Fgf is required to regulate anterior-posterior patterning in the Xenopus lateral plate mesoderm.

Abstract

Given that the lateral plate mesoderm (LPM) gives rise to the cardiovascular system, identifying the cascade of signalling events that subdivides the LPM into distinct regions during development is an important question. Retinoic acid (RA) is known to be necessary for establishing the expression boundaries of important transcription factors that demarcate distinct regions along the anterior posterior axis of the LPM. Here, we demonstrate that fibroblast growth factor (Fgf) signalling is also necessary for regulating the expression domains of the same transcription factors (nkx2.5, foxf1, hand1 and sall3) by restricting the RA responsive LPM domains. When Fgf signalling is inhibited in neurula stage embryos, the more posterior LPM expression domains are lost, while the more anterior domains are extended further posterior. The domain changes are maintained throughout development as Fgf inhibition results in similar domain changes in late stage embryos. We also demonstrate that Fgf signalling is necessary for both the initiation of heart specification, and for maintaining heart specification until overt differentiation occurs. Fgf signalling is also necessary to restrict vascular patterning and create a vascular free domain in the posterior end of the LPM that correlates with the expression of hand1. Finally, we show cross talk between the RA and Fgf signalling pathways in the patterning of the LPM. We suggest that this tissue wide patterning event, active during the neurula stage, is an initial step in regional specification of the LPM, and this process is an essential early event in LPM patterning.

Full text

Page 1

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright

Page 2

Author's personal copy
Fgf is required to regulate anterior–posterior patterning
in the Xenopus lateral plate mesoderm
Steven J. Deimlinga,b, Thomas A. Drysdalea,b,c,d,*
aChildren’s Health Research Institute, 800 Commissioners Rd. E. London, Ontario, Canada N6C 2V5
bDepartment of Biology, University of Western Ontario, Canada N6A 5B7
cDepartment of Paediatrics, University of Western Ontario, Canada N6A 5W9
dDepartment of Physiology and Pharmacology, University of Western Ontario, Canada N6A 5C1
A R T I C L E I N F O
Article history:
Received 11 February 2011
Received in revised form
21 June 2011
Accepted 23 June 2011
Available online 8 July 2011
Keywords:
Fibroblast growth factor
Retinoic acid
Lateral plate mesoderm
hand1
foxf1
Heart
A B S T R A C T
Given that the lateral plate mesoderm (LPM) gives rise to the cardiovascular system,
identifying the cascade of signalling events that subdivides the LPM into distinct regions
during development is an important question. Retinoic acid (RA) is known to be necessary
for establishing the expression boundaries of important transcription factors that demar-
cate distinct regions along the anterior posterior axis of the LPM. Here, we demonstrate
that fibroblast growth factor (Fgf) signalling is also necessary for regulating the expression
domains of the same transcription factors (nkx2.5, foxf1, hand1 and sall3) by restricting the
RA responsive LPM domains. When Fgf signalling is inhibited in neurula stage embryos, the
more posterior LPM expression domains are lost, while the more anterior domains are
extended further posterior. The domain changes are maintained throughout development
as Fgf inhibition results in similar domain changes in late stage embryos. We also demon-
strate that Fgf signalling is necessary for both the initiation of heart specification, and for
maintaining heart specification until overt differentiation occurs. Fgf signalling is also nec-
essary to restrict vascular patterning and create a vascular free domain in the posterior end
of the LPM that correlates with the expression of hand1. Finally, we show cross talk between
the RA and Fgf signalling pathways in the patterning of the LPM. We suggest that this tissue
wide patterning event, active during the neurula stage, is an initial step in regional speci-
fication of the LPM, and this process is an essential early event in LPM patterning.
? 2011 Elsevier Ireland Ltd. All rights reserved.
1.Introduction
The development of organ systems requires signals that
specify cell fate and the patterning of cells into distinct re-
gions of the embryo. The coupling of cell fate specification
with the coordinated patterning events that define the body
plan is fundamental to our understanding of organogenesis.
This question is increasingly important, as much of the prom-
ise of regenerative medicine depends on differentiating
unspecified progenitors into particular cell types and that
process generally means cell fate specification without the
associated patterning.
Three cell types that are of particular interest in regenera-
tive medicine are cardiomyocytes, endothelial, and smooth
muscle cells because these have the potential for improving
treatments for cardiovascular disease (Chien et al., 2008;
Frontini et al., 2011) and these cell types are derived from
the lateral plate mesoderm (LPM). Differentiation of the
0925-4773/$ - see front matter ? 2011 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.mod.2011.06.002
* Corresponding author at: Children’s Health Research Institute, 800 Commissioners Rd. E. London, Ontario, Canada N6C 2V5.
Tel.: +1 519 685 8500x55072; fax: +1 519 685 8186.
E-mail address: tadrysda@uwo.ca (T.A. Drysdale).
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/modo

Page 3

Author's personal copy
lateral plate mesoderm derivatives can be viewed in two
ways. Looking at stem cells to see how these tissues differen-
tiate has pointed out potential precursors that can give rise to
one or more of the derivatives. One example of such a precur-
sor is the hemangioblast that can give rise to both hematopoi-
etic and endothelial cells (Xiong, 2008). In mammals,
individual stem cells that can give rise to cardiomyocytes,
endothelial cells and smooth muscle cells have also been
identifiedin theearlyembryo
Martinez-Estrada et al., 2010; Moretti et al., 2006; Yang et al.,
2008). However, when viewed in vivo, these different cell types
arise from distinct regions of the embryo implying that spatial
cues are important in determining which cell type will form.
In the Xenopus embryo, after gastrulation, the LPM can be
divided into several distinct regions along the ventral midline
that can be lined up along the rostral/caudal axis. These re-
gions from most rostral to caudal are: the secondary heart
field (Brade et al., 2007), the primary heart field (Tonissen
et al., 1994), an anterior, ventral blood island (Smith et al.,
2002), and a posterior, ventral blood island (Ciau-Uitz et al.,
2000). In addition to these distinct regions, the LPM is also
essential for patterning the underlying endoderm during the
tailbud stage (Horb and Slack, 2001), suggesting that it must
have intrinsic pattern. We have recently described an addi-
tional, distinct anterior–posterior pattern in the Xenopus
LPM based on the expression of nkx2.5, foxf1, hand1 and sall3
(formerly Xsal1) (Deimling and Drysdale, 2009).
Retinoic acid signalling is essential for proper patterning
within the LPM, including the pattern exhibited by nkx2.5,
foxf1, hand1 and sall3 expression (Collop et al., 2006; Deimling
and Drysdale, 2009; Keegan et al., 2005; Waxman et al., 2008).
RA signalling clearly alters, but does not eliminate the pattern
indicating that additional signalling events are required for
this pattern. To better understand how the pattern is gener-
ated, we chose to investigate fibroblast growth factor (Fgf) sig-
nalling because of the well documented interplay between RA
and Fgf signalling in a variety of tissues (Diez del Corral et al.,
2003; Moreno and Kintner, 2004; Shiotsugu et al., 2004). In the
early heart field (anterior LPM), both RA (Collop et al., 2006;
Hochgreb et al., 2003; Keegan et al., 2005) and Fgf (Alsan and
Schultheiss, 2002; Keren-Politansky et al., 2009; Marques
et al., 2008; Samuel and Latinkic, 2009) have been implicated
in proper cardiac specification and patterning. Furthermore,
in the posterior pole of the embryo, Fgf signals are also re-
quired for proper body axis extension by maintaining the tail
bud domain (Griffin et al., 1995; Pownall et al., 1996). In the
mouse, RA acts in opposition to Fgf to pattern both the ante-
rior–posterior axis at the atrial (posterior) end of the heart
tube (Sirbu et al., 2008) as well as in the trunk during body axis
extension in mouse (Zhao et al., 2009). These results led us to
hypothesize a dynamic interaction between RA and Fgf sig-
nals in conferring coordinate information to cells across the
anterior–posterior axis of the early lateral plate mesoderm.
The expression of the enzymes that regulate the levels of
retinoic acid and the location of in the embryo support the
concept of these two signalling pathways having opposing
centres of action that could pattern the LPM. Aldh1a2 (raldh2),
the enzyme primarily responsible for RA synthesis during
early organogenesis, is expressed at the dorso-anterior shoul-
der of the LPM (Chen et al., 2001; Lynch et al., 2011) and cyp26
(Ferdous etal.,2009;
is expressed at the posterior end of the embryo and a region
corresponding to the cement gland at the anterior end (de
Roos et al., 1999; Hollemann et al., 1998). Anterior to the ald-
h1a2 expression domain there is an anterior crescent of fgf8
expression and an additional fgf8 expression domain at the
posterior end of the embryo (Christen and Slack, 1997;
Shiotsugu et al., 2004) and both correspond to regions of ac-
tive extracellular signal-regulated protein kinase (ERK) signal-
ling (Christen and Slack, 1999). Thus, there is potential for two
different regions with high RA signalling at one end and high
Fgf signalling at the other. First, over the bulk of the LPM there
is high RA signalling at the anterior end of the LPM whereas
Fgf signalling is high at the posterior end. However, the form-
ing first and second heart fields have high RA signalling at the
posterior end and high Fgf signalling on the anterior side.
Using an Fgf receptor inhibitor, that allowed us to bypass
the requirements for fgf during early embryogenesis (Amaya
et al., 1991; Bottcher and Niehrs, 2005), we demonstrate that
loss of Fgf signalling, after gastrulation, results in a loss of
specific anterior-ventral and posterior LPM domains. This
alteration of the early patterning has direct consequences
for LPM derivatives, including the heart and vascular system.
We demonstrate that Fgf signalling is required for both initi-
ation and maintenance of the heart and that myocardial dif-
ferentiation is able to partially recover if Fgf signalling is
restored before heart tube closure is complete. A posterior
LPM domain that is normally free of vasculature is also absent
when Fgf signalling is lost and the expression domains of two
vascular markers, etv2 and aplnr are extended to the posterior
limits of the LPM. Therefore, correct patterning of the early
Xenopus LPM, and LPM derived cardiovascular cell lineages
requires a complex input of Fgf signals in the post gastrula
embryo. We propose a model whereby Fgf signals are required
to restrict anterior RA responsive domains and to maintain a
domain of unspecified mesoderm in the posterior end of the
embryo.
2.Results
2.1.
LPM
Fgf signalling is required for proper patterning of the
We have previously described four distinct domains
throughouttheLPMalong
(Deimling and Drysdale, 2009): nkx-2.5 (anterior-ventral LPM;
early heart field marker), foxf1 (anterior-dorsal LPM), hand1
(middle LPM) and sall3 (posterior LPM). In this study, we
extend the mesoderm domains to include the expression of
isl1 (anterior LPM; marker of the secondary heart field in
Xenopus (Brade et al., 2007), and Xbra marking the posterior
tail bud domain (Smith et al., 1991). To test whether Fgf
signalling plays a role in the patterning these expression do-
mains, we treated embryos with SU5402, a well established
inhibitor of that pathway, at the end of gastrulation (stage
12.5) and examined the expression domains of the described
marker genes at time the neural folds close (stage 20).
When Fgf signalling was blocked with SU5402, the expres-
sion of foxf1, normally localized to the anterior-dorsal LPM,
was expanded toward the ventral pole (Fig. 1A and B). In
theanterior–posterioraxis
328
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 4

Author's personal copy
addition, foxf1 is normally restricted to the anterior flank of
the embryo at stage 20, but when Fgf signalling is inhibited,
foxf1 expression was observed in a narrow band at the poster-
ior end of the LPM near the closed blastopore (Fig. 1C and D).
Hand1, is normally expressed in an inverted saddle shape in
the middle of the LPM. Blocking Fgf signalling resulted in a
posterior expansion of the hand1 domain, particularly evident
in the dorsal half of the expression domain (Fig. 1E–H). As
with the other LPM markers, the posterior LPM domain
marked by the expression of sall3 was displaced even further
posterior when Fgf signalling was inhibited (Fig. 1I–L). The
sall3 expression domain was observed in the region immedi-
ately adjacent to the blastopore that normally expresses Xbra.
The Xbra expression domain, normally present adjacent to
the blastopore, was completely undetectable after treatment
with 10 lM SU5402 (Fig. 1M and N). Previous studies have
demonstrated an absolute requirement for Fgf signalling for
Xbra expression (Fletcher and Harland, 2008; Isaacs et al.,
1994) and these results extend that requirement to post gas-
trulation in intact embryos. Therefore, over the entire LPM
there is a shift of all domain markers towards the posterior
end of the embryo when FGF signalling is lost.
The early LPM pattern is transient as the sall3 expression is
rapidly lost after stage 20 (Deimling and Drysdale, 2009). How-
ever, we have previously shown that altered RA signalling,
even if altered for a few hours after gastrulation, results in
shifts in the expression domains of genes at later stages of
LPM development and that these shifts reflect the earlier ob-
served changes. To determine if later LPM patterning was
dependent on the early Fgf signalling, SU5402 was applied
from stage 12.5 to stage 28–32 and then fixed for analysis. A
loss of Fgf signalling during that time resulted in a stunted
embryo phenotype (Fig. 2). Blocking Fgf signalling before or
during gastrulation is known to result in shortened embryos
due to inhibition of cellular movements during gastrulation
and a lack of tail formation (Amaya et al., 1991; Fletcher and
Harland, 2008; Isaacs et al., 1994). As gastrulation was com-
pleted when we treated the embryos, we only observed the
shortened tail.
At stage 28–30, both hand1 and foxf1 are normally ex-
pressed in the anterior two thirds of the LPM, with a defined
area free of staining in the posterior LPM in both cases
(Fig. 2A and C). However, the expression of both markers
was detectable in the dorsal half of the LPM and extended
all the way to the closed blastopore when Fgf signalling was
blocked (Fig. 2B and D). Conversely, hoxc10 expression, which
is normally restricted to the posterior half of the LPM, was en-
tirely absent in the LPM, and severely restricted in the somitic
Fig. 1 – Fgf signalling is necessary for proper LPM pattern.
Foxf1, hand1, sall3 and Xbra were assayed for a response in
expression domain when Fgf signalling was inhibited with
SU5402. The anterior domain, marked by foxf1 expression
(A–D), was expanded toward the ventral side of the embryo
(arrowheads in A and B), and was upregulated at the
posterior end of the LPM just ventral to the blastopore
(arrowheads in C and D). The posterior border of hand1 (E–H)
is displaced further posterior, particularly at the dorsal
edges of the domain (arrowheads in E–H). The LPM domain
of sall3 (I–L) is displaced posterior with a loss of Fgf
signalling (white arrowheads in K and L) while the posterior
neural tube domain is undetectable (black arrowheads in K
and L). Expression of Xbra in the tailbud domain is
completely lost in the absence of Fgf signalling (arrowheads
in M and N). llv: left lateral view, embryos oriented with the
anterior pole toward the left of the image, dorsal at top. pos:
posterior view, embryos oriented with dorsal at top of
image. The total number of embryos examined for each
panel is indicated in the lower left hand corner.
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
329

Page 5

Author's personal copy
mesoderm, when Fgf signalling is lost (Fig. 2E and F). Lastly,
differentiation of the heart, an anterior LPM derivative, was
completely lost when the embryo was treated with SU5402
(Fig. 2G and H).
2.2.
differentiation of the heart field
Fgf signalling is necessary for both initiating and
The loss of cardiac markers when Fgf signalling was
blocked was expected as nkx2.5 expression has been shown
to require Fgf signalling in recent studies of Xenopus heart
development (Keren-Politansky et al., 2009; Samuel and Lat-
inkic, 2009). To further characterize the requirement for Fgf
in cardiogenesis, we examined whether it was required for
development of the secondary heart field, characterized by
expression of isl1 (Brade et al., 2007). When Fgf signalling is
blocked after gastrulation, the expression of nkx2.5 was
essentially below detection and although expression was still
detectable, isl1 expression was severely restricted (Fig. 3).
We next wished to relate these results with our overall
LPMeffects bydeterminingwhenFgf signalling was
necessary during heart development. Embryos were treated
with SU5402 at successively later stages and assayed for ef-
fects on gross heart morphology and differentiation using
the expression of nkx2.5 and tnni3 as markers. If Fgf signalling
is inhibited immediately after heart specification (stage 12.5),
early expression of nkx2.5 is lost (Fig. 4A and B) and later dif-
ferentiation is blocked as assessed by tnni3 expression
(Figs. 2G,H and 4F). If Fgf signalling was blocked after stage
20 (Fig. 4C and G), a time point after nkx2.5 expression is ini-
tiated, neither nkx2.5 or tnni3 expression were detectable at
stage 32. When embryos were examined shortly after treat-
ment (approximately 3 h), nkx2.5 expression was already lost
(data not shown). If Fgf signalling was inhibited after stage 24,
expression of both nkx2.5 and tnni3 was detected (Fig. 4), but
the heart did not form a tube.
We next took advantage of the fact that SU5402 can be suc-
cessfully washed from a Xenopus embryo (Delaune et al., 2005;
Fletcher and Harland, 2008) to narrow the window when Fgf
signalling is required. If embryos were treated with SU5402 at
stage 12.5, and then had Fgf signalling restored at stage 20,
expression of both nkx2.5 and tnni3 (Fig. 4J and N) was detect-
ablebystage32.However,heartmorphologywasclearlyabnor-
mal with two bilateral patches of either nkx2.5 or tnni3
expressionandnoindicationoftubeformation.Iftheinhibitor
was removed at the early tail bud stage (stage 22) both nkx2.5
and tnni3 (Fig. 4K and O) domainswerepresent in similar bilat-
eral domains but the size of these domains was further re-
duced. However, if SU5402 was not removed until stage 26,
expression of nkx2.5 and tnni3 (Fig. 4L and P) was lost.
2.3.
vascular system
FGF signalling is necessary for patterning the trunk
Since Fgf signalling was required for patterning the LPM
and for differentiation of the heart, we examined its role in
Fig. 3 – Fgf signalling is required for the initiation of heart
specification. Expression of both nkx2.5 and isl1 is readily
detectable in the cardiac crescent of stage 20 embryos (A
and C). Embryos treated with SU5402 to inhibit fgfr1 activity
show a loss of nkx2.5 expression (B), and a down regulation
of isl1 expression (D). In both cases, the anterior pole of the
embryo is viewed, with dorsal folds visible toward the top of
the image. The total number of embryos examined for each
panel is indicated in the lower left hand corner.
Fig. 2 – Fgf is required for the patterning of late LPM
markers. Hand1 (A and B) and foxf1 (C and D) are normally
restricted in the anterior and middle LPM with a clear
domain free of expression in both cases. When embryos are
treated with SU5402, both hand1 (B) and foxf1 (D) are
expressed along the entire anterior–posterior axis.
Conversely, the LPM expression domains of both hoxc10 (E
and F; normally expressed in the posterior half of the LPM)
and tnni3 (G and H; marker of cardiac differentiation) are
completely undetectable when Fgf signalling is inhibited.
(A–F): lateral view of the embryos is shown, with anterior
toward the left, dorsal at top. (G–H); ventral view of the heart
region is shown, with anterior toward left. The total number
of embryos examined for each panel is indicated in the
lower left hand corner.
330
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 6

Author's personal copy
patterning the vascular system; another derivative of the
LPM. Fgf signalling was blocked after gastrulation (stage
12.5), and embryos were allowed to develop until the tail
bud stage (stage 32) when they were fixed and assayed for
vascular markers etv2 (Fig. 5A–D) and aplnr (Fig. 5E–H). There
exists a vascular free region in the posterior trunk of the
embryo where the vascular plexus does not appear to form
(Salanga et al., 2010) (Fig. 5B and F). This vascular free zone
was lost when Fgf signalling was inhibited, as both etv2
(Fig. 5D) and aplnr (Fig. 5H) expression was extended to the
posterior end of the body. In addition, the gap in staining be-
tween the rostral lymph sac and the trunk vasculature, seen
in the controls and normally occupied at least in part by the
heart, is absent. In the SU5402 treated embryos, the trunk
vasculature is continuous with staining in the rostral lymph
sac (Fig. 5C and G). It should be noted that vascular precursors
that are forming the posterior cardinal veins do extend to the
tail in all embryos.
Since SU5402 is known to cross react with the Vegf recep-
tor (Mohammadi et al., 1997), we tested whether the effects of
SU5402 on the early cardiovascular system were due to a loss
of Vegf signalling. Embryos were treated with a Vegfr specific
inhibitor, KRN633 (Nakamura et al., 2004) after gastrulation
and assayed for both cardiac and vascular marker expression.
Doses of 10 lM, and 25 lM KRN633 were sufficient to block
angiogenesis, however there was no detectable effect on
either tnni3 expression, or on patterning the posterior limits
of the vascular system (Suppl. Fig. 1). This suggests that the
effects of SU5402 on the heart field, as well as those on the
posterior vascular free zone are due to specific interactions
with the Fgfrs, and not due to cross reactivity of SU5402 on
the Vegf signalling pathway.
We have previously demonstrated that size of the LPM
expression domains is dependent on RA signalling (Deimling
and Drysdale, 2009) and that RA can block differentiation of
the heart (Collop et al., 2006; Drysdale et al., 1997). As the
expansion of the middle LPM domains corresponds to the
expansion of the vascular system when Fgf signalling is
blocked, we predicted that the altered RA signalling treat-
ments that affect the LPM expression domains would also al-
ter the size of the vascular plexus. When RA signalling was
blocked with an RA antagonist, expression of either etv2 or
aplnr was essentially the same as observed in control embryos
(Fig. 5I,J and O,P). However, increasing RA signalling caused a
posterior expansion of the vascular plexus as defined by the
expression domains of both etv2 and aplnr (Fig. 5M,N and
S,T). At the anterior end of the embryo, expression in the ros-
tral lymph sac is also significantly decreased with increased
RA signalling (Fig. 5). This suggests that the extent of vascular
patterning in the LPM is limited and that the boundaries are
determined, at least in part, by Fgf and RA signalling.
A close correspondence was noted between the size of the
vascular free zone and the posterior region of the LPM that
does not express hand1 in both normal conditions and under
all treatments tested. To demonstrate this correlation in indi-
vidual embryos, we performed double in situs with probes
against hand1 and etv2 using control embryos and embryos
that had either Fgf or RA signalling altered. In control em-
bryos and in RA antagonist treated embryos, the vascular free
zone at the posterior end of the embryo did not express hand1
(Fig. 6A, B, D and E). When embryos were treated with SU5402,
or RA, the vascular free region was eliminated and the hand1
domain extended to the end of the body (Fig. 6C, F, H and J).
Despite this tight correlation along the rostral caudal axis,
we did observe differences between the expression of hand1
and the pattern of the vascular plexus. When embryos are
treated with either RA or SU5402, the correspondence be-
tween the hand1 domain and the vasculature was lost when
viewed in the dorsal–ventral axis. There is clearly forming
Fig. 4 – Fgf signalling is necessary for heart patterning and
morphogenesis. Embryos were treated with SU5402 during
different developmental windows to determine when Fgf
signalling is necessary for heart development and assayed
for expression of nkx2.5 (A–D and I–L) or tnni3 (E–H and M–P).
Fgf signalling was inhibited at successively later stages (top
panel: A–H) and compared to control embryos (A and E).
Embryos treated with SU5402 at either stage 12.5 (B and F) or
stage 20 (C and G) demonstrate a complete loss of heart
marker expression by stage 32, while embryos treated at
stage 24 (D and H) show expression of both nkx2.5 and tnni3
but no discernable heart tube. Fgf signalling was also
inhibited at stage 12.5 and restored at later stages by
removing the inhibitor (bottom panel: I–P) and compared to
control embryos (I and M). When Fgf signalling is restored
by either stage 20 (J and N) or stage 22 (K and O), both heart
markers are expressed however a normal heart tube is not
formed. If signalling is not restored until stage 26 (L and P),
neither nkx2.5 or tnni3 are detectable. White arrows mark
the heart region. The total number of embryos examined for
each panel is indicated in the lower left hand corner.
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
331

Page 7

Author's personal copy
vasculature in ventral regions that do not appear to express
hand1 (Fig. 6C, F, H and J).
To further examine potential LPM derivates, we also exam-
ined the anterior and ventral blood islands. In the chick em-
bryo, inhibition of Fgf signalling causes ectopic blood
formation and inhibits the expression of endothelial cell
markers after gastrulation (Nakazawa et al., 2006). In Xenopus,
treatment with SU5402 during gastrulation is known to cause
an expansion of ventral blood island markers towards the
posterior end of the embryo but little effect was seen when
embryos were treated during gastrulation (Walmsley et al.,
2008). In addition, retinoic acid has been shown to inhibit dif-
ferentiation of haematopoietic precursors but not their for-
mation whentheembryos
gastrulation (Bertwistle et al., 1996). We observed inhibition
of haematopoietic differentiation, as assayed by hba1 expres-
sion when embryos were either exposed to SU5402 after gas-
trulation but we did not observe any clear changes in the
expression of mpo and spib (markers of primitive myeloid line-
age), etv2 (early marker of hematopoietic and endothelial pre-
cursors) or scl (marker of the early ventral blood islands),
are exposedduringmid
when RA or Fgf signalling was altered after gastrulation (data
not shown).
2.4. FGF and RA signalling directly interact
Since both RA and Fgf are necessary to pattern separate
areas of the LPM, and have been previously shown to form
mutually antagonistic gradients in other tissues, we wished
to test if the two pathways directly interact in the LPM at
the time of our treatments. To determine the effect of the
RA signalling pathway on Fgf signalling, embryos were trea-
ted with 1 lM RA, RAA or a DMSO control at stage 14 and as-
sayed for both fgf4 and fgf8 expression at the end of
neurulation. When RA signalling was antagonized, the ante-
rior trunk domain of fgf8, located just posterior to the cement
gland, was decreased in intensity in comparison to staining in
the pituitary anlagen when compared to control DMSO trea-
ted embryos (Fig. 7D and E). However, the posterior domains
of both fgf4 and fgf8 remained unaffected (Fig. 7A,B and
G,H). Increasing RA signalling by treatment with all-trans RA
resulted in an increase in the fgf8 expression domain in ante-
Fig. 5 – Fgf and RA signalling are required for proper patterning of the trunk vascular system. Embryos were treated with
DMSO or SU5402 at stage 12 (top panel: A–H) and allowed to develop until stage 32 at which point vascular pattern was
assessed by expression of etv2 (A–D) and aplnr (E–H). Both etv2 (A and B) and aplnr (E and F) are normally restricted from the
posterior end of the LPM. However, when Fgf signalling is lost, the expression domains of both etv2 (C and D) and aplnr (G and
H) are extended posterior to the end of the trunk. Also, the anterior gap in vasculature between the rostral lymph sac and
trunk vasculature, corresponding to the location of developing heart, is absent in SU5402 treated embryos. Embryos treated at
stage 14 with a synthetic RA antagonist (RAA), all-trans RA or a DMSO control (bottom panel: I–T) were assayed for etv2 (A–F)
or aplnr (G–L) expression. While no obvious changes are present after treating embryos with RAA (I,J and O,P), the posterior
limit of the LPM vasculature (white arrowheads) is extended further posterior in RA treated embryos (M,N and S,T) when
compared with control embryos (K,L and Q,R). The staining seen in the rostral lymph sac is also absent in embryos treated
with RA. White arrows: posterior limit of trunk vascular pattern. Red arrows: rostral lymph sac. Black arrows: gap between
rostral lymph sac and trunk vasculature. The total number of embryos examined for each panel is indicated in the lower left
hand corner.
332
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 8

Author's personal copy
rior trunk (Fig. 7E and F), becoming much thicker and sur-
rounding the entire cement gland rather than residing poster-
ior to it. In the posterior neural tube fgf8 expression, was also
extended further anterior along the neural folds (Fig. 7H and I)
in agreement with previous studies (Moreno and Kintner,
2004). Conversely, the expression domain of fgf4 was de-
creased in the posterior neural tube to essentially background
levels (Fig. 7B and C). The increased levels of fgf8 was reflected
in increased levels of sprouty2, a direct target of Fgf signalling
(Fig. 7J–O).
To confirm these results we also treated embryos with
inhibitors to aldh1a2 (DEAB) and cyp26 (ketoconazole), the
endogenous enzymes responsible for RA synthesis and catab-
olism, respectively. A loss of RA by treatment with DEAB
caused a similar reduction in fgf8 staining in the anterior do-
main similar to RAA treatments (Suppl. Fig. 2). We also
wished to test the net effect of altering endogenous RA levels
on FGF signalling, again using sprouty2 expression as an assay.
Sprouty2 is expressed in both anterior and posterior domains
overlapping with the fgf8 expression domains (Suppl. Fig. 2N
and Q). In either case there are no obvious changes with re-
spect to the size or position of the sprouty2 domain.
To test if Fgf signalling was necessary for proper RA signal-
ling, we treated embryos with 1 lM SU5402 and assayed for
expression of both aldh1a2, the enzyme predominantly
responsible for in vivo RA synthesis, and cyp26, the enzyme
responsible for RA catabolism. Aldh1a2 is normally expressed
in the anterior half of the somites and dorsal LPM, as well as
the anterior trunk along the anterior border of the LPM. Inhib-
iting Fgf signalling caused a posterior expansion of the dorsal
aldh1a2 domain (Fig. 7P and Q) much further posterior than in
control embryos. Cyp26 however, is normally expressed in the
posterior neural tube and tail bud domain in a domain remi-
niscent of the Fgf4 and Fgf8 expression domains, and the Xbra
domain (Fig. 7R). Inhibiting Fgf signalling caused a complete
loss of the posterior cyp26 domain in the posterior neural
folds and the adjacent tail bud domain (Fig. 7S), which is con-
sistent with previous studies (Moreno and Kintner, 2004).
To directly test if up-regulating RA signalling and inhibit-
ing Fgf signalling would accentuate the effects on LPM pat-
terning, we treated stage 12.5 embryos with RA, SU5402 or
both and assayed for the expression of LPM markers at stage
20. Treating with RA reduced the expression domain of nkx2.5
when compared to control embryos (Fig. 8). Treating with
SU5402 also significantly reduced nkx2.5 expression to a small
patch underlying the cement gland. However, when embryos
were treated with both RA and SU5402, nkx2.5 expression was
essentially undetectable (Fig. 8A–D). The isl1 expression do-
main was also smaller when treated with either RA or
SU5402 as compared to controls, but was undetectable in
Fig. 6 – The extent of the vascular plexus is related to the size of the hand1 domain. To more directly assess the relationship
between the forming vasculature and hand1 expression double in situ hybridizations were done with hand1 being visualized
in light blue and etv2, marking the forming vasculature, in purple. The forming vascular plexus (emphasized in the lower
panels of each example) along the side of the embryo does not extend to the end of the embryo in control (DMSO, ATP)
embryos nor does the hand1 staining. When the extent of the vascular plexus is shifted towards the anterior end by
treatment with RAA (A and D) this is reflected in the changes to the hand1 expression domain. When the vascular plexus is
shifted towards the end of the embryo by addition of RA (C and F) or SU5402 (H and J) this is again reflected in the changes to
the extent of hand1 expression domain along the anterior–posterior axis. Note that the close correspondence appears to be
lost in the dorsal ventral axis in the SU5402-treated embryos (H and J).
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
333

Page 9

Author's personal copy
the heart domain when treated with RA and SU5402 together
(Fig. 8E–F). Conversely, the foxf1 expression domain was ex-
panded when in both the RA and SU5402 treatments. When
embryos were treated with both RA and SU5402, foxf1 was ex-
pressed over almost the entire LPM (Fig. 8I–L). The hand1, sall3
and Xbra expression domains were also assayed, but changes
are indistinguishable between the RA, SU5402 and RA and
SU5402 treatments (data not shown); the sall3 domain is
completely undetectable when treated with either RA or RA
and SU5402 (data not shown); and the Xbra domain is
undetectable when treated with SU5402 or RA and SU5402.
Thus any additive effect on these expression domains could
not be assessed.
Finally, if the two systems are directly interacting, it would
be predicted that the altered pattern caused by reduced Fgf
signalling could be rescued by simultaneously blocking RA
signalling. To test this, embryos were treated with both
SU5402 and the RA antagonist simultaneously from stage 12
to stage 20 and the resulting embryos were assessed for the
expression pattern of nkx2.5, isl1, hand1, and Xbra. In regions
that would be expected to have high Fgf signalling (presump-
tive heart and tail region) there did appear to be partial recov-
ery of the specific expression patterns when embryos were
treated with both SU5402 and the RA antagonist. In particular,
the expression of isl1 was greater when the retinoic acid
antagonist was added in addition to SU5402 (Fig. 9E–H).
Expression of Xbra is essentially eliminated by the addition
of SU5402 but staining can be detected when the RA antago-
nist was also added (Fig. 9Q–T). The loss of nkx2.5 staining
in embryos treated with SU5402 was not rescued by blocking
RA signalling (Fig. 9A–D). Also, the expression of hand1 was re-
stricted by the RA antagonist effect but that restriction was
not alleviated by the inhibition of Fgf signalling (Fig. 9I–P)
Fig. 7 – The RA and Fgf pathways regulate each other. The
levels of RA signalling (A–I) were altered by addition of a
synthetic RA antagonist (left column) or all-trans RA (right
column) and compared to a DMSO control (centre column).
Embryos were assayed for fgf4 (A–C) and fgf8 (D–I)
expression. The posterior domain of fgf4 (a) is lost in RA
treated embryos (C) when compared to the control (B), but
unaffected in embryos treated with RAA. Expression of fgf8
is expanded both anteriorly (E and F) and posteriorly (H and
I; compare distance between arrowheads (d) marking the
anterior limits of domain, and (e) marking posterior limits of
domain) under treatment with RA. Decreasing RA signalling
also reduces the anterior domain of fgf8 underlying the
heart region (compare ratio of staining intensity between (b)
marking the pituitary anlagen to (c)). A similar effect is seen
with sprouty2 expression (J–O), as its domain is increased
with RA in both the anterior heart region (L; arrowhead f)
and it extends further anterior (g) in the dorsal neural tube
(O) when compared to controls (K–N). Conversely, embryos
were treated with SU5402 and assayed for expression of
aldh1a2 (P and Q) or cyp26 (R and S) to determine the effect of
a loss of Fgf signalling on the RA signalling pathway. The
expression domain of aldh1a2 was expanded posterior (Q)
(arrowhead: h – marking posterior limit of expression
domain) as compared to control embryos (P). Cyp26,
normally present in the posterior LPM tailbud domain (R;
arrowhead i) is undetectable when Fgf signalling is
inhibited (S). Ant: anterior view with dorsal at top of image.
Dor: dorsal view with anterior at top. Llv: left lateral view
with anterior toward left, dorsal at top of image. Pos:
posterior view with dorsal at top. The total number of
embryos examined for each panel is indicated in the lower
left hand corner.
334
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 10

Author's personal copy
suggesting that direct interactions may have greater impor-
tance at the poles of embryo than in the centre of the LPM
pattern where RA signalling is predicted to be highest.
3.Discussion
3.1. FGF and LPM patterning
We have previously described an early pattern in the LPM
of Xenopus defined by the expression domains of three tran-
scription factors: foxf1, hand1, and sall3 (Fig. 1). This pattern
can be altered by increased retinoic acid signalling that
causes the anterior domains (foxf1, hand1) to shift towards
the posterior end of the embryo with the posterior domain
(sall3) being lost (Deimling and Drysdale, 2009).
Here, we show that blocking Fgf signalling can alter the
same pattern. Blocking Fgf signalling causes an expansion
of the anterior markers towards the posterior end of the em-
bryo (Fig. 1), as does addition of retinoic acid. This opposite ef-
fect of RA and Fgf signalling suggests that these two systems
are acting in an opposing fashion along the anterior–posterior
axis. This is, at least in part, due to one system directly alter-
ing components of the other (Fig. 6). The one result that ap-
pears contradictory is that addition of RA increases the
levels of fgf8 expression and signalling (Fig. 7H and I). The re-
sult also appears to contradict the observation that the loss of
RA signalling results in an expansion of Fgf signalling in
mouse (Ryckebusch et al., 2008; Sirbu et al., 2008; Zhao and
Duester, 2009). Previous studies have shown the same expan-
sion at the posterior end of the Xenopus embryo (Moreno and
Kintner, 2004) and suggested that the differences may be due
to the relatively acute nature (hours) of RA treatments as
compared to the chronic loss of RA signalling due to loss of
aldh1a2 in knockout mice. We have also shown that treating
embryos with RA accentuates the effect of a blocking Fgf sig-
nalling in the heart field, as well as the posterior expansion of
the anterior-dorsal foxf1 domain (Fig. 8). This synergistic ef-
fect suggests that although the two systems clearly interact,
they also have effects that are independent from one another.
Similar reciprocal interactions between these two pathways
have been described in many embryonic patterning events
(Moreno and Kintner, 2004; Shiotsugu et al., 2004; Sirbu
et al., 2008) and these results extend that interrelationship
to the LPM.
Expression of several Fgfs supports a role for Fgf signalling
in regulating the LPM. Fgf8 is expressed at the anterior end of
the heart field, supporting a role for fgf8 in heart specification
as has previously been shown (Lea et al., 2009). SU5402, the
inhibitor used in this study, is specific for fgfr1, and fgf8 has
been shown to act through fgfr1 (Chung et al., 2008; Scholpp
et al., 2004). Fgfr1 is widely expressed throughout the neurula
stage Xenopus embryo suggesting that this signalling system
is present in the time and place to have a significant role in
patterning of the LPM (Lea et al., 2009). It is worth noting that
fgf4 is also expressed in the posterior neural folds and could
contribute to signalling through fgfr1 in the tail bud domain
Fig. 8 – Retinoic acid and fgf are opposing signalling molecules in patterning the LPM. Embryos were treated at stage 12 with
either RA, SU5402 or both RA and SU5402 and assayed by whole mount in situ hybridization for expression of nkx2.5
(arrowhead a), isl1 (arrowhead b) and foxf1. Treating embryos with RA reduces the expression domain of both nkx2.5 (B) and
isl1 (F) compared with controls (A and E). Significantly reduced domains of nkx2.5 and isl1 were also present in SU5402 treated
embryos (C), however when embryos are treated with both RA and SU5402 neither marker is detectable (D and H). The
expression domain of foxf1 was expanded in both the RA (J) and SU5402 (K) treatments when compared to controls (I),
however when embryos were treated with both RA and SU5402, foxf1 expression was detectible across the entire LPM
although expression was still graded. Ant: anterior view with dorsal at top of image. Pos: posterior view with dorsal at top of
image. Llv is left lateral view. The total number of embryos examined for each panel is indicated in the lower left hand corner.
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
335

Page 11

Author's personal copy
(Fig. 7). These expression patterns, coupled with the expres-
sion patterns of aldh1a2 and cyp26, the primary regulators of
RA availability, suggest a model where these systems contrib-
ute to the size regulation of key expression domains within
the LPM (Fig. 10).
Our results indicate that Fgf signalling is at least in part
responsible for maintaining the posterior Xbra domain and
that requirement continues after gastrulation is complete.
This is similar to several studies in Xenopus that show the
importance of Fgf signalling in maintaining the tailbud do-
main (Moreno and Kintner, 2004; Pownall et al., 1996). In zeb-
rafish, a recent study has pointed out the importance of an
autoregulatory loop where brachyury directly activates the
expression of cyp26 in order to maintain the tail progenitor
zone (Martin and Kimelman, 2010). Fgf does not play a role
in that autoregulatory loop as blocking Fgf signalling, after
gastrulation, had no effect on either cyp26 or brachyury (no tail)
expression. Our results suggest that, in Xenopus, Fgf signalling
does play a role in the maintenance of this progenitor zone,
although a complete understanding of the dynamics of this
system will require further characterization of other potential
interacting systems such as notch (Moreno and Kintner, 2004)
and wnt (Martin and Kimelman, 2010) signalling. We do note
that exogenous RA and SU5402 treatments may be inhibiting
formation of the tail in different ways as there was a clear dif-
ference in the response of the posterior marker sall3 and foxf1.
With the addition of RA, the expression of sall3 in the poster-
ior LPM is lost (Deimling and Drysdale, 2009) whereas when
FGF signalling is blocked, expression of sall3 and foxf1 can
be detected in the region that normally expresses Xbra.
Fig. 9 – Loss of retinoic acid signalling partially rescues effects of loss of Fgf signalling on the domains of isl1 and Xbra.
Decreasing RA signalling in SU5402 treated embryos had little effect on the nkx2.5 domain (D) as it was still present in a
highly restricted domain similar to the SU5402 treatment (C). However, a loss of Fgf signalling leads to a loss of isl1 (H) and
Xbra (T), while neither domain is changed when RA signalling is lost (F and R). However, when Fgf signalling is decreased in
conjunction with reduced RA signalling both isl1 and Xbra expression is detectable in their normal domains (H and T).
However, the restriction of the hand1 expression domain under reduced RA condition (J and N) seems to be dominant to the
extended domain seen in the SU5402 treated embryos (K and O) as losing both RA and Fgf signalling (L) leads to a restricted
domain similar to the RA antagonist alone.
336
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 12

Author's personal copy
As we observed with the changes caused by altering RA
signalling (Deimling and Drysdale, 2009), the changes in
LPM pattern appear to be permanent. The changes in LPM
patterning resulting from a loss of Fgf signalling remain
detectable in the expression domains of foxf1, hand1, and
hoxc10 at later stages (Fig. 2), consistent with changes ob-
served at earlier stages. Both of the markers present in the
anterior-middle LPM, foxf1 and hand1, are expanded towards
the posterior end and are detectable throughout the LPM.
Our posterior marker, the hoxc10 domain, which is usually
present in the posterior half of the LPM, is completely lost
when FGF signalling is inhibited. Thus, Fgf signalling is re-
quired to define the boundaries between the tail bud domain
and the more anterior LPM, and this signalling event is clearly
necessary in early development.
3.2.Significance of the LPM pattern
Although we have documented an anterior–posterior pat-
tern in the LPM, and shown that these domains are shifted
by RA (Deimling and Drysdale, 2009) and Fgf (Fig. 2), it is not
immediately apparent as to what these domains represent,
with exception of the anterior-ventral nkx2.5 domain. We
have now correlated the changes of the LPM boundaries with
one derivative of the LPM, the developing vasculature. An
endothelial cell free domain can normally be seen in the pos-
terior LPM at the mid to late tailbud stage by the lack of etv2
and aplnr expression (Fig. 5B and F). When Fgf signalling is
inhibited, that vascular free zone is lost, and both etv2 and
aplnr are expressed along the entire axis. Furthermore, there
is a second domain which is both etv2 and aplnr free at the
anterior end of the LPM between the rostral lymph sac and
the body vasculature which corresponds to the heart field
(Fig. 5). When Fgf signalling is lost, there is no gap between
the rostral lymph sac and the trunk vasculature. These results
together suggest that Fgf signalling is necessary to restrict
vascular patterning both in the anterior-ventral and posterior
ends of the LPM. However, the expansion of vascular markers
into the anterior end may be a secondary effect as a result of
losing the differentiated heart in between the rostral lymph
sac and the body vasculature. A posterior expansion of vascu-
lature markers is clearly seen when RA signalling is increased
(Fig. 5).
The tight correlation with the expression of hand1 and the
forming vasculature suggests that hand1 may play a role in
defining that tissue. In all conditions tested, the extent of
hand1 expression corresponded with the extent of the form-
ing vascular plexus (Fig. 6) strongly suggesting a role for hand1
Fig. 10 – Model of signalling input into the LPM of the neurula stage Xenopus embryo. A high level of RA signalling input is
suggested by the expression domain of raldh2 in the anterior and dorsal LPM (A). Conversely, Fgf signalling is proposed in the
anterior-ventral and posterior-ventral LPM, suggested by the expression domains of fgf8 in the anterior, and fgf4 and fgf8 in
the posterior pole of the embryo. Dark blue and yellow bars represent the expression of raldh2 and fgf ligands respectively,
while blue and yellow arrows depict the proposed area which RA (blue) and fgf (yellow) are required for normal patterning.
The model depicts a left lateral view of the LPM. Ant: anterior, Dor: dorsal, Lat: lateral view, Pos: posterior, Ven: ventral. (B)
Diagrammatical representation of LPM expression domain response to decrease in either RA (left) or Fgf (right) signalling
when compared with the DMSO control (centre). The anterior-dorsal and middle LPM domains require RA signalling for their
full expression domain. When embryos are treated with the RA antagonist the anterior-dorsal domain is restricted (a) and the
middle LPM domain is contracted (b). When Fgf signalling is inhibited the anterior-dorsal domain is expanded ventrally (c)
and the posterior domains are severely restricted (d), indicating that the anterior-ventral and posterior LPM domains are
dependant on Fgf signalling. Red: nkx2.5, blue: foxf1, yellow: hand1, green: sall3, purple: Xbra.
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
337

Page 13

Author's personal copy
in regulating the vascular plexus. In zebrafish, recent studies
have suggested that hand2 expression in the LPM is essential
for remodelling the extracellular matrix that is required for
gut-looping (Yin et al., 2010) and is also necessary for inhibit-
ing fibronectin expression, allowing for normal cardiac fusion
(Garavito-Aguilar et al., 2010). Hand1 has been shown to be
important in the development of the yolk sac vasculature in
the mouse (Morikawa and Cserjesi, 2004), at least in part
through its regulation of the actin binding protein, Thymosin
b4 (Smart et al., 2010). Thus, hand1 could be regulating the
extracellular matrix of the LPM that then controls the extent
of the vascular bed or hand1 may be important for regulating
the endothelial cell lineage itself. It should be noted that the
expression of hand1 does not correspond to all of the vascula-
ture as the posterior cardinal vein and intersomitic vascula-
ture forms in regions that
addition, although vascular progenitor markers appear in a
patchy manner, hand1 expression is uniform across the lateral
plate suggesting that hand1 is not a marker of early vascular
progenitors. At early stages, foxf1 shows a more restricted
expression pattern than hand1 but does expand at later stages
to closely reflect the hand1 expression domain. Interestingly,
foxf1 is also required for formation of the vasculature in
mouse (Astorga and Carlsson, 2007).
In the chick, inhibition of Fgf signalling leads to expanded
blood formation while down regulating endothelial cell differ-
entiation (Nakazawa et al., 2006). This inverse relationship
does not appear to exist in Xenopus as we do not see any clear
change in the size of the scl domain that marks blood progen-
itors of the ventral blood island in response to either SU5402,
RA or RAA when applied after gastrulation (data not shown).
Fgf does play a role in the timing of scl expression and block-
ing Fgf signalling with SU5402 causes an expansion of the scl
domain but this is only observed when SU5402 is applied be-
fore gastrulation (Walmsley et al., 2008), as we found when we
added SU5402 after gastrulation to no effect.
lackhand1 expression. In
3.3.Fgf signalling and cardiac development
Fgf8 is expressed in a domain underlying the heart field
that suggests it may be an important secreted molecule dur-
ing Xenopus cardiogenesis and two recent studies have dem-
onstrated a requirement for Fgf signalling during early
Xenopus early heart development (Keren-Politansky et al.,
2009; Samuel and Latinkic, 2009). We now extend those re-
sults by demonstrating that Fgf signalling is required for the
early expression of the secondary heart marker isl1 in addi-
tion to its role in regulating nkx2.5 (Fig. 3). In addition, we
demonstrate that Fgf is required throughout early myocardial
development to maintain expression of these heart field
markers. When embryos are treated with SU5402 at stage
20, a time well after nkx2.5 is first detectable by in situ hybrid-
ization, nkx2.5 expression is not maintained and expression
of the cardiac differentiation marker, tnni3, is eliminated
(Fig. 4). While this effect can be partially rescued by removing
the inhibitor and restoring active signalling, heart morphol-
ogy is never normal if Fgf signalling is restored at any point
after a five hour treatment at stage 20. It is interesting to note
that this is the same window of time during which retinoic
acid cansuppress differentiationofthe myocardium
(Drysdale et al., 1997) possibly indicating an interaction be-
tween these two systems during heart development as is
found in cardiac patterning in mouse (Sirbu et al., 2008).
Coordinated patterning between mesodermal regions is
well established. Such reciprocal patterning has been de-
scribed for adjacent LPM fields including: myocardial and
endocardial (Ferdous et al., 2009; Misfeldt et al., 2009), epicar-
dial and myocardial (van Wijk et al., 2009) lineages within the
heart field, and the heart field and limb bud (Waxman et al.,
2008), and between the heart and vasculature (Schoenebeck
et al., 2007). In addition, LPM derivatives and adjacent non-
LPM mesoderm demonstrate reciprocal interactions in both
the embryo (Mudumana et al., 2008) and in extraembryonic
tissues (Shin et al., 2009). In each of these examples, cells in
close proximity are co-ordinately patterned leading to distinct
cell lineages as a result of encountering different extracellular
signals. Here, we propose a tissue wide patterning event
occurring upstream of these developmental decision points
regulated, at least in part, by opposing inputs of RA and Fgf
signalling into the LPM. Defects in this process would have
significant consequences on the size and position of mesoder-
mally derived structures, such as the heart and vasculature.
After this early patterning event, the LPM continues to be sub-
divided into successively smaller developmental fields lead-
ing to proper size and position of organ progenitors. The
information presented here may be of particular use in devel-
oping methods to differentiate progenitor cells into cell types
normally derived from the LPM including myocardial and
endothelial cells.
4.Experimental procedures
4.1.Embryo collection
Female Xenopus laevis frogs were injected with 600–700 IU
of human chorionic gonadotropin. Eggs were obtained and
fertilized in vitro using minced testis in 80% Steinberg’s solu-
tion. Embryos were dejellied in 2.5% cysteine, pH 8.0, and cul-
tured in 20% Steinberg’s solution. Staging of embryos was
done according to (Faber and Nieuwkoop, 1994).
4.2.Embryo treatments
The Fgf inhibitor experiments were performed by adding
10 lM SU5402 (Calbiochem) to embryos in conjunction with
0.1 mM ATP, or DMSO with 0.1 mM ATP as a control. The addi-
tion of ATP is rarely reported and we found that it greatly en-
hanced the ability of SU5402 to block Fgf signalling (Suppl.
Fig. 3). With the addition of 0.1 M ATP, we found that treating
embryos with a dose of 10 lM SU5402 was sufficient to yield a
full loss of Fgf signalling phenotype as judged by a loss of tail
bud outgrowth or cardiac differentiation. The ability to block
Fgf signalling was also verified by assaying for the expression
of sprouty2, a direct target of Fgf signalling (Nutt et al., 2001)
and 10 lM SU5402 was able to essentially eliminate sprouty2
expression(Suppl.Fig.4).TorestricttheinhibitionofFgfsignal-
ling to specificwindows of time, SU5402-treated embryoswere
moved to a clean dish with 20% Steinberg’s without SU5402 or
ATP, and the solution was changed every 20 min for 2 h. With
338
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41

Page 14

Author's personal copy
this regime Fgf signalling was restored, as assayed by sprouty2
expression, after approximately 2 h (data not shown), consis-
tent with previous reports (Crump et al., 2004; Maroon et al.,
2002; Marques et al., 2008; Nechiporuk et al., 2005).
Retinoic acid signalling was altered by treatment with
1 lM all-trans RA (Sigma), or 1 lM of a pan retinoic acid recep-
tor antagonist (RAA) (Allergan #193109) (Collop et al., 2006;
Teng et al., 1997) in 20% Steinberg’s solution at stage 14. Stock
solutions for both RA and RAA were 1mM dissolved in DMSO,
and therefore a control treatment was performed with 0.1%
DMSO in 20% Steinberg’s solution.
4.3.In situ hybridization
Whole-mount
according to (Harland, 1991) with modifications small modifi-
cations (Deimling and Drysdale, 2009). Antisense riboprobes
for hand1 (Sparrow et al., 1998), sall3 (Hollemann et al., 1996),
foxf1 (Koster et al., 1999), nkx2.5 (Tonissen et al., 1994), hoxc10
(Christen et al., 2003), tnni3 (cardiac troponin I) (Drysdale et al.,
1994), sprouty2 (Genbank Accession# AF369901), isl1 (Brade
et al., 2007), Xbra (Smith et al., 1991), aplnr (Devic et al., 1996),
etv2 (Salanga et al., 2010), fgf4 (Isaacset al., 1992), fgf8 (Christen
andSlack, 1997), cyp26 (Hollemannet al., 1998), aldh1a2(raldh2)
(Chen et al., 2001), hba1 (globin) (Knochel et al., 1987), scl (Mead
et al., 1998), spib (Costa et al., 2008) and mpo (Smith et al., 2002)
werelabelledwithdigoxygenin(Dig)-labelledUTP(RocheDiag-
nostics) following the protocol by (Harland, 1991) except that
incorporating P32labelled nucleotides was omitted. BM Purple
(RocheDiagnostics)was usedas thealkalinephosphatasesub-
strate.Afterthecolourreaction,embryoswerefixedfortwenty
minutes in MEMPFA and endogenous pigment was bleached
with 0.5% hydrogen peroxide, 5% formamide, and 0.5% SSC
forseveralhours.EmbryoswerevisualizedonaLeicaMZ12dis-
secting microscope and images were captured using Northern
Eclipsesoftware(EmpixImaging;Mississauga, ON, Canada). In
each case, more than 20 embryos were assayed over at least
three separate replicates. For each conclusion drawn, a mini-
mum of 80% of embryos must have displayed each phenotype.
Double whole mount in situ hybridizations were accom-
plished according to Koga et al. (2007) with the following mod-
ifications. The probes were synthesized separately, one probe
was labelled with Dig-11-UTP as described above, while a sec-
ond probe was labelled with fluorescein-12-UTP (Roche Diag-
nostics). The same in situ hybridization protocol, described
above,wasusedexceptthatthedoubleprobe(probecontaining
the 1.5· concentrated mixture each of the Dig-labelled and
fluorescein-labelled probes) was added at the end of the first
day in place of a single in situ hybridization probe. Following
the first colour reaction the antibody was inactivated in 0.1 M
glycine pH 2.0 as previously described (Sive et al., 2000). The
firstcolourreactionwascarriedoutusingBMPurpleasthesub-
strate as above and the second colour reaction used 0.5 mg/ml
BCIP in alkaline phosphatase buffer as the substrate.
in situ hybridizations were performed
Acknowledgments
We would like to thank Dr. C. Pin, Dr. G. DiMattia, Dr. T.
Shepherd, Dr. B. Deroo, Dr. P. Kreig, and S. Grover for fruitful
discussions. We also thank Dr. T. Pieler, Dr. J. Slack, Dr. R. Har-
land, Dr. B. Pownall, Dr. P. Krieg, Dr. R. Friesel, Dr. M. Horb, and
Dr. W. Kno ¨chel for generously providing plasmids, and Jean
Wangforprovidingtechnical
Chandraratna for providing the retinoic acid antagonist and
advice on its use. We also thank the anonymous reviewers
for their suggestions that improved the manuscript. This
project was supported by a grant from the Canadian Institutes
of Health Research (MOP-74663) awarded to TAD and a CIHR
doctoral fellowship to S.J.D.
help.WethankDr.R.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.mod.2011.06.002.
R E F E R E N C E S
Alsan, B.H., Schultheiss, T.M., 2002. Regulation of avian
cardiogenesis by Fgf8 signaling. Development 129, 1935–1943.
Amaya, E., Musci, T.J., Kirschner, M.W., 1991. Expression of a
dominant negative mutant of the FGF receptor disrupts
mesoderm formation in Xenopus embryos. Cell 66, 257–270.
Astorga, J., Carlsson, P., 2007. Hedgehog induction of murine
vasculogenesis is mediated by Foxf1 and Bmp4. Development
134, 3753–3761.
Bertwistle, D., Walmsley, M.E., Read, E.M., Pizzey, J.A., Patient,
R.K., 1996. GATA factors and the origins of adult and
embryonic blood in Xenopus: responses to retinoic acid. Mech.
Dev. 57, 199–214.
Bottcher, R.T., Niehrs, C., 2005. Fibroblast growth factor signaling
during early vertebrate development. Endocr. Rev. 26, 63–77.
Brade, T., Gessert, S., Kuhl, M., Pandur, P., 2007. The amphibian
second heart field: Xenopus islet-1 is required for
cardiovascular development. Dev. Biol. 311, 297–310.
Chen, Y., Pollet, N., Niehrs, C., Pieler, T., 2001. Increased XRALDH2
activity has a posteriorizing effect on the central nervous
system of Xenopus embryos. Mech. Dev. 101, 91–103.
Chien, K.R., Domian, I.J., Parker, K.K., 2008. Cardiogenesis and the
complex biology of regenerative cardiovascular medicine.
Science 322, 1494–1497.
Christen, B., Beck, C.W., Lombardo, A., Slack, J.M., 2003.
Regeneration-specific expression pattern of three posterior
Hox genes. Dev. Dyn. 226, 349–355.
Christen, B., Slack, J.M., 1997. FGF-8 is associated with
anteroposterior patterning and limb regeneration in Xenopus.
Dev. Biol. 192, 455–466.
Christen, B., Slack, J.M., 1999. Spatial response to fibroblast growth
factor signalling in Xenopus embryos. Development 126, 119–
125.
Chung, W.C., Moyle, S.S., Tsai, P.S., 2008. Fibroblast growth factor 8
signaling through fibroblast growth factor receptor 1 is
required for the emergence of gonadotropin-releasing
hormone neurons. Endocrinology 149, 4997–5003.
Ciau-Uitz, A., Walmsley, M., Patient, R., 2000. Distinct origins of
adult and embryonic blood in Xenopus. Cell 102, 787–796.
Collop, A.H., Broomfield, J.A., Chandraratna, R.A., Yong, Z.,
Deimling, S.J., Kolker, S.J., Weeks, D.L., Drysdale, T.A., 2006.
Retinoic acid signaling is essential for formation of the heart
tube in Xenopus. Dev. Biol. 291, 96–109.
Costa, R.M., Soto, X., Chen, Y., Zorn, A.M., Amaya, E., 2008. Spib is
required for primitive myeloid development in Xenopus. Blood
112, 2287–2296.
M E C H A N I S M S O F D E V E L O P M E N T 1 2 8 ( 2 0 11 ) 3 27 –3 4 1
339

Page 15

Author's personal copy
Crump, J.G., Maves, L., Lawson, N.D., Weinstein, B.M., Kimmel,
C.B., 2004. An essential role for Fgfs in endodermal pouch
formation influences later craniofacial skeletal patterning.
Development 131, 5703–5716.
de Roos, K., Sonneveld, E., Compaan, B., ten Berge, D., Durston,
A.J., van der Saag, P.T., 1999. Expression of retinoic acid 4-
hydroxylase (CYP26) during mouse and Xenopus laevis
embryogenesis. Mech. Dev. 82, 205–211.
Deimling, S.J., Drysdale, T.A., 2009. Retinoic acid regulates
anterior–posterior patterning within the lateral plate
mesoderm of Xenopus. Mech. Dev. 126, 913–923.
Delaune, E., Lemaire, P., Kodjabachian, L., 2005. Neural induction
in Xenopus requires early FGF signalling in addition to BMP
inhibition. Development 132, 299–310.
Devic, E., Paquereau, L., Vernier, P., Knibiehler, B., Audigier, Y.,
1996. Expression of a new G protein-coupled receptor X-msr is
associated with an endothelial lineage in Xenopus laevis. Mech.
Dev. 59, 129–140.
Diez del Corral, R., Olivera-Martinez, I., Goriely, A., Gale, E.,
Maden, M., Storey, K., 2003. Opposing FGF and retinoid
pathways control ventral neural pattern, neuronal
differentiation, and segmentation during body axis extension.
Neuron 40, 65–79.
Drysdale, T.A., Patterson, K.D., Saha, M., Krieg, P.A., 1997. Retinoic
acid can block differentiation of the myocardium after heart
specification. Dev. Biol. 188, 205–215.
Drysdale, T.A., Tonissen, K.F., Patterson, K.D., Crawford, M.J.,
Krieg, P.A., 1994. Cardiac troponin I is a heart-specific marker
in the Xenopus embryo: expression during abnormal heart
morphogenesis. Dev. Biol. 165, 432–441.
Faber, J., Nieuwkoop, P.D., (1994). Normal table of Xenopus laevis
(Daudin):asystematicalandchronologicalsurveyofthedevelop-
ment from the fertilized egg till the end of meta morphosis.
Garland Pub., New York, pp. 252 (p., 10 leaves of plates).
Ferdous, A., Caprioli, A., Iacovino, M., Martin, C.M., Morris, J.,
Richardson, J.A., Latif, S., Hammer, R.E., Harvey, R.P., Olson,
E.N., Kyba, M., Garry, D.J., 2009. Nkx2–5 transactivates the Ets-
related protein 71 gene and specifies an endothelial/
endocardial fate in the developing embryo. Proc Natl Acad Sci
U S A 106, 814–819.
Fletcher, R.B., Harland, R.M., 2008. The role of FGF signaling in the
establishment and maintenance of mesodermal gene
expression in Xenopus. Dev. Dyn. 237, 1243–1254.
Frontini, M.J., Nong, Z., Gros, R., Drangova, M., O’Neil, C., Rahman,
M.N., Akawi, O., Yin, H., Ellis, C.G., Pickering, J.G., 2011.
Fibroblast growth factor 9 delivery during angiogenesis
produces durable, vasoresponsive microvessels wrapped by
smooth muscle cells. Nat. Biotechnol. 29, 421–427.
Garavito-Aguilar, Z.V., Riley, H.E., Yelon, D., 2010. Hand2 ensures
an appropriate environment for cardiac fusion by limiting
Fibronectin function. Development 137, 3215–3220.
Griffin, K., Patient, R., Holder, N., 1995. Analysis of FGF function in
normal and no tail zebrafish embryos reveals separate
mechanisms for formation of the trunk and the tail.
Development 121, 2983–2994.
Harland, R.M., 1991. In situ hybridization: an improved whole-
mount method for Xenopus embryos. Methods Cell Biol. 36,
685–695.
Hochgreb, T., Linhares, V.L., Menezes, D.C., Sampaio, A.C., Yan,
C.Y., Cardoso, W.V., Rosenthal, N., Xavier-Neto, J., 2003. A
caudorostral wave of RALDH2 conveys anteroposterior
information to the cardiac field. Development 130, 5363–5374.
Hollemann, T., Chen, Y., Grunz, H., Pieler, T., 1998. Regionalized
metabolic activity establishes boundaries of retinoic acid
signalling. EMBO J. 17, 7361–7372.
Hollemann, T., Schuh, R., Pieler, T., Stick, R., 1996. Xenopus Xsal-1,
a vertebrate homolog of the region specific homeotic gene
spalt of Drosophila. Mech. Dev. 55, 19–32.
Horb, M.E., Slack, J.M., 2001. Endoderm specification and
differentiation in Xenopus embryos. Dev. Biol. 236, 330–343.
Isaacs, H.V., Pownall, M.E., Slack, J.M., 1994. EFGF regulates Xbra
expression during Xenopus gastrulation. EMBO J. 13, 4469–4481.
Isaacs, H.V., Tannahill, D., Slack, J.M., 1992. Expression of a novel
FGF in the Xenopus embryo. A new candidate inducing factor
for mesoderm formation and anteroposterior specification.
Development 114, 711–720.
Keegan, B.R., Feldman, J.L., Begemann, G., Ingham, P.W., Yelon, D.,
2005. Retinoic acid signaling restricts the cardiac progenitor
pool. Science 307, 247–249.
Keren-Politansky, A., Keren, A., Bengal, E., 2009. Neural ectoderm-
secreted FGF initiates the expression of Nkx2.5 in cardiac
progenitors via a p38 MAPK/CREB pathway. Dev. Biol. 335, 374–
384.
Knochel, W., Beck, J., Meyerhof, W., 1987. Nucleotide sequence of
the Xenopus tropicalis larval beta globin gene. Nucleic Acids
Res. 15, 10062.
Koga, M., Kudoh, T., Hamada, Y., Watanabe, M., Kageura, H., 2007.
A new triple staining method for double in situ hybridization
in combination with cell lineage tracing in whole-mount
Xenopus embryos. Dev. Growth Differ. 49, 635–645.
Koster, M., Dillinger, K., Knochel, W., 1999. Genomic structure and
embryonic expression of the Xenopus winged helix factors
XFD-13/130. Mech. Dev. 88, 89–93.
Lea, R., Papalopulu, N., Amaya, E., Dorey, K., 2009. Temporal and
spatial expression of FGF ligands and receptors during Xenopus
development. Dev. Dyn. 238, 1467–1479.
Lynch, J., McEwan, J., Beck, C.W., 2011. Analysis of the expression
of retinoic acid metabolising genes during Xenopus laevis
organogenesis. Gene Expr. Patterns. 11, 112–117.
Maroon, H., Walshe, J., Mahmood, R., Kiefer, P., Dickson, C.,
Mason, I., 2002. Fgf3 and Fgf8 are required together for
formation of the otic placode and vesicle. Development 129,
2099–2108.
Marques, S.R., Lee, Y., Poss, K.D., Yelon, D., 2008. Reiterative roles
for FGF signaling in the establishment of size and proportion
of the zebrafish heart. Dev. Biol. 321, 397–406.
Martin, B.L., Kimelman, D., 2010. Brachyury establishes the
embry onic mesodermal progenitor niche. Genes Dev. 24, 2778–
2783.
Martinez-Estrada, O.M., Lettice, L.A., Essafi, A., Guadix, J.A., Slight,
J., Velecela, V., Hall, E., Reichmann, J., Devenney, P.S.,
Hohenstein, P., Hosen, N., Hill, R.E., Munoz-Chapuli, R., Hastie,
N.D., 2010. Wt1 is required for cardiovascular progenitor cell
formation through transcriptional control of Snail and E-
cadherin. Nat. Genet. 42, 89–93.
Mead, P.E., Kelley, C.M., Hahn, P.S., Piedad, O., Zon, L.I., 1998. SCL
specifies hematopoietic mesoderm in Xenopus embryos.
Development 125, 2611–2620.
Misfeldt, A.M., Boyle, S.C., Tompkins, K.L., Bautch, V.L., Labosky,
P.A., Baldwin, H.S., 2009. Endocardial cells are a distinct
endothelial lineage derived from Flk1+ multipotent
cardiovascular progenitors. Dev. Biol. 333, 78–89.
Mohammadi, M., McMahon, G., Sun, L., Tang, C., Hirth, P., Yeh,
B.K., Hubbard, S.R., Schlessinger, J., 1997. Structures of the
tyrosine kinase domain of fibroblast growth factor receptor in
complex with inhibitors. Science 276, 955–960.
Moreno, T.A., Kintner, C., 2004. Regulation of segmental
patterning by retinoic acid signaling during Xenopus
somitogenesis. Dev. Cell 6, 205–218.
Moretti, A., Caron, L., Nakano, A., Lam, J.T., Bernshausen, A., Chen,
Y., Qyang, Y., Bu, L., Sasaki, M., Martin-Puig, S., Sun, Y., Evans,
S.M., Laugwitz, K.L., Chien, K.R., 2006. Multipotent embryonic
isl1+ progenitor cells lead to cardiac, smooth muscle, and
endothelial cell diversification. Cell 127, 1151–1165.
340
M E C H A N I S M S O F D E V E L O P M E N T 1 28 ( 2 01 1 ) 32 7 –3 41