Heart Defects in Connexin43-Deficient Mice

Jing Ya, Erna B.H.W. Erdtsieck-Ernste, Piet A.J. de Boer, Marjan J.A. van Kempen, Habo Jongsma,
Daniel Gros, Antoon F.M. Moorman, Wouter H. Lamers

Abstract—Cardiac malformation in connexin43 (CX43)-disrupted mice is restricted to the junction between right ventricle
and outflow tract, even though CX43 is also expressed abundantly elsewhere. We analyzed cardiac morphogenesis in
immunohistochemically and hybridohistochemically stained and three-dimensionally reconstructed serial sections of
CX43-deficient embryos between embryonic day (ED) 10 and birth. The establishment of the D configuration in the
ascending loop of CX43-deficient hearts is markedly retarded, so that the right ventricle retains a craniomedial position
and is connected with the outflow tract by a more acute bend in ED10 and ED11 embryos. Because of the subsequent
growth of the right ventricle, this condition usually evolves into a D loop, but when it persists, a “crisscross” configuration
develops, with the atrioventricular cushions rotated 90°, a horizontal muscular ventricular septum, and a parallel course of
the endocardial ridges of the outflow tract. After ED12, large intertrabecular pouches develop at the ventricular side of both
shelflike myocardial structures that support the endocardial ridges of the outflow tract, ie, at the location that was earlier
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017

characterized by the acute bend between the right ventricle and the outflow tract and that subsequently develops into the
anterosuperior leaflet of the tricuspid valve. Retarded development of the D configuration in the ascending loop of the
embryonic heart predisposes the myocardium at the junction of the right ventricle and outflow tract to excessive
development of intertrabecular pouches during subsequent development. (Circ Res. 1998;82:360-366.)
Key Words: connexin n cardiac malformation n cardiac development

C ongenital malformations of the heart are among the most

common birth defects in humans. The systematic analysis of
these defects almost entirely rests on findings in human neonates
with these structural anomalies. As a consequence, little direct
information is available about the development of these malfor-
mations. In particular, it is usually not known to what extent a
particular malformation can be traced to a single morphogenetic
event or process. A case in point is the development of the
malformation of the subpulmonary outflow tract of the heart in
mice that are deficient for gap junctions of the CX43 type that
was reported by Reaume et al.1
Many cells in the body are interconnected by gap junctions.
A gap junction is a membrane specialization that can be
recognized as a polygonal lattice of connexons. Connexons are
composed of a family of proteins known as connexins. Each
connexon is made up of six connexin molecules in each of the
two apposed membranes, which together surround a central
pore. These pores provide a direct low-resistance intercellular
pathway for the passage of ions and small molecules (,1 kD).
At present, at least 13 connexin genes that belong to two
subfamilies have been identified.2 Different connexins confer
distinct physiological properties on gap junctions.2,3 The func-
tional role of a protein can sometimes be deduced from the
phenotype of a congenital deficiency of its corresponding
gene. Thus, CX32 deficiency was shown to be the underlying
cause of X-linked Charcot-Marie-Tooth neuropathy.4 How-
ever, congenital gene deficiencies usually remain undetected if
the protein plays a crucial role during early embryonic devel-
opment. CX43, for example, is already expressed in the
eight-cell embryo.5 Furthermore, CX43 is abundantly ex-
pressed in the developing and adult heart,6 –12 and gap junctions
of this type are thought to be responsible for the behavior of
the myocardium as an electrical syncytium.13 In this connec-
tion, it has recently been shown that the spread of the impulse
in animals heterozygous for the CX43-null mutation is mark-
edly impaired.14 The survival to birth of embryonic mice
homozygous for the CX43-null mutation is therefore remark-
able.1 The death of CX43-deficient mice has been associated
with an obstruction of the subpulmonary outflow tract. How-
ever, the underlying cause of this abnormality and, in partic-
ular, why the malformation is restricted to the subpulmonary
outflow tract, whereas other parts of the heart in which CX43
is also expressed abundantly are not affected, remain to be
established. In the present study, we analyzed serially sectioned
CX43 gene– disrupted mouse embryos between ED10 and
birth. The sections were stained immunohistochemically or
hybridohistochemically to aid in delineating the myocardial
and fibrous structures and were subsequently reconstructed to

Received March 25, 1997; accepted November 7, 1997.

From the Department of Anatomy and Embryology (J.Y., E.B.H.W.E.-E., P.A.J. d B., A.F.M.M., W.H.L.), Academic Medical Center, University of
Amsterdam (the Netherlands); the Department of Medical Physiology (M.J.A. v K., H.J.), University of Utrecht (the Netherlands); and the Institut de
Biologie du Développement de Marseille (D.G.), Faculté des Sciences de Luminy, Marseille, France.
Correspondence to Wouter H. Lamers, MD, PhD, Department of Anatomy and Embryology, University of Amsterdam, Academic Medical Center,
Meibergdreef 15, 1105 AZ, Amsterdam, Netherlands.
E-mail w.h.lamers@amc.uva.nl
© 1998 American Heart Association, Inc.

360
 Ya et al
Selected Abbreviations and Acronyms
ANF 5 atrial natriuretic factor
CMV 5 cytomegalovirus
CX (with number) 5 connexin
ED (with number) 5 embryonic day
MHC 5 myosin heavy chain
neor 5 neomycin phosphotransferase
PCR 5 polymerase chain reaction
SERCA2 5 sarcoplasmic/endoplasmic reticulum
Ca21-ATPase 2
SMA 5 smooth muscle actin
visualize structural anomalies. Cardiac looping was found to be
impaired during early organogenesis, resulting in a more acute
bend between the embryonic right ventricle and outflow tract.
Subsequently, an abnormal delamination of the anterosuperior
leaflet of the tricuspid valve develops at this location.
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017
Materials and Methods
Animals
A heterozygous pair of the Sv109 strain of CX43 gene– disrupted
mice1 was purchased from the Jackson Laboratory, Bar Harbor, Me.
Matings were overnight, and the day of finding a plug was designated
day 0. The offspring were screened by PCR using primers specific for
neor and CX43. The sequence of both primers was as follows: neor,
plus 59-CAAGATGGATTGCACGCAGG-39 and minus 59-TAT
CACGGGTAGCCAACGC-39; CX43, plus 59-CCCATCCAAA
GACTGCGGATC-39 and minus 59-TGAGTACCACCTC
CACGGGAAC-39. A CX43-negative, neor-positive PCR signal was
used to identify embryos with both CX43 alleles disrupted. Embryos
were collected every day between ED10 and ED18 and staged
according to Theiler.15 Neonates were collected within 2 hours of
birth. Embryos were fixed either in a mixture of ice-cold methanol/
acetone/water (2:2:1) for immunohistochemical analysis or in 4%
(para)formaldehyde (wt/vol) in PBS (150 mmol/L NaCl and
10 mmol/L sodium phosphate, pH 7.4) for in situ hybridization
studies. The embryos were fixed overnight, dehydrated in a graded
series of ethanol, embedded in Paraplast Plus (Sherwood), serially
sectioned at 7-mm thickness, and mounted on poly-L-lysine– coated
or 3-aminopropyltriethoxysilane– coated slides for immunohisto-
chemistry or in situ hybridization, respectively. The mounted sections
were stored at 4°C until use.
Immunohistochemistry
After the paraffin was removed, the sections were pretreated, stained,
and mounted as described.16 Serial sections were incubated overnight
with alternately monoclonal antibodies against a-MHC (1:10),17
b-MHC (1:10),18 a-SMA (1:1000, Sigma IMMH-2), desmin (1:50,
Monosan, Mon 3001, clone 33), and a polyclonal antibody against
fibronectin (1:6000, AB 1942, Brunschwig Chemie). Sections incu-
bated with monoclonal antibodies were stained with rabbit-antimouse
IgG (1:7500, noncommercial) followed by incubation with goat
anti-rabbit IgG (1:250, noncommercial); the sections incubated with
the polyclonal antifibronectin antiserum were stained directly with
goat anti-rabbit IgG. Antibody binding was demonstrated with rabbit
peroxidase–anti-peroxidase complex (1:750, Nordic) and 3,3-diami-
nobenzidine tetrahydrochloride and H2O2 as substrates.
Immunofluorescence
Hearts were isolated and directly frozen in liquid nitrogen. Unfixed
cryosections (10 mm) were collected on slides coated with aminopro-
pyltriethoxysilane (Sigma) and stored at 280°C until used. Slides were
equilibrated to room temperature and subsequently incubated with
0.2% Triton X-100 in PBS for 1 hour, followed by 30 minutes of
blocking with 2% BSA. Sections were incubated overnight with
361
appropriate dilutions of primary antibodies in PBS supplemented with
2% BSA and 10% normal goat serum. All incubation steps were
performed at room temperature, and between all incubation steps the
slides were thoroughly washed with PBS. The anti-CX40 and
anti-CX37 antibodies were raised in rabbits against a synthetic peptide
containing residues 335 to 356 of rat CX40 and residues 315 to 331
of mouse CX37, affinity-purified, and diluted to 5 to 15 mg/mL and
2 to 4 mg/mL, respectively. The specificity of the CX40 antibody was
reported,19 whereas that of the CX37 antibody was demonstrated in
communication-deficient cells transfected with mouse CX37.20 The
next morning cells were preincubated with 2% BSA in PBS. Immu-
nolabeling was carried out with fluorescein isothiocyanate– conjugated
secondary antibodies (Jackson immunostaining) against rabbit immu-
noglobulins. Sections were mounted in Vectashield mounting me-
dium (Vector Laboratories) and examined and photographed with an
epifluorescence microscope equipped with the appropriate filter.
In Situ Hybridization
In situ hybridization was carried out on the sections of ED14, ED16,
and ED18 embryos to visualize the expression of CX40, CX43,
CX45, SERCA2, and ANF, as described in detail previously.21
Single-stranded antisense RNA probes were made by in vitro RNA
transcription. All clones were inserted into pBluescript. The rat CX43
clone (1.5 kb)22 was linearized with Sal I and transcribed with T7
RNA polymerase, with both 35S-UTP and 35S-CTP as labeled
substrates. The mouse CX40 (1.35 kb)23 and CX45 (1.68 kb)24 clones
were linearized with Asp718 and transcribed with T3 RNA polymer-
ase. The rat ANF clone (0.6 kb)25 was linearized with BamHI and
transcribed with T7 RNA polymerase. The SERCA2 clone (1.6
kb)21,26 was linearized with Xho I and transcribed with T3 RNA
polymerase. cRNA probes were degraded to fragments of '100 bp
by hydrolysis in 80 mmol/L NaHCO3 and 120 mmol/L Na2CO3 at
60°C for 10 to 20 minutes. The hybridization was performed with 30
pg of 35S-labeled antisense RNA (16 to 18 hours, 54°C) in 6 to 10 mL
of hybridization mixture per section (50% formamide, 10% dextran
sulfate, 300 mmol/L NaCl, 30 mmol/L sodium citrate, 23 Denhardt’s
solution, 0.1% Triton X-100, 10 mmol/L dithiothreitol, and 200
ng/mL herring sperm DNA; pH 7.4). The specific activity of the
cRNA was 1.53109 cpm/mg for CX40, CX45, ANF, and SERCA2
cRNA and 33109 cpm/mg for CX43 (double-labeled). After hybrid-
ization, the sections were washed and treated with RNase (30
minutes, 37°C) to reduce nonspecific binding. Under these condi-
tions, the hybridization signal in sections is quantitative.27
No CX43 protein is expressed in CX43-null mutants (not shown;
see Reference 1). However, because the promoterless neor gene was
placed in frame after the second amino acid of CX43 and replaced the
transmembrane regions of the CX43, but not its 39 tail in the
CX43-targeting vector,1 the pattern of expression of the chimeric
neor-CX43 mRNA could be visualized with labeled antisense CX43
probe.
Three-Dimensional Reconstruction
Selected hearts of CX43-mutant mice and their heterozygous litter-
mates were reconstructed three-dimensionally as described28 and
rendered with the help of a medical artist.
Results
We have investigated the architecture and a number of gene
expression patterns in the hearts of homozygous and heterozy-
gous CX43-disrupted mice between ED10 and birth and
compared them with wild-type littermates. A total of 44
homozygous mice, obtained from matings of heterozygotes,
were available for study. These homozygotes represented
16.2% of the live animals, which is significantly below the
expected 25% (P,.01) and confirms the earlier finding of
Reaumé et al.1 The reduced percentage of homozygotes was
observed at all ages investigated. Apart from the malformations
in CX43-deficient hearts described below, we observed a left
 362 Heart Defects in Connexin43-Deficient Mice

coronary artery that ended in the right ventricle in an ED14

heart (not shown). No abnormalities were found in the
conduction system of affected hearts. The distribution of
cardiac nerves was not studied. In addition to the homozygous
CX43 knockout embryos and neonates, 20 heterozygotes and
20 wild-type littermates were analyzed. No abnormalities in
architecture or gene expression pattern were observed in
wild-type or heterozygous animals, in agreement with another
recent study.14

Development of the Heart Between ED10

and ED12
In the mouse embryo, cardiac looping is initiated at ED8.5
(Theiler’s stage 13)15 and is virtually completed at ED9
(Theiler’s stage 14), ie, before the respective segments of the
embryonic heart have become identifiable by their lateral
expansion. At this stage, the heart loop has a sagittally oriented
component with inflow and outflow tract dorsal and the
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017

ventricle ventral and a transversely oriented component with

the left ventricle left and the right ventricle right (the so-called
dextro or D loop29,30). When we compared homozygous and
heterozygous CX43-disrupted ED10 hearts (Fig 1A to 1F), we
observed that in the homozygous animals, the transverse
component of the loop had only incompletely developed. The
right ventricle had retained a more cranial and medial position
compared with that of control animals (so-called A loop),29,30
so that the developing interventricular septum occupied a
transverse rather than a sagittal position and the connection of
the outflow tract with the right ventricle was more acute than
in control animals. This configuration was also seen in ED11
animals (Fig 1G to 1J). Subsequently, the rightward expansion
of the right ventricle changed its distinct cranial position into
a more lateral one, so that the looping abnormality became less
obvious and most CX43 knockout hearts could no longer be
identified externally at ED12. Although the pattern, as de-
scribed, was seen in all homozygous embryos, some variation
in the severity of the looping problems was observed.
Histological examination of ED10, ED11 (Fig 1A, 1D, 1I,
and 1J), and ED12 hearts (not shown) of CX43 knockout
animals did not reveal structural abnormalities. The sinus
venosus, left and right atria, and atrial septum were positioned
normally. These structures showed high a-MHC expression,
whereas in the outflow tract and ventricles, a-MHC expres-
sion was low (not shown). b-MHC expression in the hearts
between ED10 and ED12 was mainly restricted to the outflow
tract and ventricles (Fig 1). A strong expression of a-SMA and
desmin was detected throughout the heart between ED10 and
ED12, especially in the outflow tract, where fibronectin
delineated the endocardial cushions/ridges (Fig 2A to 2D).
Apart from the looping abnormality and the cases to be
described in the next paragraph, no structural abnormalities
were observed in the ventricles and outflow tract, in particular
not at the junction of the right ventricle and the outflow tract.
Nevertheless, it was observed that the embryos with the more
severe looping problems showed some developmental
retardation.
Especially interesting cases that illustrate the looping abnor-
malities observed in CX43 knockout mice were provided by
the ED11 and the ED12 embryo illustrated in Figs 1H to 1J
Figure 1. Abnormal formation of the heart loop in CX43-defi-
cient mouse embryos. AVC indicates atrioventricular canal; LV,
embryonic left ventricle; OFT, outflow tract; RA, right atrium;
and RV, embryonic right ventricle. The orientations are indicated
as follows: a, anterior; p, posterior; r, right; l, left; d, dorsal; and
v, ventral). Panels A to F show hearts of ED10 embryos; panels
G to J, hearts of ED11 embryos. The hearts shown in panels A
to C and in panel G are from heterozygous animals, whereas
those in panels D to F and H to J are from homozygous CX43-
deficient animals. The transverse sections shown were stained
for the presence of b-MHC; the luminal boundary of the endo-
cardial cushions/ridges is stippled for clarity. The reconstruc-
tions show the external shape of the myocardium of the ventri-
cles and OFT, based on b-MHC staining of serial sections
(panels A to C, D to F, and H to J are grouped together; the
position of each of the sections is indicated by its panel letter
[encircled] in the respective reconstructions). Panels B and E are
viewed from the right lateral side; panels C and F, from the cau-
dal aspect; and panels G and H, from the cranial aspect.
Bar5100 mm.
and 2, respectively. The ED11 embryo shown in Fig 1G to 1J
is considered to be the most seriously affected of 4 CX43-
deficient ED11 embryos studied because of its pronounced
cranial position of the right ventricle and its leftward convex
bend midway in the outflow tract and is thought to be a likely
precursor of the condition seen in the ED12 embryo shown in
Fig 2. In this latter embryo, which is one of four knockouts of
this age group studied, the inflow tract and the atrium appeared
normal. The systemic veins drained into the right atrium,
which was substantially larger than the left. The developing
 Ya et al
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017
Figure 2. CX43-deficient ED12 embryo with persisting cranial
position of the right ventricle (RV). IV indicates 4th-arch artery;
VI, 6th-arch artery; AVC, atrioventricular canal; IVS, interventric-
ular septum; LV, embryonic left ventricle; M, mitral valve orifice;
OFT, outflow tract; LA, left atrium; RA, right atrium; SCV, supe-
rior caval vein; and T, tricuspid valve orifice. The orientations of
the reconstructions are indicated as follows: a, anterior; p, pos-
terior; r, right; l, left; d, dorsal; v, ventral; dr, dextrodorsal; and
vl, sinistroventral. Panels A to D show transverse sections to
illustrate the topographic relation between RV, OFT, and arterial
pole (panels A and B, serial sections stained for the presence of
b-MHC and fibronectin, respectively) and between RV, LV, LA,
and RA (panels C and D, stained for the presence of a-SMA
and desmin, respectively). The luminal boundary of the endocar-
dial cushions/ridges is stippled for clarity. Panels E to H show a
reconstruction of this embryo, based on the staining pattern of
b-MHC (myocardium) and fibronectin (endocardial cushions/
ridges). Panel E shows a dextrocranial view, and panels F to H
show a dorsal view, with the atrium removed in panels E, G,
and H and the endocardial tissue inserted in panel G. The posi-
tion of each of the sections (panel B is at same level as panel A)
is indicated by its panel letter (encircled) in the respective
reconstructions. Bar5200 mm.
sinoatrial node could be identified at the entrance of the
superior cardinal vein into the right atrium. The atrioventric-
ular canal was still located over the embryonic left ventricle.
However, the fusion line of both atrioventricular cushions was
oriented craniocaudally instead of laterolaterally, with the
tricuspid opening positioned cranially and the mitral opening
positioned caudally. The right ventricle had remained posi-
tioned cranially to the left ventricle, so that the interventricular
septum retained a transverse orientation and so that a pro-
nounced leftward convex bend developed midway in the
outflow tract (Fig 2A and 2B). The parietal and septal
endocardial ridges of the outflow tract followed a parallel
rather than the usual spiral course, with the parietal ridge
located cranially and the septal ridge caudally. In addition, both
ridges had developed asymmetrically and were apposed in the
363
Figure 3. Initial development of intertrabecular pouches at the
ventricular base of the endocardial outflow-tract (OFT) ridges in
CX43-deficient hearts. AVC indicates atrioventricular canal; IVF,
interventricular foramen; IVS, interventricular septum; LV, embry-
onic left ventricle; and RV, embryonic right ventricle. The orien-
tations of the reconstructions are indicated as follows: a, ante-
rior; p, posterior; d, dorsal; v, ventral; vr, dextroventral; and dl,
sinistrodorsal. Panels A and D were stained for the presence of
b-MHC, and panels B and E were stained for the presence of
a-SMA. Panels A and B show histology of the junction of the
RV and OFT of a heterozygous ED13 heart, and panels D and E
show the corresponding sections of a CX43-deficient littermate.
Panel C shows the reconstruction of the ventricles and OFT of
the morphologically normal heart as seen in a right-lateral view,
and panels F to H show a similar view of the CX43-deficient
heart. The position of each of the sections is indicated by its
corresponding panel letter (encircled) in the respective recon-
structions. Panel G shows almost the same right-sided view as
panel F, with the cranial portion of the RV and the right half of
the OFT removed to expose the left myocardial shelf. Panel H
shows a left-sided view with the cranial portion of the LV, the
medial portion of the RV, and the left half of the OFT removed
to expose the right myocardial shelf. The intracardiac location of
the myocardial shelves that form the ventricular base of both
OFT ridges is shown in panels G and H. The transparent struc-
tures, delineated with the white dashed line, represent the endo-
cardial ridges that “rest” on the shelves (indicated by arrows).
The white area represents the portion of both endocardial ridges
that has fused. The intertrabecular pouches (asterisk in panel D)
and excessive delamination (arrows in panel E) develop at the
ventricular side of these shelves and result in the outward bulg-
ing of wall of the RV at its junction with the OFT (arrows in
panel F; compare with panel C). Bar5100 mm.
inner curvature of the outflow tract. Since the outflow tract
had not (yet) septated, the single outlet channel split, on its
emergence from the myocardial sleeve, into bilateral sixth
arches (the left being approximately twice the size of the right
one) and an initially single, cranially running aortic channel.
This ED12 embryo had, therefore, retained the A-loop con-
figuration that is normally only found in CX43-deficient ED10
 364 Heart Defects in Connexin43-Deficient Mice
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017
Figure 4. Development of intertrabecular pouches in the right
ventricle (RV) at its junction with the outflow tract in fetal CX43-
deficient mice. AO indicates aorta; LV, left ventricle; M, mitral
valve orifice; OS, outlet septum; P, pulmonary trunk; SV, semilu-
nar valve; and T, tricuspid valve orifice. The orientations of the
reconstructions are indicated as follows: a, anterior; p, posterior;
r, right; l, left; d, dorsal; v, ventral; dr, dextrodorsal; and vl, sinis-
troventral. Panels A and C were stained for the presence of
a-SMA; panel B, for the presence of desmin; and panels D to F,
for the presence of b-MHC. The panels show transverse sec-
tions of CX43-deficient ED14 (panels A and B), ED15 (panel C),
and ED17 (panels D to F) embryos and reconstructions of the
ventricles and outflow tract of a normal (panel G) and CX43-
deficient (panels H and I) ED17 heart. Panels G and H represent
dorsal (d) views, whereas panel I represents a right-lateral view.
The position of each of the sections is indicated by its corre-
sponding panel letter (encircled) in the respective reconstruc-
tions. The lumen of the subpulmonary portion of the outflow
tract is recognizable as a centrally located, narrow channel
(asterisk in panels B and E). Bar5200 mm.
and ED11 embryos. In summary, this ED12 heart presented as
situs solitus, with the tricuspid orifice and right ventricle cranial
instead of lateral and a leftward-bent outflow tract.
Development of the Heart Between ED13
and Birth
Three-dimensional reconstruction of ED13 hearts showed
that, externally, homozygous CX43 knockout animals could
be distinguished from their heterozygous littermates by a slight
dilatation of the right ventricle at its junction with the outflow
tract (compare Fig 3C with 3F). In sections, the structural
abnormality that is characteristic for CX43-disrupted hearts,
namely, the excessive development of intertrabecular pouches
from the lumen of distal portion of the right ventricle, had
become identifiable. These pouches develop at the ventricular
side of two shelflike myocardial structures, which support the
base of both endocardial ridges of the outflow tract (Fig 3D and
3E). Therefore, the pouches develop at the location that in
Figure 5. The trabecular pouches of CX43-deficient hearts
develop in the right ventricle (RV) at its junction with the outflow
tract (OFT). AO indicates aorta; P, pulmonary trunk. Panels A
and B show the expression of the chimeric neor-CX43 mRNA;
panels C and D show the expression of SERCA2 mRNA in
hearts of CX43-deficient mice at ED14 (panels A and C) and
ED16 (panels B and D). Bar5200 mm.
previous stages was characterized by an acute bend between the
right ventricle and the outflow tract as a result of the persisting
A-loop configuration. The medial shelf is located on the right
ventromedial wall of the interventricular septum, and the lateral
shelf is located on the lateral free wall of the right ventricle (Fig 3G
and 3H, respectively). These shelves, which we observed in both
normal and abnormal embryos, were not yet seen in ED12 hearts
and arise as a result of a local ingrowth of myocytes into the base
of the endocardial ridges. In the CX43 knockout embryos, the
extended intertrabecular pouches that develop underneath the
shelves emphasized the presence and location of these shelves.
The endocardial and myocardial components of these shelves will
evolve into the smooth atrial and rough ventricular surface of the
anterosuperior leaflet of the tricuspid valve, respectively.31 The
development of the intertrabecular pouches at the site where, in
normal development, the anterosuperior valve leaflet delaminates
explains the involvement of this valve leaflet in CX43 deficiency.
External inspection of ED14 hearts revealed that the myo-
cardium at the junction of right ventricle and outflow tract
continued to dilate symmetrically (not shown). In the embry-
onic sections, we observed that the intertrabecular spaces near
the junction of the right ventricle with the outflow tract that
we first observed at ED13 had already become so enlarged at
ED14 that sizable cavities had developed (Fig 4A and 4B). On
the basis of the staining pattern of SERCA2 and CX43 (Fig 5),
which are lower and absent, respectively, in the outflow
tract,6,21 we concluded that the cavities developed within the
confines of the right ventricle but at its junction with the
outflow tract. In older animals, the pouches continued to
increase in diameter and to expand dorsolaterally, thereby
surrounding the pulmonary semilunar valves (Fig 4C and 4D).
As a result, the bulging wall of the right ventricle gradually
came to surround and enclose the root of the pulmonary trunk
(compare the normal heart in Fig 4G with the CX43-deficient
one in Fig 4H and 4I). No such process was seen to occur near
the aortic outlet. The trabeculae traversing the pouches clearly
had a hypertrophic appearance. In particular, the developing
medial (septal) papillary muscle, which attaches the anterosu-
perior leaflet of the tricuspid valve to the interventricular
septum and which can be recognized as medially flanking the
central channel near its transition into the subpulmonary
portion of the outflow tract, was markedly hypertrophied (Fig
4B). As a consequence of the strong development of the
trabeculae, the lumen of the original endocardial tube re-
mained an easily identifiable, centrally located, tortuous, and
 Ya et al
Figure 6. CX43-deficient hearts express ANF. AO indicates
aorta; LA, left atrium; P, pulmonary trunk; RA, right atrium; and
RV, right ventricle. Panels A and B show ANF mRNA in a con-
trol and a CX43-deficient ED18 heart, respectively.
Bar5200 mm.
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017
narrow channel (asterisk in Fig 4B and 4E), to which the
intertrabecular pouches were connected by sievelike openings.
The compact myocardium that borders the pouches externally
remained relatively thin. The region involved in this patho-
logical remodeling of the right ventricle remained confined to
its distal portion, ie, the myocardium supporting the develop-
ing subpulmonary outlet portion of the outflow tract. In a
caudal direction, the pouches rapidly decreased in size and did
not extend beyond (caudal to) the myocardial component of
the developing anterosuperior leaflet and the anterior papillary
muscle; ie, the caudal part of the right ventricle and the entire
left ventricle were not visibly affected (Fig 4F).
The expressions of a- and b-MHC, a-SMA, desmin,
SERCA2, and fibronectin, which highlight the structural and
functional development of the embryonic myocardium and
the endocardial cushions/ridges, were found to be comparable
in homozygous and heterozygous animals (not shown). The
staining intensity of a-MHC in the ventricles and outflow tract
continued to decrease with age, whereas the expression of
b-MHC was restricted to the ventricles from ED10 onward
(not shown). The intensity of expression of a-SMA started to
decrease gradually after ED13 and that of desmin began to
decrease after ED14, with the decrease beginning in the
interventricular septum, followed successively by the left
ventricle, right ventricle, and the central atrium (not shown).
These findings show that the developmental changes in ex-
pression of these genes proceed similarly to those previously
observed in rat embryos.16 To find out if the hypertrophic
embryonic trabeculae resemble postnatal hypertrophic myo-
cardium, we also investigated ANF expression. The difference
in staining between affected and control embryos gradually
became more pronounced, with ANF expression being clearly
upregulated in the affected portion of the heart in CX43-
deficient ED18 mice (Fig 6).
To determine whether the absence of CX43 expression was
accompanied by changes in the expression of CX37, CX40, or
CX45, these connexins were studied by immunofluorescence
(CX37 and CX40) or in situ hybridization (CX40 and CX45).
However, the expression of these connexins was similar in control
365
and mutant embryos. The vessel walls were positive for CX37 and
CX40, with CX40 also being abundantly expressed in the atrial
and the ventricular trabeculae (not shown). The marked differ-
ence in the developmental timing of the disappearance of CX40
mRNA from the trabeculae of left and right ventricles that we
previously noted in the rat6 was also seen in control and mutant
mice (not shown). The myocardium of the shelves at the base of
the endocardial ridges of the outflow tract, ie, the structures below
which the intertrabecular pouches develop in the CX43-deficient
hearts, clearly expressed CX45 in both control and affected
embryos (not shown). Elsewhere in the heart, CX45 expression
was very low or absent.
Discussion
The major abnormalities of cardiac development in CX43-
deficient embryos are a delay in the normal looping of the
ascending limb of the heart tube and, subsequently, the
progressive development of intertrabecular pouches at the
junction of the right ventricle and the outflow tract, resulting
in an abnormal delamination of the anterosuperior leaflet of the
tricuspid valve. In normal mouse embryos, the ascending limb
of the heart tube, which includes the right ventricle and the
outflow tract, will adopt a rightward position (D loop) during
ED9, but in CX43-deficient embryos, it temporarily retains a
more symmetric middle position (A loop29,30). This A-loop
configuration is metastable and usually changes into a D loop
in CX43 knockout embryos after ED11. However, if the
A-loop configuration persists, a heart develops in which both
atria retain their respective left- and right-sided position but in
which the position of the more distal parts of the heart has
rotated by 90°. Thus, the atrioventricular cushions occupy
lateral positions in the atrioventricular canal, while the right
ventricle occupies a cranial position relative to the left ventri-
cle. Accordingly, the left atrioventricular connection is caudal
and the right atrioventricular connection is cranial, while the
ventricular septum occupies a transverse plane. This configu-
ration is reminiscent of that found in “crisscross” hearts,32,33
although the development of such malformed hearts as a result
of a persisting A loop remains to be shown.
The increasingly overt structural malformation that is typical
for CX43-deficient hearts is confined to the myocardium
supporting the ventricular base of the endocardial outflow tract
ridges. The markers CX43 and SERCA2 show that these
shelf-shaped myocardial structures, which are involved in the
formation of the anterosuperior leaflet of the tricuspid valve,31
develop at the boundary of the right ventricle with the outflow
tract.6,21 Topographically, therefore, the malformation devel-
ops where, in preceding stages, the junction of the embryonic
right ventricle and the outflow tract made a more acute bend
in CX43-deficient embryos than in morphologically normal
embryos, but we do not know at present how such a
temporary geometric distortion contributes to the subsequent
developmental abnormalities at this location.
Both CX43-deficient mice and mice in which local over-
expression of CX43 is driven by the CMV promoter/enhanc-
er34 suffer from an obstruction of the subpulmonary outlet and
neonatal death. In this respect, mouse neonates are apparently
more vulnerable than human neonates. However, both mal-
formations are structurally different. The hearts of neonatal
 366 Heart Defects in Connexin43-Deficient Mice
CMV-CX43 transgenes are characterized by a narrowing of
the subpulmonary outlet as a result of hypertrophy of the
compact myocardium of the right ventricle and the interven-
tricular septum. In CX43-deficient hearts, on the other hand,
it seems likely that contraction of the markedly hypertrophied
trabeculae, including the papillary muscles of the tricuspid
valve, squeezes the narrow and tortuous outlet channel and
prevents the ejection of the blood. Similarly, the bilateral
dilatation of the right ventricular conus probably results from
the squeezing of the narrow muscular openings of the pouches
into the outlet channel of the right ventricle during contrac-
tion. In accordance with this scenario, we observed a locally
increased expression of ANF, a sensitive marker for cardiac
overload in adult heart.35 Furthermore, CX43-deficient ani-
mals that express the CMV-CX43 transgene are temporarily
rescued after birth,34 suggesting that the locally failing myocar-
dium of CX43-deficient hearts is shored up by the hypertro-
phic stimulus originating from the CMV-CX43 transgene.
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017
Despite the obstruction of the right ventricular outlet, the
more caudal part of the right ventricle is morphologically
normal, probably because blood can be diverted to the left
ventricle via the oval foramen.
Acknowledgments
The drawings of the three-dimensional reconstructions were skill-
fully prepared by I.E.M. Oosterling; C. Gravemeijer is credited for
the photographs.
References
1. Reaume AG, de Sousa PA, Kulkarni S, Langille BL, Zhu D, Davies TC,
Juneja SC, Kidder GM, Rossant J. Cardiac malformation in neonatal mice
lacking connexin43. Science. 1995;267:1831–1834.
2. Goodenough DA, Goliger JA, Paul DL. Connexins, connexons, and inter-
cellular communication. Annu Rev Biochem. 1996;65:475–502.
3. Bruzzone R, White TW, Paul DL. Connections with connexins: the
molecular basis of direct intercellular signaling. Eur J Biochem. 1996;
238:1–27.
4. Bergoffen J, Scherer SS, Wang S, Scott MO, Bone LJ, Paul DL, Chen K,
Lensch MW, Chance PF, Fischbeck KH. Connexin mutations in X-linked
Charcot-Marie-Tooth Disease. Science. 1993;262:2039 –2042.
5. Yancey SB, Biswal S, Revel JP. Spatial and temporal patterns of distri-
bution of the gap junction protein connexin43 during mouse gastrulation
and organogenesis. Development. 1992;114:203–212.
6. van Kempen MJA, Vermeulen JLM, Moorman AFM, Gros DB, Paul DL,
Lamers WH. Developmental changes of connexin40 and connexin43
mRNA-distribution patterns in the rat heart. Cardiovasc Res. 1996;32:
886 –900.
7. van Kempen MJA, Fromaget C, Gros D, Moorman AFM, Lamers WH.
Spatial distribution of connexin43, the major cardiac gap junction protein,
in the developing and adult rat heart. Circ Res. 1991;68:1638 –1651.
8. Dahl E, Winterhager E, Traub O, Willecke K. Expression of gap junction
genes, connexin40 and connexin43, during fetal mouse development. Anat
Embryol. 1995;191:267–278.
9. Gourdie RG, Green CR, Severs NJ, Anderson RH, Thompson RP.
Evidence for a distinct gap-junctional phenotype in ventricular conduction
tissues of the developing and mature avian heart. Circ Res. 1993;72:
278 –289.
10. Gourdie RG, Green RC, Severs NJ, Thompson RP. Immunolabeling
patterns of gap junction connexins in the developing and mature rat heart.
Anat Embryol. 1992;185:363–378.
11. Fromaget C, El Aoumari A, Dupont E, Briand JP, Gros D. Changes in the
expression of connexin 43, a cardiac gap junctional protein, during mouse
heart development. J Mol Cell Cardiol. 1990;22:1245–1258.
12. Ruangvoravat CP, Lo CW. Connexin43 expression in the mouse embryo:
localization of transcripts within developmentally significant domains. Dev
Dyn. 1992;194:261–281.
13. Gros DB, Jongsma HJ. Connexins in mammalian heart function. Bioessays.
1996;18:719 –730.
14. Guerrero PA, Schuessler RB, Davis LM, Beyer EC, Johnson CM, Yamada
KA, Saffitz JE. Slow ventricular conduction in mice heterozygous for a
connexin 43 null mutation. J Clin Invest. 1997;99:1991–1998.
15. Theiler K. The House Mouse, Development and Normal Stages From Fertili-
zation to 4 Weeks of Age. Berlin, Germany: Springer-Verlag; 1972.
16. Jing Y, Markman MWM, Wagenaar GTM, Blommaart PJE, Moorman
AFM, Lamers WH. Expression of the smooth-muscle proteins alpha
smooth-muscle actin and calponin, and of the intermediate filament
protein desmin are parameters of cardiomyocyte maturation in the prenatal
rat heart. Anat Rec. 1997;249:495–505.
17. de Groot IJM, Lamers WH, Moorman AFM. Isomyosin expression pattern
during rat heart morphogenesis: an immunohistochemical study. Anat Rec.
1989;224:365–373.
18. Sant’Ana Pereira JAA, Wessels A, Nytmans L, Moorman AFM, Sargeant AJ.
New method for the accurate characterization of single human skeletal muscle
fibres demonstrates a relation between mATPase and MyHC expression in
pure and hybrid fiber types. J Muscle Res Cell Motil. 1995;16:21–34.
19. Gros D, Jarry-Guichard T, Ten Velde I, de Mazière A, van Kempen MJA,
Davoust J, Briand JP, Moorman AFM, Jongsma HJ. Restricted distribution
of connexin40, a gap junctional protein, in mammalian heart. Circ Res.
1994;74:839 – 851.
20. Delorme B, Dahl E, Jarry-Guichard T, Briand J-P, Willecke K, Gros D,
Théveniau-Ruissy M. Expression pattern of connexin gene products at the
early developmental stages of the mouse cardiovascular system. Circ Res.
1997;81:423– 437.
21. Moorman AFM, Vermeulen JLM, Koban MU, Schwartz K, Lamers WH,
Boheler KR. Patterns of expression of sarcoplasmic reticulum Ca21-
ATPase and phospholamban mRNAs during rat heart development. Circ
Res. 1995;76:616 – 625.
22. Beyer EC, Paul DL, Goodenough DA. Connexin43: a protein from rat
heart homologous to a gap junction protein from liver. J Cell Biol. 1987;
105:2621–2629.
23. Henneman H, Suchyna T, Lichtenberg-Fraté H, Jungbluth S, Dahl E,
Schwarz J, Nicholson BJ, Willecke K. Molecular cloning and functional
expression of mouse connexin40, a second gap junction gene preferentially
expressed in lung. J Cell Biol. 1992;117:1299 –1310.
24. Henneman H, Schwarz HJ, Willecke K. Characterization of gap junction
genes expressed in F9 embryonic carcinoma cells: molecular cloning of
mouse connexin31 and -45 cDNAs. Eur J Cell Biol. 1992;57:51–58.
25. Seidman CE, Duby AD, Choi E, Graham RM, Haber E, Homcy C, Smith
JA, Seidman JG. The structure of rat preproatrial natriuretic factor as
defined by a complementary DNA clone. Science. 1984;225:324 –326.
26. Lompré AM, de la Bastie D, Boheler KR, Schwartz K. Characterization
and expression of the rat heart sarcoplasmic reticulum Ca21-ATPase
mRNA. FEBS Lett. 1989;249:35– 41.
27. Jonker A, de Boer PAJ, van den Hoff MJB, Lamers WH, Moorman AFM.
Towards quantitative in situ hybridisation. J Histochem Cytochem. 1997;45:
413– 423.
28. Verbeek FJ, Huysmans DP, Baeten RJAM, Schoutsen NJC, Lamers WH.
Design and implementation of a database and program for
3D-reconstruction from serial sections: a data-driven approach. Microsc Res
Tech. 1995;30:496 –512.
29. van Praagh R, van Praagh S. Embryology and anatomy: keys to the
understanding of complex congenital heart disease. Coeur. 1982;13:
315–337.
30. Okamoto N, Satow Y, Hidaka N, Akimoto N, Miyabara S. Morpho-
genesis of congenital heart anomaly: bulboventricular malformations. Jpn
Circ J. 1978;42:1105–1120.
31. Lamers WH, Virágh S, Wessels A, Moorman AFM, Anderson RH. For-
mation of the tricuspid valve in the human heart. Circulation. 1995;91:
111–121.
32. Anderson RH. Criss-cross hearts revisited. Pediatr Cardiol. 1982;3:305–313.
33. Thilenius OG, Bharati S, Lev M, Karp RB, Arcilla RA. Horizontal
ventricular septum with dextroversion: hearts with and without aortic
atresia. Pediatr Cardiol. 1987;8:187–193.
34. Ewart JL, Cohen MF, Meyer RA, Huang GY, Wessels A, Gourdie RG,
Chin AJ, Park SMJ, Lazatin BO, Villabon S, Lo CW. Heart and neural tube
defects in transgenic mice overexpressing the Cx43 gap junction gene.
Development. 1997;124:1281–1292.
35. Jougasaki M, Yasue H, Okumura K, Mukoyama M, Saito Y, Nakao K,
Takahashi K. Atrial natriuretic peptide in the ventricles of patients with
dilated cardiomyopathy and human foetuses. Histochem J. 1989;21:
715–720.
 Heart Defects in Connexin43-Deficient Mice
Jing Ya, Erna B. H. W. Erdtsieck-Ernste, Piet A. J. de Boer, Marjan J. A. van Kempen, Habo
Jongsma, Daniel Gros, Antoon F. M. Moorman and Wouter H. Lamers
Downloaded from http://circres.ahajournals.org/ by guest on April 6, 2017

Circ Res. 1998;82:360-366

doi: 10.1161/01.RES.82.3.360
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1998 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/82/3/360

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Circulation Research is online at:

http://circres.ahajournals.org//subscriptions/