Cardiovascular Research 56 (2002) 312–322
www.elsevier.com / locate / cardiores
Altered apoptosis pattern during pharyngeal arch artery remodelling is
associated with aortic arch malformations in Tgfb2 knock-out mice
¨ G.M. Molin a , Marco C. DeRuiter a , Lambertus J. Wisse a , Mohamad Azhar b ,
Daniel
Thomas Doetschman b , Robert E. Poelmann a , Adriana C. Gittenberger-de Groot a , *
a
Department of Anatomy and Embryology, Leiden University Medical Centre, P.O. Box 9602, 2300 RC Leiden, The Netherlands
b
Department of Chemistry and Microbiology, University of Cincinnati, Cincinnati, OH, USA
Received 18 December 2001; accepted 13 June 2002
Abstract
Objective: The morphogenetic process underlying the remodelling of the embryonic mammalian pharyngeal arch artery system is
unknown. Within this process, the right sixth, carotid ducts and the distal part of the dorsal aorta (right a-segment) regress. In order to
unravel the underlying mechanism we studied the role of apoptosis in the normal regression of pharyngeal arch artery segments and in a
mouse model that develops aortic arch malformations. Methods: Normal remodelling was studied in wild-type Swiss (CPBS) and altered
remodelling in the Tgfb2 2 / 2 compared to the Tgfb2 1 / 1 (Swiss / Bl6) strain using immunohistochemistry and morphometric analysis.
Results: During normal remodelling, apoptosis occurs in the mesenchyme surrounding pharyngeal arch arteries before regression starts.
With the onset of regression, apoptosis spreads from the mesenchyme to the media. Morphometric evaluation confirms the increase in
apoptosis in the actin-positive media of the disappearing segments. In Tgfb2 2 / 2, aberrant apoptosis was found in both fourth arch
arteries, whereas the right dorsal aorta lacks apoptosis associated with normal regression. Fourth arch hypoplasia is the main arch
abnormality. In the most severe case, the fourth arch is interrupted and the right dorsal aorta a-segment persists, giving rise to aortic arch
interruption type-B and an aberrant right subclavian artery. Conclusions: We have shown for the first time that specific vascular apoptosis
patterns accompany normal regression and that the incidence of apoptosis is selectively altered in the case of arch artery abnormalities in
Tgfb2 knock-out mice.
 2002 Elsevier Science B.V. All rights reserved.
Keywords: Apoptosis; Arteries; Congenital defects; Developmental biology; Growth factors
1. Introduction
The morphogenetic processes underlying pharyngeal
arch artery (PAA) remodelling from the symmetrical
configuration towards the unilateral left-sided aortic arch
have not been fully unravelled. The early embryonic
mammalian system consists of five paired arch arteries,
numbered I to VI from cranial to caudal. The fifth artery is
considered to be rudimentary or absent [1]. Congdons
schematic time scale models, representing a general overview of the origin, persistence and regression of specific
*Corresponding author. Tel.: 131-71-527-6691 / 6660; fax: 131-71527-6680.
E-mail address: [email protected] (A.C. Gittenberger-de Groot).
PAA segments, can be regarded as the foundation for the
descriptions used today [2].
Several mechanisms have been postulated to play a role
in the remodelling process. Hemodynamic factors, especially flow reduction, are considered to regulate the
regression of specific PAA segments [3,4]. The same
counts for morphogenetic factors. The arterial wall of the
fourth arches exclusively expresses the deformed paralogous group of Hox genes (Hox4 A–D) [5], whereas the sixth
arch specifically expresses Hox5B in the surrounding
mesenchyme [6]. Also, the cellular composition of the
PAA, especially neural crest cell (NCC) derived smooth
muscle cells (SMC) [7], can contribute, as NCC disturbance resulted in PAA abnormalities [8,9].
Time for primary review 22 days.
0008-6363 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved.
PII: S0008-6363( 02 )00542-4
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
We decided to take a novel approach and studied the
presence of vascular apoptosis during PAA remodelling.
Apoptosis is associated with both normal and defective
cardiogenesis [10,11], but data on apoptosis and PAA
remodelling are lacking [12,13].
We evaluated vascular apoptosis under normal and
abnormal conditions using wild-type (CPBS Leiden) and
Tgfb2 mutant mice (Swiss / Bl6), which present aortic arch
abnormalities and intra-cardiac defects [14,15].
2. Methods
313
ethanol, transferred to 100% xylene and embedded in
paraffin.
Consecutive transverse sections (5 mm) were stained
with Mayer’s hematoxylin (HE), the primary mouse
antibody against alpha-smooth muscle actin (a-SM-actin:
1A4 / M851; Dako, Denmark), or used for TdT-mediated
dUTP-X nick end labeling (TUNEL; Boehringer, Mannheim, Germany). Both a-SM-actin and TUNEL protocols
have been described elsewhere [11,16]. Apoptotic cells, as
determined by cell shrinkage, chromatin condensation and
DNA fragmentation, were studied in parallel HE- and
TUNEL-stained sections. The sections were studied by
light microscopy.
2.1. Immunohistochemical procedures
2.2. Incidence of apoptosis and morphometry
Wild-type Swiss CPBS Leiden mice were used to study
normal remodelling. Altered PAA remodelling was studied
using Tgfb2 mutant mice, which are derived from succeeding F-generations of Tgfb2 129 / Ola male chimeras
bred to Swiss / Bl6 females [14].
Pregnant mice were killed by cervical dislocation.
Detection of the vaginal plug was designated as day E0.5
of development. The procedures conform to the Guide for
the Care and Use of Laboratory Animals published by the
NIH.
Twenty-four wild-type Swiss CPBS Leiden embryos
were analysed for normal PAA remodelling. For abnormal
PAA development, 32 Tgfb2 2 / 2 and 16 Tgfb2 1 / 1
littermates were used. Developmental day E11.0–18.0
embryos were fixed in 4% paraformaldehyde / phosphatebuffered saline (0.1 mol / l, pH 7.2), dehydrated in graded
To describe apoptosis in regressing segments during
PAA remodelling, a morphometric analysis was used, as
general approaches using, for example, CASPASE-3 are
not applicable because separate segments cannot be isolated for biochemical evaluation [17].
The PAA system was divided into segments, as outlined
in Fig. 1. Dorsal aorta (DAo) segmentation in the a-, band g-segments was defined by their boundaries. The
proximal boundary of the b-segment is marked by the
sixth and fourth arches; the latter can be recognised by its
weaker a-SM-actin staining [16]. The distal boundary of
the b-segment is delineated by the subclavian artery. The
a-segment is proximally bordered by the subclavian artery
and distally by the fusion point with the other DAo. The
right sixth arch artery was defined as the vessel connecting
Fig. 1. Schematic representation of the normal mouse pharyngeal arch artery (PAA) system remodelling between E11.5 and E15.5. (a) The paired arterial
system at E11.5, the aortic arches III, IV and VI, the dorsal aorta (DAo) segments a, b and g and various connecting arteries. (b) At E12.5, the aortic sac
(AoSac) has remodelled into a separate ascending aorta (AAo) and pulmonary trunk (PT). (c) Slightly after E14.0 the coronary arteries (CoA) connect to
the AAo. (d) After E15.5 there is a mature configuration of the PAA. Note the presence of a double arch configuration until E12.5. Abbreviations: DesAo,
descending aorta; DA, ductus arteriosus; PA, pulmonary artery; RCA / LCA, right / left carotid artery; RSA / LSA, right / left subclavian artery.
314
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
the pulmonary trunk to the right DAo. At the stages
studied the pulmonary arteries connected directly to the
pulmonary trunk.
To analyse the difference in apoptosis between the right
(regressing) and left (persisting) segment of the sixth arch
artery and the dorsal aorta a-segment, respectively, the
number of apoptotic cells per a-SM-actin-positive vessel
wall volume (apoptose incidence: apoptotic cells / mm 3 ) of
each segment was estimated. Apoptotic cells, determined
by cell shrinkage and chromatin condensation, were scored
in HE-stained sections. Only apoptotic cells located in the
vessel wall of the segments were counted; apoptosis in the
mesenchymal compartment surrounding the vessels was
only described quantitatively as objective boundaries were
lacking. The volume of the vessel wall was estimated by
the method proposed by Cavalieri [18]. Between 12 and 23
(depending on developmental stage) a-SM-actin-stained
sections, taken systematically at equal distances, were used
to determine the vessel wall volume of the segments as
described by Bouman et al. [19]. The apoptotic cells were
counted in the adjacent HE-stained slides. At all stages the
selected sections enclosed the complete vascular segments
of the PAA system, and the counting analysis was repeated
three times.
As PAA remodelling is tightly regulated, a landmark for
vascular development was introduced that was more
applicable than the time point of conception or somite
stages. The ratio between the vessel wall volume of the
right regressing and the left persisting artery was used as
the landmark of vascular remodelling. For analysis of
normal PAA remodelling, a-SM-actin and HE sections of
nine E12.0–14.0 (distribution: 23 E12.0, 23 E12.5, 23
E13.0, 13 E13.5 and 23 E14.0) wild-type Swiss CPBS
Leiden embryos were used. To outline the difference in
apoptosis between the regressing and the persisting segments during remodelling, the ratio of apoptosis incidence
was plotted versus the ratio of vessel wall volume.
To estimate the number of apoptotic cells per vessel
wall volume of different PAA segments under abnormal
circumstances, four E14.5 embryos of both Tgfb2 2 / 2
and Tgfb2 1 / 1 genotypes were analysed morphometrically as described above. The mean apoptosis incidence and
standard deviation of each segment was obtained for all
embryos analysed. To analyse the apoptosis incidence per
segment between Tgfb2 2 / 2 and Tgfb2 1 / 1 embryonic
mice, a Mann–Whitney test with a P-value of 0.05 was
used.
3. Results
3.1. Remodelling of the normal pharyngeal arch artery
system
At developmental day 11.5 the PAA system (Fig. 1a)
consisted of a left and right third (III), fourth (IV) and
sixth (VI) arch artery, connecting both continuous DAo
with the ventrally located aortic sac. The latter had
separated into an ascending aorta and pulmonary trunk
around E12.0 (Fig. 1a and b). Between E11.5 and E14.0
the arterial system had developed towards the mature
left-sided configuration, due to regression of the rightsided sixth arch artery (R-VI), the right DAo a-segment
(R-a), and both the left and right carotid duct (L / R-g)
(Figs. 1a–d, 2 and 3).
The vascular segments revealed a comparable regression
process, showing a progressive decline of lumen diameter,
accompanied by a reduction of vessel wall thickness of the
right segment as compared to the left (Figs. 2 and 3). The
regression resulted in the formation of an a-SM-actinpositive strand (Figs. 2g and 3e) that soon became
interrupted at its midpoint. Finally, the proximal and distal
ends disappeared completely. Despite the morphologic
resemblance, the regression of the R-VI progressed faster
than the R-a segment
3.2. Location and incidence of apoptosis during
remodelling
At E10.5, before the start of regression, there is no
marked apoptosis in the mesenchyme surrounding the
vascular segments. As early as E11.5, apoptotic cells,
found in both the TUNEL- and HE-stained sections, were
preferentially located in the mesenchyme surrounding the
vascular segments that will regress (Figs. 2d and 3b).
Between E12.0 and E14.0 the localisation and number of
apoptotic cells changed for the regressing segments, i.e.
spreading from the surrounding mesenchyme (Figs. 2c,d
and 3a,b) towards the outer border of the a-SM-actinpositive media (Figs. 2e,f and 3c,d) and eventually into the
media. At the strand stage of regression (Figs. 2g,h and
3e,f), only a non-luminised a-SM-actin-positive cord of
cells was seen as a remnant of the former vessel.
Both TUNEL- and HE-stained sections gave comparable
results for the morphometric analysis; for merely practical
reasons, HE sections were applied. The difference between
Fig. 2. Transverse sections of the right (RDAo-a) and left (LDAo-a) dorsal aorta a-segment as found during developmental stages E10.5–11.0 (a,b),
E11.5–12.0 (c,d), E12.0–12.5 (e,f) and E12.5–13.0 (g,h). Sections (a), (c), (e) and (g) are stained for a-SM-actin, and adjacent sections (b), (d), (f) and (h)
are TUNEL-stained. (a,b) At E10.5–11.0 the size of the RDAo-a is at its maximum, but is nonetheless smaller than the LDAo-a. (b) No apoptosis is found
in the mesenchyme surrounding both dorsal aortae. (c,d) At E11.5–12.0, apoptotic cells are present in the mesenchyme surrounding the RDAo-a (arrow).
Note the apparent reduction of the lumen and the number of smooth muscle cells of the vessel between (e) and (g), showing an increase in left / right
differences. The actin-positive strands at the bifurcation level of both dorsal aortae, a-segments being visible in (c) and (e), are almost absent in (g). Note
the difference in wall thickness and lumen diameter, which is accompanied by a shift in apoptosis from the mesenchyme (arrows) surrounding the vessel
(d) towards the media (arrowheads) (f,h). Abbreviation: Oe, oesophagus. Bar 200 mm.
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
315
316
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
Fig. 3. Transverse sections of the right sixth arch artery (R-VI) segment as found during developmental stages E11.5–12.0 (a,b), E12.0–12.5 (c,d) and
E12.5–13.0 (e,f) and the left sixth arch artery (L-VI) at E12.5–13.0 (g,h). Sections (a), (c), (e) and (g) are stained for a-SM-actin and the adjacent sections
(b), (d), (f) and (g) for TUNEL. Note the spreading of apoptotic cells located in the mesenchyme surrounding the R-VI (arrows) at the onset of regression
(b), towards the outer border (d) into the media of the segment (f) (arrowheads). Section (f) shows the high incidence of apoptosis at the most distal part of
the R-VI strand and section (h) the low incidence for the corresponding L-VI, both at the level of the dorsal aorta b-segment (RDAo-b and LDAo-b,
respectively). Abbreviation: NX, nervus vagus. Bar 100 mm.
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
the triple countings was negligible (data not shown).
Morphometrical analysis of the apoptotic cells located in
the media (apoptosis incidence) of the regressing R-a and
R-VI segment in E12.0–14.0 embryos revealed a substantial increase as compared to the left counter part (Fig. 4).
This difference was greatest during the end stage of
remodelling and is comparable to the stage at which the
regressing vessel segments were remodelled into an a-SMactin-positive strand without a continuous lumen (see Figs.
2g,h and 3e,f).
The timing and apoptosis patterning of PAA remodelling
was comparable between wild-type CPBS Leiden and
Tgfb2 1 / 1 (data not shown).
3.3. Influence of Tgfb2 depletion on PAA remodelling
and apoptosis patterning
PAA abnormalities were found in 24 of 32 (75%)
Tgfb2 2 / 2 mouse embryos. Intra-cardiac malformations,
consisting of outflow tract and inflow tract septation
abnormalities, were encountered in all embryos from E13.5
and no isolated PAA malformations were observed in this
group. Within the time window E12.0–15.5, Tgfb2 2 / 2
mice had developed a spectrum of PAA anomalies ranging
from aortic arch hypoplasia to interruption. Before E12.0
(E11.0–11.5) no vascular differences were found between
Tgfb2 1 / 1 and Tgfb2 2 / 2 mice. At E12.0–13.0 and
E13.5–14.5, two of six and eight of 10 embryos, respectively, had developed tubular hypoplasia of the proximal
aortic arch and / or the more distally located fourth arch
artery segments (Fig. 5). Additionally, a substantial delay
of R-a regression was found in both E12.0–13.0 cases and
in one E13.5–14.5 embryo (shown schematically in Fig. 6,
317
panels 1 and 2), as shown by the lack of apoptotic cells
normally found in the media of this segment. The other
embryos (4 / 6 and 2 / 10) had developed mild vascular
hypoplasia. A marked number of apoptotic cells was
found, predominantly within the mesenchyme and media at
the basis of the fourth arch artery segment (Fig. 5c and d).
This eccentric patterning of apoptosis was never observed
in Tgfb2 1 / 1 littermates (Fig. 5a and b). Noteworthy is
the increased condensation of the mesenchyme surrounding the trachea and oesophagus, seen in the Tgfb2 2 / 2
phenotype (Fig. 5c).
To substantiate our findings, four Tgfb2 2 / 2 E14.5
mice with a variable degree of tubular hypoplasia of the
aortic arch were analysed morphometrically for their PAA
apoptosis incidence. The measurements confirmed a higher
apoptosis incidence for all PAA segments, being most
marked for the fourth arch arteries and the proximal aortic
arch as compared to four Tgfb2 1 / 1 embryos (Fig. 7). Of
the PAA segments, only the R-IV and L-IV arch arteries
revealed a significantly higher apoptosis incidence (both
P50.029) for the Tgfb2 2 / 2 mice (Fig. 7). All other
segments, with the exception of the proximal part of the
aortic arch (P50.057), were far above P50.05. The
apoptosis patterning characteristic for the normal regression of the R-a, with a spread from the surrounding
mesenchyme to the media, was absent in all cases in which
the R-a persisted (not shown).
At E15.5, three of six embryos had developed double
aortic arch interruption (L- and R-IV) and persistence of
R-a, resulting in a type-B aortic arch interruption accompanied by an aberrant right subclavian artery (Fig. 6c and
d). One embryo still possessed extremely thin remnants of
the former L- and R-IV pharyngeal arch (Fig. 6a and b).
All three E15.5 embryos with a R- and L-IV arch artery
interruption showed a left-sided system. Regression of the
R-VI segment had taken place normally in all cases. The
remaining three E15.5 embryos revealed a variable degree
of hypoplasia of the fourth arch artery segment.
All six E16.0–18.0 embryos showed vascular hypoplasia
at comparable locations as described for younger embryos.
No PAA-associated apoptosis could be discerned at these
stages.
4. Discussion
Fig. 4. Morphometric analysis of the R-a (n) and R-VI (j) regression in
nine E12.0–14.0 mouse embryos, given as the ratio right / left apoptosis
incidence versus the ratio right / left vessel wall volume (in percentage).
The increased apoptosis incidence parallels the enhanced decline of vessel
volume at the final stage of regression. Three embryos for each segment
were analysed at this stage.
The formation of the aortic arch and its tributaries from
the paired PAA system has been described for mammals
[1,2,16] and avian embryos [7,20]. We have shown for the
first time that apoptosis accompanies normal PAA remodelling and that alterations in this process coincide with
PAA malformations. It should be kept in mind that PAA
remodelling also depends on the proliferation, migration
and differentiation of multiple cell types (e.g. NCC,
318
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
Fig. 5. Comparable TUNEL-stained transverse sections of an E12.0–12.5 Tgfb2 1 / 1 (a,b) and an E12.0–12.5 Tgfb2 2 / 2 embryo (c,d). (a) Normal left
fourth (IV) arch segment with a low apoptosis incidence (arrows) in the surrounding mesenchyme (boxed area and (b)). Apoptosis ventrally (arrowheads)
of the trachea (T) is associated with regression of the right sixth arch artery (a). (c) Abnormal vascular development (compare dimensions RDAo, AAo and
IV between (a) and (c)) is apparent and most extreme for the left-IV arch artery. The left-IV in the knockout shows higher mesenchymal and vascular
apoptosis (arrowheads) (boxed area in (d)) than (b). A schematic representation of the PAA configuration of (c) can be found in Fig. 6 (panel 1).
Abbreviations: Oe, oesophagus; NX, nervus vagus; AAo, ascending aorta. Bars 100 mm.
endothelial and mesenchymal cells), which are not addressed in this study.
4.1. Normal pharyngeal arch artery remodelling and
apoptosis
The development of the mammalian PAA system into
the left-sided aortic arch configuration requires a tightly
regulated remodelling process involving regression of
various segments. The presence of mesenchymal and
media-located apoptosis, predominantly in specific rightsided segments, associated with the development of the
left-sided aortic arch system is evident. Intriguing is the
spreading of apoptosis with the onset of regression from
the mesenchyme towards the media at the final stages of
regression. Accompanying regression, we find apoptosis in
the surrounding mesenchyme and in the media. This outer
mesenchymal area has been referred to as an important
cellular source for investment of mesenchymal cells to the
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
319
Fig. 6. Panels 1–3: Schematic representation of the potential development of the aortic arch (AoA) interruption type-B and aberrant right subclavian artery
(ARSA) in Tgfb2 2 / 2. The lines correspond to the plain of sectioning in Figs. 5c and 6a–d, respectively. The double arch configuration (consisting of the
left and right dorsal aorta a-segments and both left- and right-IV segments) persists until E14.5 (1). Interruption of the right- and left-IV arch segment
occurs between E14.5 and E15.5, showing the eccentric regression of the left-IV arch segment (2 and 3). Transverse HE-stained sections (a–d) reflect two
E15.5 Tgfb2 2 / 2 mice. Remnants of the R- and L-IV arch artery segment are still present at early E15.5 (a,b). (a) The arrow indicates the caudal absence
of the L-IV, whereas in (a) and (b) a small remnant of the R-IV and L-IV segment is still found (overview 2). At late E15.5, no fourth remnants were found
(c, overview 3). The L-IV has regressed completely (arrow in c). (d) ARSA at the level of the hypoplastic right dorsal aorta a-segment (a) and its fusion
point with the left (LDAo), showing a clear retro-oesophageal course. Abbreviations: AAo, ascending aorta; DA, ductus arteriosus; Oe, oesphagus;
LCA / RCA, left / right carotid artery; T, trachea; NX, nervus vagus. Bars (d)–(f) 200 mm.
320
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
Fig. 7. Morphometric analysis of the incidence of apoptosis of the great arteries in four E14.5 Tgfb2 2 / 2 (grey bars) and four Tgfb2 1 / 1 (black bars)
mice. The mean apoptosis incidence (solid bars) and the standard deviation (error bars) per segment of both groups are depicted. The mean apoptosis
incidence per segment in the Tgfb2 2 / 2 mice is higher than for the Tgfb2 1 / 1 mice. A significant difference in apoptosis incidence (*P,0.05) was
found for the right and left fourth arch artery segments (R-IV and L-IV). R-a is not shown, as this segment was already absent at the stage presented.
Abbreviations: AAo, ascending aorta; PAoA, proximal part of the aortic arch; PT, pulmonary trunk; DA, ductus arteriosus; DesAo, left dorsal aorta
a,b-segment and descending aorta.
media [21]. The morphometrically evaluated spatio-temporal apoptosis incidence is considerable as several
pharyngeal arch arteries present different progressions of
the remodelling process; e.g. the R-VI being slightly faster
than R-a. The sequence of events starts at the midpoint of
the vessel and extends in the proximal and distal direction
during vessel regression.
The time in which apoptotic cells can be detected is
limited, and is often associated with the execution phase of
apoptosis [13,22], which is 6 h for TUNEL-positive cells
in vitro [22]. Moreover, there is in vivo apoptotic clearance, which in E11.0–13.0 mouse embryos takes 15–30
min [23,24]. Besides apoptosis, mitosis is a key factor in
cell dynamics. A small percentage of the total cell population of an embryo is in G1 and cell division for one cell
typically requires 8–16 h [25,26]. We regard vascular
regression to occur when apoptosis is not counterbalanced
by mitosis. The limited detection time of apoptotic cells
implies that, during the period of one cell cycle, 16–64
times more cells can be removed by apoptosis than added
by mitosis. Therefore, it is reasonable to assume that even
a small number of apoptotic cells can account for the
regression of a vessel.
In general, apoptosis during development is considered
to be a mechanism by which superfluous cells are removed. The mechanism driving differential apoptosis
patterns during PAA remodelling still remains elusive. So
far, a relation between apoptosis and vascular development
has only been reported for the development of the embryonic endothelial network [27]. For normal intra-cardiac
development, specifically the endocardial outflow tract and
atrio-ventricular cushions, numerous reports exist on
spatio-temporal apoptosis patterning, as reviewed by Poel-
mann et al. [13]. Apoptosis-related knockout models, e.g.
Caspase-8 and Fadd [28], manifest cardiac abnormalities,
underlining a functional relationship between apoptosis
and cardiovascular development.
Potential inductive and regulatory mechanisms of PAA
apoptosis could be flow-regulated, as found in programmed capillary regression [3]. The relation between
endothelial-mediated signals and PAA regression remains
to be discovered. With regard to the right-sided dominance
in PAA remodelling, we considered a relation with genes
that orchestrate left / right asymmetry. Left / right patterning
genes, such as Nodal, Lefty, and Sonic Hedgehog, are
promising candidates [29], but no link with apoptosis has
been proven. Several transcription factors, MSX2 in particular, do correlate with apoptosis and patterning, however
they reflect an anterior–posterior rather than left–right
asymmetry [30].
Cellular heterogeneity of the PAA system might contribute to the remodelling differences found. The most appealing example of cellular heterogeneity is the NCCderived SMC composition of the PAA system, revealing
strong boundaries between the pharyngeal arch arteries
(fourth and sixth) and the dorsal aorta [7,31]. An in vitro
study by Topouzis and Majesky showed a difference in
growth, apoptosis and differentiation between NCC and
mesodermally derived SMCs [32].
There is also heterogeneity within the NCC-derived
PAA as exemplified by the fourth arch artery, revealing a
poor a-SM-actin and elastin make-up [16], and extended
NCC-related CX43 expression [33] as compared to the
adjacent segments, but clear data that link these morphologic differences to a higher susceptibility for malformations is lacking.
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
4.2. Tgfb2 depletion and aortic arch remodelling
PAA development is hampered by Tgfb2 depletion,
giving rise to fourth arch artery and right dorsal aorta
a-segment anomalies. All other segments, i.e. the R-VI and
both g-segments, regress normally. The onset of these
vascular defects can be observed before intra-cardiac
defects are detectable [14]. This difference in Tgfb2
dependency is surprising, as TGFb ligands have been
reported to influence the cellular behaviour of virtually all
cellular participants of vascular development. Striking is
the opposite effect on the fourth arch and the R-a segment,
inducing regression of the former and persistence of the
latter. Both events correlated with an aberrant vascular
apoptosis incidence, being enhanced for the fourth and
reduced for the R-a.
This differential effect could be related to the vascular
composition, being predominantly NCC (fourth) or mesodermally (R-a) derived. In this context, TGFb has been
reported to exert opposite effects on the cellular fate of
SMC precursors, stimulating the proliferation and differentiation of NCC-SMC derivatives and inhibiting non-NCCrelated SMCs [32,34]. A relation between TGFb2 and
SMC apoptosis is lacking. Intriguing is the overlap between the eccentric apoptosis patterning found during
abnormal fourth arch remodelling and the signet ringshaped discontinuous a-SM-actin expression as described
by Bergwerff [16].
Future research will have to elucidate if TGFb2 acts
upon PAA remodelling in an instructive (differentiation)
and / or selective (apoptosis) order.
4.3. Relation to clinically and experimentally reported
aortic arch malformations
Kutsche and Van Mierop [35] associated aortic arch
interruptions with the anomalous origin of the right
subclavian artery. Their clinical study showed that 14 / 21
aortic arch interruption type-B (AAI-B) cases were associated with an aberrant right subclavian artery (ARSA).
Almost 50% of these patients carry a deletion of chromosome 22q11 [36], a genetic disorder that can give rise to
the DiGeorge and Velocardiofacial syndrome.
The AAI-B /ARSA and aortic arch hypoplasia found in
our Tgfb2 2 / 2 mice resulted from abnormal pharyngeal
arch remodelling of the R-a and fourth arch arteries
between E11.5 and E15.5. The coincidence of these
anomalies with a changed apoptosis pattern might point
towards an apoptosis-related process in the aetiology of
aortic arch defects.
Interruption and fourth arch artery hypoplasia are not
restricted to the Tgfb2 knock-out model, as they were also
present in mesenchyme fork head-1 (Mfh-1 ) [37], endothelin converting enzyme-1 (Ece-1 ), endothelin-A receptor
(EtA ) [38] and human 22q11 deletion syndrome homologous Df1 knockout mice [39]. Future research will have to
321
elucidate the downstream cellular and genetic targets of
TGFb2 during cardiovascular development.
In conclusion, asymmetric PAA remodelling coincides
with a highly spatio-temporal apoptosis pattern. Alterations
in this pattern are associated with PAA anomalies as
exemplified by the Tgfb2 mutant, giving rise to fourth arch
artery and R-a-related defects comparable to aortic arch
interruptions (type-B) and aberrant right subclavian arteries in men. The involvement of apoptosis in normal and
abnormal PAA development provides a new focus in the
research field of vascular development and in understanding the aetiology of PAA anomalies.
Acknowledgements
This research was supported by NIH grants HL58511
and HD26471 and Netherlands Heart Foundation grant
NHS46.014. The authors would like to thank Ron Slagter
of Inter Medics and Jan Lens for the graphics and layout.
References
[1] DeRuiter MC, Gittenberger-de Groot AC, Poelmann RE. The
special status of the pulmonary arch artery in the branchial arch
system of the rat. Anat Embryol 1989;179:319–325.
[2] Congdon ED. Transformation of the aortic arch system during the
development of the human embryo. Carnegie Inst Contr Embryol
1922;14:47–110.
[3] Meeson A, Palmer M, Calfon M, Lang R. A relationship between
apoptosis and flow during programmed capillary regression is
revealed by vital analysis. Development 1996;122:3929–3938.
[4] Moene RJ, Gittenberger-de Groot AC, Oppenheimer-Dekker A,
Bartelings MM. Anatomic characteristics of ventricular septal defect
associated with coarctation of the aorta. Am J Cardiol 1987;59:952–
955.
[5] Krumlauf R. Hox genes and pattern formation in the branchial
region of the vertebrate head. Trends Genet 1993;9:106–112.
[6] Kuratani SC, Wall NA. Expression of Hox 2.1 protein in restricted
populations of neural crest cells and pharyngeal ectoderm. Dev Dyn
1992;195:15–28.
[7] Bergwerff M, Verberne ME, DeRuiter MC, Poelmann RE, Gittenberger-de Groot AC. Neural crest cell contribution to the developing
circulatory system. Implications for vascular morphology? Circ Res
1998;82:221–231.
[8] Bergwerff M, DeRuiter MC, Gittenberger-de Groot AC. Comparative anatomy and ontogeny of the ductus arteriosus, a vascular
outsider. Anat Embryol 1999;200:559–571.
[9] Kirby ML, Hunt P, Wallis K, Thorogood P. Abnormal patterning of
the aortic arch arteries does not evoke cardiac malformations. Dev
Dyn 1997;208:34–47.
[10] Poelmann RE, Mikawa T, Gittenberger-de Groot AC. Neural crest
cells in outflow tract septation of the embryonic chicken heart:
differentiation and apoptosis. Dev Dyn 1998;212:373–384.
[11] Poelmann RE, Gittenberger-de Groot AC. A subpopulation of
apoptosis-prone cardiac neural crest cells targets to the venous pole:
multiple functions in heart development? Dev Biol 1999;207:271–
286.
´ S, Moorman AFM.
[12] van den Hoff MJB, Van den Eijnde SM, Viragh
Programmed cell death in the developing heart. Cardiovasc Res
2000;45:603–620.
322
D.G.M. Molin et al. / Cardiovascular Research 56 (2002) 312–322
[13] Poelmann RE, Molin D, Wisse LJ, Gittenberger-de Groot AC.
Apoptosis in cardiac development. Cell Tissue Res 2000;301:43–52.
[14] Sanford LP, Ormsby I, Gittenberger-de Groot AC et al. TGFb2
knockout mice have multiple developmental defects that are nonoverlapping with other TGFb knockout phenotypes. Development
1997;124:2659–2670.
[15] Bartram U, Molin DGM, Wisse LJ et al. Double-outlet right
ventricle and overriding tricuspid valve reflect disturbances of
looping, myocardialization, endocardial cushion differentiation, and
apoptosis in TGFb2-knockout mice. Circulation 2001;:2745–2752.
[16] Bergwerff M, DeRuiter MC, Hall S, Poelmann RE, Gittenberger-de
Groot AC. Unique vascular morphology of the fourth aortic arches:
possible implications for pathogenesis of type-B aortic arch interruption and anomalous right subclavian artery. Cardiovasc Res
1999;44:185–196.
[17] Watanabe M, Choudhry A, Berlan M et al. Developmental remodelling and shortening of the cardiac outflow tract involves myocyte
programmed cell death. Development 1999;125:3809–3820.
[18] Gundersen HJ, Jensen EB. The efficiency of systematic sampling in
stereology and its prediction. J Microsc 1987;147:229–263.
[19] Bouman HGA, Broekhuizen MLA, Baasten AMJ, Gittenberger-de
Groot AC, Wenink ACG. Stereological study of stage 34 chicken
hearts with looping disturbances after retinoic acid treatment:
disturbed growth of myocardium and atrioventricular cushion tissue.
Anat Rec 1997;248:242–250.
[20] DeRuiter MC, Gittenberger-de Groot AC, Poelmann RE, van Iperen
L, Mentink MMT. Development of the pharyngeal arch system
related to the pulmonary and bronchial vessels in the avian embryo.
Circulation 1993;87:1306–1319.
[21] Thayer JM, Meyers K, Giachelli CM, Schwartz SM. Formation of
the arterial media during vascular development. Cell Mol Biol Res
1995;41:251–262.
[22] Kockx MM. Apoptosis in the atherosclerotic plaque. Quantitative
and qualitative aspects. Arterioscler Thromb Vasc Biol
1998;18:1519–1522.
[23] Van den Eijnde SM, Boshart L, Reutelingsperger CPM, De Zeeuw
CI, Vermeij-Keers C. Phosphatidyl serine plasma membrane
asymmetry in vivo: a pancellular phenomenon which alters during
apoptosis. Cell Death Differ 1997;4:311–316.
[24] Van den Eijnde SM, Luijsterberg AJM, Boshart L et al. In situ
detection of apoptosis during embryogenesis with annexin V: from
whole mount to ultrastructure. Cytometry 1997;29:313–320.
[25] Goedbloed JF. The embryonic and postnatal growth of the rat and
mouse. Acta Anat 1974;87:209–247.
[26] Poelmann RE. Differential mitosis and degeneration patterns in
relation to the alterations in the shape of the embryonic ectoderm of
early postimplantation mouse embryos. J Embryol Exp Morphol
1980;55:33–51.
[27] Pollmann MJ, Naumovski L, Gibbons GH. Endothelial cell apoptosis in capillary network remodeling. J Cell Physiol 1999;178:359–
370.
[28] Vaux DL, Korsmeyer SJ. Cell death in development. Cell
1999;96:245–254.
[29] Capdevila J, Vogan KJ, Tabin CJ, Belmonte JCI. Mechanisms of
left–right determination in vertebrates. Cell 2000;101:9–21.
[30] Lumsden A, Graham A. Death in the neural crest: implications for
pattern formation. Semin Cell Dev Biol 1996;7:169–174.
[31] Jiang X, Rowitch DH, Soriano P, McMahon AP, Sucov HM. Fate of
the mammalian cardiac neural crest. Development 2000;127:1607–
1616.
[32] Topouzis S, Majesky MW. Smooth muscle lineage diversity in the
chick embryo. Two types of aortic smooth muscle cell differ in
growth and receptor-mediated transcriptional responses to transforming growth factor-b. Dev Biol 1996;178:430–445.
[33] Waldo KL, Lo CW, Kirby ML. Connexin 43 expression reflects
neural crest patterns during cardiovascular development. Dev Biol
1999;208:307–323.
[34] Garcia-Castro M, Bronner-Fraser M. Induction and differentiation of
the neural crest. Curr Opin Cell Biol 1999;11:695–698.
[35] Kutsche LM, van Mierop LHS. Cervical origin of the right
subclavian artery in aortic arch interruption: pathogenesis and
significance. Am J Cardiol 1984;53:892–895.
[36] Lewin MB, Lindsay EA, Jurecic V et al. A genetic etiology for
interruption of the aortic arch type B. Am J Cardiol 1997;80:493–
497.
[37] Iida K, Koseki H, Kakinuma H et al. Essential roles of the winged
helix transcription factor MFH-1 in aortic arch patterning and
skeletogenesis. Development 1997;124:4627–4638.
[38] Yanagisawa H, Hammer RE, Richardson JA et al. Role of
endothelin-1 / endothelin-A receptor-mediated signaling pathway in
the aortic arch patterning in mice. J Clin Invest 1998;102:22–33.
[39] Lindsay EA, Botta A, Jurecic V et al. Congenital heart disease in
mice deficient for the DiGeorge syndrome region. Nature
1999;401:379–383.