55
Signaling Pathways in AP Patterning
1997; Koide et al., 2001; Kolm et al., 1997; van der Wees et
al., 1998). The hindbrain, but not the forebrain or posterior
spinal cord, was particularly sensitive to manipulated RA
signaling levels (Godsave et al., 1998; Kolm et al., 1997;
van der Wees et al., 1998). However, the interpretation of
the results might be complicated by the discovery that the
unliganded RA receptors were not inactive but could bind
to their targets and recruit co-repressors to act as a transcription repressor in the absence of RA (Koide et al., 2001;
Weston et al., 2003). Examination of the distribution of RAR
and RXR receptors (Ellinger-Ziegelbauer and Dreyer, 1991;
Pfeffer and De Robertis, 1994) revealed that they were not
only expressed in the posterior mesodermal and neural cells
but also in the anterior mesendoderm (RARγ) or posterior
to the hindbrain regions (RXRα). Thus, the pattern of RA
receptor distribution was consistent with its role in regulating hindbrain development and posterior patterning.
The level of available RA in cells is controlled by multiple enzymes and RA binding proteins that inf uence the
synthesis, degradation, and activity of RA. Several of the
RA biogenesis enzymes, such as Aldh1a2/Raldh2, Rdh10,
Sdr16C5/Rdhe2, and Cyp26c1, were analyzed for their functions in AP neural specifcation. Consistent with the studies
manipulating RA or RA receptor activities, increased levels
of aldh1a2/raldh2, rdh10, or sdr16c5/rdhe2, which promote
RA biosynthesis, caused posteriorization of neural domains
(Belyaeva et al., 2012; Chen et al., 2001; Strate et al., 2009).
In contrast, increased levels of cyp26c1, which is involved
in RA neutralization, anteriorized hindbrain and rescued
defects induced by RA (Hollemann et al., 1998; Tanibe et
al., 2008). aldh1a2/raldh2 and cyp26c1 are expressed in
non-overlapping domains in Xenopus gastrula- and neurulastage embryos and have antagonistic functions in regulating
AP neural gene expression. In addition, ectopic expression
of a cellular retinoic acid binding protein (Crabp2) induced
embryonic AP defects and enhanced expression posterior
hox genes, an effect similar to providing exogenous RA
(Dekker et al., 1994). Taken together, the data strongly support a role of RA signaling in transforming anterior to more
posterior neural tissues of Xenopus embryos, especially in
the hindbrain region.
5.3.2. FIBROBLAST GROWTH FACTOR SIGNALING
Interest in fbroblast growth factor (FGF) signaling in
early Xenopus development was stimulated in the late
1980s by the discovery that FGF could induce mesoderm
formation in naïve ectodermal explants, referred to as animal caps (Kimelman et al., 1988; Slack et al., 1989, 1987,
1988). Molecular cloning of mammalian FGF homologs in
Xenopus revealed that various FGF ligands and their receptors were expressed in early embryos. In situ hybridization
on embryonic sections or whole embryos showed that the
mRNAs encoding many FGF ligands, including fgf4/efgf,
fgf3/int-2, and fgf8, were expressed in the marginal region
above the blastopore at the gastrula stages and had dynamic
distribution patterns including a posterior domain during
neurula and tailbud stages (Christen and Slack, 1997; Isaacs
et al., 1995, 1992; Lea et al., 2009; Lombardo et al., 1998;
Tannahill et al., 1992). The seemingly graded expression of
the FGF ligands and one of the FGF receptors, fgfr1 ( Friesel
and Dawid, 1991; Golub et al., 2000; Lea et al., 2009), in the
posterior region of late gastrula- and neurula-stage embryos
was particularly interesting, as it implied a function of FGF
signaling in posterior development.
A role of the FGF pathway in forming posterior structures was supported directly by both gain- and loss-offunction studies. Ectopic expression of FGF ligands, such
as fgf3, fgf4/efgf, fgf8, or fgf9 caused head reduction in the
manipulated tadpoles (Christen and Slack, 1997; Lombardo
et al., 1998; Pownall et al., 1996; Song and Slack, 1996 ),
whereas expression of a dominant negative FGF receptor 1
(DN-Fgfr1 or XFD) led to trunk and tail truncation (Amaya
et al., 1991; Godsave and Durston, 1997 ). As FGF was
shown to regulate both mesodermal formation and posterior
development, the direct posteriorization activity of FGF signaling was tested using explants. Neuralized animal caps or
explants from anterior neural plate cultured with FGF could
express hindbrain (egr2/krox20) and spinal cord (hoxb9)
markers, whereas control explants lacked expression of these
posterior neural genes (Cox and Hemmati-Brivanlou, 1995;
Fletcher et al., 2006). The induction of posterior neural gene
expression occurred in the absence of mesodermal markers, suggesting that FGF signaling could transform anterior
into posterior neural tissues, supporting the transformation
or caudalization hypothesis. When XFD or DN-Ras was
expressed in neuralized animal caps treated with Fgf2/
bFGF or conjugated with dorsal mesoderm, posterior, but
not anterior, neural marker expression was blocked (Cox
and Hemmati-Brivanlou, 1995; Holowacz and Sokol, 1999;
Ribisi et al., 2000).
Although the experiments described above point to a crucial role of FGF signaling in caudalization of the nervous
system, interpretation of the results may be complicated by
several issues. One issue, as mentioned, is that FGF signaling is involved in both mesodermal induction and posterior
neural development. It can be debated whether impairment of posterior structures when the pathway is inhibited
in vivo is due to the activities of FGF on early mesoderm
induction and not subsequent patterning of the neural tissue. Several approaches were used to demonstrate that FGF
could caudalize neural tissues directly. The explant experiments described before took advantage of the ease of providing the Fgf2/bFGF protein at the desired temporal points of
development and demonstrated that FGF signaling at gastrula to early neurula stages could induce posterior neural
genes without concurrent induction of mesodermal markers
(Cox and Hemmati-Brivanlou, 1995; Fletcher et al., 2006).
FGF-soaked beads were also used for implantation into the
neural plate directly, hence avoiding an early effect of stimulating FGF signaling (Lombardo and Slack, 1998; Pownall
et al., 1996 ). In addition, a synthetic Fgfr1 receptor that can
be activated by a dimerizing agent but not FGF ligands was
used in embryos to induce FGF signaling with the drug at
Signaling Pathways in AP Patterning
1997; Koide et al., 2001; Kolm et al., 1997; van der Wees et
al., 1998). The hindbrain, but not the forebrain or posterior
spinal cord, was particularly sensitive to manipulated RA
signaling levels (Godsave et al., 1998; Kolm et al., 1997;
van der Wees et al., 1998). However, the interpretation of
the results might be complicated by the discovery that the
unliganded RA receptors were not inactive but could bind
to their targets and recruit co-repressors to act as a transcription repressor in the absence of RA (Koide et al., 2001;
Weston et al., 2003). Examination of the distribution of RAR
and RXR receptors (Ellinger-Ziegelbauer and Dreyer, 1991;
Pfeffer and De Robertis, 1994) revealed that they were not
only expressed in the posterior mesodermal and neural cells
but also in the anterior mesendoderm (RARγ) or posterior
to the hindbrain regions (RXRα). Thus, the pattern of RA
receptor distribution was consistent with its role in regulating hindbrain development and posterior patterning.
The level of available RA in cells is controlled by multiple enzymes and RA binding proteins that inf uence the
synthesis, degradation, and activity of RA. Several of the
RA biogenesis enzymes, such as Aldh1a2/Raldh2, Rdh10,
Sdr16C5/Rdhe2, and Cyp26c1, were analyzed for their functions in AP neural specifcation. Consistent with the studies
manipulating RA or RA receptor activities, increased levels
of aldh1a2/raldh2, rdh10, or sdr16c5/rdhe2, which promote
RA biosynthesis, caused posteriorization of neural domains
(Belyaeva et al., 2012; Chen et al., 2001; Strate et al., 2009).
In contrast, increased levels of cyp26c1, which is involved
in RA neutralization, anteriorized hindbrain and rescued
defects induced by RA (Hollemann et al., 1998; Tanibe et
al., 2008). aldh1a2/raldh2 and cyp26c1 are expressed in
non-overlapping domains in Xenopus gastrula- and neurulastage embryos and have antagonistic functions in regulating
AP neural gene expression. In addition, ectopic expression
of a cellular retinoic acid binding protein (Crabp2) induced
embryonic AP defects and enhanced expression posterior
hox genes, an effect similar to providing exogenous RA
(Dekker et al., 1994). Taken together, the data strongly support a role of RA signaling in transforming anterior to more
posterior neural tissues of Xenopus embryos, especially in
the hindbrain region.
5.3.2. FIBROBLAST GROWTH FACTOR SIGNALING
Interest in fbroblast growth factor (FGF) signaling in
early Xenopus development was stimulated in the late
1980s by the discovery that FGF could induce mesoderm
formation in naïve ectodermal explants, referred to as animal caps (Kimelman et al., 1988; Slack et al., 1989, 1987,
1988). Molecular cloning of mammalian FGF homologs in
Xenopus revealed that various FGF ligands and their receptors were expressed in early embryos. In situ hybridization
on embryonic sections or whole embryos showed that the
mRNAs encoding many FGF ligands, including fgf4/efgf,
fgf3/int-2, and fgf8, were expressed in the marginal region
above the blastopore at the gastrula stages and had dynamic
distribution patterns including a posterior domain during
neurula and tailbud stages (Christen and Slack, 1997; Isaacs
et al., 1995, 1992; Lea et al., 2009; Lombardo et al., 1998;
Tannahill et al., 1992). The seemingly graded expression of
the FGF ligands and one of the FGF receptors, fgfr1 ( Friesel
and Dawid, 1991; Golub et al., 2000; Lea et al., 2009), in the
posterior region of late gastrula- and neurula-stage embryos
was particularly interesting, as it implied a function of FGF
signaling in posterior development.
A role of the FGF pathway in forming posterior structures was supported directly by both gain- and loss-offunction studies. Ectopic expression of FGF ligands, such
as fgf3, fgf4/efgf, fgf8, or fgf9 caused head reduction in the
manipulated tadpoles (Christen and Slack, 1997; Lombardo
et al., 1998; Pownall et al., 1996; Song and Slack, 1996 ),
whereas expression of a dominant negative FGF receptor 1
(DN-Fgfr1 or XFD) led to trunk and tail truncation (Amaya
et al., 1991; Godsave and Durston, 1997 ). As FGF was
shown to regulate both mesodermal formation and posterior
development, the direct posteriorization activity of FGF signaling was tested using explants. Neuralized animal caps or
explants from anterior neural plate cultured with FGF could
express hindbrain (egr2/krox20) and spinal cord (hoxb9)
markers, whereas control explants lacked expression of these
posterior neural genes (Cox and Hemmati-Brivanlou, 1995;
Fletcher et al., 2006). The induction of posterior neural gene
expression occurred in the absence of mesodermal markers, suggesting that FGF signaling could transform anterior
into posterior neural tissues, supporting the transformation
or caudalization hypothesis. When XFD or DN-Ras was
expressed in neuralized animal caps treated with Fgf2/
bFGF or conjugated with dorsal mesoderm, posterior, but
not anterior, neural marker expression was blocked (Cox
and Hemmati-Brivanlou, 1995; Holowacz and Sokol, 1999;
Ribisi et al., 2000).
Although the experiments described above point to a crucial role of FGF signaling in caudalization of the nervous
system, interpretation of the results may be complicated by
several issues. One issue, as mentioned, is that FGF signaling is involved in both mesodermal induction and posterior
neural development. It can be debated whether impairment of posterior structures when the pathway is inhibited
in vivo is due to the activities of FGF on early mesoderm
induction and not subsequent patterning of the neural tissue. Several approaches were used to demonstrate that FGF
could caudalize neural tissues directly. The explant experiments described before took advantage of the ease of providing the Fgf2/bFGF protein at the desired temporal points of
development and demonstrated that FGF signaling at gastrula to early neurula stages could induce posterior neural
genes without concurrent induction of mesodermal markers
(Cox and Hemmati-Brivanlou, 1995; Fletcher et al., 2006).
FGF-soaked beads were also used for implantation into the
neural plate directly, hence avoiding an early effect of stimulating FGF signaling (Lombardo and Slack, 1998; Pownall
et al., 1996 ). In addition, a synthetic Fgfr1 receptor that can
be activated by a dimerizing agent but not FGF ligands was
used in embryos to induce FGF signaling with the drug at