THE ANATOMICAL RECORD 297:1414–1429 (2014)

The Development of Septation in the

Four-Chambered Heart
ROBERT H. ANDERSON,1* DIANE E. SPICER,2,3 NIGEL A. BROWN,4
5
AND TIMOTHY J. MOHUN
1
Institute of Genetic Medicine, Newcastle University, Newcastle, United Kingdom
2
Department of Pediatric Cardiology, University of Florida, Gainesville, Florida
3
Congenital Heart Institute of Florida, St. Petersburg, Florida
4
Division of Biomedical Sciences, St. George’s University of London, London, United
Kingdom
5
Division of Developmental Biology, MRC National Institute for Medical Research, London,
United Kingdom

ABSTRACT
The past decades have seen immense progress in the understand-
ing of cardiac development. Appreciation of precise details of cardiac
anatomy, however, has yet to be fully translated into the more general
understanding of the changing structure of the developing heart, par-
ticularly with regard to formation of the septal structures. In this
review, using images obtained with episcopic microscopy together with
scanning electron microscopy, we show that the newly acquired infor-
mation concerning the anatomic changes occurring during separation
of the cardiac chambers in the mouse is able to provide a basis for
understanding the morphogenesis of septal defects in the human
heart. It is now established that as part of the changes seen when the
heart tube changes from a short linear structure to the looped
arrangement presaging formation of the ventricles, new material is
added at both its venous and arterial poles. The details of these early
changes, however, are beyond the scope of our current review. It is
during E10.5 in the mouse that the first anatomic features of septa-
tion are seen, with formation of the primary atrial septum. This mus-
cular structure grows toward the cushions formed within the
atrioventricular canal, carrying on its leading edge a mesenchymal
cap. Its cranial attachment breaks down to form the secondary fora-
men by the time the mesenchymal cap has used with the atrioventric-
ular endocardial cushions, the latter fusion obliterating the primary
foramen. Then the cap, along with a mesenchymal protrusion that
grows from the mediastinal mesenchyme, muscularizes to form the
base of the definitive atrial septum, the primary septum itself forming
the floor of the oval foramen. The cranial margin of the foramen is a
fold between the attachments of the pulmonary veins to the left
atrium and the roof of the right atrium. The apical muscular ventricu-
lar septum develops concomitant with the ballooning of the apical
components from the inlet and outlet of the ventricular loop. Its apical
part is initially trabeculated. The membranous part of the septum is
derived from the rightward margins of the atrioventricular cushions,
with the muscularizing proximal outflow cushions fusing with the
muscular septum and becoming the subpulmonary infundibulum as
the aorta is committed to the left ventricle. Perturbations of these

*Correspondence to: Robert H. Anderson, Institute of Genetic Revised 13 July 2013; Accepted 30 August 2013.
Medicine, Newcastle University, Newcastle, United Kingdom. DOI 10.1002/ar.22949
Fax: 44-20-8870-4368. E-mail: sejjran@ucl.ac.uk Published online 27 May 2014 in Wiley Online Library
(wileyonlinelibrary.com).

C 2014 WILEY PERIODICALS, INC.

V
 DEVELOPMENT OF CARDIAC SEPTATION
processes explain well the phenotypic variants of deficient atrial and
ventricular septation. Anat Rec, 297:1414–1429, 2014. V C 2014 Wiley
Periodicals, Inc.
Key words: atrial septum; interatrial communications; atrioven-
tricular septum; atrioventricular septal defects; ven-
tricular septum; ventricular septal defects
INTRODUCTION
The past decades have seen immense progress in the
understanding of cardiac development. Appreciation of
precise details of cardiac anatomy, however, has yet to
be fully translated into the more general understanding
of the changing structure of the developing heart. For
example, many, if not most, current textbooks of embry-
ology have still to provide an accurate account of the
development of the atrial septum (Cole-Jeffrey et al.,
2012). Uncertainty remains with regard to the key
processes underscoring postnatal persistence of a com-
mon atrioventricular junction, as opposed to the normal
formation of separate right and left atrioventricular
junctions (Briggs et al., 2012). It has to be still recog-
nized that, in the normal heart, although there is an
extensive embryonic outflow septum, barely any of the
septal structures persist between the definitive ventric-
ular ouflow tracts (Webb et al., 2003). Some of the rea-
sons for this situation reflect the difficulties in
appreciating the four-dimensional changes that take
place during cardiac development. It is hard enough to
understand the three-dimensional changes that occur
during formation of the septal structures without
access to complex and time-consuming reconstructions
(Soufan et al., 2003). The added problem of temporal
morphological changes during the process of develop-
ment compounds the situation. With the exception of
the temporal aspect, all of this has potentially been
solved with the advent of episcopic microscopy. This
technique now permits rapid analysis, in high resolu-
tion and in three dimensions, of the morphologic
changes occurring during embryological development
(Mohun and Weninger, 2011). In addition, the ability to
analyze large series of developing embryos, as is now
possible when using experimental models such as the
mouse, provides sufficient material to ensure the accu-
racy and uniformity of the changes observed, thus over-
coming some of the temporal problems. In this review,
we show that the newly acquired information concern-
ing the anatomic changes occurring during separation
of the cardiac chambers in the mouse is able to provide
a basis for understanding the morphogenesis of septal
defects in the human heart.
MATERIALS AND METHODS
We had previously used scanning electron microscopy
to demonstrate the anatomic changes seen during car-
diac development in the mouse (Webb et al., 1996). We
have revisited the images collected to show the changing
morphologic features of the normal heart, comparing
these additionally to the changes induced by knocking
1415
out the gene Pitx2c (Martin et al., 2010). We have then
extended the initial findings obtained using scanning
electron microscopy by analyzing, using high-resolution
episcopic microscopy (Mohun and Weninger, 2011), a
large series of mouse embryos obtained from day E9.5 to
term.
Initial Development of the Cardiac Chambers
It is now established that, as part of the changes
seen when the heart tube changes from a short linear
structure to the looped arrangement presaging forma-
tion of the ventricles (Fig. 1), new material is added at
both its venous and arterial poles (Rana et al., 2007).
The details of these early changes are beyond the scope
of our current review. Acceptance that the linear tube
does not contain all the component parts of the defini-
tive heart, nonetheless, has revolutionized our under-
standing of cardiac development (Moorman et al.,
2007). By E10.5 in the mouse, the stage at which the
first anatomic features of septation are seen, the sys-
temic venous tributaries have already been remodeled
from their initial state of symmetry (Fig. 2) so that
they open exclusively to the right side of the still com-
mon atrial chamber (Fig. 3).
By E10.5, the atrial orifice of the pulmonary vein
has become visible within the left side of the common
atrium (Fig. 3B). The vein opens to the atrial chamber
within the pulmonary pit, this being the area initially
bounded by the reflections of the dorsal mesocardium.
The origin of the pulmonary vein has been the topic of
discussion for more than a century. In a recent exhaus-
tive review, combined with elegant experimental tech-
niques, Jahr and Manner (2011) have shown
conclusively that the pulmonary vein connects directly
with the left side of the developing atrial chamber and
has never had any connection within the confines of
the systemic venous sinus. The experiments of Jahr
and Manner were carried out in the amphibian Xeno-
pus; however, their findings endorse the recent
accounts of pulmonary venous development provided
for man, mouse, and chicken (Webb et al., 1998a,b,
2000, 2001). The lateralization of the systemic venous
return (Figs. 2 and 3), such that the venoatrial junc-
tion is committed exclusively to the right side of the
developing common atrial chamber, sets the scene for
septation of the cardiac chambers. The first sign of this
septation is the appearance in the roof of the common
atrial chamber, during E10.5, of the primary atrial sep-
tum (Fig. 3).
By the time the primary atrial septum has become
visible in the roof of the common atrial chamber,
endocardial cushions have also formed superiorly and
1416 ANDERSON ET AL.
Fig. 1. The scanning electron micrographs show the changes in
structure of the developing heart during E9. The heart has been
revealed by removing the ventral wall of the pericardial cavity. Initially,
as shown in Panel A, the tube retains its initial linear structure,
although already material is being added to the tube at both the arte-
rial and venous poles. As shown in Panel B, the continuing addition of
inferiorly within the atrioventricular canal, providing
the basis for septation of this part of the developing
heart. The canal itself at this stage has significant
length, interposing between the walls of the common
atrial chamber and the proximal walls of the newly
formed ventricular loop (Fig. 4A). Initially, the opening
between the atrioventricular cushions is positioned so as
to open exclusively into the cavity of the developing left
ventricle. The apical component of the left ventricle will
form from the proximal, or inlet, component of the loop
(Fig. 4B). From the outset of formation of the ventricular
loop, nonetheless, the right wall of the canal is in direct
continuity with the cranial wall of the distal, or outlet,
component of the heart tube, from which will grow
the apical trabecular component of the right ventricle
(Fig. 4A).
Lateralization of the Chambers
By E10.5, it is already possible to recognize the com-
ponents which, in the postnatal heart, will permit dis-
tinction of the morphologically right as opposed to the
morphologically left chambers. For the atrial chambers,
the key components are the appendages, which from the
outset have discrete and different morphologies. For the
ventricles, it is the apical components that, in man, best
permit distinction of the morphologically left and right
structures when the heart is congenitally malformed, as
they, too, eventually have discrete trabecular patterns
(Anderson et al., 2012). These distinctive components of
new material is associated with lengthening of the tube, with the ven-
tricular component looping as the primordium of the right ventricle
becomes evident within the pericardial cavity, supporting the develop-
ing outflow tract. The initial linear component of the tube eventually
forms little more than the apical part of the left ventricle.
the developing chambers are formed by the process that
has become known as ballooning (Moorman and Chris-
toffels, 2003). The location at which these components
balloon from the initial primary heart tube provides a
good explanation for the abnormal changes seen in the
syndromes described as visceral heterotaxy (Cohen
et al., 2007). The essence of these syndromes is the bilat-
eral presence of the anatomic features that, in the nor-
mally developed individual, are lateralized. Thus, in the
normal individual, the right lung has three lobes and is
supplied by a short bronchus, whereas the left lung has
two lobes and a long bronchus. In the subsets of visceral
heterotaxy, individuals have either trilobed lungs sup-
plied by a short bronchus bilaterally or bilobed lungs
with a long bronchus bilaterally. In other words, the tho-
racic organs are arranged in isomeric rather than later-
alized fashion. Within the heart, however, at least in
humans, it is only the appendages of the atrial chambers
that are truly isomeric (Uemura et al., 1995a). As can be
seen in Fig. 4, the primordiums of the appendages bal-
loon in parallel from the atrial component of the primary
heart tube. As the atrial component itself is a midline
structure, the right-sided appendage grows under the
influence of the genes responsible for morphological
right-sidedness. The left-sided appendage, in contrast,
develops under the influence of genes such as Pitx2, and
Cited2, which produce morphologically leftness (Mom-
mersteeg et al., 2007; Martin et al., 2010). When Pitx2c
is knocked out in the mouse, both appendages are
formed with morphologically right identity, in other
 DEVELOPMENT OF CARDIAC SEPTATION
Fig. 2. The scanning electron micrograph, prepared using a devel-
oping mouse heart at embryonic day 9.5, shows evidence of the initial
symmetry of the systemic venoatrial connections, which drain to the
common atrial chamber through the right and left sinus horns,
although already the left horn has become smaller than the right. The
section is taken through the dorsal mesocardium and shows the pul-
monary pit (arrow) between the reflections of the myocardial walls into
the mediastinal mesenchymal tissues. At this stage, there are no dis-
crete structures marking the boundaries between the venous connec-
tions and the atrial cavity (compare with Fig. 3).
words, the atrial component of the heart exhibits right
isomerism (Fig. 5; Martin et al., 2010). The apical ven-
tricular components, rather than taking origin from the
same component of the heart tube, develop in series
from the primary linear heart tube, ballooning from the
inlet and outlet components of the ventricular loop,
respectively (Fig. 4A). Subsequent to looping, although
the ventricular part of the initial tube is no longer a
midline structure, its entire left side is developed under
the influence of Pitx2c. Each of its components, there-
fore, develops under the control of the same genes that
determine morphologically rightness and leftness. Nei-
ther of the developing ventricles shows evidence of iso-
merism subsequent to knocking out genes such as
Pitx2c. It is often stated that the initial evidence of loss
of symmetry is provided by ventricular looping.
Although this may be true, the arrangement of the ven-
tricular loop does not prove to be a discriminating fea-
ture in patients with visceral heterotaxy (Cohen et al.,
2007), and Pitx2c does not control the process of looping
(Martin et al., 2010). It behooves those seeking the cues
to isomeric as opposed to lateralized development within
the heart, therefore, to direct their attention to the atrial
chambers, ideally the appendages, rather than the ven-
tricular loop. When seeking markers of morphologically
leftness, it should be remembered that the coronary
sinus, derived from the left sinus horn, is one such
structure. In patients with right isomerism, the coronary
sinus is universally absent (Uemura et al., 1995b). If
caval veins develop bilaterally in the setting of right iso-
1417
merism, then the left-sided vein always drains directly
to the roof of the left-sided atrium and is never incorpo-
rated within the left-sided atrioventricular junction, as
is the case in normal development (Fig. 6). Hence, as the
presence of morphologically right appendages is bilat-
eral, the absence of a coronary sinus in the left-sided
atrioventricular groove is an excellent anatomic marker
of right isomerism.
Septation of the Atrial Chambers
We have already shown that the first sign of separa-
tion of right and left halves of the common atrium is the
appearance in the atrial roof, during E10.5, of a distinct
ridge that is the beginnings of the primary atrial septum
(Fig. 3). During E11.5, this muscular septum extends
through the cavity of the initially common atrial cham-
ber, growing toward the cushions already present within
the atrioventricular canal. As it extends toward the atri-
oventricular junction, it carries a mesenchymal cap on
its leading edge (Fig. 7A). In addition to the mesenchy-
mal cap, a second mesenchymal protrusion contributes
to the eventual formation of the basal part of the devel-
oping atrial septum. This structure, directly continuous
dorsally with the mediastinal mesenchyme, protrudes
into the atrial cavity at the caudal end of commissure
between the valves of the systemic venous sinus (Fig.
7A). The entity was initially observed by His in the nine-
teenth century (His, 1885) and described by him as the
“spina vestibuli.” We describe it as the vestibular spine.
Normal growth of the spine is essential for separate for-
mation of the right and left atrioventricular junctions
(Webb et al., 1997). When development proceeds nor-
mally, it grows so as to reinforce the right side of the
area over which the mesenchymal cap on the primary
atrial septum fuses with the atrial surface of atrioven-
tricular endocardial cushions (Fig. 7B). It is fusion of the
mesenchymal tissues with the atrioventricular endocar-
dial cushions that obliterates the primary atrial fora-
men. This process serves to anchor the primary atrial
septum to the atrioventricular cushions. The cushions
themselves, by this stage, have fused to separate the
right and left atrioventricular junctions. At the same
time, expansion of the rightward wall of the atrioventric-
ular canal has brought the cavity of the right atrium
into direct continuity with that of the developing right
ventricle (Fig. 8). Moreover, even before the primary
atrial foramen has closed, a linear communication
between the atriums has appeared due to breakdown of
the cranial attachment of the primary septum to the
atrial roof. It is this space that becomes the secondary
atrial foramen (Fig. 7A). With closure of the primary
foramen, it is the secondary foramen that provides the
essential interatrial communication that is part of the
fetal circulation. It might be thought that the primary
atrial foramen, or “ostium primum,” was an atrial septal
defect. This is not strictly the case, as the foramen is
bounded on the one side by the leading edge of the
developing atrial septum, and on the other side by the
atrial surfaces of the atrioventricular endocardial cush-
ions. As it is not within the developing atrial septum, it
cannot be an “atrial septal defect”. It is an atrioventricu-
lar septal defect, as it provided, prior to fusion of the
atrioventricular cushions, a potential communication
between the cavities of the atrial and ventricular
 Fig. 3. The panels show a frontal section through an episcopic data-
set (A) and a scanning electron micrograph illustrating the atrial cavity
as seen from the amputated cardiac apex (B). Both are prepared from
mouse embryos from early day 10.5. The episcopic section (A) shows
how the systemic venous tributaries now drain exclusively to the right
side of the common atrial chamber, with the primordium of the primary
atrial septum seen as a ridge in the atrial roof. The scanning electron
Fig. 4. The sections are from episcopic datasets of developing
mouse embryos at embryonic day 11.5. The left-hand panel (A) shows
a four-chamber section through the atrioventricular canal (double-
headed black arrows). By this stage of development, the atrial appen-
dages are beginning to balloon in parallel fashion from the common
atrial chamber (upper white arrows), whereas the apical components
of the developing ventricles are ballooning in series from the ventricu-
lar loop (lower white arrows). As the apical components balloon out,
with the left ventricular apical part growing from the inlet of the loop
and the apical component of the right ventricle from the outlet, so is
the primordium of the muscular ventricular septum formed between
micrograph, orientated so as to focus on the dorsal wall of the com-
mon atrium, reveals the opening of the pulmonary vein, which is adja-
cent to the developing atrioventricular junction. Both images show
how the opening of the systemic venous sinus to the developing right
atrium is now guarded by venous valves. It is the appearance of the
valves that permits definition of the boundary between the atrial wall
and the walls of the systemic venous tributaries.
them (star). The right-hand panel (B) shows a short-axis cut through
the ventricular loop equivalent to the black line shown in Panel A. The
ventricular mass is then viewed from the aspect of the transected api-
cal components. The star again shows the developing ventricular sep-
tum. The opening between the atrioventricular cushions (double-
headed white arrow), the cushions themselves develop so as eventu-
ally to septate the atrioventricular canal, at this stage opens exclu-
sively into the cavity of the developing left ventricle. From the outset,
however, the right side of the atrioventricular canal is in direct continu-
ity with the superior wall of the inlet of the developing right ventricle
(curved arrow in Panel A).
 DEVELOPMENT OF CARDIAC SEPTATION
chambers (Becker and Anderson, 1982). It is the second-
ary foramen, therefore, that represents the first hole to
be found within the developing atrial septum. Subse-
Fig. 5. The scanning electron microscopic image shows the atrial cham-
bers, viewed from the ventricular aspect having cut the heart in its short
axis, from a Pitx2c knockout mouse. There is isomerism of the right atrial
appendages and the sinuatrial junctions. There is also absence of a supe-
rior caval vein in the left-sided atrioventricular junction, failure of formation
of the primary atrial septum, and a midline location for the pulmonary vein.
Fig. 6. The scanning electron microscopic images show views of
the atrial chambers, seen from the aspect of the removed ventricular
chambers, from a developing mouse heart obtained late at E10.5.
Panel A is viewed from beneath, and Panel B is shown obliquely from
the right side. The primary atrial septum is seen in both panels, divid-
ing the common atrial chamber into its morphologically right and left
components. The opening of the pulmonary pit into the left atrium is
seen in Panel A. Panel B shows how the upper part of the primary
1419
quent to formation of this communication, a small mus-
cular ridge can be seen in the atrial roof. It marks the
cranial margin of the space that will become the oval
foramen. The primary atrial septum itself becomes the
flap valve of the foramen, which closes postnatally when
interatrial shunting is no longer a necessary part of the
circulation (Vettukattil et al., 2013). Standard textbooks
of cardiac development usually illustrate the right atrial
margins of the oval foramen as being formed by down-
growth from the atrial roof of a secondary atrial septum.
There is no evidence from study of the developing
mouse, or human embryonic hearts, to support the
notion that such a second septum grows into the atrial
cavities from their roof. The mesenchymal cap, along
with the vestibular spine, nonetheless, become trans-
formed by muscularization to form the ventral and cau-
dal margins of the rim of the oval fossa, which then
buttress the septum to the atrioventricular junctions
(Fig. 9). These structures that form the buttress of the
oval fossa, therefore, can justifiably be considered to rep-
resent a secondary atrial septum. With ongoing develop-
ment, the cranial margin of the fossa continues to be
formed by the muscular ridge that initially marked the
junction between the walls of the developing right and
left atrial chambers. Just prior to birth in the mouse,
however, the dorsal margin of the fossa is formed by a
more pronounced infolding of the atrial walls (Fig. 9).
The left component of this fold is formed by the attach-
ment of the pulmonary vein to the left atrium, whereas
right atrial myocardium forms the right wall between
septum has already broken down to produce the secondary atrial
foramen. The dissections show how the left sinus horn, which in the
mouse persists as the left superior caval vein, has become incorpo-
rated into the developing left atrioventricular junction. It possesses its
own walls discrete from those of the left atrium. The presence of this
venous channel within the atrioventricular groove is an excellent ana-
tomic marker of morphologically leftness.
 Fig. 7. The images are frontal sections obtained from an episcopic
datasets from a developing mice hearts at late E11.5 (A) and early on
E13.5 (B). They show (A) how the muscular primary atrial septum
grows from the atrial roof toward the cushions present in the atrioven-
tricular canal, breaking down at its original origin from the roof to pro-
duce the secondary atrial foramen. The primary atrial foramen, or
“ostium primum,” is the space between the mesenchymal cap carried
on the leading edge of the primary septum and the atrial surface of
the atrioventricular (AV) endocardial cushions. Note that there is also
an obvious protrusion at the caudal extent of the venous valves mark-
ing the boundaries of the systemic venous sinus. This is the vestibular
spine, which is emphasized by the white dotted lines. In addition,
Fig. 8. The images are four-chamber sections taken from episcopic
datasets from developing mice at E10.5 (Panel A) and early on E11.5
(Panel B). Panel A shows that, initially, when the apical ventricular septum
is beginning to form (red star), the right atrioventricular orifice (long white
arrow) opens the cavity of the developing left ventricle. Already, however,
the right wall of the atrioventricular canal is in continuity with the parietal
there is a cranial muscular ridge (arrow) that marks the cranial bound-
ary between the right and left atrial chambers. Panel B shows how, by
E13.5, the mesenchymal cap has fused with the atrioventricular cush-
ions to obliterate the primary atrial foramen. The rightward margin of
the zone of fusion has been reinforced by continuing growth of the
vestibular spine (white dotted lines). The rightward margin of the
developing septum is also now recognizable as the oval fossa (dou-
ble-headed white arrow), with the interatrial muscular ridge forming its
cranial margin (arrow) and the vestibular spine its caudal border. The
more obvious muscular ridge to the right is the attachment of the left
venous valve.
wall of the developing right ventricle (see also Fig. 4A). Early on E11.5
(Panel B), rightward expansion of the atrioventricular canal has permitted
the right atrioventricular orifice to communicate directly with the cavity of
the right ventricle (red arrow), the orifice now opening to the right side of
the muscular ventricular septum (red star). Note the beginning of forma-
tion of the primary atrial septum (red arrow) in Panel A.
 DEVELOPMENT OF CARDIAC SEPTATION 1421

Fig. 10. The image, from a human heart, is prepared to simulate the
Fig. 9. The four-chamber section is from a developing mouse at echocardiographic four-chamber section of the atrial septum. It shows
E18.5. The opening of the oval fossa is again shown by the double- how the cranial margin of the oval fossa (double-headed white arrow),
headed arrow, white in this image. Its dorsal rim is formed by an seen to the left-hand in this figure, is formed by a deep infolding, filled
infolding of the atrial walls, which is continuous cranially with the mus- by extracardiac fat, between the atrial walls. Note the location of the
cular ridge initially seen in the atrial roof. The right wall extends membranous septum, which forms the right wall of the aortic outflow
between the leftward attachment of the systemic venous sinus to the tract from the left ventricle.
right atrium and the apex of the fold, whereas the left wall is formed
by the attachments of the pulmonary veins to the left atrium. Even at
this late stage, the right and left pulmonary veins still drain to the left
atrium through a common venous orifice. The ventral and caudal mar-
gin of the fossa is formed by the muscularized mesenchymal struc-
tures, which provide a buttress than anchors the primary septum to
the atrioventricular junctional tissues. Note the discrete walls of the
left superior caval vein (LSCV), which is the persisting left sinus horn.

the attachment of the left venous valve and the apex of

the fold (Fig. 9). In the mouse, therefore, the caudal
component of the right atrial margin of the oval fossa,
traditionally considered to represent the secondary atrial
septum, or the “septum secundum,” is formed by the
infolding between the right and left atrial walls. This
fold is itself continuous cranially with the muscular
ridge that, from the outset, marks the boundary between
the developing right and left atrial chambers. Ventrocau-
dally, the rim of the fossa is formed by the buttress
derived by muscularization of the tissues derived from
the vestibular spine and the mesenchymal cap of the pri-
mary atrial septum. As yet, we are unable to determine
the precise mechanisms of this muscularization. In the
human heart, in contrast, it is the cranial margin,
rather than the dorsal margin of the fossa, which is
formed by the extensive interatrial fold (Fig. 10). This is Fig. 11. The four-chamber section through a human heart shows
the left atrial opening of a vestibular defect. The communication tracks
because, in man, the pulmonary venous walls, with
through the caudal buttress of the oval fossa (double-headed white
ongoing development, become incorporated into the roof arrow).
of the left atrium. In man, therefore, one pulmonary
vein eventually drains at each corner of the left atrium
(Webb et al., 2001). tered in man exist because of failure of anatomic fusion
Appreciation of these features of development facili- of the flap valve of the oval foramen with its right atrial
tates understanding of the morphogenesis of the holes rims. Such persistent patency exists throughout life in
that, in man, provide the potential for postnatal inter- up to one-third of the normal human population (Hagen
atrial shunting. By far, the commonest defects encoun- et al., 1984). In postnatal life, however, interatrial
1422 ANDERSON ET AL.
shunting occurs only when right atrial pressure exceeds
that in the left atrium. In man, when compared with
mouse, and as shown in Fig. 10, there is a much more
Fig. 12. The image is a frontal section through an episcopic dataset
prepared from a developing mouse early on E12.5. The atrioventricular
(AV) canal has expanded so that the cavity of the developing right atrium
is now in direct continuity with the cavity of the right ventricle. The supe-
rior and inferior cushions in the atrioventricular canal have yet to fuse.
Fig. 13. The images show the ventricular aspect (A) and a four-
chamber section through (B) the atrioventricular junction later on
E12.5 in the developing mouse heart. The major atrioventricular (AV)
endocardial cushions have now fused to divide the junction into right
and left orifices, which are slit like. It is the accompanying fusion of
the vestibular spine with the atrial aspects of the cushions (Panel B)
that subsequently permits the right atrioventricular junction to expand
extensive cranial rim to the fossa, against which the flap
valve closes. In the murine heart, in which the pulmo-
nary veins retain a solitary dorsal opening (Fig. 9), the
oval foramen is closed by adherence of the flap valve
derived from the primary septum with the atrial roof,
with the point of closure extending well to the left of the
cranial margin of the foramen (Cole-Jeffrey et al., 2012).
It is when the flap valve is of deficient size to cover the
margins of the foramen, or else when it is multifenes-
trated, that there are true defects within the oval fossa
(Anderson et al., 1999). Often called as secundum
defects, the holes exist because of deficiencies of the pri-
mary atrial septum. Hence, they are strictly “ostium
secundum” defects. There is then another true atrial
septal defect to be found in man. When present, this
lesion takes the form of a small channel that extends
through the muscular rim of the oval foramen that but-
tresses its attachment to the atrioventricular junctions
(Fig. 11). Well described as the vestibular defect (Shar-
ratt et al., 2003), but not thus far commonly recognized,
the defect is almost certainly due to failure of union of
the muscularizing mesenchymal structures that produce,
in man, the normal ventroinferior rim of the oval
foramen.
The remaining defects that produce the potential for
interatrial shunting are all outside the confines of the
oval fossa and its muscular boundaries, and hence are
not true septal defects (Anderson and Brown, 1996). The
“primum defect,” like the “ostium primum” itself, as we
will describe in our next section, is an atrioventricular
septal defect, but one which permits shunting only at
atrial level. The “sinus venosus defect” is the
caudally, overlapping the crest of the muscular ventricular septum
[white arrow in (A)]. The margins of the atrioventricular cushions have
been emphasized by dotted white lines in Panel B. Note that at this
stage of development, the proximal ends of the outflow cushions
(stars in A) have yet to fuse, although the intermediate parts of the
cushions have already fused centrally to separate the aortic root from
the pulmonary root.
 DEVELOPMENT OF CARDIAC SEPTATION
Fig. 14. The images show a frontal section (A) and a short-axis sec-
tion (B) from episcopic datasets prepared from separate mice at
E13.5. Panel A shows how, subsequent to expansion of the right side
of the atrioventricular canal, the base of the atrial septum, formed by
muscularization of the vestibular spine, is in line with the muscular
ventricular septum. The inferior atrioventricular cushion (white dotted
lines) at this stage is positioned so as to be a septum between the
Fig. 15. The sections in short axis (A) and frontal (B) projections are
from separate datasets from developing mice at E14.5. The aortic root
has been transferred into the left ventricle, and the interventricular
communication closed. Note the separation of the mitral and tricuspid
valvar orifices, with the aortic root wedged between the mitral valve
and the septum (Panel A). The inferior atrioventricular cushion (double-
headed black arrow) now forms the primordium of the membranous
1423
right atrium and the left ventricle, in other words an atrioventricular
septum (double-headed arrow). In addition, note the continuity from
the mediastinal tissues alongside the pulmonary vein to the vestibular
spine (white arrow in Panel A). The short-axis section (Panel B) shows
how the bulk of the fused atrioventricular cushions are within the left
ventricle, forming its roof (star).
atrioventricular septum at the cranial margin of the newly developed
left ventricular outflow tract (Panel B). Note the location of the cranial
rim of the oval foramen (white arrow) relative to the cranial attachment
of the flap valve of the foramen (red arrow), the latter derived from the
primary atrial septum. The septal leaflet of the developing tricuspid
valve has still to begin its delamination from the ventricular septum.
1424 ANDERSON ET AL.
Fig. 16. The images, from postnatal human hearts, replicate the
four-chamber echocardiographic sections showing the separating atri-
oventricular structures. Panel A is through the membranous septum,
which forms the right wall of the left ventricular outflow tract. The
location of the hinge of the septal leaflet of the tricuspid valve sepa-
rates this fibrous part of the septum into interventricular (double-
headed blue arrow) and atrioventricular (red arrow) components. Panel
B is a section taken more dorsally from a different heart. It shows
Fig. 17. The images are frontal sections from episcopic datasets
prepared from developing mice at E13.5 (A) and E14.5 (B). They show
how fusion of the proximal outflow cushions separates the dorsal part
of the outflow tract as the aortic root (Panel A), and then how remod-
eling of the inner heart curvature brings the root from its position
above the developing right ventricle (Panel A) to its definitive location
how, in the limited area of the atrioventricular junctions where the
hinges of the atrioventricular valvar leaflets are offset, a continuation
of the inferior atrioventricular groove, filled with epicardial fat, and
labeled as such, interposes between the atrial and ventricular muscle
masses. The area (double-headed white arrow), while interposing
between the atrial and ventricular cavities, is a sandwich rather than a
muscular septum.
within the left ventricle (Panel B). The arrows show the reorientation of
the dorsal wall. Panel B also shows how the rightward extensions of
the atrioventricular cushions fuse so as to complete ventricular septa-
tion. The cushions will subsequently form the interventricular compo-
nent of the membranous septum.
 DEVELOPMENT OF CARDIAC SEPTATION
Fig. 18. The episcopic sections show the right sides of the develop-
ing right atrium and ventricle. Panel A is from a dataset from E13.5,
and Panel B is from a dataset from E14.5. The images show how the
surface of the proximal outflow cushions, having fused to form the
ventral margin of the embryonic interventricular (IV) communication
(Panel A), then muscularize (Panel B). The central part of the cushion
consequence of anomalous connection of one or more of
the the right pulmonary veins to a caval vein, but with
the pulmonary veins retaining their left atrial connec-
tions (Butts et al., 2011). The hole described as the
“coronary sinus defect” provides interatrial shunting
through the right atrial orifice of the left sinus horn.
The shunting becomes possible when there is either
fenetrations of the walls that normal separate the cav-
ities of the left atrium and the left sinus horn, or else
total absence of the walls. In this setting, the right atrial
orifice of the systemic venous channel then persists as
the site of interatrial shunting, but is again outside the
confines of the normal atrial septum (Knauth et al.,
2002). It is an interatrial communication rather than an
atrial septal defect.
Septation and Separation of the
Atrioventricular Junctions
At the beginning of E12.5, although the atrioventricu-
lar canal has expanded, the cushions within it remain as
separate entities (Fig. 12), with shunting potentially pos-
sible from ventricular to atrium between the cushions
and through the ostium primum atrioventricular septal
defect. By the end of E12.5, the cushions have fused
along their facing margins (Fig. 13A), and the mesenchy-
mal atrial components have fused to their atrial surfa-
ces, thus separating the cavity of the canal into discrete
right and left atrioventricular orifices (Fig. 13B). A key
feature that ensures subsequent separation of the mus-
cular right and left atrioventricular junctions is the
growth of the vestibular spine (Fig. 13B). It is subse-
quent to the anchorage of the spine and the developing
1425
mass is then removed, presumably by a process of apoptosis, so that
the muscularized surface becomes transformed into the free-standing
muscular subpulmonary infundibulum (Panel B). The rightward margins
of the atrioventricular (AV) endocardial cushions fuse to close the right
ventricular entrance to the subaortic vestibule, completing the process
of ventricular septation.
primary atrial septum to the atrial aspect of the atrio-
ventricular cushions (Fig. 13B) that the newly formed
right atrioventricular junction expands ventrally and
caudally. These processes also bring the dorsal end of
the developing muscular ventricular septum in line with
the atrial septum (Fig. 14A). Even after expansion of the
atrioventricular canal, and fusion of the cushions, the
larger part of the fused cushion mass remains within
the left ventricle (Fig. 14B). A plane of separation can
then be seen between the left ventricular components of
the cushions and the developing muscular ventricular
septum. The inferior cushion, which roofs this plane, is
now positioned so as to produce an atrioventricular sep-
tum (Fig. 14A). It is then over the period of E13.5, and
into E14.5, that the aortic root is transferred, by remod-
eling of the part of the ventricles derived from the origi-
nal primary heart tube, into the cranial part of the left
ventricular cavity, thus wedging the aortic vestibule
between the mitral valvar orifice and the muscular ven-
tricular septum (Fig. 15).
Subsequent to the aortic root achieving its definitive
position between the mitral valvar orifice and the ven-
tricular septum, the leaflets of the mitral valve have a
limited area of attachment to the muscular ventricular
septum (Fig. 15). After the developing aortic root has
become committed to the left ventricle, its right-sided
wall is formed by the membranous septum. As we will
describe, this structure itself is derived from the right-
ward margins of the atrioventricular endocardial cush-
ions, which close the embryonic interventricular
communication. As we have already shown, the inferior
cushion itself separates the cavities of the left ventricu-
lar outflow tract and the right atrium. In the postnatal
1426 ANDERSON ET AL.
Fig. 19. The images compare a developing mouse heart at E13.5 as
shown by a long-axis section through the right atrium and ventricle
(upper panel) with a perimembranous ventricular septal defect in a
human heart sectioned to replicate the subcostal oblique echocardio-
graphic section (lower panel). The phenotypic feature of the perimem-
branous defect, which opens to the right ventricle between the limbs
of the septomarginal trabeculation, or septal band (yellow, Y) is the
presence of fibrous continuity between the leaflets of the aortic and
tricuspid valves. This feature is well explained by the fact that the dor-
sal wall of the embryonic interventricular communication is formed by
continuity between the atrioventricular (AV) and outflow cushions
(arrows). The ventral trabeculations in the developing heart (stars in
upper panel) compact to form the septoparietal trabeculations of the
postnatal heart (stars in lower panel).
heart, both in man and mouse, it is this component that
becomes the atrioventricular membranous septum. The
delamination of the septal leaflet of the tricuspid valve
then eventually produces the division of the overall
membranous septal component into its fibrous atrioven-
tricular and interventricular parts (Fig. 16A). The
fibrous component of the membranous septum is the
only true atrioventricular septal structure found within
the human heart. Previously, the area where the cavity
of the right atrium overlaps the cavity of the left ventri-
cle by virtue of the off-setting of the hinges of the mitral
and tricuspid valves had been considered to represent a
muscular atrioventricular septum. As shown in Fig.
16B, a superior extension of the inferior atrioventricular
groove interposes in this area between the atrial and
ventricular muscular walls (Fig. 16B). The combined
area, therefore, is a muscular atrioventricular sandwich
(Anderson and Brown, 1996), rather than a solitary
muscular septum interposed between the cavities of the
right atrium and left venticle.
Appreciation of the growth of the vestibular spine into
the heart from the mediastinal myocardium, contribut-
ing as it does to the anterior–inferior muscular buttress
of the atrial septum, makes it easier to understand how
the cranial extension from the inferior atrioventricular
groove reaches toward the middle of the separated atrio-
ventricular junctions. The possible role of the vestibular
spine in ensuring normal separation of the right and left
atrioventricular junctions was initially suggested at the
latter end of the twentieth century (Webb et al., 1997,
1998a). At the same time, deficiency of the spine was
proposed as playing a role in the morphogenesis of atrio-
ventricular septal defects (Webb et al., 1997). Further
evidence supporting the significance of the spine then
emerged during the first decade of the twenty-first cen-
tury, with the protrusion being shown to originate from
the second heart field (Snarr et al., 2007). Most investi-
gators seeking to unravel the genetic and molecular
abnormalities underscoring deficient atrioventricular
septation, however, have tended to concentrate their
attention on the developing endocardial cushions (Briggs
et al., 2012). The very fact that patients are found with
unequivocal evidence of deficient atrioventricular septa-
tion, but with obvious continuity of the atrioventricular
valvar tissues to produce separate right and left valvar
orifices, is already sufficient to cast doubt on the pre-
sumption that it is failure of fusion of the endocardial
cushions that is the prime mover in retention of a com-
mon atrioventricular junction (Anderson et al., 2010).
This suggests that future investigations should devote
just as much, if not more, attention to the spine as to
the endocardial cushions. In this respect, it should be
noted that although patients with atrioventricular septal
defects and common atrioventricular junctions can have
the potential for shunting at either atrial or ventricular
levels through the defect or at both levels, with the holes
permitting the shunting are part of the atrioventricular
septal defect (Becker and Anderson, 1982). It is a mis-
take to consider them as representing “atrial septal
defects” or “ventricular septal defects,” as they are not
within either the atrial or the ventricular septal struc-
tures, but rather represent parts of the atrioventricular
septal defect.
Septation of the Ventricular Chambers
We have already demonstrated that the key features
ensuring normal atrioventricular septation are expan-
sion rightward of the atrioventricular canal and leftward
expansion of the ventricular outflow tract, the latter per-
mitting the aortic root to achieve its required position in
the roof of the developing left ventricle. The latter fea-
ture, along with fusion and muscularization of the proxi-
mal outflow cushions, is also essential for the completion
of ventricular septation. The morphological changes
required for closure of the embryonic interventricular
communication and the subsequent formation of the
 DEVELOPMENT OF CARDIAC SEPTATION
Fig. 20. The images show a dataset from a developing mouse heart
at E14.5 in which the furin enzyme has been perturbed (left-hand
panel) when compared with a doubly committed and juxta-arterial
ventricular septal defect as seen in the human (right-hand panel). The
image from the heart of the genetically modified mouse shows both
arterial trunks arising from the right ventricle, but with failure of forma-
tion of the subpulmonary infundibulum. The distal outflow cushions
have fused to produce separate aortic and pulmonary valvar orifices;
membranous part of the ventricular septum were
described in the middle part of the twentieth century
(Odgers, 1937/1938). Our episcopic images endorse the
accuracy of these observations and interpretations. As
we have shown, the expansion of the atrioventricular
canal brings the rightward margins of the central cush-
ions into the cavity of the right ventricle. Subsequent to
fusion, it is the cushion mass itself that forms the dorsal
wall of the embryonic interventricular communication.
At this stage, with fusion of the atrioventricular cush-
ions occurring during embryonic days 12.5 and 13.5, the
cavity of the proximal outflow tract remains undivided
and is supported exclusively by the musculature of the
right ventricle. From the outset, nonetheless, the dorsal
part of the ouflow tract, destined to become the aortic
vestibule, is in morphological continuity with the roof of
the developing left ventricle (Fig. 17). As the aortic root
becomes transferred to the left ventricle, the fused proxi-
mal components of the outflow cushions are brought into
line with the crest of the muscular septum, reducing
markedly the size of the persisting communication
between the cavity of the right ventricle and the aortic
root (Fig. 18).
The muscular ventricular septum itself was initially
formed concomitant with the process of ballooning of the
ventricular apical components (Figs. 4, 7, and 12). As it
forms, the crest of the developing septum is compact.
Even as late as E14.5, however, the episcopic sections
show that the texture of the apical part of the septum is
more comparable with the trabecular parts of the ven-
tricular walls than their compact components (Figs. 12A,
1427
however, the proximal cushions have failed to fuse. The fused interme-
diate components of the cushions form the roof of the interventricular
communication, which opens beneath both arterial roots (yellow
dashed line). The features of the doubly committed defect shown in
the right-hand panel illustrate how the defect, as in the mouse heart,
represents failure of formation of the subpulmonary infundibulum. The
cranial roof of the defect is formed by a fibrous raphe (yellow dashed
line) between the leaflets of the aortic and pulmonary valves.
15B, and 18B). It is also the case that multiple muscular
ventricular septal defects are associated with the process
described as “ventricular noncompaction.” The evidence
suggests, therefore, that the muscular ventricular sep-
tum is itself formed by a process of compaction of
trabeculations.
Abnormalities in these various processes involved in
normal ventricular septation provide rational explana-
tions for the different types of communication seen in
humans with deficient ventricular septation. Defects can
exist in any part of the muscular ventricular septum
and can be multiple. Indeed, the worst variant of multi-
ple defects is typically described as forming the “swiss-
cheese” septum. This is well explained on the basis of
incomplete compaction of the trabeculations that form
the apical component of the developing septum. It has
long been known that such compaction is responsible for
formation of the muscular ventricular septum in chicken
hearts (Ben-Shachar et al., 1985). The episcopic images
certainly show a likeness between the texture of the sep-
tum and that of the trabecular walls; however, the sug-
gested process of compaction needs to be confirmed. The
commonest type of ventricular septal defect requiring
surgical closure is the one directly related to the aortic
root. This variant is well explained simply on the basis
of failure of closure of that part of the embryonic inter-
ventricular communication that normally exists between
the cavity of the right ventricle and the aortic root (Fig.
19). The rarest type of ventricular septal defect found in
human hearts is characterized by fibrous continuity
between the leaflets of the aortic and pulmonary valves.
1428 ANDERSON ET AL.
This defect cannot exist in the normally developed heart,
as its prime feature is failure of formation of the muscu-
lar subpulmonary infunbibulum. It is closely related to
the type of defect found in the setting of common arte-
rial trunk. In consequence of the absence of formation of
a subpulmonary infundibulum in both these lesions, the
defect itself is roofed cranially by the leaflets of either
the truncal valve or the conjoined leaflets of the aortic
and pulmonary valves (Fig. 20, right-hand panel). In the
heart with separate aortic and pulmonary roots, the
lesion is well explained on the basis of failure of fusion
and muscularization of the proximal outflow cushions.
Indeed, we have now discovered such lesions, albeit
associated with double outlet from the right ventricle, in
mice with perturbation of the furin enzyme (Fig. 20, left-
hand panel). In these mice, some also show common
arterial trunk due to complete failure of fusion of the
outflow cushions, endorsing the long-established concept
for the morphogenesis of this lesion (Van Mierop et al.,
1978).
DISCUSSION
The advances made over the past two decades in
understanding cardiac development, along with those
made in the fields of molecular and cellular biology,
have been truly remarkable. The current knowledge of
development of the septal components now permits cred-
ible explanations to be offered for the morphogenesis of
the different phenotypic variants existing in the human
heart that permit interatrial, atrioventricular, or inter-
ventricular communications. Not all of these communica-
tions are simply holes within the various septal
components. At the atrial level, it is only the holes exist-
ing within the oval fossa, or the channels that percolate
through its anterior–inferior buttress, that are true sep-
tal defects. At the atrioventricular level, the essential
feature of the deficient atrioventricular septation is the
commonality of the atrioventricular junction. Within the
ventricles, the rarest type of defect, which is doubly com-
mitted and juxta-arterial because of failure to form the
muscular subpulmonary infundibulum, cannot obviously
exist when there has been normal formation of the
infundibulum. Patients with this defect are more closely
related to those with common arterial trunk than to
those having muscular or perimembranous defects.
Indeed, both of these lesions have been seen in our mice
programmed to perturb the furin enzyme. All of these
features point to the need for developmental biologists to
take note of the morphologic features when constructing
concepts to explain cardiac development, and for pediat-
ric cardiologists to take similar notice when considering
the basis of genetic counseling. The current notions for
atrial septation, for example, which typically are based
on formation of a secondary atrial septum growing in
caudal fashion from the atrial roof, no longer stand rig-
orous scrutiny. The atrial septum does arise from multi-
ple components, but the secondary component, rather
than extending toward the atrioventricular cushions
from an origin in the atrial roof, as depicted in most cur-
rent textbooks of embryology, grows ventrally into the
atrial cavity from the mediastinal mesenchyme. It is
derived from the so-called dorsal mesenchymal protru-
sion (Snarr et al., 2007); this structure initially being
described as the vestibular spine (His, 1885; Webb et al.,
1998b). Many of the anatomic features emphasized in
our review have been known for quite some time. The
illustrations provided by His (1885), and the account
given by Odgers (1937/1938) for development of the
membranous part of the ventricular septum, are remark-
ably accurate. The fact that the superior rim of the oval
fossa in the human heart is an infolding, rather than a
solid muscular septum, was also stressed in the latter
part of the nineteenth century (Rose, 1889). It is difficult
to explain why these facts did not make their way into
standard accounts of cardiac development. Perhaps this
deficiency will now be rectified, as the evidence is more
readily seen from the episcopic reconstructions. Infer-
ences made on the basis of anatomic findings, nonethe-
less, remain speculative. It is now incumbent on those
experts in developmental biology to unravel the molecu-
lar and cellular mechanisms that underscore the
observed morphologic changes.
LITERATURE CITED
Anderson RH, Brown NA. 1996. The anatomy of the heart revisited.
Anat Rec 246:1–7.
Anderson RH, Mohun TJ, Moorman AFM. 2012. What is a ventri-
cle? Cardiol Young 21:14–22
Anderson RH, Webb S, Brown NA. 1999. Clinical anatomy of the
atrial septum with reference to its developmental components.
Clin Anat 12:362–374.
Anderson RH, Wessels A, Vettukattil JJ. 2010. Morphology and
morphogenesis of atrioventricular septal defect with common atri-
oventricular junction. World J Ped Cong Heart Surg 1:59–67.
Becker AE, Anderson RH. 1982. Atrioventricular septal defects.
What’s in a name? J Thorac Cardiovasc Surg 83:461–469.
Ben-Shachar G, Arcilla RA, Lucas RV, Manasek FJ. 1985. Ventricu-
lar trabeculations in the chick embryo heart and their contribu-
tion to ventricular and muscular septal development. Circ Res 57:
759–766.
Briggs LE, Kakarla J, Wessels A. 2012. The pathogenesis of atrial
and atrioventricular septal defects with special emphasis on the
dorsal mesenchymal protrusion. Differentiation 84:117–130.
Butts RJ, Crean AM, Hlavacek AM, Spicer DE, Cook AC, Oechslin
EN, Anderson RH. 2011. Veno-venous bridges: the forerunners of
the sinus venosus defect. Cardiol Young 21:623–630.
Cohen MS, Anderson RH, Cohen MI, Atz AM, Fogel M, Gruber PJ,
Lopez L, Rome JJ, Weinberg PM. 2007. Controversies, genetics,
diagnostic assessment, and outcomes relating to the heterotaxy
syndrome. Cardiol Young 17 (Suppl 2):29–43.
Cole-Jeffrey CT, Terada R, Neth MR, Wessels A, Kasahara H. 2012.
Progressive anatomical closure of foramen ovale in normal neona-
tal mouse hearts. Anat Rec 295:764–768.
Hagen PT, Scholz DG, Edwards WD. 1984. Incidence and size of
patent foramen ovale during the first decades of life: an autopsy
study of 965 normal hearts. Mayo Clin Proc 59:17–20.
His W. 1885. Das Herz. In: His W, editor. Anatomie menschlicher embry-
onen. Vol. 3: Z€
ur Geschichte der Organe. Leipzig: Vogel. p 129–184.
Jahr M, Manner J. 2011. Development of the venous pole of the
frog Xenopus laevis: a morphological study with special focus on
the development of the venoatrial connections. Dev Dynamics
240:1518–1527.
Knauth A, McCarthy KP, Webb S, Ho SY, Allwork SP, Cook AC,
Anderson RH. 2002. Interatrial communication through the
mouth of the coronary sinus. Cardiol Young 12:364–372.
Martin JF, Arendt BA, Brown NA. 2010. Pitx2 in cardiac left–right
asymmetry and human disease. In: Rosenthal N, Harvey RP, edi-
tors. Heart development and regeneration. Amsterdam: Academic
Press. p 307–322.
Mohun TJ, Weninger WJ. 2011. Imaging heart development using
high-resolution episcopic microscopy. Curr Opin Genet Dev 21:
573–578.
 DEVELOPMENT OF CARDIAC SEPTATION
Mommersteeg MTM, Brown NA, Prall OWJ, Gier-de Vries C,
Harvey RP, Moorman AFM, Christoffels VM. 2007. Pitx2c and
Nkx2–5 are required for the formation and identity of the pulmo-
nary myocardium. Circ Res 101:902–909.
Moorman AFM, Christoffels VM. 2003. Cardiac chamber formation:
development, genes, and evolution. Physiol Rev 83:1223–1267.
Moorman AFM, Christoffels VM, Anderson RH, van den Hoff MJB.
2007. The heart-forming fields: one or multiple? Phil Trans R Soc
Biol 362:1257–1265.
Odgers PNB. 1937/1938. The development of the pars membranacea
septi in the human heart. J Anat 72:247–259.
Rana MS, Horsten NCA, Tesink-Taekema S, Lamers WH, Moorman
AFM, van den Hoff MJB. 2007. Trabeculated right ventricular free
wall in the chicken heart forms by ventricularization of the myocar-
dium initially forming the outflow tract. Circ Res 100:1000–1007.
Rose C. 1889. Zue entwickelungsgeschichte des Suagetierherzens.
Morphol Jahrb 15:436–456.
Sharratt GP, Webb S, Anderson RH. 2003. The vestibular defect: an
interatrial communication due to a deficiency in the atrial septal
component derived from the vestibular spine. Cardiol Young 13:
184–190.
Snarr BS, O’Neal JL, Chintalapudi MR, Wirrig EE, Phelps AL,
Kubalak SW, Wessels A. 2007. Isl1 expression at the venous pole
identifies a novel role for the second heart field in cardiac devel-
opment. Circ Res 101:971–974.
Soufan AT, Ruijter JM, van den Hoff MJ, de Boer PA, Hagoort J,
Moorman AF. 2003. Three-dimensional reconstruction of gene
expression patterns during cardiac development. Physiol
Genomics 13:187–195.
Uemura H, Ho SY, Devine WA, Anderson RH. 1995a. Analysis of
visceral heterotaxy according to splenic status, appendage mor-
phology, or both. Am J Cardiol 76:846–849.
1429
Uemura H, Ho SY, Devine WA, Kilpatrick LL, Anderson RH. 1995b.
Atrial appendages and venoatrial connections in hearts with
patients with visceral heterotaxy. Ann Thorac Surg 60:561–569.
Van Mierop LH, Patterson DF, Schnarr WR. 1978. Pathogenesis of
persistent truncus arteriosus in light of observations made in a
dog embryo with the anomaly. Am J Cardiol 41:755–762.
Vettukattil JJ, Ahmed Z, Salmon AP, Mohun T, Anderson RH. 2013.
Defects in the oval fossa: morphologic variations and impact on
transcatheter closure. J Am Soc Echocardiogr 26:192–199.
Webb S, Brown NA, Anderson RH. 1996. The structure of the
mouse heart in late fetal stages. Anat Embryol 194:37–47.
Webb S, Brown NA, Anderson RH. 1997. Cardiac morphology at
late fetal stages in the mouse with trisomy 16: consequences
for different formation of the atrioventricular junction when
compared to humans with trisomy 21. Cardiovasc Res 34:
515–524.
Webb S, Brown NA, Anderson RH. 1998a. Formation of the atrio-
ventricular septal structures in the normal mouse. Circ Res 1998;
82:645–656.
Webb S, Brown NA, Anderson RH, Richardson MK. 2000. Relation-
ship in the chick of the developing pulmonary vein to the embry-
onic systemic venous sinus. Anat Rec 259:67–75.
Webb S, Brown NA, Wessels A, Anderson RH. 1998b. Development
of the murine pulmonary vein and its relationship to the embry-
onic venous sinus. Anat Rec 250:325–334.
Webb S, Kanani M, Anderson RH, Richardson MK, Brown NA.
2001. Development of the human pulmonary vein and its incorpo-
ration in the morphologically left atrium. Cardiol Young 11:
632–642.
Webb S, Qayyum SR, Anderson RH, Lamers WH, Richardson MK.
2003. Septation and separation within the outflow tract of the
developing heart. J Anat 202:327–342.