Developmental anatomy of the heart: a tale of mice … · Developmental anatomy of the heart: a...

review

Developmental anatomy of the heart: a tale of mice and man1

Andy Wessels and David SedmeraDepartment of Cell Biology and Anatomy, Cardiovascular Developmental Biology Center,Medical University of South Carolina, Charleston, South Carolina 29425

Submitted 12 March 2003; accepted in final form 26 August 2003

Wessels, Andy, and David Sedmera. Developmental anatomy ofthe heart: a tale of mice and man. Physiol Genomics 15: 165–176, 2003;10.1152/physiolgenomics.00033.2003.—Because of the increasing avail-ability of tools for genetic manipulation, the mouse has become the mostpopular animal model for studying normal and abnormal cardiac devel-opment. However, despite the enormous advances in mouse genetics,which have led to the production of numerous mutants with cardiacabnormalities resembling those seen in human congenital heart disease,relatively little comparative work has been published to demonstrate thesimilarities and differences in the developmental cardiac anatomy inboth species. In this review we discuss some aspects of the comparativeanatomy, with emphasis on the atrial anatomy, the valvuloseptal com-plex, and ventricular myocardial development. From the data presentedit can be concluded that, apart from the obvious differences in size, themouse and human heart are anatomically remarkably similar through-out development. The partitioning of the cardiac chambers (septation)follows the same sequence of events, while also the maturation of thecardiac valves and myocardium is quite similar in both species. Themajor anatomical differences are seen in the venous pole of the heart. Weconclude that, taking note of the few anatomical “variations,” the use ofthe mouse as a model system for the human heart is warranted. Thus theanalysis of mouse mutants with impaired septation will provide valuableinformation on cellular mechanisms involved in valvuloseptal morpho-genesis (a process often disrupted in congenital heart disease), while thestudy of embryonic lethal mouse mutants that present with lack ofcompaction of ventricular trabeculae will ultimately provide clues on theetiology of this abnormality in humans.

mouse; human; cardiac development; embryo

A FAST-INCREASING NUMBER of genetically modified mousemodels with structural and functional abnormalities inthe cardiovascular system undoubtedly will contributeto an improved understanding of molecular and mor-phological mechanisms that regulate human heart de-velopment in health and disease. It becomes more andmore obvious, however, that for proper extrapolation offindings in the murine model systems it is crucial tohave detailed knowledge of the specific anatomical/morphological details of developing and adult heart inmice and humans (2, 33, 35).

Before discussing some of the specific anatomicalfeatures in both species, it is important to consider the

dimensions and properties of the full-grown hearts.The human heart weighs about 250–300 g and beats,on average, 60–70 times per minute. In contrast, themouse heart weighs only about 0.2 g, and has a heart-beat of around 500–600 times per minute. Apart fromtheir size, the most pronounced difference in externalfeatures between the adult mouse and human heart isfound in their overall shape. The most important factordetermining this shape is the anatomical context of theadult heart within in thoracic cavity. In human, theheart “rests” on the diaphragm. This is reflected in itsmore pyramidal shape and a flat dorsal (or inferior)surface. In comparison, the mouse heart, which in thefour-legged mouse does not rest on the diaphragm andhas more room to move freely within the pericardialcavity, has a more ellipsoidal (“rugby ball” shape).Another difference in external features is that in thehuman heart the atria are very prominent structures,whereas in the mouse heart the atrial chambers andtheir appendages are relatively small.

Developmentally, it is interesting to note that thegestational window during which the heart develops isquite different in the mouse and human. In the human

Article published online before print. See web site for date ofpublication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: A. Wes-sels, Dept. of Cell Biology and Anatomy, Medical Univ. of SouthCarolina, 173 Ashley Ave., Basic Science Bldg., Rm. 648, PO Box250508, Charleston, SC 29425 (E-mail: [email protected]).

1This review article was based on work originally presented at the“NHLBI Symposium on Phenotyping: Mouse Cardiovascular Func-tion and Development” held at the Natcher Conference Center, NIH,Bethesda, MD, on October 10–11, 2002.

Physiol Genomics 15: 165–176, 2003;10.1152/physiolgenomics.00033.2003.

1094-8341/03 $5.00 Copyright © 2003 the American Physiological Society 165

it takes about 2 mo (from conception) for the heart tocomplete septation, followed by another 7 mo to furthermature until the baby is born and the pulmonarycirculation kicks in. In the mouse, however, it takesonly 2 wk from the time of conception for cardiacseptation to complete. After that, the mouse fetus hasless than 1 wk of prenatal life before birth. Withoutgoing into any detail, it suffices to say that some of thedevelopmental events that in the human are more orless completed at birth are still in progress in theneonatal mouse. There are several publications in lit-erature in which the comparative developmentalstages in mouse and human are listed (28, 39). Thesetables are helpful when studying developmentalevents. It is important to realize, however, that thestaging protocol used to determine the developmentalstage of mouse embryos is not always the same. Forinstance, when using the “vaginal plug” method, mostinvestigators will mark the morning on which a spermplug is observed as embryonic day 0.5 (ED0.5 ; Ref. 34).Others, however, will consider this ED1.0. This obvi-ously can lead to confusion, and more importantly,misinterpretation of data. Thus, when describing andcomparing developmental events, it is very importantto be very explicit about the staging process whenpresenting developmental data. Moreover, when work-ing with tissues generated by others (previously col-lected materials, existing collections, collaborations)where the exact staging protocol that has been followedis not known, it is equally important to point that outwhen reporting on data based on these tissues. As moreand more clinical imaging techniques are now beingadapted for use in studies on mouse models [e.g., MRI,ultrasound (10)], it is imperative to keep the issuesmentioned above in mind. Thus, although the murineand human heart are anatomically very similar, it iscrucial to be very careful when extrapolating informa-tion obtained by studying the mouse into the humancontext.

In this review we will first discuss some of the ana-tomical characteristics of the human and murine car-diovascular system. It will be demonstrated that thesecharacteristics do not differ significantly. In the re-mainder of the review we will, therefore, focus on thegeneral development of the cardiovascular system inboth species by providing some insight in the mecha-nisms that lead to the formation of the functional, fullyseptated, four-chambered heart. As cardiac develop-ment is a very complex process, and it is beyond thescope of this review to discuss all the aspects thatrelate to this event, we will limit ourselves to thediscussion of two topics that relate to congenital mal-formations frequently seen in humans. First, as manyof the congenital heart defects found in the newbornhuman heart include abnormalities of the valvulosep-tal complex, we will discuss some of the developmentalevents that lead to proper valvuloseptal morphogene-sis. We will also, albeit less extensively, describe someof the events that are involved in myocardial differen-tiation and maturation.

MATERIAL AND METHODS

Most of the images of human and murine specimens usedin the illustrations of this review were generated in thecontext of studies published previously (26, 27, 31, 37, 38,40–43). Some of the scanning electron microscope (SEM)images are from the collection of late Tomas Pexieder. De-tails regarding tissue collection, preparation, histology, im-munohistochemical staining procedures, characterizationand properties of antibodies, and the preparation of SEMsamples can be found in papers mentioned above and in theremainder of the text. As most of the photographs in therespective histological and SEM collections did not containscale bars, we could unfortunately not provide scale bars forthe figures in this paper.

THE ANATOMY OF THE POSTNATAL HEART IN MOUSEAND HUMAN

The basic anatomical features of the postnatal heartin the human and mouse are very similar (Fig. 1). Thusin both species the heart has four chambers; two atria,separated by an interatrial septum (IAS), and twoventricles, separated by an interventricular septum(IVS). In addition, located between the IAS and IVSthere is a small “septal segment,” which, as a result ofthe offset of the atrioventricular (AV) valves, is knownas the atrioventricular septum (AVS), as it is basicallysituated in between the subaortic outlet segment of theleft ventricle from the right atrium. In the human, thisAVS is a thin fibrous structure, known as the membra-nous septum. In the mouse this structure is relativelythick and mostly muscular, partly as a result of de-layed delamination of the septal leaflet of the tricuspidvalve (35), and partly as a result of myocardializationof the mesenchymal tissues (13). In the junction situ-ated between the atria and ventricles, i.e., the AVjunction (AVJ), we find two AV valves. The arrange-ment of the AV valves in mouse and human is compa-rable (35). In the left AVJ we find a mitral valve, whichhas two distinct leaflets (a bicuspid valve), whereas inthe right AVJ a tricuspid valve is located, which hasthree distinct leaflets. In the human heart the valvesare at their “tips” in continuity with very pronouncedpapillary muscles via relatively long and numeroustendinous chords (chordae tendineae; Fig. 1A). In themouse these chordae are far less prominent (Fig. 2, Aand C). The inner lining of the ventricles is character-ized by the presence of numerous myocardial protru-sions better known as trabeculae (trabeculae carneae).In the human, but not so in the mouse, there is apronounced difference in the morphology of the trabec-ulae in the right vs. the left ventricle (35). In the leftventricle of the human heart, the trabeculae are rela-tively thin, whereas the trabeculae in the right ventri-cle have been described as being coarse. In addition tothe trabeculation and the chordae tendineae, we can,within the apical cavity of the ventricles in both spe-cies, also discriminate thin, cordlike, structures thatresemble the tendinous chords attached to the papil-lary muscle. Because of their tendonlike appearance, inolder anatomy textbooks (e.g., Told-Hochstetter,Anatomischer Atlas, 1948) these structures are often

166 REVIEW: ANATOMY OF MOUSE AND HUMAN HEART

Physiol Genomics • VOL 15 • www.physiolgenomics.org

described as trabeculae tendineae. However, despite ofthis resemblance, these thin structures are not reallytendons, but are actually extensions of the subendocar-dial ventricular network of the cardiac conduction sys-tem, and are therefore nowadays generally referred toas “false tendons” (Fig. 2D). As they consist almostexclusively of Purkinje cells (i.e., specialized myocytesof the conduction system) within a fibrous sheet, theycan be used to obtain enriched preparations of conduc-tion system related genes (20) or to perform physiolog-ical experiments (9). A slight morphological differencein the overall ventricular anatomy between mouse andhuman is found in the relative size and shape of themuscular ventricular septum and the position of theaortic outlet in relation to the IVS (cf. Figs. 1 and 2). Inthe human heart the muscular IVS is a massive struc-ture, its thickness approaching or exceeding that of theleft ventricular free wall (Fig. 1, A and B). The humanmuscular IVS has a very broad base just below the AVvalve attachments. In the mouse, the IVS is not quiteas massive and compact (Fig. 2, A and B), and at thebase it gradually tapers toward the AV septum. Thedifference in the angle of the aortic outlet relative tothe axis of the IVS is another remarkable feature (cf.Fig. 1B and Fig. 2A). The most prominent anatomicalvariations between the cardiovascular systems inmouse and human are probably those found in thevenous components of the atria (Fig. 3). All of thesevariations reflect the species-specific differences in de-velopment of the venous tributaries connecting to theinferior atrial wall. Thus, whereas in the human heartthe left atrium receives four pulmonary veins (37), inthe mouse heart, the pulmonary veins join in a pulmo-nary confluence behind the left atrium (Fig. 3, G andH), which in turn empties via a single foramen into thedorsal wall of the left atrium (13, 36). Another anatom-ical difference in the atrial anatomy relates to thevenous drainage into the right atrium. During theearly stages of cardiac morphogenesis, in both the

human and murine embryo, the left and right superiorcaval veins (LSCV and RSCV) initially drain into thesinus venosus. The sinus venosus itself opens via thesinoatrial foramen into the right atrium (36, 37). Asdevelopment progresses, in the human heart the leftcaval vein regresses and the remaining proximal por-tion (with part of the left sinus horn) becomes thecoronary sinus, the remnant of the LSCV becoming theso-called ligament of Marshall and oblique vein (Fig. 3,A–C). In the mouse, however, the LSCV does not re-gress and persists into postnatal life (Fig. 3H). In thehuman, persistent LSCV is considered a congenitalmalformation (incidence �1:100) frequently associatedwith other congenital malformations. It is likely thatthese differences in development of the venous tribu-taries into the atrial chambers are the underlyingreasons for the relatively small size of the atrial com-ponents in the murine heart.

Valvuloseptal Development

The first step in cardiac development in the verte-brate heart involves the formation of two bilateralheart fields of precardiac mesoderm. These heart fieldsare situated on opposite sides of the embryonic midline(6, 24) and contain endocardial as well as myocardialprecursor cells. As the embryo develops, the heartfields fuse, resulting in the formation of a primaryheart tube. This tubular heart consists of a myocardialouter mantle with an endocardial inner lining. Be-tween these two concentric epithelial cell layers, anacellular matrix is found which is generally referred toas the cardiac jelly (Fig. 4A). During cardiac looping,the cardiac jelly basically disappears from the futuremajor chambers of the heart (i.e., atria and ventricles),whereas in the junction between the atria and ventri-cles, the atrioventricular junction (AVJ), as well as inthe developing outflow tract (OFT), the cardiac jellystarts to accumulate. This results in the formation of

Fig. 1. Anatomy of the postnatal humanheart. A: a postnatal human heart in 4-cham-ber view (from Ref. 1, with permission). B:histological section of a 5-wk old neonatalheart, sectioned in the same orientation, isshown. The section was immunohistochemi-cally stained for the presence of atrial myosinheavy chain (aMHC) as described in Ref. 40.In addition to the expression of aMHC in theatria, expression of this MHC isoform is alsoobserved in the left and right bundlebranches of the atrioventricular conductionsystem (arrows). Note the orientation andthickness of the interventricular septum(IVS) in relation to the left and right ventric-ular free wall. Ao, aorta; AVS, atrioventricu-lar septum; IAS, interatrial septum; LA, leftatrium; LBB, left bundle branch; LV, leftventricle; RA, right atrium; RBB, right bun-dle branch; RV, right ventricle.

167REVIEW: ANATOMY OF MOUSE AND HUMAN HEART


Fig. 2. Scanning electron micrographs(SEM) of the postnatal mouse heart. Aand B: SEM images of the posterior (A)and anterior (B) half of an adult mouseheart. C: an enlargement of the boxedarea in A, showing the relatively smallleaflets of the murine mitral valve (cf.Fig. 1A). D: an enlargement of theboxed area in B and demonstratessome of the false tendons in the apicalportion of the right ventricle. See leg-end to Fig. 1 for definitions of abbrevi-ations.

Fig. 3. Development of the venous pole of the heart. A–C: a schematic representation of the development of thevenous pole in human heart (modified from Fig. 7-10 in Human Embryology, William J. Larsen, ChurchillLivingstone, 1993). D–F: immunohistochemically stained section of human embryos stained for �-MHC at,respectively, 5.5 (D), 7 (E), and 8 wk (F) of development. G: a histological section of a nonperfused neonatal mouseheart. H: is a schematic representation of G (color coding according to A–C). A: in early development (�4–5 wk)the left and right horns of the sinus venosus communicate with the developing (unseptated) atrium. The left sinushorn (LSH) receives blood from several vessels including the left superior caval vein (LSCV), and the right sinushorn (RSH) is connected to the right superior caval vein (RSCV). As development progresses, the LSCV regresses(see B), resulting in the formation of the oblique vein (see C), with the remnant of the LSH becoming the coronarysinus (CS) emptying into the right atrium (RA; see C). The cartoon also illustrates how during human cardiacdevelopment the confluence of the pulmonary veins (PuV; see A) becomes incorporated into the left atrium (see Band C). D: at 5.5 wk of development the pulmonary veins connect to the LA through a single communication (notethat the walls of the pulmonary veins are not myocardial at this point). At 7 wk the walls of the confluence of thepulmonary veins are becoming myocardial (E). The process of incorporation is nearly completed at 8 wk; the wallsof the pulmonary veins are now myocardialized up to the pericardial reflection (arrows). The images in G and Hdemonstrate that in the mouse the pulmonary confluence does not become incorporated into the LA. Instead, thepulmonary veins drain into the LA via a single channel (arrow). G and H also illustrate that, unlike in the human,persisting LSCV is a normal anatomical situation in the postnatal mouse. ICV, inferior caval vein; PA, pulmonaryartery; SCV, superior caval vein. See legends of previous figures for other abbreviations.



the endocardial cushion tissues in the AVJ (see Figs. 4,E–H, for examples in the mouse and Figs. 6, E and Ffor an example in the human) and OFT. An importantdevelopmental event in the subsequent maturation ofthese cushions is the endocardial epithelial-to-mesen-chymal transformation (EMT; see Ref. 19). During thisEMT, a subset of the endocardial cells delaminate from

their epithelial context, transform, thereby assuming amesenchymal phenotype, and start to migrate into theextracellular matrix of the cushions (Fig. 4B). Thisprocess, which is, at least partly, regulated by growthfactors that are produced in the underlying myocar-dium (for a review, see Ref. 7), results in the mesen-chymalization of the cushions. The endocardial cushion



tissues are extremely important for cardiac morpho-genesis. They are the major “building blocks” of, andprovide the “glue” for, virtually all of the septal struc-tures in the heart (38). Dysmorphogenesis of the endo-cardial cushions and/or failure of proper fusion aregenerally thought to play a major role in the etiology ofcongenital heart disease in humans, for instance, inthe setting of common AV defects. Gross abnormalitiesin formation of endocardial cushion tissue is observedin a variety of genetically modified mouse models suchas the neurofibromin-1 (16), hyaluronan synthase-2(Has2) (4), and RXR-� knockout mice (8) and is usuallyassociated with fetal lethality. The spectrum of cushionabnormalities is very broad. In some genetically al-tered mice, hardly any cushion tissue is formed,whereas others develop hypercellularized and/or en-larged cushions. In many cases this appears to be aresult of perturbed endocardial-to-mesenchymal trans-formation leading to either too many (16) or too few(14) endocardially derived cells in the cushions tissues.Although, in general, mice with cushion abnormalitiesmanifest defects in both the AV as well as the OFTregion, sometimes this is not the case and only the OFTcushions are affected (NFAT-C and SOX-4; Ref. 44). Asindicated above, endocardial cushions develop in theAVJs as well as in the OFT. There are many similari-ties in the fate of endocardial cushions in AVJ andOFT. The cushion tissues in the AVJ contribute to theformation of AV septal structures and AV valves, andthe cushion tissues in the OFT take part in septation ofthe OFT and in the formation of the semilunar valvesof aorta and pulmonary artery (see, e.g., Ref. 23). In thenext section we will discuss some of the aspects of AVdevelopment in a little more detail.

The AV junction. As briefly outlined above, the en-docardial jelly in the AVJ forms the base material forthe AV cushions (21, 38). Initially only two endocardialcushions develop which face each other on opposingsides of the common AV canal. These cushions, whichwe will refer to as the “major” AV cushions, are knownas the inferior (also known as posteroinferior or dorsal)AV cushion and the superior (also known as ventral oranterosuperior) AV cushion (Fig. 4, B and E). Themajor AV cushions are the most prominent endocardialcushion tissues in the heart, and, as development pro-ceeds, the leading edges of the cushions fuse (Fig. 4, C,F, and G), thereby separating the common AV canalinto a left and right AV orifice (Fig. 4, C, D, and F). Atthis point, another set of, relatively small, endocardial

cushions starts to develop in the lateral aspects of theAVJ. These cushions are generally referred to as thelateral cushions (Fig. 4, C and F). Interestingly, al-though the lateral cushions are very important forvalvuloseptal morphogenesis, as they contribute to theformation of anterosuperior leaflet of the tricuspidvalve (17) and the mural leaflet of the mitral valve, themechanisms that are responsible for inducing theirdevelopment and differentiation have thus far beenvery poorly studied.

Fate of the AV cushions. After fusion, the (major) AVcushion-derived tissue basically forms a large mesen-chymal “bridge” spanning from the posterior wall of theAV canal to the anterior wall of the heart (37). Postero-inferiorly, this mesenchymal component is in continu-ity with the dorsal mesenchymal protrusion/dorsal me-socardium and the mesenchymal cap (also endocardialcushion material) that covers the leading edge of theforming primary interatrial septum (pIAS) septum (de-scribed in detail in Ref. 37). This mesenchymal cap isanterosuperiorly also in continuity with the remnant ofthe superior AVC cushion. The pIAS at this point hasnot closed the communication between left and rightatrium yet, and as a result the mesenchymal bridgespans the primary atrial foramen (interatrial commu-nication) between the dorsal and ventral aspects of thefused cushions. As development progresses, the mes-enchymal cap on the primary septum and the fusedcushions merge completely, thereby closing this pri-mary foramen. The remodeling of this mesenchymal“crux of the heart,” which at this point basically con-sists of the fused major AV cushions and the mesen-chymal cap of the pIAS and which is situated betweenthe muscular IVS and the IAS (Fig. 5A), eventuallyleads to the formation of the membranous AV septum,the septal leaflet of the tricuspid valve (17), and theaortic leaflet of the mitral valve (Fig. 5B). Although theendocardial cushion tissues are important in the for-mation of the valves, AV valve morphogenesis alsoinvolves a number of myocardial remodeling steps (Fig.6). Below, we have summarized (and simplified) thesteps that are involved in the formation of the valves. Itis important to realize that not all the leaflets of the AVvalves develop exactly the same (see Refs. 17 and 21).However, in the context of this paper it suffices todescribe the general principle, focusing on the fate ofthe lateral cushions. First, it is important to determinethe relationship between the cushions and the adjacenttissues. The cushions have an upper (atrial) boundary

Fig. 4. Development of the atrioventricular cushions in the mouse. A–D: schematic of the development of the AVcushions in the mouse/human heart. A: the tubular heart stage. The outer myocardial layer and the innerendocardial cell layer sandwich the cardiac jelly (CJ). B: the stage in which the cardiac jelly has given rise to theformation of the major AV cushions. The cells within the subendocardial space indicate that endocardial-to-mesenchymal transformation has been initiated. C: the mesenchymalized major cushions are fusing, while at thelateral aspects of the AV canal the lateral AV cushions have formed. In D, the fusion has been completed, and awedge of mesenchyme separates the left from the right AV orifice. The specimen shown in E is representative forthe cartoon in B, F shows a specimen that correlates to C, while H demonstrates the fused mesenchymal tissuesas depicted in D. G: a higher magnification of the boxed area in F, showing the fusion of the endocardial layers ofthe two major AV cushions. “cap,” mesenchymal cap on leading edge of IAS; endo, endocardium; epi, epicardium;iAVC, inferior AV cushion; llAVC, left lateral AV cushion; OFT, outflow tract; sAVC, superior AV cushion; myo,myocardium; rlAVC, right lateral AV cushion. See legends of previous figures for other abbreviations.



and a lower (ventricular) boundary. The bulk of thecushion tissue is plastered against the AV junctionaland ventricular myocardium (Fig. 6, A, E, and F). Thefirst myocardial remodeling event leads to the inter-ruption of the continuity between atrial and myocar-dial ventricular myocardium (Fig. 6, B and C) and is aresult of the fusion between cushion-derived tissuesand epicardially derived mesenchyme (38). Specifi-cally, this process takes place at the lower end of theAV canal myocardium resulting in the incorporation ofmost of the AV canal myocardium into the lower rim ofthe atrial chambers (Fig. 6C). This process establishesthe functional separation, and electrical insulation,between atrial and ventricular working myocardium

(for a detailed description of this process see Refs. 18,38, and 41). There is only one place where this separa-tion does not take place, and that is where the special-ized myocardium of the AV conduction axis (AV nodeand proximal His bundle) develops (38). These compo-nents of the AV conduction system (AVCS) relay thecardiac impulse from atria to ventricles. The secondmyocardial remodeling event that plays an importantrole in the formation of the freely moveable leaflets ofthe AV valves is myocardial delamination (17, 21).During this process the lower part of the AV cushions,with associated myocardial layer (Fig. 6, B, C, G, andH), separates from adjacent ventricular myocardiumby a hitherto undetermined process. This results in the

Fig. 5. Contributions of the major atrioventricular cush-ions to valvuloseptal morphogenesis in the human heart.These panels show two serial sections of a human heart at�6 wk of development stained for the presence of atrialMHC (A) and ventricular MHC (B) as described before(42). A: this is how the mesenchymal tissues of the twomajor cushions and the mesenchymal cap on the IAS havefused to form the mesenchymal “crux” of the heart. B: thisindicates how the structures indicated in A contribute tovalvuloseptal morphogenesis in the AV junction (AVJ).epi, epicardially derived mesenchyme. See legends of pre-vious figures for other abbreviations.

Fig. 6. Atrioventricular valve development in the embryonic human heart. A–D: schematic representations of AVvalves in the developing human heart. E–N: a series of immunohistochemically stained sections of representativestages in this process. Sections shown in E, J, K, L, and N were stained with the “cushion-tissue antigen” antibody249-9G9 (anti-CTA; Ref. 38), the sections in F, G, and I were stained with anti-ventricular MHC (43), the sectionin H (sister section to G) with anti-mesenchymal cell antigen (anti-MCA; Ref. 38), and the section in M (sistersection to N) was stained with an antibody that recognizes the atrial as well as the ventricular isoform of MHC (38).The sections shown in E and F are from a heart at �5 wk of development (38). G and H: sister sections from a9-wk-old heart. I–K: sister sections from a heart at 12–13 wk of development. L–N: sister sections from a heart at17 wk of development. The panels demonstrate (represented schematically in A–D) how during development theleaflets of the valves delaminate and detach from the myocardial wall. Initially, the leaflets contain a considerableamount of myocardium at their ventricular aspect (best seen in G), but as development progresses this myocardiumgradually disappears (cf. G, I, and M). AV myo, atrioventricular myocardium. See legends of previous figures forother abbreviations.



formation of “immature” or “prevalvar” bilayered leaf-lets, with the inferior (or “ventricular”) layer consistingof myocardium and the superior (“atrial”) layer beingcushion derived (Fig. 6, C and G). Toward the ventric-ular apex, these leaflets are attached to the ventricularwalls and/or IVS through the developing papillarymuscles (Fig. 6, G and H). The myocardium graduallydisappears from the developing leaflets between 10and 17 wk of development (Fig. 6, I–N). Interestingly,in the right AVJ of the chick heart, the initial develop-ment of the lateral AV leaflet resembles that of themouse and human. However, instead of the regressionof muscular tissue, leading to the formation of a fibrousleaflet, the mesenchymal tissue gets replaced by myo-cardium, resulting in the formation of the characteris-tic muscular flap valve in the right AVJ of the chick.Immunohistochemical studies indicate that the mech-anism that is responsible for this process is myocardi-alization, i.e., the ingrowth of existing myocardiuminto flanking mesenchymal tissues. Experimentalstudies have revealed that this myocardialization pro-cess in the avian right AV valve can be disturbed by

surgical left atrial ligation (25). This procedure leads tothe formation of a (albeit dysmorphic) fibrous leaflet,reminiscent of the situation in mouse and human.In-depth studies of the molecular mechanisms involvedin the molecular regulation of this myocardializationprocess might provide important clues regarding val-vuloseptal morphogenesis in general, as myocardial-to-fibrous and fibrous-to-myocardial “transformations”are common steps in normal and perturbed cardiovas-cular development in the chick, mouse, and humanheart (13, 29, 30).

Myocardial Compaction

Ventricular trabeculation starts to develop in theapical region of the ventricles soon after looping in bothmouse and human. The trabeculation serves primarilyas a means to increase myocardial oxygenation in ab-sence of coronary circulation. At its peak prior to com-pletion of ventricular septation, the trabeculae canform as much as 80% of the myocardial mass in theembryonic human heart (3) and provide most of the

Fig. 7. Development of the compactventricular myocardium in mouse de-velopment in health and disease. Com-paction of the ventricular trabeculaeadds to proportion and thickness of thecompact myocardium. In the human,this takes place between 12 wk (A) and18 wk (B). Sagittally dissected parietalhalves of left ventricles are shown.Scale bars 100 �m. (Images are fromthe Pexieder collection: B first pub-lished in Ref. 26, and A appeared inRef. 31). Compaction occurs in themouse between ED13 and 14, resultingin well-formed multilayered compactlayer by ED14.5 in the wild-type ani-mal (black arrows in C and white ar-row in E). This process is perturbed inseveral knockouts, such RXR-� (blackarrows in D and white arrow in F),resulting in “thin compact myocardiumsyndrome” and embryonic lethality.



ventricular wall thickness between ED11 and ED16 inthe mouse. Concomitant with ventricular septation,the trabeculae start to compact at their base adjacentto the outer compact myocardium, adding substan-tially to its thickness (reviewed in Ref. 26). This pro-cess is fairly rapid (in mouse, between ED13 and ED14,in human between 10 and 12 wk; Ref. 26), coincidingwith establishment of the coronary circulation. FromED16 in the mouse and the 4th mo of gestation inhuman, the compact layer forms most of the ventricu-lar myocardium, although its structural complexitycontinues to develop (12). Noncompaction of the myo-cardium presents serious functional consequences forthe heart. Several mouse null mutants present with acomplete lack of compaction [e.g., RXR-� knockoutmouse (8); see Fig. 7], which results in lethality aroundED14 (reviewed in Ref. 26), emphasizing the impor-tance of the compact myocardium for force generationin later fetal stages. In humans, ventricular noncom-paction is usually localized. Although it can occur inisolation, it was found in association with heart failureand sudden cardiac death (32). Even if only a relativelysmall proportion of ventricular wall is affected, ventric-ular dynamics is perturbed (11). This condition cannow be diagnosed by ultrasound and is classified as adistinct cardiomyopathy. There is a growing body ofevidence that the epicardium and the epicardially de-rived cells play a crucial role in the development of thecompact layer of the ventricular myocardial walls. Notonly is the “thin-myocardium syndrome” observed inseveral mouse models with perturbed epicardial devel-opment [e.g., �4-integrin (45) and VCAM-1 (15) knock-out mice], recent experimental studies in which epicar-dial development was perturbed in avian embryos alsodemonstrated inhibition of ventricular compaction(22). The mechanism(s) that are responsible for theregulation of ventricular myocardial differentiation,growth, and compaction are at present poorly under-stood, but it is to be expected that further analysis ofmutant mice with “thin compact myocardium,” such asthe RXR-� knockout mouse (5, 14), can provide clues onthe cascade of genetically regulated events that lead totrabecular compaction and hence molecular etiology ofthis condition in humans.

CONCLUSIONS

In this review we have demonstrated that, given theconsiderable differences in cardiac size and heart rate,the cardiac anatomy in mouse and human is, with theexception of some small variations (mainly in the ve-nous pole), remarkably similar. We have also shownthat, at least within the context of the topics discussed,the developmental events that lead to the formation ofthe four-chambered heart are also very comparable.Consequently, the mouse can serve as a good modelwith which to study the human heart. The geneticmodification of the murine genome is resulting in agrowing numbers of mouse models presenting withcardiac malformations that are very reminiscent of theheart anomalies seen in patients with congenital heart

disease. We conclude that mouse molecular geneticscombined with detailed morphological assessment ofgenerated cardiac malformations in the geneticallymodified mouse models will lead to important break-throughs in the identification of a series of (candidate)genes, and the elucidation of pathways, involved in theetiology of human congenital heart disease.

ACKNOWLEDGMENTS

We thank Dr. Tom Trusk for preparing the informative cartoonsin Figs. 3 and 4, and we thank Dr. Steve Kubalak for providing theillustrations of the RXR-� knockout mouse (Fig. 7).

GRANTS

A. Wessels and D. Sedmera are financially supported by NationalInstitutes of Health Grants PO1-HL-52813 (to A. Wessels), P20-RR-16434 (to D. Sedmera), and PO1-HD-39946 (to A. Wessels and D.Sedmera).

REFERENCES

1. Anderson RH and Becker AE. Cardiac Anatomy: An Inte-grated Text and Colour Atlas. London: Gower, 1980.

2. Anderson RH, Webb S, and Brown NA. The mouse withtrisomy 16 as a model of human hearts with common atrioven-tricular junction. Cardiovasc Res 39: 155–164, 1998.

3. Blausen BE, Johannes RS, and Hutchins GM. Computer-based reconstructions of the cardiac ventricles of human em-bryos. Am J Cardiovasc Pathol 3: 37–43, 1990.

4. Camenisch TD, Spicer AP, Brehm-Gibson T, BiesterfeldtJ, Augustine ML, Calabro A Jr, Kubalak S, Klewer SE, andMcDonald JA. Disruption of hyaluronan synthase-2 abrogatesnormal cardiac morphogenesis and hyaluronan-mediated trans-formation of epithelium to mesenchyme. J Clin Invest 106: 349–360, 2000.

5. Chen TH, Chang TC, Kang JO, Choudhary B, Makita T,Tran CM, Burch JB, Eid H, and Sucov HM. Epicardialinduction of fetal cardiomyocyte proliferation via a retinoic acid-inducible trophic factor. Dev Biol 250: 198–207, 2002.

6. De Haan RL. Morphogenesis of the Vertebrate Heart. New York:Holt, Rinehart and Winston, 1965.

7. Eisenberg L and Markwald R. Molecular regulation of atrio-ventricular valvuloseptal morphogenesis. Circ Res 77: 1–6, 1995.

8. Gruber P, Kubalak S, Pexieder T, Sucov H, Evans R, andChien K. RXR alpha deficiency confers genetic susceptibility foraortic sac, conotruncal, atrioventricular cushion, and ventricularmuscle defects in mice. J Clin Invest 98: 1332–1343, 1996.

9. Han W, Wang Z, and Nattel S. Slow delayed rectifier currentand repolarization in canine cardiac Purkinje cells. Am J PhysiolHeart Circ Physiol 280: H1075–H1080, 2001.

10. Huang GY, Cooper ES, Waldo K, Kirby ML, Gilula NB, andLo CW. Gap junction-mediated cell-cell communication modu-lates mouse neural. J Cell Biol 143: 1725–1734, 1998.

11. Jenni R, Oechslin E, Schneider J, Jost CA, and KaufmannPA. Echocardiographic and patho-anatomical characteristics ofisolated left ventricular non-compaction: a step towards classifi-cation as a distinct cardiomyopathy. Heart 86: 666–671, 2001.

12. Jouk PS, Usson Y, Michalowicz G, and Grossi L. Three-dimensional cartography of the pattern of the myofibres in thesecond trimester fetal human heart. Anat Embryol (Berl) 202:103–118, 2000.

13. Kruithof BPT, Van den Hoff MJB, Wessels A, and Moor-man AFM. Cardiac muscle cell formation after development ofthe linear heart tube. Dev Dyn 227: 1–13, 2003.

14. Kubalak SW, Hutson DR, Scott KK, and Shannon RA.Elevated transforming growth factor beta2 enhances apoptosisand contributes to abnormal outflow tract and aortic sac devel-opment in retinoic X receptor alpha knockout embryos. Develop-ment 129: 733–746, 2002.

15. Kwee L, Baldwin HS, Shen HM, Stewart CL, Buck C, BuckCA, and Labow MA. Defective development of the embryonic



and extraembryonic circulatory systems in vascular cell adhe-sion molecule (VCAM-1) deficient mice. Development 121: 489–503, 1995.

16. Lakkis ML and Epstein JA. Neurofibromin modulation of rasactivity is required for normal endocardial-mesenchymal trans-formation in the developing heart. Dev Biol 125: 4359–4367,1998.

17. Lamers WH, Viragh S, Wessels A, Moorman AF, andAnderson RH. Formation of the tricuspid valve in the humanheart. Circulation 91: 111–121, 1995.

18. Lamers WH, Wessels A, Verbeek FJ, Moorman AF, ViraghS, Wenink AC, Gittenberger-de Groot AC, and AndersonRH. New findings concerning ventricular septation in the hu-man heart. Implications for maldevelopment. Circulation 86:1194–1205, 1992.

19. Markwald RR, Fitzharris TP, and Manasek FJ. Structuraldevelopment of endocardial cushions. Am J Anat 148: 85–119,1977.

20. Oosthoek PW, Viragh S, Mayen AE, van Kempen MJ, Lam-ers WH, and Moorman AF. Immunohistochemical delineationof the conduction system. I. The sinoatrial node. Circ Res 73:473–481, 1993.

21. Oosthoek PW, Wenink AC, Vrolijk BC, Wisse LJ, DeRuiterMC, Poelmann RE, and Gittenberger-de Groot AC. Devel-opment of the atrioventricular valve tension apparatus in thehuman heart. Anat Embryol (Berl) 198: 317–329, 1998.

22. Perez-Pomares JM, Phelps A, Sedmerova M, Carmona R,Gonzalez-Iriarte M, Munoz-Chapuli R, and Wessels A. Ex-perimental studies on the spatiotemporal expression of WT1 andRALDH2 in the embryonic avian heart: a model for the regula-tion of myocardial and valvuloseptal development by epicardi-ally derived cells (EPDCs). Dev Biol 247: 307–326, 2002.

23. Qayyum SR, Webb S, Anderson RH, Verbeek FJ, BrownNA, and Richardson MK. Septation and valvar formation inthe outflow tract of the embryonic chick heart. Anat Rec 264:273–283, 2001.

24. Redkar A, Montgomery M, and Litvin J. Fate map of earlyavian cardiac progenitor cells. Development 128: 2269–2279,2001.

25. Sedmera D, Pexieder T, Rychterova V, Hu N, and ClarkEB. Remodeling of chick embryonic ventricular myoarchitectureunder experimentally changed loading conditions. Anat Rec 254:238–252, 1999.

26. Sedmera D, Pexieder T, Vuillemin M, Thompson RP, andAnderson RH. Developmental patterning of the myocardium.Anat Rec 258: 319–337, 2000.

27. Seo JW, Kim EK, Brown NA, and Wessels A. Section directedcryosectioning of specimens for scanning electron microscopy: anew method to study cardiac development. Microsc Res Tech 30:491–495, 1995.

28. Sissman NJ. Developmental landmarks in cardiac morphogen-esis: comparative chronology. Am J Cardiol 25: 141–148, 1970.

29. van den Hoff MJ, Kruithof BP, Moorman AF, MarkwaldRR, and Wessels A. Formation of myocardium after the initialdevelopment of the linear heart tube. Dev Biol 240: 61–76, 2001.

30. van den Hoff MJ, Moorman AF, Ruijter JM, Lamers WH,Bennington RW, Markwald RR, and Wessels A. Myocardi-alization of the cardiac outflow tract. Dev Biol 212: 477–490,1999.

31. Varnava AM. Isolated left ventricular non-compaction: a dis-tinct cardiomyopathy? Heart 86: 599–600, 2001.

32. Waller BF, Smith ER, Blackbourne BD, Arce FP, SarkarNN, and Roberts WC. Congenital hypoplasia of portions of bothright and left ventricular myocardial walls. Clinical and nec-ropsy observations in two patients with parchment heart syn-drome. Am J Cardiol 46: 885–891, 1980.

33. Waller BR, McQuinn T, Phelps AL, Markwald RR, Lo CW,Thompson RP, and Wessels A. Conotruncal anomalies in thetrisomy 16 mouse: an immunohistochemical analysis with em-phasis on the involvement of the neural crest. Anat Rec 260:279–293, 2000.

34. Waller RB III and Wessels A. Cardiac morphogenesis anddysmorphogenesis II. An immunohistochemical approach. In:Developmental Biology Protocols, edited by Lo CW. Totowa, NJ:Humana, 2000, p. 151–161.

35. Webb S, Brown NA, and Anderson RH. The structure of themouse heart in late fetal stages. Anat Embryol (Berl) 194: 37–47,1996.

36. Webb S, Brown NA, Wessels A, and Anderson RH. Develop-ment of the murine pulmonary vein and its relationship to theembryonic venous sinus. Anat Rec 250: 325–334, 1998.

37. Wessels A, Anderson RH, Markwald RR, Webb S, BrownNA, Viragh S, Moorman AF, and Lamers WH. Atrial devel-opment in the human heart: an immunohistochemical studywith emphasis on the role of mesenchymal tissues. Anat Rec 259:288–300, 2000.

38. Wessels A, Markman MW, Vermeulen JL, Anderson RH,Moorman AF, and Lamers WH. The development of the atrio-ventricular junction in the human heart. Circ Res 78: 110–117,1996.

39. Wessels A and Markwald R. Cardiac morphogenesis and dys-morphogenesis. I. Normal development. Methods Mol Biol 136:239–259, 2000.

40. Wessels A, Mijnders TA, de Gier-de Vries C, VermeulenJLM, Viragh S, Lamers WH, and Moorman AFM. Expres-sion of myosin heavy chain in neonatal human hearts. CardiolYoung 2: 318–334, 1992.

41. Wessels A, Vermeulen JL, Verbeek FJ, Viragh S, KalmanF, Lamers WH, and Moorman AF. Spatial distribution of“tissue-specific” antigens in the developing human heart andskeletal muscle. III. An immunohistochemical analysis of thedistribution of the neural tissue antigen G1N2 in the embryonicheart; implications for the development of the atrioventricularconduction system. Anat Rec 232: 97–111, 1992.

42. Wessels A, Vermeulen JL, Viragh S, Kalman F, LamersWH, and Moorman AF. Spatial distribution of “tissue-specific”antigens in the developing human heart and skeletal muscle. II.An immunohistochemical analysis of myosin heavy chain iso-form expression patterns in the embryonic heart. Anat Rec 229:355–368, 1991.

43. Wessels A, Vermeulen JL, Viragh S, Kalman F, Morris GE,Man NT, Lamers WH, and Moorman AF. Spatial distributionof “tissue-specific” antigens in the developing human heart andskeletal muscle. I. An immunohistochemical analysis of creatinekinase isoenzyme expression patterns. Anat Rec 228: 163–176,1990.

44. Ya J, Schilham MW, de Boer PA, Moorman AF, Clevers H,and Lamers WH. Sox-4 deficiency syndrome in mice is ananimal model for common trunk. Circ Res 83: 986–994, 1998.

45. Yang JT, Rayburn H, and Hynes RO. Cell adhesion eventsmediated by alpha 4 integrins are essential in placental andcardiac development. Development 121: 549–560, 1995.



Developmental anatomy of the heart: a tale of mice … · Developmental anatomy of the heart: a...

Documents

Transcript of Developmental anatomy of the heart: a tale of mice … · Developmental anatomy of the heart: a...