Portal de Periódicos da CAPES

Molecular Regulation of Atrioventricular Valvuloseptal Morphogenesis

1995; Lippincott Williams & Wilkins; Volume: 77; Issue: 1 Linguagem: Inglês

10.1161/01.res.77.1.1

1524-4571

Leonard M. Eisenberg, Roger R. Markwald,

Cardiomyopathy and Myosin Studies

HomeCirculation ResearchVol. 77, No. 1Molecular Regulation of Atrioventricular Valvuloseptal Morphogenesis Free AccessResearch ArticlePDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessResearch ArticlePDF/EPUBMolecular Regulation of Atrioventricular Valvuloseptal Morphogenesis Leonard M. Eisenberg and Roger R. Markwald Leonard M. EisenbergLeonard M. Eisenberg From the Department of Cell Biology and Anatomy, Medical University of South Carolina, Charleston. and Roger R. MarkwaldRoger R. Markwald From the Department of Cell Biology and Anatomy, Medical University of South Carolina, Charleston. Originally published1 Jul 1995https://doi.org/10.1161/01.RES.77.1.1Circulation Research. 1995;77:1–6Malformations of the cardiovascular system are the most frequently occurring type of birth defect, appearing in nearly one percent of newborn infants. The majority of congenital heart defects are due to abnormal development of the valves and membranous septa.1 These tissues arise from mesenchymal outgrowths, referred to as the cardiac cushions, that develop within two regions of the tubular embryonic heart: the atrioventricular (AV) canal and the ventricular outflow tract (conotruncus).2 The formation of the cushions within only these regions of the developing heart may be due to segmental patterning of the primitive heart tube. The cushions initially appear as regionally restricted thickenings of the cardiac jelly, the extracellular matrix (ECM) that resides between the myocardium and endocardium of the primitive heart tube. Cells migrate into the cushions as endothelial cells of the endocardium transform to mesenchyme.2 With regard to the formation of the endocardium, there is evidence for dual cellular origins, both within and outside the heart-forming fields.345 This may explain, at least in part, the heterogeneity of the endocardium; that is, some cells transform to mesenchyme, whereas other endothelial cells remain vascular.467 Furthermore, those endocardial and myocardial cells that arise from the heart-forming regions may share a common bipotential precursor. In this paper, we will give an overview of AV valvuloseptal morphogenesis (Fig 1) from the commitment of stem cells to this lineage during the onset of gastrulation, through the establishment of the AV cushions and their subsequent development into the valves and membranous septa. Moreover, we will present what is known about the molecular regulation of these developmental processes. It should be noted that the developmental biology of the outflow tract cushions contains many similarities to that of the AV canal. However, as the molecular biology of the AV cushions has been analyzed in much greater detail, we will emphasize AV cushion development in this review. Readers interested in biological details unique to the outflow tract region are encouraged to consult a recent review on outflow tract morphogenesis.8Origins of the EndocardiumThe hearts of all vertebrates are derived from bilateral paired fields of primary mesodermal cells that are specified to the cardiac lineage during gastrulation. These precardiac cells coalesce into an epithelium and become situated within the splanchnic layer, as lateral plate mesoderm separates into two distinct layers. Thereafter, the two heart-forming fields fuse at the ventral midline to form the primary heart tube, which consists of an outer myocardium surrounding an inner endocardium.9While there is general consensus in this description of early heart development in reference to the myocardium, the specifics regarding the endocardium are less clear. That the endocardial endothelium arises from cells that are situated between the cardiogenic splanchnic mesoderm and the endoderm is not in dispute.34510 However, the source of these pre-endocardial cells has not been resolved satisfactorily. Some researchers have shown that the endocardial precursors originate from the heart-forming fields themselves.46710 The presumptive endocardial cells would undergo an epithelial-to-mesenchymal transformation as they move in proximity to the anterior endoderm. Others have suggested that these cells emerge from mesodermal regions that are anterior and/or lateral to the primary heart-forming fields.3 These cells would then migrate into the cardiogenic regions on the endodermal side of the presumptive myocardium. As evidence for both points of view is compelling, the resolution of this conflict may simply be that both schools of thought are correct; that is, the endocardium consists of cells that arise from distinct locations. This hypothesis is supported by recent studies with the endothelial cell antibody marker JB3,6 which during early gastrulation stains only the precardiac mesoderm and later stains the endocardium in a heterogeneous pattern. Interestingly, positive reactivity to this antibody in the endocardium is exhibited only by those cells that will subsequently differentiate into cushion mesenchyme.Since it has been shown that at least some endocardial cells (JB3+) arise within the same mesodermal region as do the cells of the myocardium, another controversy that has developed concerns whether they share a progenitor. In support of this idea are data suggesting that a cell differentiation intermediate may exist that expresses both myocardial and endothelial markers.4 Additional support for the existence of the common progenitor comes from studies with the cardiac mesoderm cell line QCE-6.7 This cell line was derived from precardiac mesoderm of the Japanese quail and exhibits a phenotype consistent with its being a cardiac stem cell. If treated with a combination of retinoic acid, basic fibroblast growth factor (bFGF), transforming growth factor (TGF)–β2, and TGF-β3, the QCE-6 cells will differentiate into two distinct phenotypes—myocardial and endocardial endothelial—within the same culture. Conversely, data that conflict with the common progenitor hypothesis have been generated by labeling cells from the heart-forming fields with a β-galactosidase–expressing replication-defective virus.11 In these experiments, the viral label appears only among myocardial cells, implying that no common progenitor exists. However, alternative interpretations may be that divergence occurs either prior to retroviral labeling or as a function of positional information within the heart-forming fields. Recent experiments in zebrafish have demonstrated that positionally the entire heart-forming region is not identical.12 These experiments involved the injection of individual blastomeres, labeled with dye, into different positions within the heart-forming region. Depending on the exact position of injection, the labeled cells gave rise to either both myocardial and endocardial cells or exclusively one or the other phenotype. Although these experiments involved the injection of noncommitted cells, they do point out that identical multi-potential stem cells might have radically different fates due to slight differences in position within the early embryo.On the heels of the skeletal muscle biologists, many in the cardiac development field have tried to identify the transcriptional factors that determine cardiac commitment. Since cardiac and skeletal muscle express many of the same proteins, the initial attempts involved searches for cardiac homologues of the MyoD family. However, the possible common precursor of both myocardial and endocardial cells would imply a different regulatory mechanism for cardiac versus skeletal muscle differentiation. Regardless, the effort to discover "CardioD" has been largely unsuccessful. Other investigators have begun to identify transcription factors that are expressed by early myocardial cells and thus may be important for the specification of cell lineage. However, the expression of these transcriptional proteins in the early endocardial cells has either been negative or not analyzed. The only exception is GATA-4, which is a tissue-restricted transcription factor that is expressed in both early myocardial and endocardial cells.13 To date, no endocardial-specific transcription factors have been identified.Patterning of the Heart Tube and Positioning of the CushionsAs the primary heart forms, the cardiac jelly expands within the AV canal and outflow tract regions. These regions of matrix swelling, the cushion primordia, are regionally restricted not only along the longitudinal axis of the primary heart tube but also with reference to the cross-sectional axis.2 Specifically, the cushion swellings within the AV canal extend from the dorsal and ventral quadrants. In contrast, they extend from the right and left in the conus portion of the ventricular outflow tract. The mechanisms that regulate cushion placement within the developing heart are little understood. In some manner, the positioning of the cushions may be causally related to the looping of the primary heart tube.14 It is known that aberrant looping results in congenital defects developmentally linked to cushion positioning or fusion. Therefore, candidate genes for regulating cushion patterning may be those associated with looping abnormalities. Examples include the iv and inv genes or genes regulated by retinoic acid, as looping is modified in mice homozygous for the iv and inv mutations,1415 as well as among normal embryos treated with retinoic acid.16Insights into the molecular mechanism of cushion placement may be provided by a model of cardiac development that considers the primary heart tube to be segmented along the longitudinal axis. As a simplified model, the early heart could be subdivided into four segments: the ventricular outflow tract (conotruncus),17 ventricle, AV canal, and atrium. Accordingly, the alternating pattern of matrix swelling among these four regions would be the result of segmentation. In this respect, heart development might be homologous with that of the rhombomeres of the developing brain. These hindbrain structures are morphologically distinct units composed of regions of alternating patterns of gene and protein expression.18 That patterning of the primitive heart tube—including cushion placement along the longitudinal axis—is the morphological manifestation of such segmentation is supported by the expression patterns of several potential regulatory molecules within the heart. Most intriguing is the restricted expression of both bone morphogenetic protein (BMP)–419 and msx-220 within the myocardium of the AV canal and ventricular outflow tract. These molecules are members of the TGF-β and the homeobox gene families, respectively. Interestingly, both BMP-4 and msx-2 are also expressed together within alternating rhombomeres of the developing hindbrain.18Several other molecules have been shown to localize to the AV canal and outflow tract regions of the developing heart. Both TGF-β1 and TGF-β2 are initially expressed by cardiogenic mesoderm at high levels in a ubiquitous pattern.21 However, by the time cell migration into the cushions begins, their expression becomes restricted to either the endocardium or myocardium, respectively, of the AV canal and ventricular outflow tract. The homeobox genes msx-120 and mox-122 demonstrate expression that is restricted to the endocardium and cushion mesenchyme of these two regions of the heart.Finally, immediately prior to cushion tissue formation, particulates of 0.1 to 0.5 μm in size accumulate within the cardiac jelly of only the AV canal and ventricular outflow tract regions of the heart.23 These molecular complexes are referred to as adherons, and their components have been named the ES antigens. These particles can be visualized using antibodies to fibronectin, which is one of the constituent proteins of the particles. The ES proteins seem to be expressed in areas where epithelial-mesenchymal interactions occur, including the developing limb bud, neural tube, and cranial facial bones.24 Moreover, all these areas coexpress BMP-4, msx-1, and msx-2. The ES proteins are important for the formation of cushion mesenchyme and will therefore be discussed in greater detail in the next section.Cellularization of the CushionsFollowing the expansion of the cardiac jelly within the AV canal and ventricular outflow tract regions, these swellings are invaded with mesenchymal cells. The cushion cells arise from endocardial endothelial cells that transform into mesenchyme. Why does this mesenchymal transformation occur only within the AV canal and outflow tract? As discussed above, this may be explained, in part, by the existence of distinct populations of endothelial cells. However, this can be only part of the story, as the swelling of the cardiac jelly occurs prior to transformation. Since the cardiac jelly is composed of matrix molecules produced primarily by the myocardium, it is likely that the myocardium plays an important role in regulating the transformation of endothelial cells into the cushion mesenchyme. A representation of the molecular events of mesenchymal transformation within the cushions is shown in Fig 2.Understanding of the biology of cushion tissue formation has been greatly aided by the development of an in vitro model system.2324 The transformation of the endothelium to mesenchyme can be replicated by culturing explants—isolated from the embryonic heart just prior to the time of cushion cell seeding—on a collagen substrate. Endothelial cells that transform into mesenchyme will migrate away from the explant atop the gel and subsequently invade the collagen gel. Mesenchymal cells will be produced only if AV (or outflow tract) endothelium is cultured in the presence of AV (or outflow tract) myocardium. That is, mesenchyme will not form if AV endothelial cells are cultured in the absence of myocardium. Importantly, ventricular myocardium cannot induce the transformation of AV endothelium in vitro. Moreover, ventricular endothelium will not form mesenchyme, even in the presence of AV myocardium. These data correlate well with what occurs in the embryo; namely, that AV but not ventricular tissue will produce mesenchyme. Furthermore, it demonstrates that the regionally restricted formation of mesenchyme within the heart tube is due to a localization of both a myocardial-inducing activity and an endothelial-cell competence. Importantly, conditioned medium from cardiac cell cultures can replace the myocardial requirement for AV endothelial cell transformation.2324 This finding suggests that the myocardium elicits this endocardial cell differentiation via a secreted molecule(s).The first evidence for a secreted myocardial product unique to the cushion-forming regions was the identification of the ES protein particulates (adherons).23 Since the cardiac jelly is principally an expanded myocardial basement membrane, the localization of these particles to this matrix implied that they were produced by the myocardium. These particulates, isolated by EDTA extraction and ultracentrifugation, were shown by electron microscopy to be composed of smaller 30 to 60 nm particles. An antiserum prepared to these ES complexes stained the heart in a similar pattern as the fibronectin antibodies, with particles apparent within the cardiac jelly of the AV canal and outflow tract. The particulates have been shown to contain ≤10 proteins, which include fibronectin, transferrin, and a novel protein termed ES/130.24Do the ES proteins play a role in the transformation of endothelial cells? Again, the collagen gel assay has provided insights into this question. The addition of EDTA extracts to explants of AV endothelium allowed mesenchymal cells to form and invade the collagen substrate.23 As the extraction of hearts with EDTA was used to identify the regionally expressed particles of the cardiac jelly, it implied that these particles may be involved in the transformation process. Furthermore, the ability of cardiocyte-conditioned medium to promote mesenchymal cell transformation could be inhibited by the ES antiserum.Another group of proteins implicated in the formation of cushion mesenchyme is the TGF-β family. As discussed above, the expression of various TGF-β proteins is localized to the AV canal and ventricular outflow tract. While the addition of TGF-β by itself will not elicit endothelial transformation in vitro, the treatment of AV explants with TGF-β–specific antibodies will block the formation of mesenchyme.25 Furthermore, AV endothelial explants treated with cardiocyte-conditioned medium not only will produce mesenchymal cells but also will increase their expression of TGF-β.26 Accordingly, the addition of ES antiserum to these treated cultures will block both events; that is, the transformation into mesenchyme and the increase in TGF-β expression. An interpretation of these results is that TGF-β is an endocardial elaboration of the myocardial signal for transformation. As TGF-β expression by the endocardium is under regulation of the ES proteins, we believe that an understanding of the role the myocardium plays in cushion tissue formation rests with a further characterization of the ES proteins.As the transformation of endothelial cells to mesenchyme involves changes in both cell-cell and cell-substrate adhesion, an understanding of cushion mesenchyme would be incomplete without a consideration of the molecules that mediate these interactions (Fig 2). The cell adhesion molecule N-CAM is expressed throughout the endocardium prior to the formation of the cushions.27 As cells from the cushion-forming regions detach from the endothelial layer, their levels of cell surface N-CAM decrease dramatically in prelude to their migration into the cardiac jelly. As N-CAM is downregulated, endothelial cells begin to express the substrate adhesion molecule (SAM) cytotactin, also known as tenascin.27 This molecule, which is expressed transiently by cushion cells, is presumably involved during the initial phases of the mesenchymal transformation. Cytotactin has been shown in other cell systems to disrupt cell-substrate adhesion, therefore allowing cells to migrate more freely through matrix. Another class of molecules that are probably required for mesenchymal cell migration is the serine proteases, which are secreted by cells in the cushions as they transform to mesenchyme.28Differentiation of Mesenchyme Into Valvuloseptal TissueThe stage of valvuloseptal development for which the least is known concerns the transition from endocardial cushions to embryonic valves and septa. Contrary to initial hypotheses,29 it now seems clear that cushion mesenchyme does differentiate into the fibrous connective tissue of both inlet and outlet valvular leaflets.30 However, the inlet valves—particularly the tricuspid valve as well as the septa formed by the fusion of the cushions (eg, the conal septum of the proximal outflow tract)—are partially "myocardialized." This outcome is due to the invasion of cushions by myocardial cells to give a muscular component to the terminally differentiated tissues. Thus, myocardialization appears to be a normal morphogenetic event occurring in most vertebrates, including humans, to give specific cushion structures their mature phenotype. While virtually nothing is known about the processes regulating the differential myocardialization of cushion tissue, the process, if abnormal, may explain complex congenital anomalies such as Ebstein's anomaly.30The differentiation of the cushion mesenchymal cell into a valvuloseptal fibroblast correlates with the expression of the two microfibrillar proteins, fibrillin and fibulin.631 A possible role of these two ECM molecules may be to serve as a scaffolding for cell adhesion as these cells differentiate. Presumably, the transitional ECM that is laid down as the cushions develop into the valve leaflets and membranous septa is necessary for proper tissue remodeling. Other matrix proteins that may be involved in the remodeling of the cushions into valvuloseptal tissue include sulfated proteoglycans.32 These molecules accumulate at the cell surface or pericellular matrix of the migrating cushion cells that secrete them and appear to play a role in the final positioning and spacing of cells within the differentiated tissue.Additional molecules that may be important for formation of valvuloseptal tissue include TGF-β1, which maintains high expression levels throughout the formation of the cardiac valves.21 Recently, a novel protein has been described, referred to as GbHLH1.4, which is a member of the helix-loop-helix (HLH) family of transcription factors.33 GbHLH1.4 is expressed in cushions from their initial cellularization and is maintained at high levels throughout valve development. Interestingly, there is an overlapping expression pattern with another HLH protein, Id.34 Unlike other HLH proteins, Id lacks a DNA binding region. Id is thought to exert its effects on differentiation by forming heterodimers with other HLH proteins. Presumably, this renders the other HLH proteins inactive by inhibiting their ability to bind DNA. Both GbHLH1.4 and Id are expressed in the early cushions. However, as cardiac valve development proceeds, the expression of Id drops significantly without a concomitant decrease in GbHLH1.4. Possibly, the drop in Id expression may have effects on the activity of the latter protein during cardiac development and on the development of valvuloseptal tissue. Regardless, the regulation of this last phase of valvuloseptal development is little understood and in need of further study.SummaryThe majority of congenital heart defects arise from abnormal development of valvuloseptal tissue. The primordia of the valve leaflets and membranous septa of the heart are the cardiac cushions. Remodeling of the cushions is associated with a transitional extracellular matrix that includes sulfated proteoglycans and the microfibrillar proteins fibulin and fibrillin. Cushion formation is restricted to the AV canal and ventricular outflow tract regions of the primary heart tube. The proper placement of the cushions may be the result of the development of the primary heart tube as a segmented organ, as well as the subsequent looping of the heart. Segmentation of the heart tube may be demonstrated by the alternating molecular expression pattern along the longitudinal axis. In support of this hypothesis is the restricted expression of BMP-4 and msx-2 to the AV canal and ventricular outflow tract. The importance of looping for cushion positioning may imply that the iv and inv genes and retinoic acid are important for the proper patterning of the heart. The cells of the cushions evolve from endocardial cells that undergo an epithelial-to-mesenchymal transformation. This developmental event is regulated by the myocardium and is probably due to the production of protein complexes, present within the cardiac jelly of the cushion-forming regions, that consist of fibronectin and the ES proteins. Both the cushion mesenchyme and its endocardial cell antecedents express JB3, an ECM protein. JB3 expression is also featured within the heart-forming fields of the primary mesoderm, from which the endocardial progenitors of the cushion cells originate. Although not proven definitively, it is believed that these endothelial cells are descendants of a bipotential stem cell that is also the progenitor of the myocardium.Download figureDownload PowerPoint Figure 1. Schematic diagram depicting valvuloseptal morphogenesis of the developing heart. The primordia of valvuloseptal tissue are the endocardial cushions. The JB3+ progenitors of the cushion cells originate from the heart-forming fields and may also give rise to the myocardial lineage. Following their differentiation into pre-endocardial mesenchyme, these cells coalesce—possibly along with vascular mesenchyme that originates from outside the heart-forming fields—to form the endocardial endothelium. The cushions first develop as regionalized swellings of the cardiac jelly and contain protein complexes of fibronectin and ES proteins (adherons). These ES proteins, produced by the myocardium, elicit the transformation of endothelial cells to cushion mesenchyme. Differentiation of the cushions into valvuloseptal tissue is mediated, in part, by the transitional matrix, which includes the microfibrillar proteins fibulin and fibrillin.Download figureDownload PowerPoint Figure 2. Depiction of the molecular events involved in the transformation of endothelial cells into cushion mesenchyme. The myocardium of the primitive heart tube demonstrates a segmental pattern of molecular expression. Molecules expressed uniquely within the AV canal and ventricular outflow tract regions include transforming growth factor (TGF)–β2, bone morphogenetic protein (BMP)–4, msx-2, and the ES proteins. These latter proteins (which vary in size from 28 to 130 kD) form complexes with fibronectin and appear to regulate endothelial cell transformation. As endothelial cells are activated to transform, cell surface neural cell adhesion molecule (N-CAM) is downregulated as the cells begin secreting higher levels of both substrate adhesion molecules (SAM; ie, cytotactin) and proteases. Also, cells undergoing transformation increase their expression of TGF-β and the msx-1 homeobox gene. Subsequently, cells that have converted to mesenchyme express the helix-loop-helix protein Id and begin to produce the extracellular matrix proteins fibulin and fibrillin.This paper was supported by National Institutes of Health grant HL-K37-33756. We are grateful to Robert Thompson, Charlie Little, and Carol Eisenberg for critical reading of this manuscript.FootnotesCorrespondence to Roger R. Markwald, PhD, Department of Cell Biology and Anatomy, Medical University of South Carolina, Charleston, SC 29425. References 1 Clark EB. Mechanisms in the pathogenesis of congenital cardiac malformations. In: Pierpont ME, Moller JH, eds. Genetics of Cardiovascular Disease. Boston, Mass: Martinus-Nijhoff; 1987:3-11. Google Scholar2 Markwald RR, Fitzharris TP, Manasek FJ. Structure and development of endocardial cushions. Am J Anat.1977; 148:85-120. CrossrefMedlineGoogle Scholar3 Drake CJ, Jacobson AG. A survey by scanning electron microscopy of the extracellular matrix and endothelial components of the primordial chick heart. Anat Rec.1988; 222:391-400. CrossrefMedlineGoogle Scholar4 Linask KK, Lash JW. Early heart development: dynamics of endocardial cell sorting suggests a common origin with cardiomyocytes. Dev Dyn.1993; 195:62-69. CrossrefGoogle Scholar5 Virágh S, Szabó E, Challice CE. Formation of the primitive myo- and endocardial tubes in the chicken embryo. J Mol Cell Cardiol.1989; 21:123-137. CrossrefMedlineGoogle Scholar6 Wunsch A, Little CD, Markwald RR. Cardiac endothelial heterogeneity is demonstrated by the diverse expression of JB3 antigen, a fibrillin-like protein of the endocardial cushion tissue. Dev Biol.1994; 165:585-601. CrossrefMedlineGoogle Scholar7 Eisenberg CA, Bader D. QCE-6: a clonal cell line with cardiac myogenic and endothelial cell potentials. Dev Biol.1995; 167:469-481. CrossrefMedlineGoogle Scholar8 Pexieder T. Conotruncus and its septation at the advent of the molecular biology era. In: Clark EB, Markwald RR, Takao A, eds. Developmental Mechanisms of Heart Disease. Armonk, NY: Futura Publishing Company, Inc; 1995:227-247. Google Scholar9 Stalsberg H, DeHaan RL. The precardiac areas and formation of the tubular heart in the chick embryo. Dev Biol.1969; 19:128-159. CrossrefMedlineGoogle Scholar10 DeRuiter MC, Poelmann RE, Mentink MMT, Vaniperen L, Gittenberger-De Groot AC. Early formation of the vascular system in quail embryos. Anat Rec.1993; 235:261-274. CrossrefMedlineGoogle Scholar11 Mikawa T, Borisov A, Brown AMC, Fischman DA. Clonal analysis of cardiac morphogenesis in the chicken embryo using a replication-defective retrovirus, I: formation of the ventricular myocardium. Dev Dyn.1992; 193:11-23. CrossrefMedlineGoogle Scholar12 Lee RKK, Stainier DYR, Weinstein BM, Fishman MC. Cardiovascular development in the zebrafish, II: endocardial progenitors are sequestered within the heart field. Development.1994; 120:3361-3366. CrossrefMedlineGoogle Scholar13 Heikinheimo M, Scandrett JM, Wilson DB. Localization of transcription factor GATA-4 to regions of the mouse embryo involved in cardiac development. Dev Biol.1994; 164:361-373. CrossrefMedlineGoogle Scholar14 Icardo JM, Arrechedera H, Colvee E. Atrioventricular endocardial cushion in the pathogenesis of common atrioventricular canal: morphological study in the iv/iv mouse. In: Clark EB, Markwald RR, Takao A, eds. Developmental Mechanisms of Heart Disease. Armonk, NY: Futura Publishing Company, Inc; 1995:529-544. Google Scholar15 Yokoyama T, Copeland NG, Jenkins NA, Montgomery CA, Elder FFB, Overbeek PA. Reversal of left-right asymmetry: a situs inversus mutation. Science.1993; 260:679-682. CrossrefMedlineGoogle Scholar16 Chen Y, Solursh M. Comparison of Hensen's node and retinoic acid in secondary axis induction in the early chick embryo. Dev Dyn.1992; 195:142-151. CrossrefMedlineGoogle Scholar17 Gittenberger-de Groot AC, Bartelings MM, Poelmann RE. Overview: cardiac morphogenesis. In: Clark EB, Markwald RR, Takao A, eds. Developmental Mechanisms of Heart Disease. Armonk, NY: Futura Publishing Company, Inc; 1995:157-168. Google Scholar18 Graham A, Francis-West P, Brickell P, Lumsden A. The signalling molecule BMP4 mediates apoptosis in the rhombencephalic neural crest. Nature.1994; 372:684-686. CrossrefMedlineGoogle Scholar19 Jones CM, Lyons KM, Hogan BL. Involvement of bone morphogenetic protein-4 (BMP-4) and Vgr-1 in morphogenesis and neurogenesis in the mouse. Development.1991; 111:531-542.