has not been explored. Here, we show that cardiac-specific RNA-sequencing studies reveal a disrupted embryonic Nkx2.5 transcriptional profile in the adult Nkx2.5 loss-of-function myocardium. Furthermore, nkx2.5-/- fish exhibit an impaired ability to recover following ventricular apex amputation as illustrated by retention of fibrin and collagen in the injured area. Complex network analyses illuminate that Nkx2.5 is required to mount a proliferative response for cardiomyocyte renewal and to provoke proteolytic pathways necessary for sarcomere disassembly. Moreover, direct Nkx2.5 targets embedded in these distinct gene regulatory modules coordinate appropriate, multi-faceted injury responses. Altogether, our findings support a previously unrecognized, Nkx2.5-dependent regenerative circuit that invokes myocardial proliferation and dedifferentiation to ensure effective regeneration in the teleost heart.

Cardiomyocyte (CM) maturation is a dynamic process that occurs from mid-gestation to early postnatal stages and is accompanied by structural, functional, and metabolic changes. Recent studies indicate that ACTN2 might be important for CM maturation and ACTN2 mutations, typically result in cardiomyopathies. However, the link between disrupted CM maturation and heart disease is unknown. Through largescale GWAS analysis, we recently identified a heart failure (HF) associated variant within an evolutionary conserved ACTN2 enhancer region. Engineered hPSC-CM carrying an ACTN2 enhancer deletion, showed mildly reduced ACTN2 expression, and CMs developed myofibrillar disarray, hypertrophy, and suppressed calcium transients and mechanical force. Moreover, mitochondrial respiration and oxidative phosphorylation was reduced and glycolysis was increased. Consistently, single cell transcriptomic analysis suggested disrupted CM maturation. Manipulation of ACTN2 enhancer using a novel CRISPRa line, resulted in more adult-like hPSC-CMs. Additional analyses using ChIP and CRISPRi suggested temporal regulation of ACTN2 expression through enhancer binding both in vitro and in vivo. Interestingly, heterozygous ACTN2 enhancer KO mice, developed hypertrophy, reduced cardiac function, disrupted sarcomeres, and increased glycolysis with upregulation of insulin/PI3K/mTOR signaling pathway suggestive of disrupted in vivo maturation. Our results demonstrate that the conserved and clinically relevant ACTN2 enhancer can regulate CM maturation, and can be leveraged to promote hPSC-CM maturation and study the association between myocyte maturation and HF.

Oxygen is crucial for appropriate embryonic and fetal development, including cardiogenesis. The heart is the first organ formed in the embryo and is required to provide oxygen and nutrients to all cells in the body. Embryonic cardiogenesis is a complex process finely regulated and prone to congenital malformations. It takes place in a hypoxic environment that activates the HIF-1a signaling pathway which mediates cellular and systemic adaptations to low oxygen levels. Since inhibition or overactivation of the HIF-1a signaling pathway in the myocardium lead to severe cardiac malformations and embryonic lethality, it is important that the cellular response to hypoxia be precisely regulated. While many gene regulatory networks involved in embryonic cardiogenesis have been characterized in detail, the modulation of the response of cardiomyocytes (CM) to hypoxia has remained less studied. 

In vivo and Ex vivo approaches were used to investigate the role of LRRFIP2 during cardiogenesis and to determine the cell autonomous contribution to the phenotype. Wild type and Mutant embryos were generated by crossing Lrrfip2^+/- mice and then analyzed.  

We identified LRRFIP2 as a novel negative cofactor of HIF-1a. Indeed, we showed that the absence of Lrrfip2 expression in a mouse KI model led to an enhance of many HIF-1a target genes including Aldoc, Bnip3 and Ndufa4l2 in embryonic CM during development. Overactivation of HIF-1a in Lrrfip2-null mice led to structural and functional mitochondrial dysfunction, a decreased ROS levels and ultimately to a precocious maturation of CM. 

Overall, these defects led to the formation of a smaller heart unable to supply sufficient oxygen to the embryo and ultimately to severe hypoxia and early lethality.  

Glucose is central to the metabolic environment of the embryonic heart, and hyperglycemia is the most common non-genetic cause of heart anomalies.  Using chemically defined in vitro induction of hPSC-derived cardiomyocytes, we previously found that a high level of glucose negatively impacts the maturation of cardiomyocytes through excess nucleotide biosynthesis.  Despite the mounting evidence, little is known about the impact of high glucose on the in vivo formation of embryonic organs.  Here, we present the first comparative in vivo analysis of the metabolic state of over 500 samples of embryonic organs including the heart and three other organs (liver, brain, and placenta) using a diabetic mouse model at mid-to-late gestational stages. Consistent with previous reports, metabolomics analysis revealed that the fetal heart of diabetic pregnancies is more glycolytic than that of normal pregnancies throughout the various stages. scRNA-seq showed that glycolytic enzymes were downregulated in the cardiomyocytes, suggesting a possible feedback downregulation due to metabolic changes. Interestingly, di- and tri-phosphate nucleotides are markedly elevated in the embryonic heart at later stages of diabetic pregnancy, resulting in a significant reduction in the AMP-to-ATP ratio.  Accordingly, AMPKɑ was found less phosphorylated in the fetal heart of diabetic pregnancy.  AMPK signaling is an evolutionarily conserved pathway that maintains cellular catabolism and mitochondrial homeostasis.  While there was no change in mitochondrial content, mitochondrial genes (Ppargc1a, Ulk1, Pink1, etc.) were downregulated in the fetal hearts of diabetic pregnancy.  Furthermore, the oxygen consumption reserve was significantly lower in the fetal heart of diabetic pregnancy. These data suggest that a disproportionate nucleotide pool due to excess glucose leads to mitochondrial dysfunction via AMPK signaling in the late embryonic heart development of diabetic pregnancy.

The nucleus acts as a main mechanosensing organelle and its deformations in response to mechanical stimuli have been associated with the modulation of gene expression. During heart formation, tissue tension forces drive remodeling of the linear heart tube (LHT) to the looped and chambered heart. How the tissue tension forces generated during this process affect the nuclear mechanosensing and what signaling pathways are involved in its regulation remain unresolved. Here, we described the major changes in nuclear morphology during the LHT remodeling in both, zebrafish and mouse models. We found morphogenetic Planar Cell Polarity (PCP) signaling directs nuclear mechanosensing of the developing cardiomyocytes through the function of its core component, Vangl2. We discovered the Vangl2-dependent nuclear mechanosensing is linked to transcriptional changes in gene programs regulating muscle differentiation. Functionally, the perturbations in Vangl2-dependent signaling led to impaired mechanical coupling, muscle contractility, and reduced cardiac output in zebrafish as well as impaired cardiomyocyte differentiation in human iPS-CM model. At the molecular level, we linked the activity of Vangl2 to YAP/TAZ nuclear translocation. Altogether, we discovered Vang2/YAP/TAZ signaling axis contributes to muscle differentiation and maturation.

Cardiomyocyte (CM) polyploidy is prevalent across mammals.  Shortly after birth, rodent CMs fail cytokinesis and become multinucleated.  This transition coincides with CM cell cycle exit and is associated with the poor regenerative potential of the heart.  However, the regulatory mechanisms of CM polyploidization remain largely unknown.  Using primary postnatal CM cultures as a model, our previous work uncovered a key role for the cardiac extracellular matrix and identified specific matrisome proteins in modulating CM polyploidy.  One of the candidates, SLIT2, is primarily a secreted protein that is important for axon guidance, angiogenesis, and embryonic heart development; how it regulates cell division is currently unknown.  Intriguingly, in addition to expression in cardiac fibroblasts, we found that Slit2 and its receptor Robo1 are expressed in a subset of mouse CMs, particularly the mononuclear population at postnatal day 7, suggestive of a cell-autonomous function.  Functionally, Slit2 knockdown significantly promoted primary CM polyploidization, which could be rescued by recombinant SLIT2.  Furthermore, injection of recombinant SLIT2 was able to promote CM cytokinesis in postnatal mice.  Unexpectedly, we observed that knockdown of Robo1, 2, and 4 (major receptors of SLIT) do not affect primary CM polyploidization, suggesting that Robo-independent signaling is required for CM cytokinesis.  To elucidate the underlying mechanisms, we analyzed the transcriptome of Slit2-knockdown primary postnatal CMs.  We found that Cep55, a centrosome- and midbody-associated protein, is downregulated upon the loss of Slit2 expression.  Consistent with this observation, knockdown of Cep55 significantly inhibited cytokinesis in primary CMs.  Altogether, we found that SLIT2 is required for postnatal CM cytokinesis, potentially through the regulation of Cep55.  These findings further support a key role for extracellular matrix proteins in CM ploidy regulation.

 The endothelium of the 3^rd, 4^th, and 6^th pharyngeal arch arteries (PAAs) arises from Isl1-derived progenitors, which remodel into bilaterally-symmetrical PAAs by embryonic day E10.5.  PAAs 3-6 then undergo asymmetric remodeling giving rise to the aortic arch and its major branches. Aberration during these steps of development lead to sever congenital heart defects, emphasizing the need for understand the mechanisms of PAA formation and remodeling.

It is often observed that in mutant embryos with aberrant PAA formation, PAA development is disrupted asymmetrically, with one side affected more than the other. Furthermore, either the left or right side can be affected stochastically. Mutations of integrin-ɑ5 in the Isl1 lineage belong to such a category of mutations. To investigate mechanisms, we analyzed the pharyngeal mesenchyme of control and integrin-ɑ5^flox/-;Isl1^Cre mutant embryos at E9.5, before the vascular plexus remodels into patent PAAs. In control embryos, the vascular plexus was symmetrically distributed between the left and right sides. Additionally, the numbers of Isl1-derived cells on each side of the pharyngeal mesenchyme were statistically equal. In integrin-ɑ5^flox/-;Isl1^Cre embryos, the vascular plexus formed asymmetrically, and the allocation of Isl1-derived cells was no longer symmetrical between the left and right pharyngeal mesenchyme. Disruption in the ECM structure has been shown to allow cells to aberrantly migrate contralaterally. Preliminary data show disruption of Laminin-gamma1 expression in integrin ɑ5^flox/-;Isl1^Cre mutants, suggesting that integrin-ɑ5 signaling in the Isl1-lienage regulates ECM integrity needed for the symmetrical allocation of Isl1-derived endothelial progenitor of the pharyngeal arches.

It is widely known that hemodynamic perturbations dramatically alter valvuloseptal morphogenesis, but specific mechanobiological mechanisms controlling clinically critical valve remodeling are lacking. We first profiled the in vivo expression of YAP, a key mechanotransduction mediator, during murine fetal valvular remodeling, and associated it with local mechanical forces. We determined YAP activation increased with stage, concomitant with increasing hemodynamic loading, but became restricted to the outflow (arterial) surface that experiences oscillatory shear stress (OSS). To directly test the remodeling mechanisms of specific loading modes, we directly exposed ex vivo chick endocardial cushions to hydrostatic stress (compressive vs tensile) or shear stress (steady laminar vs oscillatory) in organ culture. We determined that oscillatory shear stress (OSS) and compressive stress induce cushioned proliferation and biomechanical softening, supporting a pro-growth phenotype. In contrast, unidirectional shear stress (USS) and tensile hydrostatic stress promoted cushion compaction and stiffening, supporting a pro-maturation behavior. YAP nuclear localization was enhanced in OSS and compressive stress conditions but was suppressed in USS and tensile stress conditions, confirming the in vivo conditions. Further, YAP inhibition in either pro-growth stress conditions reduced proliferation, increased stiffness, and highly compacted morphology in cushions. To further test the role of YAP mechanotransduction in vivo, we applied partial left atrial ligations in embryonic chicks to de-load the ventricle. This resulted in both smaller mitral valves and reduced YAP activity. Together, our evidence supports that local shear and hydrostatic stress orchestrate local growth vs maturation decisions in endocardial cushions by controlling YAP mediated mechanotransduction.

Valvular heart disease is a major source of morbidity and mortality with an anticipated increase in prevalence secondary to an aging population and increase in survivorship for patients with congenital disease. Further developmental studies are required to advance novel therapeutic development. In this study, we have interrogated the later stages of valvulogenesis to understand the molecular mechanisms of valve formation and how these mechanisms are disrupted in the context of disease. Leveraging a combination of single-cell RNA/Chromatin Accessibility sequencing in the developing mouse heart, we identified a novel, rare cell population in the developing valve with a unique transcriptional profile comprised of highly specific developmental signaling pathway genes. These cells are first detectable after valve primordia formation at embryonic day (E) 12.5 and are spatially localized at the leading edge of the developing leaflets. Ablation of this rare subpopulation during development results in highly dysplastic valves, characterized by hyperplastic, redundant, immature leaflets associated with valvular stenosis and regurgitation. These dysplastic features are consistent with the features of several congenital valvulopathies including Ebstein’s Anomaly, and pulmonary or aortic valve stenosis. Single-cell RNA sequencing analysis of a human fetal heart with hypoplastic left heart syndrome and critical aortic stenosis demonstrated a depletion of this cell population in the diseased aortic valve, suggesting these cells may be required for normal human valvular development as well. This study establishes the existence of a novel, rare subpopulation of cardiac cells that are critical to valve development and may contribute to the pathogenesis of congenital valvulopathies.

During cardiac development, a blood-flow independent early phase, during which bilateral populations of cardiac progenitor cells assemble the nascent heart tube, is followed by a blood flow-dependent late phase, when cardiac chambers are shaped and cardiac valve leaflets are formed. We previously showed that blood flow-dependent Erk5-Klf2a signaling induces the expression of wnt9a and that this is essential for atrioventricular valve formation in zebrafish. While biomechanical signaling clearly impacts late heart morphogenesis, we still lack a good understanding of how morphogenesis of the early heart is regulated independently of blood flow.  Here, we present new data on an earlier role of Wnt9a and of its paralogue Wnt9b in the assembly of the nascent heart tube prior to the onset of blood flow. In wnt9a/b double mutants, the migratory potential of cardiomyocytes progenitors is affected, as they fail to converge from positions within the anterior lateral plate mesoderm towards the embryonic midline. This prevents heart cone formation and causes a collapsed heart at the midline. These mutants also exhibit premature differentiation of cardiomyocyte progenitor cells. This is in line with the known role of canonical Wnt signaling in preventing cardiac differentiation at stages of early cardiac development.  Using a pharmacological inhibition approach, we find that the physiological effect of Wnt9 in controlling cardiomyocyte progenitor migration towards the embryonic midline is mediated by non-canonical Wnt signaling. Hence, Wnt9 has both blood flow-dependent and -independent functions during heart development.

While mechanisms of early heart valve formation are well investigated, the processes of growth and leaflet remodeling are clinically significant, yet poorly understood; genetic ablation studies have revealed limited cell lineage-specific roles in these programs. We here explore the role of local hemodynamic loading as a driver of conditional signaling in the remodeling of fetal semilunar valves.

Immunostaining of murine aortic valve leaflets revealed ventricularis-specific endocardial Notch1 and mesenchymal CXCR4, coincident with unidirectional fluid shear stress (USS). We next generated valve endocardial-specific dnMAML knockout mice and found markedly hyperplastic semilunar valve cushions with reduced mesenchymal CXCR4. In complement, we generated valve specific CXCR4 deletion mice, which demonstrated similar hyperplastic valves and perinatal lethality despite preserved coronary morphogenesis. Both mouse models exhibited elevated Wnt-BMP signaling, increased proliferation, increased nuclear Sox9, increased glycosaminoglycan deposition, and decreased aSMA, which are indicative of impaired leaflet phenotypic maturation and matrix stratification.

We then employed in vitro flow experiments with primary chick aortic valve endocardial cells to demonstrate that ventricularis-occurring USS directly activates Notch1 and upregulates CXCR4-ligand Sdf1 expression, while arterial-sided oscillatory shear stress (OSS) blocks Notch1, thereby activating Wnt-BMP. Treatment with Notch inhibitor DAPT blocked Sdf1 expression in USS treated endocardium. Further, semilunar cushion organ culture experiments showed that CXCR4 inhibition reduced aSMA expression and reduced tissue stiffness.

Together, these results identify a hemodynamic pattern-specific activation of semilunar valve growth (via OSS-Wnt/BMP) or tissue maturation (via USS-Notch1/Sdf1/Cxcr4) programming to robustly coordinate leaflet sculpting during dynamic and variable growth of the outlets.