Modulation of Bmp4 signalling in the epithelial-mesenchymal interactions that take place in early thymus and parathyroid development in avian embryos.
ABSTRACT Epithelial-mesenchymal interactions are crucial for the development of the endoderm of the pharyngeal pouches into the epithelia of thymus and parathyroid glands. Here we investigated the dynamics of epithelial-mesenchymal interactions that take place at the earliest stages of thymic and parathyroid organogenesis using the quail-chick model together with a co-culture system capable of reproducing these early events in vitro. The presumptive territories of thymus and parathyroid epithelia were identified in three-dimensionally preserved pharyngeal endoderm of embryonic day 4.5 chick embryos on the basis of the expression of Foxn1 and Gcm2, respectively: the thymic rudiment is located in the dorsal domain of the third and fourth pouches, while the parathyroid rudiment occupies a more medial/anterior pouch domain. Using in vitro quail-chick tissue associations combined with in ovo transplantations, we show that the somatopleural but not the limb bud mesenchyme, can mimic the role of neural crest-derived pharyngeal mesenchyme to sustain development of these glands up to terminal differentiation. Furthermore, mesenchymal-derived Bmp4 appears to be essential to promote early stages of endoderm development during a short window of time, irrespective of the mesenchymal source. In vivo studies using the quail-chick system and implantation of growth factor soaked-beads further showed that expression of Bmp4 by the mesenchyme is necessary during a 24 h-period of time. After this period however, Bmp4 is no longer required and another signalling factor produced by the mesenchyme, Fgf10, influences later differentiation of the pouch endoderm. These results show that morphological development and cell differentiation of thymus and parathyroid epithelia require a succession of signals emanating from the associated mesenchyme, among which Bmp4 plays a pivotal role for triggering thymic epithelium specification.
- [show abstract] [hide abstract]
ABSTRACT: The development of the thymus depends initially on epithelial-mesenchymal and subsequently on reciprocal lympho-stromal interactions. The genetic steps governing development and differentiation of the thymic microenvironment are unknown. With the use of a targeted disruption of the whn gene, which recapitulates the phenotype of the athymic nude mouse, the WHN transcription factor was shown to be the product of the nude locus. Formation of the thymic epithelial primordium before the entry of lymphocyte progenitors did not require the activity of WHN. However, subsequent differentiation of primitive precursor cells into subcapsular, cortical, and medullary epithelial cells of the postnatal thymus did depend on activity of the whn gene. These results define the first genetically separable steps during thymic epithelial differentiation.Science 06/1996; 272(5263):886-9. · 31.03 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Deletion of chromosome 22q11, the most common microdeletion detected in humans, is associated with a life-threatening array of birth defects. Although 90% of affected individuals share the same three megabase deletion, their phenotype is highly variable and includes craniofacial and cardiovascular anomalies, hypoplasia or aplasia of the thymus with associated deficiency of T cells, hypocalcemia with hypoplasia or aplasia of the parathyroids, and a variety of central nervous system abnormalities. Because ablation of neural crest in chicks produces many features of the deletion 22q11 syndrome, it has been proposed that haploinsufficiency in this region impacts neural crest function during cardiac and pharyngeal arch development. Few factors required for migration, survival, proliferation and subsequent differentiation of pharyngeal arch neural crest and mesoderm-derived mesenchyme into their respective cardiovascular, musculoskeletal, and glandular derivatives have been identified. However, the importance of epithelial-mesenchymal interactions and pharyngeal endoderm function is becoming increasingly clear. Fibroblast growth factor 8 is a signaling molecule expressed in the ectoderm and endoderm of the developing pharyngeal arches and known to play an important role in survival and patterning of first arch tissues. We demonstrate a dosage-sensitive requirement for FGF8 during development of pharyngeal arch, pharyngeal pouch and neural crest-derived tissues. We show that FGF8 deficient embryos have lethal malformations of the cardiac outflow tract, great vessels and heart due, at least in part, to failure to form the fourth pharyngeal arch arteries, altered expression of Fgf10 in the pharyngeal mesenchyme, and abnormal apoptosis in pharyngeal and cardiac neural crest. The Fgf8 mutants described herein display the complete array of cardiovascular, glandular and craniofacial phenotypes seen in human deletion 22q11 syndromes. This represents the first single gene disruption outside the typically deleted region of human chromosome 22 to fully recapitulate the deletion 22q11 phenotype. FGF8 may operate directly in molecular pathways affected by deletions in 22q11 or function in parallel pathways required for normal development of pharyngeal arch and neural crest-derived tissues. In either case, Fgf8 may function as a modifier of the 22q11 deletion and contribute to the phenotypic variability of this syndrome.Development 11/2002; 129(19):4591-603. · 6.21 Impact Factor
Modulation of Bmp4 signalling in the epithelial–mesenchymal interactions that take
place in early thymus and parathyroid development in avian embryos
Hélia Nevesa,c,⁎, Elisabeth Dupina,1, Leonor Parreirac,d, Nicole M. Le Douarina,b
aCNRS UPR3294 Laboratoire Neurobiologie et Développement, Institut de Neurobiologie Alfred Fessard, Avenue de la Terrasse 91190 Gif-sur-Yvette Cedex, France
bCollège de France, 3 Rue d'Ulm 75231 Paris Cedex 05, France
cUnidade de Biologia da Hematopoiese, Instituto de Histologia e Biologia do Desenvolvimento, Edifício Egas Moniz, Piso 3, Ala C, Faculdade de Medicina da Universidade de Lisboa, Av.
Prof. Egas Moniz, 1649-028 Lisboa, Portugal
dInstituto Gulbenkian de Ciência, Rua da Quinta Grande, 2780-156, Oeiras, Portugal
a b s t r a c ta r t i c l ei n f o
Received for publication 17 May 2011
Revised 2 October 2011
Accepted 3 October 2011
Available online 28 October 2011
Epithelial–mesenchymal interactions are crucial for the development of the endoderm of the pharyngeal
pouches into the epithelia of thymus and parathyroid glands. Here we investigated the dynamics of epithelial–
mesenchymal interactions that take place at the earliest stages of thymic and parathyroid organogenesis using
the quail-chick model together with a co-culture system capable of reproducing these early events in vitro.
The presumptive territories of thymus and parathyroid epithelia were identified in three-dimensionally
preserved pharyngeal endoderm of embryonic day 4.5 chick embryos on the basis of the expression of Foxn1
and Gcm2, respectively: the thymic rudiment is located in the dorsal domain of the third and fourth pouches,
while the parathyroid rudiment occupies a more medial/anterior pouch domain.Using in vitro quail-chick tissue
associations combined with in ovo transplantations, we show that the somatopleural but not the limb bud
mesenchyme, can mimic the role of neural crest-derived pharyngeal mesenchyme to sustain development of
these glands up to terminal differentiation. Furthermore, mesenchymal-derived Bmp4 appears to be essential
topromoteearly stagesofendodermdevelopmentduringa shortwindowoftime,irrespectiveofthemesenchy-
mal source. In vivo studies using the quail-chick system and implantation of growth factor soaked-beads further
showedthatexpression ofBmp4 bythemesenchyme isnecessary during a 24 h-period oftime.After this period
however, Bmp4 is no longer required and another signalling factor produced by the mesenchyme, Fgf10, influ-
ences later differentiation of the pouch endoderm. These results show that morphological development and
cell differentiation of thymus and parathyroid epithelia require a succession of signals emanating from the
associated mesenchyme, among which Bmp4 plays a pivotal role for triggering thymic epithelium specification.
© 2011 Elsevier Inc. All rights reserved.
The formation of thymus and parathyroid glands results from a
series of complex epithelial–mesenchymal interactions taking place in
the pharyngeal region of the early vertebrate embryo (reviewed by
Grevellecand Tucker, 2010).The developmentof theseglandsis initiat-
ed with the budding off and outgrowth of rudiments from the foregut
endoderm of the pharyngeal pouches (PP). The single, endodermal
germ layer origin, of the thymic epithelium (TE) was demonstrated
using the quail-chick chimera system (Le Douarin and Jotereau, 1975).
In chicken, thymic and parathyroid organ rudiments derive from the
third and fourth PP (PP3/4), which then separate from the pharynx at
embryonic day-5 (E5) (HH-stage 26) (Hamburger and Hamilton,
1951). During this process, a thin mesenchymal capsule formed by
neural crest-derived cells surrounds the thymic rudiment and, at E6.5
(HH-stage 29), the colonization of the TE by lymphoid progenitor cells
genic micethatthemesenchymederived fromcardiac neuralcrest cells
forms the embryonic capsule and is associated with vasculature of
both fetal and adult thymus (reviewed by Foster et al., 2008 and
The importance of establishing functional cellular interactions
between the developing endoderm and the surrounding mesenchyme
in order to initiate thymic development was well illustrated in studies
using the quail-chick model. The PP3/4 endoderm isolated from early
Developmental Biology 361 (2012) 208–219
⁎ Corresponding author at: Unidade de Biologia da Hematopoiese, Instituto de Histologia
e Biologia do Desenvolvimento, Edifício Egas Moniz, Piso 3, Ala C, Faculdade de Medicina
da Universidade de Lisboa, Av. Prof. Egas Moniz, 1649-028 Lisboa, Portugal. Fax: +351 21
E-mail addresses: email@example.com, firstname.lastname@example.org (H. Neves),
email@example.com (E. Dupin), firstname.lastname@example.org (L. Parreira),
email@example.com (N.M. Le Douarin).
1Present address: Centre de Psychiatrie et Neurosciences, Inserm U894 Equipe Plasticité
Gliale, 2 ter Rue d'Alésia 75014 Paris, France.
0012-1606/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/developmentalbiology
quail embryos was able to develop into TE with the cooperation of a
heterologous mesenchyme derived from the somatopleure or splanch-
sive” to endoderm development. Furthermore, the grafted endoderm
was capable of inducing the heterologous mesenchyme to participate
in the formation of a fully developed thymus (Le Douarin, 1967a;
Le Douarin and Jotereau, 1975; Le Douarin et al., 1968). In contrast,
mesenchyme from the somite and limb bud was “non-permissive” to
PP3/4 endoderm development (Le Douarin, 1967a; Le Douarin et al.,
1968). These data provided first evidence that epithelial–mesenchymal
reciprocal interactions are essential for early thymic development;
moreover, they revealed that some heterologous mesenchymal tissues
are able to mimic the role played by neural crest-derived mesenchyme
during normal development of the thymus in the pharyngeal region.
Sonic Hedgehog (Grevellec et al., 2011; Moore-Scott and Manley,
2005) and signalling pathways belonging to families of the Bone
morphogenetic proteins (Bmps) (Bleul and Boehm, 2005; Gordon
et al., 2010; Soza-Ried et al., 2008) and Fibroblast growth factors
(Fgfs) (Dooley et al., 2007; Erickson et al., 2002; Jenkinson et al., 2003;
Revest et al., 2001) were reported to influence early thymic and para-
thyroid development. Moreover, Bmp and Fgf signalling pathways are
mutually regulated at later stages of thymic development and in the
adult thymus (Rossi et al., 2007; Tsai et al., 2003). It is however
unknown how these signals crosstalk during endoderm–mesenchyme
interactions in early thymic and parathyroid organogenesis.
In the mouse, expression of Foxn1 transcription factor identifies the
prospective TE (Gordon et al., 2001) and is required cell-autonomously
for its differentiation and colonization by lymphoid progenitor cells
(Blackburn et al., 1996; Bleul et al., 2006; Nehls et al., 1996). The
parathyroid rudiment is defined by expression of Gcm2 (Glial Cell
Missing 2) transcription factor (Gordon et al., 2001); when Gcm2 is
deleted, no parathyroid glands are formed (Gunther et al., 2000; Liu
et al., 2007). In chicken, Gcm2 expression starts in PP3 from E2.5 (HH-
stage 18) onwards and becomes evident in PP4 at E3.5 (HH-stage 22)
(Okabe and Graham, 2004). The temporal and spatial identification of
the presumptive TE in the chicken endoderm is still unknown.
In this work, we used the quail-chick model to define the molecu-
lar crosstalk between the endoderm of the PP3/4 and the mesen-
chyme, which controls the early stages of TE and parathyroid
epithelium (PTE) development. We first identified the endodermal
presumptive territories of the PTE and TE using probes for Gcm2
and Foxn1 genes and we determined the precise developmental
time-window of thymus specification. We also analysed the dynamics
of the interactions between PP endoderm and various mesenchymal
tissues using in vitro tissue co-cultures and in vivo grafting combined
with the quail-chick marker. The data show that development of the
PP3/4 endoderm into PTE and TE requires sequential production of
Bmp4 and Fgf10 factors by the local mesenchyme.
Materials and methods
Isolation of quail and chick embryonic tissues
Fertilised Japanese quail (Coturnix coturnix japonica) and chicken
(Gallus gallus) eggs were incubated at 38 °C in a humidified incubator
and embryos were dissected at specific times of development. Staging
of embryos was according to Hamburger and Hamilton stages
HH-stages in the quail. Isolation of PP3/4 endoderm was performed in
E2.5 (25 to 30 somite-stage; HH-stage 16–17), E3 (HH-stage 21) and
E3.5 (HH-stage 22) quail embryos and in E3 (HH-stage 20) to E4.5
(HH-stage 25) chick embryos as previously described (Le Douarin and
Jotereau, 1975, and Suppl. Fig. 1A). Briefly, the wall of the embryonic
pharynx was treated with a solution of pancreatin (8 mg/ml, Sigma)
for 30–90 min on ice, which allowed separation of pure endoderm
from the pharyngeal mesenchyme. Mesenchymal tissues of E2.5–E3
(HH-stages 18–19) chick embryos were dissociated from endodermal
and ectodermal tissues by enzymatic digestion with pancreatin using
the same procedure as above. Somatopleural and posterior limb bud
mesenchymal tissues were obtained from the embryonic territories at
the level of somites 19–24 and 25–30, respectively (Suppl. Fig. 1B).
In vitro tissue culture assay
PP3/4 endoderm isolated from E2.5 to E3 quail embryos was grown
alone or in association with mesenchymal tissues isolated from E2.5 to
E3 chick embryos (Fig. 5A). In brief, 2–3 endodermal explants were
combined with 2–3 mesenchymal explants on Nucleopore membrane
filters (Millipore) supported byfinemeshedmetal grids (Goodfellows).
The grids were then placed in culture dishes and partly immersed in
RPMI-1640 (Sigma) supplemented with 10% FBS and pen/step (control
culture medium). In some experiments, the heterospecific associations
were grown in culture medium supplemented with 100 ng/ml recom-
binant mouse Noggin (R&D Systems). Associated tissues were cultured
for48 hat37 °Cinahumidifiedincubatorcontaining5%CO2.Following
the incubation period, cultured tissues were either used for RNA isola-
bryos that were left to develop for further 10 days in a humidified
incubator at 38 °C as described (Le Douarin and Jotereau, 1975, and
schematic representation in Fig. 5A). Triplicates were obtained for
each culture condition analysed by RT-PCR.
Total RNA was extracted from quail endoderm isolated at different
stages of development and from cultured tissues using TRIZOL reagent
(Invitrogen). Reverse transcription was conducted with oligo-dT
primers (Promega) and SuperScriptIIreverse transcriptase (Invitrogen)
according to the instructions from the manufacturer. PCR was carried
out in 25 μl-reaction with a final concentration of 0.5 μM primers and
using the Phusion Master Mix with HF buffer (Finnzymes) according
to instructions from the manufacturer. Amplification was performed
using the following chicken (c) and quail (q) primer sets and annealing
conditions: cFoxn1 (forward: 5′-ggctctgaaccctgccaaga-3′, reverse: 5′-
ctgggaagacgttggggatg-3′, at 65 °C with a 707 bp product); cGcm2
cacca-3′, at 65 °C with a 349 bp product); cGAPDH (forward: 5′-
caatgggcacgccatcacta-3′, reverse: 5′-ctccagacggcaggtcaggt-3′, at 65 °C
with a 544 bp product); qGAPDH (forward: 5′-tgccaacccccaatgtctct-3′,
reverse: 5′-tggcctctcacagcaggatg-3′, at 65 °C with a 412 bp product).
PCR products were resolved on 1% agarose gels and visualised by
GelRed (VWR) staining. Gels were loaded with 10 μl of each PCR
product per lane. For the analysis of GAPDH expression, 1/10 of the
PCR product was analysed by Typhoon-cell imaging system 9210
(GE Healthcare) with ImageQuant software and then used as a
reference of cDNA amount in each sample. Samples of quail origin
exhibited a second cGAPDH product of 172 bp used as a reference for
the amount of quail cDNA in heterospecific co-cultures.
In vivo Noggin-bead implantation
night at 4 °C in a solution of 10 μg/ml recombinant mouse Noggin (R&D
Systems) in PBS (Noggin-beads) whereas control beads were soaked in
PBS. Soaked beads (100–120 μm in diameter) were implanted unilater-
ally in E3-chicken embryos near the pharyngeal arches. Specifically,
beads were primarily trapped in the pharyngeal clefts. In order to
increase the chance of hitting the pharyngeal arches, in some cases,
multiple beads were placed in the embryos.
H. Neves et al. / Developmental Biology 361 (2012) 208–219
Quail-chick chimera production by endoderm grafting
PP3/4 endoderm isolated from E2.5 (25–30 somite-stage) quail
embryos was grafted into the body wall of E2.5–E3 chicken embryo.
The endoderm was introduced into a slit made in the body wall of the
embryo in the somatopleural region (between somites 20 and 25), as
described previously (Le Douarin and Jotereau, 1975). Quail-chick
chimeras were then incubated for further 24 h, 48 h, 72 h or 10 days
night at 4 °C in 250 μg/ml recombinant mouse FGF10 (R&D Systems)
reconstituted in PBS with 0.1% BSA (Fgf10-beads) whereas control
beads were soaked in reconstitution solution.
Quail PP3/4 endoderm isolated at E3 (HH-stage 21) and E3.5 (HH-
tion into a slit made in the posterior limb bud. Chimeric embryos were
incubated for 3–4 h and then, FGF10- or control beads (100–120 μm in
diameter) were implanted near the grafted quail endoderm. Quail-chick
chimeras with the implanted beads were analysed after further 36 h
and 5 days of development.
Embryo processing and in situ hybridisation
Embryo and tissue explant cultures were fixed overnight with 4%
paraformaldehyde at 4 °C and processed for whole-mount in situ hybri-
disation and immunohistochemistry. Whole-mount preparations and
paraffin sections were hybridised with Foxn1, Gcm2, Bmp4 (Francis
et al., 1994), Bmpr1b, BmprII, and Fgf10 (Ohuchi et al., 1997) probes as
previously described (Etchevers et al., 2001; Henrique et al., 1995).
The probes for avian Foxn1 and Gcm2 were produced using TOPO-
cloning (Invitrogen). Templates were PCR-amplified for Foxn1 and
Gcm2 products using primers described above and designed from
GenBank database sequences ID: XM_415816 and ID: NM_001008480,
respectively. Whole-mount preparations and paraffin sections were
treated for immunocytochemistry with QCPN antibody (for labelling
of quail cells) as described (Creuzet et al., 2002) and with HNK1
(Abo and Balch, 1981) to detect cells of the peripheral nervous system.
Thymus and parathyroid gland rudiments can be identified in the avian
embryo based on the expression of Foxn1 and Gcm2, respectively
To identify the presumptive territories of thymus and parathyroid
glands in chicken and quail embryos, we devised new gene probes for
chicken Foxn1 and Gcm2 transcription factors, which in mammals, are
the first markers for TE and PTE, respectively (Gunther et al., 2000;
Nehls et al., 1996). Using whole-mount in situ hybridisation, Foxn1
was first detected bilaterally in the pharyngeal region of E4-quail
in the PP3/4 endoderm in quail (not shown) and chicken (cE4, Fig. 1B).
In E5-chicken (cE5), Foxn1 and Gcm2 transcripts occupied mutually
in the dorsal/anterior domain (Fig. 1C) and Gcm2 in a more ventral one
(Fig. 1D). In cE6, the expression domains of Foxn1 and Gcm2 marked
thymus (Fig. 1E) and parathyroid (Fig. 1F) rudiments, respectively,
and were later maintained in differentiated glands (Figs. 1G and H).
To precisely determinethespatial location/orientation of the TE and
PTEterritories,we isolated three-dimensionally-preservedPP3/4 endo-
derm in cE4.5. After in situ hybridisation, we observed strong signals of
Foxn1 in the dorsal domain (Figs. 2A–C) and of Gcm2 in the median/
anterior domain (Figs. 2D–F) of the epithelial pouches. These data
show that both presumptive TE and PTE territories are established in
cE4.5 PP3/4 endoderm; they occupy distinct and contiguous domains
during chicken endoderm development (as schematically represented
in Fig. 2G). We also determined the early stages of the expression of
Gcm2 and Foxn1 by RT-PCR analysis of PP3/4 endoderm isolated from
Fig. 1. Expression of Foxn1 and Gcm2 during thymic and parathyroid gland development
inchick and quail embryos. Foxn1 expression (A, C, E and G) detected by insituhybridisa-
tion in the base of the neck of qE4 (arrowhead, A), in the dorsal/anterior domain of PP3
and PP4 endoderm of cE5 (C, lateral view), in the thymic rudiment of cE6 (E, lateral
view) and in TE sections of cE18 showing a differentiated thymic lobe (G). Gcm2 expres-
sion (B, D, F and H) was analysed by in situ hybridisation in PP3 and PP4 of cE4 (arrow-
heads, B) and cE5 (D, lateral view), in the bilateral parathyroid rudiments of cE6
(F, ventral view) and in parathyroid glands of cE12 (G). In D, in situ hybridisation of
Gcm2 (purple) was followed by immunocytochemistry using HNK1 (brown) for neural
cell labelling. In C–F, hybridisation was performed after removing the ectodermal layer
ofthe neckregion.InH, neckglandsandnodosesensoryganglion were isolatedbymicro-
surgery.InC,thearchesandpouchesare delimitedusingredandblack lines,respectively;
NG, nodose ganglion; P, posterior; PP, pharyngeal pouch; Pt, parathyroid gland; Th,
thymus; V, ventral). Scale bars, 500 μm (A, B), 100 μm (G) and 50 μm (C–F and H).
H. Neves et al. / Developmental Biology 361 (2012) 208–219
somite-stage) while Foxn1 was detected only from E3.5 onwards
(Fig. 2H), indicating that endoderm specification into PTE occurs earlier
than into TE.
Expression patterns of Bmp4 and Bmp-signalling genes in the PP3/4
endoderm and surrounding mesenchyme during chicken development
In this work we investigated the role of Bmp4 signalling in the
en thymic and parathyroid gland development. We first characterized
the three-dimensional expression pattern of Bmp-signalling genes in
the endodermal and mesenchymal compartments during PP3/4
morphogenesis in the chicken embryo.
We analysed PP3/4 development in cE3 and cE3.5, i.e., before the
establishment of thymic and parathyroid presumptive endoderm terri-
tories (that is, before the onset of Foxn1 and Gcm2 expression in situ;
Figs. 1 and 2). When the endoderm was isolated in cE3 at the level of
PP2 to PP4, Bmp4 was expressed in the dorsal tip of all the pouches
(Fig. 3A) whereas Noggin was faintly detected throughout the pouch
territory (Fig. 3B). At the same stage, BmpRIb transcripts were present
in the anterior/dorsal domain of PP2 and PP3 (Fig. 3C) while BmpRII
showed ubiquitous expression in the pouches and pharynx as well as
ly formed PP3/4 (Fig. 3D). In whole-mount preparations of cE3.5
embryos, Bmp4 expression was found at the dorsal tip of the pouches
(Figs. 3F–H), with the exception of a strong hybridisation signal of
BmpRIb in the anterior/dorsal domain of the most recently formed PP4
(Fig. 3G), not observed at E3 (Fig. 3C). Sections showed strong hybridi-
ly formed PP3 (Fig. 3I) and PP4 (Fig. 3J) in cE2.5 and cE3.5, respectively.
Noggin was faintly detected in the pharyngeal region during PP3/4
formation (data not shown) and strongly expressed in the sensory
nodose ganglion located close to PP3/4 in cE4 (Fig. 3K).
We next studied the expression of Bmp-signalling genes in PP3/4
endoderm isolated in cE4.5 (Fig. 4), i.e., a stage in which we have iden-
tified both Foxn1+
thymic and Gcm2+
(see Fig. 2). Bmp4 was strongly expressed in the dorsal domain of PP3
(Figs. 4A and B) and in dorsal and posterior domains of PP4 (Figs. 4A
and C). Low expression of Noggin was detected in dorsal PP3 and
posterior PP4 (Figs. 4D–F). In addition, strong Bmp4 (Fig. 4A) and
Noggin (Fig. 4D) hybridisation signals marked the prospective territory
of PP5, known to participate in ultimobranchial body formation
(Le Douarin and Le Lièvre, 1970; Le Douarin et al., 1974). Noggin was
also detected in a narrow line of cells located in ventral/lateral
was expressed strongly in the most anterior and dorsal region of PP3/4
and, more weakly in the posterior domain of PP3 (Figs. 4G–I). In
contrast, BmpRII exhibited dim and ubiquitous expression in PP3/4
and PP5, as well as in the pharynx down to the branch point between
respiratory and digestive tubes (Fig. 4J).
Therefore, in the PP endoderm, the thymic rudiment displayed
ments exhibited high levels of BmpRIb transcripts. Noticeably, BmpRIb
expression in the endoderm coincided with a strong Bmp4 expression
in the adjacent arch mesenchymal cells, suggesting that Bmp4 pro-
duced by the neighbouring mesenchyme may act on PP3/4 endoderm
to promote its development into TE and PTE.
Invitro association of endoderm and mesenchymal tissues can reproduce
early epithelial–mesenchymal interactions leading to thymic and
parathyroid gland development
namic spatial modification of gene expression in the developing pha-
ryngeal region in vivo, we set up an in vitro experimental system to
analyse epithelial–mesenchymal interactions and the influence of
Bmp4 during the early formation of thymic and parathyroid glands.
We performed co-cultures of early PP endoderm and mesenchymal tis-
sues isolated from quail and chick embryos, respectively, to examine
subsequent gene expression and endoderm-derived tissue fate
(Fig. 5A). The quail pharyngeal endoderm was isolated between E2
and E3 (from 15- to 30-somite-stage) (Suppl. Fig. 1A) and associated
pleure or the posterior limb bud (Suppl. Fig. 1B). Both types of mesen-
chyme expressed Bmp4 (Suppl. Figs. 1C–D); however, as opposed to
the “non-permissive” limb bud, only the “permissive” somatopleural
mesenchyme was capable to sustain the development of the PP endo-
derm into a fully differentiated thymus, as previously shown in grafting
experiments (Le Douarin, 1967a; Le Douarin and Jotereau, 1975;
Le Douarin et al., 1968).
To validate our in vitro approach, we first verified if the tissue
interactions in endodermal–mesenchymal co-cultures could mimic
the early developmental events needed to pursue thymus and parathy-
roid gland formation. For this purpose, 48 h-cultures of quail endoderm
and chick somatopleural mesenchyme were explanted onto the chorio-
allantoic membrane (CAM) of cE8, which provides the conditions for
long-term growth of grafted tissues in ovo (Fig. 5A). After 10 days on
Fig. 2. EarlygeneexpressionpatternofFoxn1andGcm2inisolatedPP3/4endoderm.InsituhybridisationandcorrespondingschemesshowingFoxn1(A–C)andGcm2(D–F)expressionin
Foxn1 and Gcm2 expression domains in PP3 endoderm of cE4.5 (G). RT-PCR analysis of Foxn1, Gcm2 and GAPDH transcripts in quail PP3/4 endoderm isolated at different stages of
development from E2.5 to E4 (H). (A, anterior; D, dorsal; P, posterior; PP, pharyngeal pouch; V, ventral). Scale bars, 100 μm.
H. Neves et al. / Developmental Biology 361 (2012) 208–219
host CAM, the grafted co-cultures gave rise to fully developed thymus
andparathyroid glands(Fig.5B). Theyieldofglandformationincreased
withendoderm stage of thedonor embryos (Table 1). Eventhepharyn-
geal endoderm isolated as early as 15-somite-stage (approximately
60 h before Foxn1 starts to be expressed in situ) was able to form a full
thymus when associated with the somatopleural mesenchyme. Immu-
nostaining with quail-specific marker showed that the gland structures
formed in the graft comprised PTE and TE derived from the quail endo-
derm, whereas hematopoietic cells in the ectopic thymus were of chick
host origin (Fig. 5B). In agreement with previous chimera experiments,
no gland formation but necrosis was observed after implantation on
CAM of the PP endoderm associated with non-permissive limb bud
mesenchyme (not shown).
Therefore, the interactions between the PP endoderm and the
permissive mesenchymal tissue during this in vitro co-culture provide
appropriate early signals for long-term development of thymus and
Effectsof Noggin treatmentonFoxn1 activationduring invitro interactions
between quail PP endoderm and chick mesenchyme
We used the in vitro 48 h-tissue culture assay described above to
study early endodermal–mesenchymal interactions at two distinct
stages of endoderm development (qE2.5 and qE3), before the detection
of Foxn1 transcripts (see Fig. 2H). The presence of TE and PTE was
evaluated by RT-PCR analysis of Foxn1 and Gcm2, respectively.
In the associations of qE2.5 (25–30 somite-stage) PP3/4 endoderm
with somatopleural or posterior limb bud mesenchyme (cE2.5–cE3),
Foxn1 transcripts were upregulated after 48 h in control medium
(Fig. 5C). To study the role of Bmp4 signalling in TE specification and
(100 ng/ml). In contrast to control medium, when the co-cultures
were treated with Noggin, the induction of Foxn1 expression in the
presence of the mesenchymes was abolished (Fig. 5C). Gcm2 tran-
scripts, asopposed to those of Foxn1, were present when the endoderm
Fig. 3. Spatio-temporal expression of Bmp-signalling genes in PP3/4 before the onset of Foxn1 expression. In situ hybridisation of Bmp4 (A, E), Noggin (B, F), BmpRIb (C, G) and
BmpRII (D, H) in PP3/4 endoderm isolated from cE3 (and corresponding schemes, A–D) and in whole-mount preparations of cE3.5 (magnified arch region and corresponding
schemes, E–H). Frontal sections of the pharyngeal region in cE2.5 (I) and cE3.5 (J) after hybridisation with Bmp4. Frontal section of cE4 PP3/4 region hybridised with Noggin
(K). (A, anterior; D, dorsal; NG, nodose ganglion; P, posterior; PP, pharyngeal pouch; V, ventral). Scale bars, 100 μm (A–H) and 50 μm (I–K).
H. Neves et al. / Developmental Biology 361 (2012) 208–219
Fig. 4. Expression of Bmp-signalling genes in early thymic and parathyroid rudiments. In situ hybridisation and corresponding schemes showing transcripts of Bmp4 (A–C), Noggin (D–F), BmpRIb (G–I) and BmpRII (J) in the PP endoderm
isolated from cE4.5. Magnified images and corresponding schemes of PP3 (B, E and H) and PP4 (C, F and I) endoderm. (A, anterior; D, dorsal; P, posterior; PP, pharyngeal pouch; V, ventral). Scale bars, 100 μm.
H. Neves et al. / Developmental Biology 361 (2012) 208–219
was grown in the absence of mesenchyme. Moreover, Gcm2 expression
was maintained when the endoderm was cultured with somatopleural
or limb bud mesenchyme, in both control and Noggin-supplemented
media (Fig. 5C). However, a decrease of Gcm2 expression was observed
in Noggin-treated associations of endoderm and somatopleural mesen-
chyme, suggesting that reducing the levels of Bpm4 may interfere with
early PTE development in vitro (Fig. 5C). Therefore, during the 48 h-in
vitro period, the E2.5-endoderm was capable of sustaining Gcm2
expression in the absence of mesenchymal influence; however, it
required Bmp4 signals provided by association with mesenchymal
tissues, in order to become specified into Foxn1-expressing TE.
Similar to qE2.5, qE3 PP3/4 endoderm cultured for 48 h without
mesenchyme did not show Foxn1 expression by RT-PCR, while
expressing both Gcm2 and Foxn1 transcripts after 48 h-association
with the somatopleural mesenchyme (Fig. 5D). Surprisingly, addition
of Noggin (100 ng/ml) to the cultures of associated tissues did not
change Foxn1 and Gcm2 expression, when compared to co-cultures
grown in control medium (Fig. 5D). These results suggest that signals
produced by the mesenchyme (or the combined endoderm and mes-
enchyme) are essential for TE specification in both E2.5 and E3-PP en-
doderm. However, in contrast to early (E2.5) endoderm, in which the
inhibition of Bmp signalling by Noggin impaired the induction of
Foxn1, the older (E3) PP3/4 endoderm no longer required Bmp4 sig-
nals to upregulate Foxn1. Therefore, the effect of Bmp4 signalling in
regulating the specification of the PP endoderm into TE is temporally
restricted to a short period between qE2 and qE3.
Local application of Noggin interferes with early in vivo development of
PP3 endoderm into PTE
Our in vitro analysis suggested that inhibition of Bmp signalling in
co-cultures of qE2.5 endoderm with somatopleural mesenchyme
could affect the maintenance of Gcm2 expression during early develop-
we used an in vivo approach that consisted in implanting an exogenous
source of Noggin (Noggin-soaked bead) unilaterally, in the vicinity of
the PP3/4 endoderm in cE3 (n=3), when Gcm2 expression is faintly
detected by in situ hybridisation (not shown). Twenty-four hours after
bead implantation, we observed expression of Gcm2 in the anterior do-
main of PP3 in the control (non-implanted) side of the embryos
(Figs. 6B and D). As compared with the control side, Gcm2 expression
was strongly decreased in the contralateral pouch implanted with
Noggin-soaked bead (Figs. 6A and C). Control beads (without Noggin)
similarly placedin thearchregiondidnot perturb normal development
of PP3/4 (not shown). These results therefore show that the reduction
of local Bmp4 levels by Noggin interferes with the in vivo expression
of Gcm2 and the early development of PP3/4 endoderm into PTE.
Fig. 5. EffectsofmesenchymaltissuesandmodulatingBmp4signallinglevelsonearlystages
of thymic and parathyroid development in endoderm tissue culture. Schematic representa-
tion of the in vitro experiments wherein quail PP3/4 endoderm was associated with chicken
mesenchyme and cultured for 48 h (A). The co-cultured tissues were then either explanted
and grown in ovo for further 10 days onto the CAM of cE8, before immunostaining with
QCPN (B) or they were analysed by RT-PCR for Foxn1, Gcm2 and GAPDH transcripts (C, D).
(B) Section of 48 h-associated tissues explanted on CAM for 10 days, which gave rise to
fully developed parathyroid glands and thymus comprising epithelia of quail origin (QCPN,
brown staining) and vascular endothelium and connective tissues of chick origin (Gill's he-
matoxylin, blue staining); in the thymus, hematopoietic (basophilic) cells are also of chick
origin. (C) RT-PCR analysis of qE2.5 PP3/4 endoderm, alone or associated with the somato-
pleure or the limb bud mesenchyme of cE2.5–E3, after 48 h-culture in the presence or
absence of Noggin (100 ng/ml). (D) RT-PCR analysis of qE3 PP3/4 endoderm, grown alone
or associated with the somatopleural mesenchyme for 48 h. An additional and specific
quail-derived GAPDH transcript (172 bp) was amplified when using the chicken primer
sets. No Foxn1 or Gcm2 expression was detected in cultures of mesenchymal tissues in the
absence of endoderm (data not show). (CAM, chorioallantoic membrane; End, endoderm;
Mes, mesenchyme; Pt, parathyroid gland; Somat Mes, somatopleural mesenchyme; Th,
thymus). Scale bars, 100 μm (C–F) and 50 μm (E, H).
Differentiation of parathyroid glands and thymus from associated PP endoderm and
mesenchyme after explantation onto the CAM.
Tissues implanted on CAMa
PP-derived differentiated glands (%)b
Isolated endoderm Isolated mesenchymeParathyroid glands Thymus
aQuail PP endoderm isolated between 15- and 30-somite-stage (ss) was cultured in
vitro for 48 h, either alone or associated with chick mesenchyme of the somatopleure
or limb bud and then grafted onto the CAM of cE8.
bAfter 10 days, the formation of endoderm-derived parathyroid and thymic glands
was analysed by histochemistry and immunochemistry of CAM sections. Data are
expressed as the number (and percent) of the tissue cultures that formed parathyroid
glands and thymus on CAM. (PE, pharyngeal endoderm).
Fig. 6. In vivo effect of an ectopic source of Noggin on the expression of Gcm2 in the
developing PP3/4 endoderm. A Noggin-soaked blue agarose bead was unilaterally placed in
the vicinity of PP3/4 endoderm of cE3 (A and C). After 24 h, the embryos were processed
for in situ hybridisation with Gcm2. Magnified images of the PP3/4 region in whole-mount
embryos (A, B) and corresponding post-hybridisation sections (C, D). (A, anterior; D,
dorsal; P,posterior; PP, pharyngeal pouch; V, ventral).Scale bars, 100 μm (B) and50 μm (D).
H. Neves et al. / Developmental Biology 361 (2012) 208–219
Fig. 7. EctopicthymusdevelopmentinvivoaftergraftingofquailPP3/4endoderminto chicksomatopleuralmesenchyme.Schematicrepresentationshowing thegraftingofisolatedquail
PP3/4 endoderm(qE2.5)into thesomatopleural regionofcE2.5–E3host(A). Transverse sectionsof chickhost embryos 24 h(B–D), 48 h(E–J)and72 h (K) afterendodermgrafting,were
processed for in situ hybridisation with Bmp4 (B–G), Fgf10 (H–J) and Foxn1 (K) probes. Serial sections are shown in (E–G) and (H–J). Arrowheads indicate strong hybridisation signals.
Transverse sections of chick host embryos 10 days post–grafting showing ectopically developed thymus (with clear discrimination between cortical and medullary compartments)
detected between the ribs (L, M). Only TE cells are quail endoderm-derived and positive for QCPN staining (brown); lymphoid and mesenchymal cells of host origin are evidenced by
Gill's hematoxylin contrast (blue staining) (M). (End, endoderm; Gr, graft; HE, hematoxylin–eosin; Mes, somatopleural mesenchyme; Rb, ribs). Scale bars, 100 μm.
H. Neves et al. / Developmental Biology 361 (2012) 208–219
Heterospecific grafting experiments show that endoderm–mesenchyme
interactions prior to TE specification involve sequential production of
Bmp4 and Fgf10 by the mesenchyme
In order to further understand the role of Bmp4 signalling and the
respective contribution of endodermal and mesenchymal signals in
early TE development in vivo, we performed heterospecific graft exper-
iments, as previously described (Le Douarin, 1967a; Le Douarin and
Jotereau, 1975; Le Douarin et al., 1968). As depicted in Fig. 7A, we
implanted the PP endoderm isolated from qE2.5 into the permissive
somatopleural mesenchyme of cE2.5–E3 and subsequently examined
Bmp4 expression in the developing grafted region. Host embryos 24 h
after endoderm implantation (n=3) showed high levels of Bmp4
expression and increased size of the somatopleural mesenchyme in
contact with the grafted endoderm (Figs. 7B and D), whereas mesen-
chymal cells in the control side exhibited low Bmp4 expression and
loose cellular arrangement (Fig. 7C). In contrast, when the qE2.5 PP
endoderm was implanted into the chick somite for 24 h, this
non-permissive mesenchyme was unable to upregulate Bmp4 activity
in response to the presence of the implanted endoderm (n=2)
(not shown). In recipient embryos 48 h after grafting of the qE2.5 PP
endoderm into the permissive somatopleural mesenchyme (n=4),
we found down-regulation of Bmp4 expression in the mesenchyme
surrounding the graft, accompanied by an increase of Bmp4 transcripts
in the grafted endodermal epithelium (Figs. 7E–G).
is temporally modulated in response to the presence of the implanted
endoderm, being first increased after 24 h-interaction and decreased
Fig. 8. In vivo effect of an ectopic source of Fgf10 on the expression of Foxn1 by PP3/4 endoderm grown in the limb bud territory. Transverse sections of chick host embryos 5 days
(A–I) and 36 h (J, K, M and N) after grafting quail E3 and E3.5 endoderm, respectively. Fgf10-beads (A, D, G, J and M) or control-beads (B, E, H, K and N) were implanted together
with the endoderm into the posterior limb bud (A, B, D, E, G, H, J, K, M and N) and somatopleural (C, F and I) regions of recipient embryos. Transverse serial sections were processed
for in situ hybridisation with Foxn1 probe (A–C, J and K) and staining with QCPN antibody and Gill's hematoxylin (D–I, M and N). Transverse section of the limb bud in cE4.5
hybridised with Fgf10 (L). Arrowheads indicate hybridisation signals (J–L). (c, cartilage; lv, liver; m, muscle; v, vessel; 5d, 5 days). Scale bars, 100 μm.
H. Neves et al. / Developmental Biology 361 (2012) 208–219
one day later, as the grafted PP3/4 endoderm develops. These data
raised the question of the identity of the mesenchymal signals able to
support further TE development, at stages when the mesenchyme
shows low Bmp4 activity. In the mouse, Fgf10 is expressed by mesen-
chymal cells of the pharyngeal arches (Revest et al., 2001) while Fgf
and Bmp signalling pathways are mutually regulated in late thymic or-
whether Fgf10 could be one of the mesenchymal factors able to pro-
mote thymic development of the PP endoderm in our grafting assays.
In chicken host embryos 24 h after grafting the qE2.5 PP endoderm
(n=3), Fgf10 transcripts were not detected in the graft area (data not
shown); however, 48 h post-operation, Fgf10 expression was present in
the implanted endoderm and the surrounding mesenchyme (Figs. 7H–J,
n=3). The permissive somatopleural mesenchyme therefore responded
to endodermal instructive signals by sequentially producing Bmp4 and
Fgf10, suggesting that, after initial involvement of mesenchymal Bmp4,
Fgf10 might substitute for Bmp4 to sustain later development of the
endoderm into a Foxn1+TE rudiment. Moreover, modulation of Bmp4
and Fgf10 expression in the mesenchymal compartment occurred one
day prior to endoderm specification into TE, since Foxn1 started to be
expressed by the quail endoderm only 72 h after grafting (Fig. 7K).
Further incubation of operated embryos until E13 (10 days post-
grafting) allowed development of a well-differentiated ectopic thymus
in similar grafting experiments (Le Douarin, 1967a; Le Douarin and
Jotereau, 1975; Le Douarin et al., 1968).
Heterospecific endoderm grafting combined with bead implantation
show that Fgf10 favours TE specification
Our data suggested that, following the early influence of Bmp4,
mesenchymal-derived Fgf10 might sustain later development of the
endoderm into a Foxn1+TE rudiment. To further address this question,
we investigated whether an exogenous source of Fgf10 could influence
the fate of the endoderm (older than E2.5) when grafted in a non-
permissive environment. For this purpose, qE3–E3.5 PP endoderm and
Fgf10-soaked beads were implanted together into the E3-chick posteri-
5 days after the operation.
E3-quail endoderm grafts analysed after 36 h, showed faint expres-
sion of Foxn1 in only one of the host embryos treated with Fgf10-beads
(n=5) and no expression in those implanted with control beads
(n=4) (data not shown). However, 5 days post-grafting, the quail en-
doderm showed strong Foxn1 expression in 3 of 4 host embryos
implanted with Fgf10-beads (Fig. 8A), as compared to 2 of 4 control
host embryos (Fig. 8B). Additionally, although expressing Foxn1 in this
non-permissive environment, the quail epithelium in both conditions
showed no colonization by chick lymphoid progenitor cells (Figs. 8D,
E, G and H). Conversely, colonization of Foxn1+thymic epithelium
occurred when the endoderm was grafted in the permissive mesen-
chyme of the somatopleural region (Figs. 8C, F and I).
When the quail endoderm was isolated at later stage (E3.5), Foxn1
expression could be identified 36 h after grafting into the limb bud
mesenchyme of E3-chick host embryos. In the presence of Fgf10-
beads, 5 of 6 recipient embryos exhibited strong Foxn1 hybridization
signal in the endoderm (Figs. 8J and M); in contrast, weaker Foxn1
expression occurred in only 2 of 6 embryos implanted with control
in the limb bud mesenchyme of E4.5 chick embryos (Fig. 8L), which
suggests that Fgf10 produced locally by the non-permissive mesenchy-
mal cells can support, at least in part, the maintenance of Foxn1 expres-
sion by the endoderm.
These results show that Fgf10 acts on E3.5 quail PP3/4 endoderm
to promote its development into TE.
Pharyngeal organs such as the thymus and parathyroid glands
mesenchyme, implying dynamic reciprocal signalling between the two
tissues. Here we investigated the role of mesenchymal tissues in PP3/4
endoderm development in the avian embryo, taking advantage of the
capacity to isolate three-dimensionally-preserved endoderm at early
developmental stages. Quail-chick chimeras were constructed in vivo
and in vitro by associating isolated quail endoderm with chick mesen-
chyme from different sources, such as that of the somatopleure and
posterior limb bud. We present evidencefor the role of Bmp4 produced
by the mesenchyme in the epithelial–mesenchymal interactions taking
place at early stages of thymic and parathyroid gland development.
Bmp4 produced by the mesenchyme was able to sustain Gcm2 expres-
sion in the presumptive parathyroid epithelium. Moreover, mesenchy-
mal derived Bmp4 was critical for further development of the thymic
rudiment during a short period of time, between E2 and E3 in the
quail (correspondingtoE2.5–E3.5 in the chick). Later on,the endoderm
itself took over the production of Bmp4. This later step coincided with
the onset of Fgf10 expression in the mesenchyme. Sequential produc-
tion of Bmp4 and Fgf10 by the mesenchyme thus results in Foxn1 activ-
ity in the endoderm. In the non-permissive limb bud environment, the
grafted endoderm expressed Foxn1 but exhibited a block or delay in
lymphoid cell colonization.
Identification of the presumptive endodermal territories of thymus and
parathyroid gland epithelia in the chicken embryo
In chicken, the rudiments of the thymus and parathyroid glands
develop from the PP3/4 endoderm. In this work, we determined the
spatial distribution and position of these rudiments in the three-
dimensionally-preserved PP endoderm, using combined techniques of
whole-mountinsituhybridisationand enzymatic isolationof embryon-
ic tissues. We first observed the expression of Gcm2 followed by that of
Foxn1, which identify parathyroid and thymic rudiments, respectively.
Moreover, Foxn1 expression in the prospective thymus rudiment was
preceded by Bmp4 expression in the developing PP3/4. Although these
data are similar to those reported in the mouse (Gordon et al., 2001;
Patel et al., 2006), the domain of Foxn1 expression is inverted along
the dorsal–ventral axis of the pouches in chicken and quail, when com-
pared to mouse embryos. These distinct positions during embryogene-
sis might contribute to the different anatomical locations of the adult
Fig. 9. Schematic representation of the anatomic location of foregut endoderm-derived
glands in adult chicken. (CB, carotid body; Es, esophagus; NG, nodose ganglion; Pt,
parathyroid glands; Th, thymus; Tr, trachea; Ty, thyroid; UBB, ultimobranchial body).
H. Neves et al. / Developmental Biology 361 (2012) 208–219
thymus between mammals and birds. Accordingly, in mammals, the
thymus (in a ventral early position in the pouch; Gordon et al., 2001)
is located in upper anterior position in the chest cavity just above the
heart, while in the chicken and quail, it initially forms in a dorsal posi-
tioninthepouchand, later,becomes bilaterallylocatednearthejugular
vein and along the neck (Fig. 9).
Role of the mesenchyme during cellular interactions with PP3/4 endoderm
to promote thymic and parathyroid gland development
Our tissue co-culture experiments have shown that ectopic mesen-
chymal tissues of the somatopleure or limb bud can provide the early
(qE2.5) PP endoderm with signals required for setting up the
expression of TE and PTE early markers. However, in contrast to the
limb bud, only the somatopleural mesenchyme was able to promote
full differentiation of TE and PTE leading to gland formation after trans-
plantation on CAM. Furthermore, Foxn1+epithelium grown in the
non-permissive mesenchymal environment showed no colonization
by lymphoid progenitors, suggesting that specification of TE and its
colonization by hematopoietic cells can occur independently during
Our data showingthat onlythe somatopleuralmesenchyme is capa-
ble of coordinated long-term development of endoderm indicate the
need of continuous molecular crosstalk between these tissues during
pharyngeal gland organogenesis. In the embryo, the PP3/4 endoderm
develops in close contact with Bmp4-expressing mesenchyme of the
arches, suggesting that, in the graft paradigm, the capacity of ectopic
mesenchymal tissues to sustain early development of these glands
may depend on their ability to produce Bmp4. Indeed, Bmp4 is
expressed by both somatopleural and limb bud mesenchymal tissues
when they are isolated and associated in vitro with the endoderm.
Furthermore, Bmp4 expression is upregulated in the somatopleural
mesenchyme after in vivo grafting of the PP3/4 endoderm into the
body wall. Considering that the presumptive territories of the glands
express BmpR1b and BmpRII and, that Bmp2/4/7 preferentially signal
through heterodimeric complexes of BmpR1a (or 1b) and BmpRII
(Feng and Derynck, 2005), altogether our data strongly suggest that
the establishment of the prospective domains of TE and PTE initially
depends on mesenchymal-derived Bmp4 signals.
driving specification of TE and PTE: involvement of mesenchymal-derived
Bmp4 and Fgf10
Our analysis of early endoderm–mesenchyme interactions in vitro
revealed that Bmp4 produced by mesenchymal cells is capable of
sustaining early development of PP3/4 endoderm into TE (Foxn1
expression) and PTE (Gcm2 expression). The requirement for Bmp
signalling varied with endoderm developmental stages: the Bmp
signalling inhibitor Noggin thus interfered with activation of Foxn1 in
qE2.5, but not in qE3-endoderm grown in vitro in association with
mesenchymal tissues. Although some tissue-autonomous effect of
Bmp signalling in the endoderm itself could not be excluded, the levels
of Bmp4 signalling within the PP endoderm seemed insufficient or not
properly maintained above-threshold levels, to induce Foxn1 expres-
sion. In agreement, Bmp4-conditional deletion driven by Foxg1-Cre
endoderm is dispensable for TE and PTE specification (Gordon et al.,
2010). Therefore our results argue that Bmp4 signalling of mesenchy-
mal origin acts during an early and narrow window of time in order
to control TE specification. In addition, we have shown that inhibition
of Bmp signals also interfered with early development of PTE: adminis-
tration of exogenous Noggin caused a decrease of Gcm2 expression in
the PP3/4 endoderm (qE2.5) both in vivo (after Noggin-bead implanta-
tion) and in vitro (after treatment of endoderm-somatopleural mesen-
Together, our results provide evidence that mesenchymal-derived
Bmp4 signalling controls endoderm expression of Foxn1 and Gcm2,
two master genes of thymus and parathyroid development, respective-
ly. Our results are thus in apparent discrepancy with recent data show-
ing a normal pattern of expression of Foxn1 and Gcm2 after targeted
mutation of Bpm4 in the early PP endoderm and adjacent mesenchyme
in Foxg1-Cre;Bmp4 mice (Gordon et al., 2010). However, these mice
exhibited incomplete and variable deletion of Bmp4 in the pharyngeal
mesenchyme prior to the onset of Foxn1 expression in the endoderm
(Gordon etal., 2010), whichraises thepossibilitythat residual Bmp sig-
nallingin the mesenchyme could have been sufficient to trigger normal
specification of the endoderm.
Using heterospecific grafting of the PP endoderm into permissive
mesenchyme we were, for the first time, able to follow the in vivo dy-
namics of the reciprocal signals between the two tissues. We first no-
ticed a temporal regulation of Bmp4 expression in the mesenchymal
compartment, in which Bmp4 activity increased 24 h after endoderm
grafting, then decreased one day later. This suggests that the Bmp4
levels need to be tightly regulated in the developing pouches. In
agreement, Bmp4 gene deletion in the mouse pharyngeal epithelia in-
duced apoptosis in mandibular mesenchyme (Liu et al., 2005). Addi-
tionally, the impairment of Bmp signalling inhibition led to defects
in pharyngeal arch development and hypoplasia of thymus and para-
thyroid glands (Bachiller et al., 2003; Stottmann et al., 2001).
and adjacent mesenchyme involved a sequential expression of Bmp4
and in adulthood requires Bmp4-regulated Fgf10 signalling (Tsai et al.,
nalling in thymic development occurs earlier than previously reported.
ling molecules during tissue interactions in thymic and parathyroid
development,whichemphasizes thehighlydynamic temporaland spa-
tial dialogue between the PP endoderm and mesenchyme (Fig. 10). The
signals from the early PP3/4 endoderm, which can induce the mesen-
chyme to participate in thymus and parathyroid gland formation,
remain to be identified. Our experiments clearly show that the first sig-
nal for the onset of thymic and parathyroid development arises from
the endoderm, as the grafted pharyngeal endoderm is able to recruit
the somatopleural mesenchyme in order to trigger its maturation into
a fully developed glandular tissue. Similarly, previous experiments
using the quail-chick model have shown that any type of somato-
splanchnopleuralmesenchymeis abletodevelop intodenseconnective
tissues, capsule or multilayer smooth muscle, according to the
endodermal-derived organ formed following the grafting of early
quail pharyngeal endoderm into the body wall of a chick embryo
(Le Douarin, 1967a, 1967b; Le Douarin et al., 1968). One molecular
candidate for the first instructive signal derived from the endoderm is
Fgf8, which is expressed in the posterior domain of the pouches and
excluded from TE and PTE presumptive territories in chicken
(not shown). Moreover, Fgf8 is known to be required for pharyngeal
Fig. 10. Schematic model for Bmp4 and Fgf10 signalling crosstalk during epithelial–
mesenchymal interactions in early thymic and parathyroid development. The temporal
sequence of molecular changes is depicted for both endodermal and mesenchymal
compartments. Arrows indicate putative signalling crosstalk involved in the epitheli-
al–mesenchymal dialogue (see Discussion for details).
H. Neves et al. / Developmental Biology 361 (2012) 208–219
arch development and to regulate Fgf10 expression in pharyngeal arch
mesenchyme in the mouse embryo (Abu-Issa et al., 2002; Frank et al.,
2002). Another candidate is Noggin, which is expressed, albeit faintly,
in the posterior domain of chicken PP3/4 endoderm and may synergize
with Fgf8, as described in mouse mandible development (Liu et al.,
2005; Stottmann et al., 2001). Further experiments aimed at modifying
gene activity in the endoderm will help in defining the early instructive
signals emanating from the PP endoderm.
Supplementary materials related to this article can be found online
The authors are grateful to Dr. Nohno (Department of Molecular and
Developmental Biology, Kawasaki Medical School, Japan) for the Bmpr1b
and BmprII chicken probes, to Interaves Portugal for contributing with
quail fertilised eggs, to S. Morgado and V. Proa for technical assistance
and to G. Calloni, R. Garcez and A. Cidadão for helpful discussions. This
work was supported by Centre National de la Recherche Scientifique
(CNRS), Association pour la Recherche contre le Cancer (ARC), Fondation
Bettencourt-Schueller and Fundação para a Ciência e Tecnologia (FCT),
and partially conducted at the facilities of the Instituto de Medicina
Molecular, Lisbon. H.N. received research fellowships from the Institut
Curie and the Fondation pour la Recherche Médicale (FRM).
Abo, T., Balch, C.M., 1981. A differentiation antigen of human NK and K cells identified
by a monoclonal antibody (HNK-1). J. Immunol. 127, 1024–1029.
Abu-Issa, R., Smyth, G., Smoak, I., Yamamura, K., Meyers, E.N., 2002. Fgf8 is required for
pharyngeal arch and cardiovascular development in the mouse. Development 129,
Bachiller, D., Klingensmith, J., Shneyder, N., Tran, U., Anderson, R., Rossant, J., De Robertis,
E.M., 2003. The role of chordin/Bmp signals in mammalian pharyngeal development
and DiGeorge syndrome. Development 130, 3567–3578.
Blackburn, C.C., Augustine, C.L., Li, R., Harvey, R.P., Malin, M.A., Boyd, R.L., Miller, J.F.,
Morahan,G., 1996. The nu gene acts cell-autonomously and is required for differenti-
ation of thymic epithelial progenitors. Proc. Natl. Acad. Sci. U. S. A. 93, 5742–5746.
Bleul, C.C., Boehm, T., 2005. BMP signaling is required for normal thymus development.
J. Immunol. 175, 5213–5221.
Bleul, C., Corbeaux, T., Reuter, A., Fisch, P., Mönting, J.S., Boehm, T., 2006. Formation of a func-
tional thymus initiated by a postnatal epithelial progenitor cell. Nature 441, 992–996.
Creuzet, S., Couly, G., Vincent, C., Le Douarin, N.M., 2002. Negative effect of Hox gene ex-
Dooley, J., Erickson, M., Larochelle, W.J., Gillard, G.O., Farr, A.G., 2007. FGFR2IIIb signaling
regulates thymic epithelial differentiation. Dev. Dyn. 236, 3459–3471.
Erickson, M., Morkowski, S., Lehar, S., Gillard, G., Beers, C., Dooley, J., Rubin, J.S.,
Rudensky, A., Farr, A.G., 2002. Regulation of thymic epithelium by keratinocyte
growth factor. Blood 100, 3269–3278.
Etchevers, H.C., Vincent, C., Le Douarin, N.M., Couly, G.F., 2001. The cephalic neural crest
provides pericytes and smooth muscle cells to all blood vessels of the face and fore-
brain. Development 128, 1059–1168.
Feng, X., Derynck, R., 2005. Specificity and versatility in TGF-β signaling through
Smads. Annu. Rev. Cell Dev. Biol. 21, 659–693.
Foster, K., Sheridan, J., Veiga-Fernandes, H., Roderick, K., Pachnis, V., Adams, R., Blackburn,
C., Kioussis, D., Coles, M., 2008. Contribution of neural crest-derived cells in the
embryonic and adult thymus. J. Immunol. 180, 3183–3189.
and a signalling pathway that controls patterning in the developing chick limb.
Development 120, 209–218.
Frank, D.U., Fotheringham, L.K., Brewer, J.A., Muglia, L.J., Tristani-Firouzi, M., Capecchi,
M.R., Moon, A.M., 2002. An Fgf8 mouse mutant phenocopies human 22q11 deletion
syndrome. Development 129, 4591–4603.
Gordon, J., Bennett, A.R., Blackburn, C.C., Manley, N.R., 2001. Gcm2 and Foxn1 mark early
parathyroid- and thymus-specific domains in the developing third pharyngeal
pouch. Mech. Dev. 103, 141–143.
Gordon, J., Patel, S.R., Mishina, Y., Manley, N.R., 2010. Evidence for an early role for BMP4
signaling in thymus and parathyroid morphogenesis. Dev. Biol. 339, 141–154.
Grevellec, A., Tucker, A.S., 2010. The pharyngeal pouches and clefts: development, evolu-
tion, structure and derivatives. Semin. Cell Dev. Biol. 21, 325–332.
Grevellec, A., Graham, A., Tucker, A.S., 2011. Shh signalling restricts the expression of Gcm2
and controls the position of the developing parathyroids. Dev. Biol. 353, 194–205.
Gunther, T., Chen, Z.F., Kim, J., Priemel, M., Rueger, J.M., Amling, M., Moseley, J.M., Mar-
tin, T.J., Anderson, D.J., Karsenty, G., 2000. Genetic ablation of parathyroid glands
reveals another source of parathyroid hormone. Nature 406, 199–203.
Hamburger, V., Hamilton, H., 1951. A series of normal stages in the development of the
chick embryo. J. Morphol. 88, 49–92 (Reprinted in 1992. Dev. Dyn. 195, 231–272).
Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J., Ish-Horowicz, D., 1995. Expression
of a Delta homologue in prospective neurons in the chick. Nature 375, 787–790.
Jenkinson, W.E., Jenkinson, E.J., Anderson, G., 2003. Differential requirement for mesen-
chyme in the proliferation and maturation of thymic epithelial progenitors. J. Exp.
Med. 198, 325–332.
Le Douarin, N., 1967a. Détermination précoce des ébauches de la thyroïde et du thymus
chez l'embryon de Poulet. C. R. Acad. Sci. 264, 940–942.
Le Douarin,N., 1967b.Influence de lanature etdu degré d'évolution du mésoderme sur la
différenciation de l'endoderme pharyngien chez l'embryon de Poulet. C. R. Soc. Biol.
Le Douarin, N., Bussonnet, C., Chaumont, F., 1968. Etude des capacités de différenciation
et du rôle morphogène de l'endoderme pharyngien chez l'embryon d'Oiseau. Ann.
Embryol. Morphog. 1, 29–40.
Le Douarin,N.,LeLièvre,C., 1970.Demonstration of neuralorigin ofcalcitonin cells ofulti-
mobranchial body of chick embryo. C. R. Acad. Sci. Hebd. Seances Acad. Sci. D. 270,
Le Douarin,N.,Fontaine,J.,Le Lièvre, C.,1974.Newstudiesonthe neuralcrestoriginofthe
avian ultimobranchial glandular cells; interspecific combinations and cytochemical
characterization of C cells based on the uptake of biogenic amine precursors.
Histochemistry 38, 297–305.
life in interspecific chimeras. J. Exp. Med. 142, 17–40.
Liu, W., Selever, J., Murali, D., Sun, X., Brugger, S.M., Ma, L., Schwartz, R.J., Maxson, R.,
Furuta,Y., Martin, J.F.,2005. Threshold-specific requirementsfor Bmp4inmandibular
development. Dev. Biol. 283, 282–293.
Liu, Z., Yu, S., Manley, N.R., 2007. Gcm2 is required for the differentiation and survival of
parathyroid precursor cells in the parathyroid/thymus primordia. Dev. Biol. 305,
Moore-Scott, B.A., Manley, N.R., 2005. Differential expression of sonic hedgehog along
the anterior–posterior axis regulates patterning of pharyngeal pouch endoderm
and pharyngeal endoderm-derived organs. Dev. Biol. 278, 323–335.
Nehls, M., Kyewski, B., Messerle, M., Waldschütz, R., Schüddekopf, K., Smith, A.J.,
Boehm, T., 1996. Two genetically separable steps in the differentiation of thymic
epithelium. Science 272, 886–889.
Ohuchi, H., Nakagawa, T., Yamamoto, A., Araga, A., Ohata, T., Ishimaru, Y., Yoshioka, H.,
Kuwana, T., Nohno, T., Yamasaki, M., Itoh, N., Noji, S., 1997. The mesenchymal
factor, FGF10, initiates and maintains the outgrowth of the chick limb bud through
interaction with FGF8, an apical ectodermal factor. Development 124, 2235–2244.
Okabe, M., Graham, A., 2004. The origin of the parathyroid gland. Proc. Natl. Acad. Sci.
U. S. A. 101, 17716–17719.
Patel, S.R., Gordon, J., Mahbub, F., Blackburn, C.C., Manley, N.R., 2006. Bmp4 and Noggin
expression during early thymus and parathyroid organogenesis. Gene Expr. Patterns
requires signaling through the fibroblast growth factor receptor r2-IIIb. J. Immunol.
Rossi, S.W., Jeker, L.T., Ueno, T., Kuse, S., Keller, M.P., Zuklys, S., Gudkov, A.V., Takahama,
Y., Krenger, W., Blazar, B.R., Holländer, G.A., 2007. Keratinocyte growth factor
(KGF) enhances postnatal T-cell development via enhancements in proliferation
and function of thymic epithelial cells. Blood 109, 3803–3811.
Rodewald, H.R., 2008. Thymus organogenesis. Annu. Rev. Immunol. 26, 355–388.
Soza-Ried, C., Bleul, C., Schorpp, M., Boehm, T., 2008. Maintenance of thymic epithelial
phenotype requires extrinsic signals in mouse and zebrafish. J. Immunol. 181,
Stottmann, R.W., Anderson, R.M., Klingensmith, J., 2001. The BMP antagonists Chordin
and Noggin have essential but redundant roles in mouse mandibular outgrowth.
Dev. Biol. 240, 457–473.
Tsai, P.T., Lee, R.A., Wu, H., 2003. BMP4acts upstream of FGF inmodulatingthymic stroma
and regulating thymopoiesis. Blood 102, 3947–3953.
H. Neves et al. / Developmental Biology 361 (2012) 208–219