Of form and function: early cardiac morphogenesis across classical and emerging model systems

Abstract

The heart is the earliest organ to develop during embryogenesis and is remarkable in its ability to function efficiently as it is being sculpted. Cardiac heart defects account for a high burden of childhood developmental disorders with many remaining poorly understood mechanistically. Decades of work across a multitude of model organisms has informed our understanding of early cardiac differentiation and morphogenesis and has simultaneously opened new and unanswered questions. Here we have synthesized current knowledge in the field and reviewed recent developments in the realm of imaging, bioengineering and genetic technology and ex vivo cardiac modeling that may be deployed to generate more holistic models of early cardiac morphogenesis, and by extension, new platforms to study congenital heart defects.

1. Introduction

The processes of cardiac differentiation and morphogenesis have intrigued developmental biologists for close to a century. The heart is an indispensable organ that is highly complex in its form and function. It is composed of several distinct cell types both myocardial (atrial, ventricular, conductive) and non-myocardial (endocardial, endothelial, epicardial, fibroblasts and smooth muscle). During the course of cardiac development these cell types arise sequentially and while endocardial cells are specified and differentiate within the early developing cardiac regions, fibroblast and smooth muscle cells are derived from the epicardium or from neural crest cells and integrate into the developing heart at later stages. Early cardiac morphogenesis encompasses the process of cardiac crescent and heart tube formation followed by asymmetric looping that eventually results in chamber formation. Not only do the cardiac progenitors organize into a complex structure during early vertebrate morphogenesis but the linear tube once formed begins to function as a pump as it undergoes further development towards a four chambered heart. This dynamic process provides a unique opportunity to study how an organ integrates its ability to both sense and generate tension and whether such mechanisms are required for the generation of a complex, multi-cellular and hemodynamically integrated system. Current understating of these processes has been informed by meticulous studies across different vertebrate and invertebrate models. Although the gene regulatory network underlying heart formation has been well characterized, there exists ample opportunity to uncover the link between these molecular programs and their effects on cellular behaviors and tissue structure during the process of cardiac morphogenesis. Uncovering these mechanisms is fascinating and will likely be key for a better understanding of congenital heart defects (CHDs). Here we have chosen to focus on discussing the mechanisms underlying early cardiac morphogenesis and how such knowledge across different developmental and cellular processes can be integrated with recent technological advances to begin to understand this complex phenomenon.

2. Cardiac morphogenesis across species

The process of cardiac morphogenesis has been extensively studied in different vertebrate and invertebrate models. Each of these model organisms has unique features that have allowed for inquiry into the cellular and molecular mechanisms driving morphogenesis. Additionally, comparisons across organisms with variations in the process by which the early heart tube is assembled offers the opportunity to uncover molecular mechanisms driving specific morphogenetic programs.

Among invertebrate models the fruit fly Drosophila melanogaster and tunicate Ciona intestinalis have provided invaluable insights into the processes of early cardiac morphogenesis and lineage specification. Although the invertebrate differentiated heart has a simpler form and lineage composition, there exist striking similarities in progenitor cell migration and linear heart tube morphogenesis with the vertebrate models [1–3]. The single chambered tubular heart limits the ability to study the late morphogenetic events of looping and chamber morphogenesis seen in vertebrates. However, the shared cardiac gene regulatory network and homology between the processes of progenitor migration and heart tube formation make them great models to study the early conserved processes in cardiac morphogenesis [4–6].

Within the vertebrate model organisms, relatively easy access to the developing embryo and concomitant live imaging to track cell dynamics have rendered zebrafish an excellent model to study early cardiac morphogenetic events [7–11]. This has been complemented by a greater breadth of genetic manipulation for reverse and forward screens, lineage tracing and the ability to follow the effects such perturbations as the fish embryo can survive for several days with a non-functional heart [12]. As the fish forms a two chambered heart it limits the ability to probe some of the mechanisms underlying human development and disease. In this effect the work in the chick and mouse embryos have been fundamental to understanding the tissue level processes of forming a four chambered heart. The chick embryo develops in a disc (similar to the human) and the heart is easy to image but approaches to genetic manipulation though on the rise remain limited[13,14]. Recently with the advent of enhanced imaging and ex vivo culture methods the mouse embryo has adopted new and exciting opportunities to study the early stages of mammalian cardiac morphogenesis [15–18]. Combined with the breadth of established genetic models, spatio-temporal information via imaging promises to further our understanding of the molecular mechanisms driving cardiac morphogenesis, and by extension some of the CHDs that occur in humans.

Cardiac morphogenesis across different species shows conserved mechanisms in cell movement and signaling pathways. These processes have been modified across flies, tunicates, fish, amphibians, amniotes and mammals to generate striking adaptive differences in heart tube morphology. The earliest step in this process involves formation of cardiac progenitors that originate in the primitive streak and migrate anteriorly within the lateral plate mesoderm [12,19,20]. They eventually reach the midline to form the primitive heart tube. In the fish, the cardiac progenitors reach the ventral midline to form a disc which through an intermediary cone shaped structure is pulled into an elongated tube by an asymmetric process of cardiac growth called ‘cardiac jogging’ [12,21]. The two wings of the cardiac mesoderm meet in the midline in the mouse to form the cardiac crescent but remain apart in the chick and the human [22]. The formation of the mouse heart tube resembles a purse string whereas the chick shows a zippering of two bilateral heart fields [23,24]. Eventually in vertebrates the early linear heart tube loops to orient blood flow and integrates complex signals driving tissue level processes of ballooning, septation, trabeculation and compaction of the myocardial walls which collectively gives rise to distinct cardiac chambers [25]. In the chick looping results in helical architecture which places the dorsal half of the anterior heart tube towards the inner curvature and the ventral part of the tube forms the outer curvature [26]. This proceeds in two defined stages; C-looping with ventral bending and rightward rotation of the linear heart tube followed by s-looping characterized by caudal displacement of the primitive ventricle and shortening of the outflow tract [27–30]. Similarly, in the mouse the asymmetrical placement of venous and arterial poles of the early heart tube combined with a buckling phenomenon give rise to a helical looped heart tube [15,18,31,32]. In contrast, in the fish, looping morphogenesis occurs in a single plane resulting in a final flat s-shape with the atrium and ventricle placed asymmetrically across the embryonic midline [33]. A comparison of the major events of early morphogenesis is illustrated in Figure 1.

Figure 1. Early cardiac morphogenesis across vertebrate model organisms.

Stages of cardiac morphogenesis within zebrafish, chick and mouse from progenitor migration to looping stages, with positional information within the developing embryos. Views of cardiac fields within the embryos show its location and position within the growing embryos. The illustrations of cardiac morphogenesis in the zebrafish represent dorsal views whereas those in chick and mouse depict ventral views. In zebrafish second heart field (SHF), atrial pole progenitors (AP), ventricular pole progenitors (VP) and endocardial progenitors (EP) are traced through time. Correspondingly, the first heart field (FHF), second heart field (SHF) and endocardial (EC) lineages are depicted in the chick and mouse.

Interestingly, though the model organisms follow a similar path from the formation of a midline cardiac region that folds into a liner tube, formation of the heart tube varies significantly in some non-model organisms that have been studied during these processes. Some mammalian species show bilateral endocardial tubes that are surrounded by a functional myocardium prior to fusing in the midline [34–37]. While these processes are challenging to investigate in humans, comparative studies suggest, again, that the major concepts are shared with other vertebrate models, but that difference may exist with respect to specific spatio-temporal mechanisms and their consequence for mature chamber formation [38,39]. Species specific differences extend beyond gross morphological ones to more subtle functional ones. For example, in the chick cardiac sarcomerogenesis is delayed relative to the mouse where it parallels tube formation. In conclusion, the striking similarities alongside species specific variations in the process of early cardiac morphogenesis makes it an immensely interesting process to study, both in established models as well as in emerging in vivo and in vitro systems.

3. Cardiac morphogenesis occurs alongside cardiac fate specification and differentiation

Heart development begins with cardiac progenitor specification in the gastrulating embryo, after which the progenitors migrate to the ventral midline in bilateral heart fields where they concomitantly assemble into a tubular structure and differentiate into cardiomyocytes. The gene regulatory networks underlying these events have been the focus of a large body of work and are reasonably well studied. Here we aim to integrate this information with progenitor cell differentiation and migration during early cardiac morphogenesis.

In Drosophila the heart structure (also referred to as the dorsal vessel) is a linear tube similar to the early heart tube seen in vertebrates. The cardiac precursors are specified within the dorsal wing of the mesoderm in a linear row with antero-posterior identity extending along the length of the embryo [40]. The precursors through the process of migration meet in the midline to form a linear closed tube-like structure. Several of the processes of cardiac specification discovered in the fly have shown similarities to the vertebrate systems through highly conserved gene regulatory networks and signaling pathways [4,6,41,42]. The process of cardiac development in Ciona begins with specification of two founder cells that generate bilaterally placed anterior and posterior cells, of which the anterior cells are the cardiac precursors that migrate ventrally to eventually fuse in the midline. The first heart field progenitors are situated medially, and the lateral group of cells generate the second heart precursors as well as a migrating population of atrial siphon muscle founders (similar to the cardio pharyngeal cells in the vertebrates) [43]. The early cardiac lineages can be visually traced over the process of morphogenesis, which has led to the discovery of key concepts such as the uncoupling of specification and migration events [44,45].

Cardiac mesoderm is specified in bilateral zones within the chick and mouse epiblast [46–48] which corresponds to the lateral marginal zone in the zebrafish blastoderm [49]. Two myocardial lineages (first and second heart fields) are thought to be defined sequentially as they ingress and migrate to the ventral midline within the anterior lateral plate mesoderm, resulting in the formation of the cardiac crescent or in case of fish the cardiac disc [19,50–52]. The directed migration of cardiac precursors is still a subject of investigation but is thought to be controlled by various midline factors. For example, PDGF secreted by the central developing foregut endoderm is postulated to drive the PDGFR dependent migration of precursor cells [53].

The developing cardiac regions show two timed anteroposterior waves in cardiac differentiation accompanied by functional maturation. In chick and mice, the early migrating population or the first heart field (FHF) is placed medially and forms the primitive heart tube and in the mature heart contributes predominantly to the to the left ventricle and atria and a portion of the right ventricle. These FHF progenitors lie between the yolk sac ventrally and the ventral foregut endoderm dorsally. Signaling from underlying endodermal cells promotes cardiac progenitor differentiation [54,55]. The FHF is the first to differentiate in the crescent at the early head fold stage, with a subsequent pause in differentiation as the crescent is sequentially converted to a hemi-tube and eventually fuses to form a closed linear heart tube [24,27,56,57]. The anterior and posterior (arterial and venous) ends of the crescent remain contiguous with the splanchnic mesodermal cells, in a highly proliferative cardiac mesodermal population called the second heart field (SHF) [17]. The SHF is placed laterally to the FHF and is in close contact with the underlying endoderm which is thought to signal the maintenance and deployment of SHF progenitors [32,50,58,59]. After the formation of the linear heart tube SHF progenitors are continuously added to the poles contributing to the elongation and looping of the linear heart tube [27,60,61]. They differentiate as they do so and eventually contribute to most of the right ventricle, inflow and outflow tracts [50]. The process of specification and differentiation occurs alongside the maturation of the cardiac sarcomeric structure and contractility as well as the integration and spatial organization of different cardiac cell types [62]. The identity and functional relevance of the SHF as well as its role in CHD formation are amongst the major discoveries in the field of heart development, and have been reviewed extensively elsewhere [63–65]. Similar spatial fate acquisition parallels zebrafish cardiac morphogenesis where the FHF progenitors are positioned medially to the SHF. The FHF progenitors are primarily ventricular and SHF progenitors are similarly contiguous with the linear heart tube and are sequentially added to the developing tube where they contribute to the atrioventricular myocardium and outflow tract smooth muscle [51,66,67]. Interestingly, myocardial SHF lineages have been shown to have early left right asymmetry that is important for atrial morphogenesis [51,68].

Over the course of embryonic development other cell types from surrounding tissues integrate and associate with the growing heart. For example, the migrating cardiac neural crest cells contribute extensively to the outflow tract [69]. The proepicardial organ located at the venous pole contributes to the epicardium which forms a covering over the cardiac myocardial surface and gives rise to fibroblast and smooth muscle cells that will integrate within the developing cardiac chambers to form the coronary vasculature [70]. Spatiotemporal profiling of the chick cardiac transcriptome has revealed that the epicardial progenitor cell migration and integration into the myocardium precedes their fate specification, raising the question of molecular mechanisms during the morphogenetic process signaling the incorporation of these progenitors and their differentiation [71–73].

What makes early cardiac development continuously intriguing to study is its complexity and the simultaneous processes of fate specification, cell movement, spatial organization and integration of the different cell types into the complex structure of the developing heart. Both morphogenesis and fate specification are critically impacted by surroundings cells and tissues, as well as biomechanical cues of the rapidly growing embryo. In vitro differentiation experiments indicate that early cardiac fate specification can be uncoupled from morphogenesis, however the same experiments suggest that the formation of a mature and functional cardiac organ is dependent on the spatio-temporally controlled progenitor specification, cellular heterogeneity and biomechanical influences inherent to early cardiac morphogenesis. It remains to be seen if and how signaling and transcriptional programs that generate different timed progenitor populations deployed during the process of in vivo differentiation have an impact on their cytoskeletal architecture that might in turn inform morphogenesis.

4. Role of transcriptional programs and signaling pathways in heart tube formation and looping

Signaling pathways activated either by ligands from non-cardiac tissue or intrinsically via the process of migration and fate acquisition play important roles in sculpting the early heart tube. Early on progenitors receive specific signals from the surrounding tissues as they migrate to their ventral location. Upon reaching the ventral side of the embryo their spatial location relative to the midline exposes them to differences in signaling either as anteroposterior, midline or left-right gradients. In addition, the relative timing of egression of progenitor populations from the streak is thought to drive their transcriptional identity, generating distinct gene regulatory networks which in turn are thought to control their cellular behaviors.

Embryonic anteroposterior patterning via Hox expression controlled through retinoic acid signaling affects the early SHF lineages that emerge within the cardiac mesoderm [74,75]. The memory imparted by these events (through transcriptional programs controlled by Tbx1, Six2, Mef2c, Foxc1/Foxc2, Isl1) generate distinct positional populations that are sequentially incorporated during heart tube elongation giving rise to different regions of the developing right ventricle, atria and outflow tract [59,76–81]. Not surprisingly, such patterning mechanisms at the progenitor stage are highly correlated with establishing the specific lineages and morphological structures, and their perturbations lead to distinct defects in the developing heart. For example, enriched expression of transcription factors implicated in looping defects were shown within regions of the developing outflow tract and atrioventricular canal [82]. Recent work in the mouse has uncovered a role for Hand2 in maintaining epithelial integrity in the transition zone as the SHF progenitors are sequentially integrated into the developing heart tube [83]. Single cell transcriptomic analysis of the Hand2 mutant mouse further confirmed defects in outflow tract specification and additionally showed defects in migration of the specified right ventricular cells during early cardiac morphogenesis [84].This and extensive additional work along the same lines raise interesting questions as to whether individual progenitor populations show distinct biomechanical signatures and how this process impacts overall morphogenesis of tube elongation and buckling movements during looping [85].

It is well established that signaling from adjacent tissues is essential not only for specification of cardiac progenitors but to regulate their migration and structural orientation within the bilateral cardiac fields and tube. Rho kinases, for example, play a role in cardiac tube morphogenesis by controlling precardiac migration and timed differentiation of cardiac progenitors [86]. In addition, crosstalk between Bmp, Fgf and Shh signaling via extracardiac signals coordinate the differentiation-proliferation balance thought to be important in tube morphogenesis and noncanonical Wnt signaling has been implicated in heart field fusion in zebrafish [87–89]. A more recent study leveraged single cell RNA sequencing to uncover a role for SHH signaling from the node in patterning and proper migration of the anterior mesoderm. Using genetic loss of function models the authors show that a defect in mesodermal Shh activity results in a reduction of the anterior SHF populations which in turn causes cardiac morphogenetic defects [81]. Understanding how these distinct programs are deployed in a spatio-temporal manner and how they interact remains to be further investigated for many cell types and morphological structures, and they will be critical to generating relevant in vitro models of the morphogenetic process during early heart development.

The breaking of left-right symmetry is another major process regulated by internal programs and external positional cues that modulate different cell populations in a spatio-temporal manner. In the fish, organ symmetry is broken between the disc to cardiac cone and asymmetric growth continues into the looping stage. In Xenopus, H2B monoubiquitination has been shown to control cardiac looping through ciliary motility, uncovering epigenetic drivers of asymmetrical morphogenesis [90]. The early induction of asymmetry is controlled via Nodal, Bmp and Shh signaling from adjacent tissues which set up differential transcriptional programs as early as the cardiac disc stage [91–93]. The signaling mechanisms can act in concert or in parallel with distinct mechanisms driving tube elongation and looping [94–96]. In xenopus, zebrafish and chick, the Nodal-Pitx2c-MYH10 axis has been described to inform asymmetric growth and looping, whereas in the mouse genetic manipulation of Pitx2 has shown to have no observed effects on looping but affects the asymmetrical positioning of the heart and mis regulation results in chamber wall defects [97–104]. But Nodal signaling has been shown to control the looping process in mice[105,106]. Interestingly, in the mouse an intrinsic driver of asymmetric growth has recently been shown to coordinate the establishment and amplification of Nodal signaling to set up the heart loop [107]. The authors propose a mechanism for modulation of ECM and proliferation within the cardiac progenitors that drives asymmetrical growth and looping. These findings and many additional studies along the same lines set the stage for further exploration into the dynamics of how intrinsic and extrinsic signaling cues are set up and in turn inform cell and tissue behavior during early cardiac morphogenesis.

5. Cellular processes in early cardiac morphogenesis

5.1. Cell extrinsic influences

Role of adjacent endoderm and non-cardiac mesoderm in cardiac morphogenesis:

Extracardiac tissues influence cardiac morphogenesis from the early stages of progenitor migration by depositing midline factors or indirectly through tissue deformations. For example, short range signaling via fibronectin secreted by the foregut and yolk sac endoderm is responsible for the migration and fusion of bilateral cardiac heart fields [108–110]. Cell autonomous migration of the early cardiac progenitors seems to be reduced as the cells egress from the primitive streak [111,112]. Instead, extrinsic tissue level deformation of endodermal shortening around the anterior intestinal portal has been shown to drive the movement of cardiac mesoderm to the midline [113].

Both foregut and yolk sac endoderm have been shown to contribute to the fusion and folding of the heart primordia [112,114]. A combination of molecular drivers and parallel tissue movements control this process. For example, mouse chimeras generated by injecting Gata4^−/− ES cells into WT embryos demonstrate that visceral yolk sac endodermal Gata4 is sufficient to rescue the fusion of bilateral heart fields [115–117]. In zebrafish, sphingosine-1-phosphate signaling and Na+ K+ ATPase within the yolk syncytial layer have been shown to affect cardiac progenitor migration and heart tube elongation [118,119]. Work in Xenopus investigating the role of endoderm on heart development illustrates that endodermal contractility modulates mesenchymal to epithelial transformation of cardiac progenitor cells suggesting yet another tissue level mechanism for endoderm in cardiac progenitor migration [120].

While challenging to study mechanistically there is a body of work corroborating that tissue level movements of non-cardiac mesodermal structures surrounding the cardiac zones aid the morphogenetic movements within the developing heart tube. Growing evidence from the chick suggests a coordination of heart tube elongation with the developing foregut in the ventral midline through convergent extension [23,121]. In the developing mouse embryo, the anteromedial displacement of the splanchnic mesoderm has been implicated in the closing of the primitive heart tube [24]. Interestingly, the Foxp4 mutant shows complete cardiac bifida with two intact heart tubes forming without foregut closure suggesting that cardiac tube formation and closure is not entirely dependent on tissue level movements of the embryonic endoderm [122]. Additionally, the torsional component in heart c-looping has been attributed in part to the splanchnopleuric membranes, yet it seems to be primarily accelerating the processes but is not essential to it [123–126]. The dorsal mesocardium, formed by splanchnic mesoderm and located beneath the foregut is suggested to play a role in murine heart looping by generating a physical constraint that results in tube buckling or through differences in ECM structure leading to asymmetric tension within the tissue [31,127,128]. Collectively, endoderm and non-cardiac mesoderm are well recognized for their role in shaping early heart development, however, the specific mechanisms by which they do so, in a cell type specific and temporally controlled manner remain a topic of great interest.

Endocardium:

During the process of cardiac progenitor migration and assembly of bilateral heart fields endocardial cells sort out and form an inner layer lining the developing myocardium [129,130]. Co-movement of these endocardial cells with surrounding ECM has been shown to direct early tube morphogenesis [131]. Divergent views exist on the origins of the endocardium as either pre-specified from anterior lateral plate mesoderm prior to the primitive streak phase from an endothelial like population or generated from Mesp1 labelled cardiac progenitors [132–137]. Although the role of endocardium in chamber formation and valve development has been extensively studied our understanding of its role in early tube formation and looping is limited, particularly in the 4 chambered hearts of higher vertebrates [138]. Endocardial progenitors in the fish are specified in the lateral plate mesoderm to then migrate to the midline directing the movement of the cardiac primordia. Interestingly, defective endocardial specification does not prevent the initial midline migration of the cardiac progenitors but it affects the formation of a cardiac disc and cone [139]. Of note, in the chick and quail the endocardial cells seem to share a positional identity with the cardiac progenitor populations in their immediate vicinity [133,140,141]. This could support a role for the differentiating endocardial cells in directing the closest myocardial neighbors during the serotyped processes of FHF differentiation and incorporation of SHF progenitors during tube elongation and looping.

5.2. Cell intrinsic cues

Migration, Polarity and Proliferation

Across all vertebrate species the migrating cardiac progenitors reach the ventral midline, polarize, and form a myoepithelial sheet. This process has been extensively studied and defects in polarity are known to contribute to abnormalities in early tube morphogenesis. As migrating cardiac progenitors segregate from the lateral plate mesoderm, they begin to display an apicobasal polarity with apical expression of N Cadherin and enrichment of ß1 integrin on the basal surface [142,143]. The spatial localization of N Cadherin is later required for maintaining the anterior SHF progenitors in a proliferative and undifferentiated state [144]. Further, during the process of early outflow tract morphogenesis, the Wnt-PCP pathway is involved in establishing apicobasal polarity within the SHF progenitors, acting in a comparable manner to establish polarity in other tissues or organisms [145,146]. In zebrafish, during the process of migration, the progenitors epithelialize and generate apicobasal polarity which is required for heart cone tilting and elongation [147]. Similarly, in Xenopus mesenchymal to epithelial transformation and apicobasal polarity is established during heart tube formation and elongation [120]. The FHF cells located within the mouse and chick crescent are organized in a myoepithelial sheet of cells [24]. After the formation of the ventral cardiac regions proliferation of cardiac progenitors is not only timed and regionally restricted as described previously but also stereotypically oriented across the process of morphogenesis. Asymmetric cell proliferation and ingression plays an important role in cardiac looping by generating directional tension within the developing tube [31,107,148]. In addition to proliferation the orientation of myocardial cell divisions drives planar expansion of the myocardial chamber [149]. In line with this, single cell transcriptomics of the mouse embryonic heart during chamber expansion revealed the stereotypic nature of proliferation and its association with transcriptional changes that underlie the process [150]. Together these complex processes drive the 3D structure of the looping heart tube.

Cytoskeletal-ECM dynamics

Although tissue level endodermal deformations are implicated in the migration, fusion, folding and elongation of the heart tube, other studies suggest a role for cardiac cytoskeletal organization in the same process. It has been proposed that apical constriction of a sheet of columnar cells could result in cylindrical bending [151]. This is supported by circumferential organization of the cardiac myofibrillar network and cell axis along the direction of the heart tube [152]. Differences in the SHF cellular actomyosin orientation that informs cellular responses to deformation stress have been implicated in driving a cardiac progenitor intrinsic mechanism towards heart tube elongation and ventral bending during c-loop morphogenesis [153]. Mechanistically this c-loop bending has been attributed to myocardial cell intrinsic actin polymerization changes and non-muscle myosin II localization [154–156]. Recent work has further uncovered an F-actin-dependent intrinsic remodeling of the right myocardium associated with initial heart looping in the chick [157]. Planar cell polarity signaling in fish similarly coordinates cardiac looping and atrio-ventricular chamber expansion through polarized actomyosin networks [158]. This was supported further and expanded by a recent study uncovering a role for Tbx5 in tissue patterning that generated the fish s-loop via torsion along the AV canal [159].

Multiple studies have implicated the emerging cardiac sarcomeric structures in the process of early morphogenesis. For example, a mutation in Tropomodulin affects sarcomeric organization and cardiac looping in a cardiac progenitor cell intrinsic manner [160]. Another structural protein, vinculin, is implicated in cardiac morphogenetic defects due to abnormal progenitor migration and neural crest migration [161]. These defects might be due to loss of structural orientation as loss of myosin-based contraction or absent myocardial contraction do not seem to impair cardiac looping [162,163]. Additional evidence employing focal adhesion kinase morphants illustrates that these form a linear heart tube but fail to undergo looping morphogenesis, underscoring the importance of structure within cardiac cells aiding the process of looping [164].

5.3. Intrinsic and extrinsic tension measurements

Direct observation of many dynamic processes of the vertebrate heart have been limited due to the inaccessibility of the organ or embryo and sensitivity of the methods used to sense tissue and cellular level force-tension relationships. Early studies probing viscoelastic properties of chick heart tube during looping uncovered differences in stiffness between the inner and outer curvature and documented residual stress at the outer curvature, although this could be a result of looping rather than a driver [124,165,166]. With respect to biomechanical processes during early tube formation, it is well known that the cardiac jelly lies in the myo-endocardial space and acts as a mechano-stabilizing structure along the early heart tube, however the specific mechanisms involved are poorly understood [167]. Its asymmetric distribution over the course of cardiac chamber morphogenesis and looping has led to its proposed role in maintaining the primitive tubular shape of the heart tube [168,169]. Functional relevance of its specific components for these processes remains to be investigated [63].

6. Emerging ex vivo systems to model early cardiac development and morphogenesis

As illustrated here the heart has been the subject of intense investigation for many decades and collectively these studies have produced a comprehensive understanding of many aspects of early cardiac development [8,17,91,170–172]. It seems therefore even more puzzling that many of the mechanisms underlying human congenital heart defects (CHDs) remain poorly understood, suggesting that there exist mechanisms that warrant further discovery and understanding. [173,174]. With an abundance of new technologies emerging, one question we ask is which of these approaches will generate the most relevant new information on early cardiac specification and morphogenesis, and by extension on the mechanisms of CHDs. Below we list some of the more recent opportunities and discuss their potential to answer longstanding questions for cardiac progenitor specification and the mechanisms involved in the formation of the highly complex mammalian heart.

6.1. Leveraging new technologies in ex vivo embryo culture and imaging modalities

Culturing the mouse embryo ex vivo for extended periods of time while maintaining normal development has proven challenging, particularly for stages post gastrulation. This is likely due to increasing complexity of these developmental stages and a concomitant reliance on maternal support via circulation. The first ex vivo embryo cultures using rat serum in combination with continuous rolling cultures were described as early as 1966 for rat embryos, and were subsequently expanded to the mouse. [175–177]. Many interesting findings have come from these studies, including lineage relationships and positional influence on transplanted cells [178]. The method allowed for manipulations that required access to the embryo itself and following of such manipulations at later stages of development. However, it lacked the continuous, high-resolution analysis provided by live imaging. A breakthrough from the Torres laboratory, decades after the first ex vivo cultures, now demonstrates feasibility of static culture in combination with live imaging of the mouse embryo from the crescent to the heart tube stage [24]. Recent work has extended that window to include looping and early ventricular chamber morphogenesis [179]. The power of live imaging is further complemented by major advancement of whole embryo and tissue imaging using clearing approaches and optical sectioning microscopy in combination with systematic and quantitative image analysis [180–185].

Recent development in live imaging modalities using light sheet microscopy combined with digital reconstructions enables studies of tissue level morphogenetic processes within the early mouse embryo. This allows for a complete use of the mouse genetic toolkit to understand the molecular drivers of cardiac morphogenesis, and for additional manipulations including signaling activities, metabolism, and biomechanical cues. Such approaches, while technically challenging, may allow the mouse to play some catch up with the impressive and indispensable breath of knowledge contributed by live imaging of the fish, chick and other organisms and provide new mechanistic insights on early cardiac morphogenesis of the four-chambered heart.

6.2. In vitro models of cardiac morphogenesis

The need for human models and scalable systems constitute strong motivations to develop in vitro models for heart biology, notably using pluripotent stem cell (PSC) differentiation methods. The excitement of such models was amplified by two groundbreaking new technologies: the discovery of induced PSCs in 2006, and the description of efficient gene editing using CRISPR/Cas9 in 2012, jointly opening up additional opportunities for investigating human disease and genetic diversity more broadly [186,187]. Initial efforts using PSCs were aimed at generating cardiovascular cells with high efficiency and in a cell type-specific manner. Following the paradigm of ‘recapitulating development in a dish’ differentiation protocols for the majority of all cell types in the heart have been developed, including atrial and ventricular cardiomyocytes, fibroblast cells, endothelial cells, smooth muscle cells and cells of the conduction system [188–193]. Remarkably, each of these cell types can be generated independently of one another when provided with the appropriate signaling environment, suggesting that complex biomechanical cues or cell-cell interactions are not required for early fate specification.

As similar progress was made across many other cell lineages and organ system, it became apparent that generating physiologically relevant, functional models will depend on generating 3D tissue structures representative of the in vivo organ, which initiated the rapidly expanding field of organoid biology [194–196]. Organoids, if defined as self-assembly of in vitro differentiating cells into 3D tissue structures with representative micro-anatomy have been generated from both adult tissue-specific stem cells (gut, mammary gland, prostate, pancreas, liver, stomach etc.) as well as from PSCs (brain, eye, skin, kidney, gut, lung, liver, others) [194,196–198]. The heart presents a unique challenge for organoid formation in that its early development leading to the 4-chambered morphology is extremely complex, and both its formation and function are heavily dependent on extrinsic biomechanical cues and hemodynamic forces, which are complicated to model in vitro [199,200]. Taking on this challenge, several reports have recently described cardiac organoid formation [201]. Using mouse embryonic stem cells Rossi and colleagues applied a developmental biology approach to form cardiac crescent organoids from A-P patterned gastruloids which can produce both first- and second heart field cells [202]. These structures contain both cardiac and foregut endodermal structures in close proximity, as well as vascular networks and endocardial-like cells. In a pioneering study Drakhlis and colleagues used human PSCs to generate heart-forming organoids (HFOs) composed of the majority of developmentally relevant cell and tissue structures [203]. These include myocardium lined by endocardial cells, septum-transversum anlagen, both anterior and posterior foregut endoderm tissues as well as a vascular network. Remarkably, HFOs can be applied to study genetic defects, as demonstrated by the in vitro NKX2–5 knockout phenotype resembling that in transgenic mice. Silva et al have similarly capitalized on the cooperative development of heart and gut lineages to generate human iPSC-derived multilineage cardiac organoids that appear to recapitulate processes such as myocardial compaction and fetal-like maturation [204] (Figure 2). Others have additionally incorporated neuronal contribution to heart development, investigated the role of signaling activities on organoid formation and used cardiac organoids to mimic infarction or recapitulate diseases with known genetic components [198,205–207]. Recently published protocols outline methods to generate lineage restricted FHF/SHF-like cells in vitro [208,209]. While it is early days and the pioneer models will be further tested for their ability to faithfully recapitulate essential developmental processes and to function in similar ways to in vivo structures, the concept of generating relevant cardiac tissues in a dish is highly intriguing.

Figure 2. Ex vivo systems to study cardiac morphogenesis.

The in vitro organoid and ex vivo embryo imaging systems are illustrated. The various manipulation and analysis methods have been highlighted to show the breadth of technology available.

6. 3. Multimodal manipulations of ex vivo model systems

Ex vivo models such as the ones described here are amenable to a wide range of intrinsic and extrinsic manipulations which can further be spatio-temporally controlled. Thus, they afford hypothesis testing and discovery in a complex yet scalable manner. Testing the role and requirement of signaling activities can be done similarly to that in directed differentiations, but now with a read-out of a morphologically complex system. Similarly, unbiased high-throughput screening experiments may uncover entirely new concepts of early cardiac development. As illustrated by pioneering studies, taking advantage of the genetic tools in mice (via mESCs) or with gene edited hPSCs, organoids and live embryo culture in combination with lineage-tracing or selective ablation technology can shed light on the mechanisms of fate specification, cell-cell interaction and morphogenesis [24,208–210]. Complementary to the more traditional gene-regulated Cre/Dre recombinase approach come photoactivable Cre and degron systems which allow for controlled measurements of tissue specific genes or proteins, transient gene expression via modified RNA and gene activation/inhibition via CRISPR/Cas9 [211,212]. Recent developments in laser ablation or FRET based molecular tension microscopy could inform tension dynamics within the developing cardiac structure, providing an unbiased measure of tissue level forces. Confocal Raman microscopy has previously been used to measure tension in fish and mouse embryos and may be applicable to the emerging field of more complex tissue organoids as well [213,214].

Multiple exciting avenues come to mind applying such manipulations in developing cardiac structures. Similar to for example the studies of cell and tissue movements in the early drosophila embryo, high resolution imaging and cell tracking may inform on cell movements and cell-cell interactions as well as concomitant cell fate specification events. The increasingly powerful methods for single cell omics (single cell transcriptomics, genomics, proteomics and metabolomics), which have already been applied successfully to a broad range of developmental biology questions, and spatial transcriptomics promise to further shed light on the cell intrinsic mechanisms driving the processes of early cardiac development (Figure 2) [71,84,136,150,215–220].

Discussion

Congenital heart defects (CHDs) are consist of the most frequent birth defects in the human population and many of them remain poorly understood mechanistically. Work across numerous model organisms continues to highlight the importance of the early stage of heart development in CHD formation. Not surprisingly the complex processes involved in specifying and shaping the cardiac lineages into the 4-chambered heart are highly susceptible to error. Underlying genetic mechanisms for CHD have been broadly explored but have not resulted in a complete understanding of the various causes for heart malformation. Here we have focused on a critical window of cardiac development encompassing early mesoderm specification to heart tube formation, and specifically the various intrinsic and extrinsic mechanisms at play during this time. Comparison between species highlights broadly conserved mechanisms with respect to the underlying GRN, signaling activities, cell movements, impact from non-cardiac surrounding tissues and biomechanical cues. In contrast to most other organs proper formation of the heart relies heavily on extrinsic activities such as signals from the surrounding tissues and forces shaping and anchoring the growing heart. Beyond the simple basic fascination of elucidating these complex concepts the hope is that understanding and accurately reproducing the developmental processes that occur in vivo in in vitro systems will lead to the most relevant and accurate model systems to study human heart development and disease. While challenging to model, recent reports on complex multi-lineage organoid formation and the expansive field of cardiac bioengineering (reviewed in detail elsewhere) suggest that these interactions will prove essential to fulfilling the high expectation of providing a physiologically relevant and scalable ex vivo model system.