59
Signaling Pathways in AP Patterning
do tissue interactions between mesendoderm and ectoderm
or neural and ectoderm infuence AP neural specif cation?
What is the physicochemical nature of the inducers that
control AP development? Does the induction always occur
between tissues of different germ layers (vertical induction), or can it happen within cells of the same germ layer
(planar or homeogenetic induction)? How many signals are
employed to provide positional information along the AP
axis? Once molecular biology emerged on the scene, modern developmental biologists took advantage of molecular
tools to identify, monitor the expression of, and manipulate genes in Xenopus to tackle these same key questions
at the molecular levels. What signals can affect expression
patterns of AP markers in explants or embryos? When are
the signals required? How many signals work in parallel or
sequentially? What is the regulatory function of each signal for AP development? Discoveries made using molecular
approaches helped to provide explanations to some of the
classical questions and lent support to or refuted the models
put forward decades ago.
One main question that occupied the attention of both
experimental and molecular biologists is how AP neural
patterning is executed by post-involuted mesendoderm as
it advances toward the future anterior end of the embryo
during gastrulation. Based on tissue explant and transplant
experiments, distinct models were proposed that favored
either the presence of multiple inducers along the AP
axis or the existence of two main signals with a universal
inducer of anterior neural character and a separate graded
transformer responsible for posterior neural development
( Figure 5.1 ) ( Mangold, 1933 ; Nieuwkoop, 1952a , 1952b ,
1952c). The evaluation of the two models in the modern
era was facilitated by the identifcation of organizer-localized secreted BMP inhibitors as direct neural inducers. All
these inducers, including Noggin, Chordin, and Follistatin,
which are expressed in midline dorsal mesoderm along
the entire notochord during neurulation, can only induce
anterior neural markers in animal cap explants ( HemmatiBrivanlou et al., 1994; Lamb et al., 1993; Sasai et al., 1995,
1994; Smith and Harland, 1992). When combined with activation of the signaling pathways discussed previously, posterior neural genes can be expressed as well. These results,
together with gain- and loss-of-function studies in vivo,
accelerated the acceptance of the activation-transformation
model of AP specifcation by the feld. However, the story
seems to be more complex than expected. FGF signaling
has been implicated in inducing posterior neural markers directly without any “activators,” but only particular
pairs of FGF ligands (Fgf8a) and receptors (Fgfr4) may be
involved in direct induction. This implies that the multipleinducer model of Mangold may be compatible with some
results. In addition, as discussed previously, interference
of caudalizing signals often impairs expression of markers
in the hindbrain, with less effect on markers of the spinal
cord. It is possible that these results refect redundancy of
signals in the trunk, but it is also conceivable that distinct
ratios of various active signals at different AP positions
serve to divide the neural plate into separate signaling
zones, with each specifying a particular AP cell fate. This
would be a hybrid paradigm between the multi-inducer
and activation-transformation models in that instead of the
presence of multiple position-specifc inducers, a universal
neural inducer works within multiple transformation zones
to specify AP neural tissues. Combinations of varying gradients of growth factor signals would defne the zones and
encode AP positional information. The dynamic expression patterns of signaling molecules and the presence
of a myriad of inhibitors for all three major caudalizing
pathways are consistent with this model. One additional
consideration is the timing of AP specifcation. Some conclusions from the classical experiments were based on the
studies using neurula-stage embryos when the AP axis had
already been well laid out. The signals at these stages are
not exactly the same as those at mid- to late gastrula stages
when AP tissues are being specifed. Dynamic temporal
expression of signaling molecules is often observed during
Xenopus embryogenesis so that neurula stage signals may
play more important roles in maintenance or elaboration of
posterior tissue development rather than inducing AP characteristics. Stage-specifc manipulations of gene functions
and dynamic temporal analyses of marker expression are
both required to address this issue.
While molecular biological studies have contributed novel
insights and proposed detailed mechanisms concerning AP
patterning, several issues remain unresolved. The patterned
expression of signaling molecules and their modulators in
distinct domains of the Spemann organizer precedes future
AP specifcation. However, little is known about molecular
circuitry controlling organizer subdivision. The competence
of neuralized animal caps or neural plate explants to respond
to caudalizing signals is lost during mid-neurulation, yet the
molecular machinery responsible for this loss is not understood. The activation-transformation model proposed by
Nieuwkoop may operate via gradient distribution of a substance with an apex level at the posterior end or different
durations that competent ectoderm is exposed to a constant
transforming agent at different AP positions. Most molecular studies so far tend to focus on the chemical gradients,
but the temporal gradient of signal exposure has not been
investigated in depth. Though multiple signals are known
to specify posterior tissues, how the signals are transduced
and converge to control downstream target gene expression has only been explored slightly. Several transcription
factors, such as Cdx members, Meis1/3, and Hox proteins,
are required for hindbrain or posterior neural development
(Dibner et al., 2001, 2004; Elkouby et al., 2010; Epstein et
al., 1997; Faas and Isaacs, 2009; Gutkovich et al., 2010;
Pillemer et al., 1998; Schyr et al., 2012). The bidirectional
interplay between these transcription factors and the caudalizing signals is not understood in detail. The eventual development of AP neural tissues requires not only instructions
from the underlying mesoderm but also further ref nement
within the neural plate via self-organization. This process
is not well appreciated at the molecular level. Most of the
Signaling Pathways in AP Patterning
do tissue interactions between mesendoderm and ectoderm
or neural and ectoderm infuence AP neural specif cation?
What is the physicochemical nature of the inducers that
control AP development? Does the induction always occur
between tissues of different germ layers (vertical induction), or can it happen within cells of the same germ layer
(planar or homeogenetic induction)? How many signals are
employed to provide positional information along the AP
axis? Once molecular biology emerged on the scene, modern developmental biologists took advantage of molecular
tools to identify, monitor the expression of, and manipulate genes in Xenopus to tackle these same key questions
at the molecular levels. What signals can affect expression
patterns of AP markers in explants or embryos? When are
the signals required? How many signals work in parallel or
sequentially? What is the regulatory function of each signal for AP development? Discoveries made using molecular
approaches helped to provide explanations to some of the
classical questions and lent support to or refuted the models
put forward decades ago.
One main question that occupied the attention of both
experimental and molecular biologists is how AP neural
patterning is executed by post-involuted mesendoderm as
it advances toward the future anterior end of the embryo
during gastrulation. Based on tissue explant and transplant
experiments, distinct models were proposed that favored
either the presence of multiple inducers along the AP
axis or the existence of two main signals with a universal
inducer of anterior neural character and a separate graded
transformer responsible for posterior neural development
( Figure 5.1 ) ( Mangold, 1933 ; Nieuwkoop, 1952a , 1952b ,
1952c). The evaluation of the two models in the modern
era was facilitated by the identifcation of organizer-localized secreted BMP inhibitors as direct neural inducers. All
these inducers, including Noggin, Chordin, and Follistatin,
which are expressed in midline dorsal mesoderm along
the entire notochord during neurulation, can only induce
anterior neural markers in animal cap explants ( HemmatiBrivanlou et al., 1994; Lamb et al., 1993; Sasai et al., 1995,
1994; Smith and Harland, 1992). When combined with activation of the signaling pathways discussed previously, posterior neural genes can be expressed as well. These results,
together with gain- and loss-of-function studies in vivo,
accelerated the acceptance of the activation-transformation
model of AP specifcation by the feld. However, the story
seems to be more complex than expected. FGF signaling
has been implicated in inducing posterior neural markers directly without any “activators,” but only particular
pairs of FGF ligands (Fgf8a) and receptors (Fgfr4) may be
involved in direct induction. This implies that the multipleinducer model of Mangold may be compatible with some
results. In addition, as discussed previously, interference
of caudalizing signals often impairs expression of markers
in the hindbrain, with less effect on markers of the spinal
cord. It is possible that these results refect redundancy of
signals in the trunk, but it is also conceivable that distinct
ratios of various active signals at different AP positions
serve to divide the neural plate into separate signaling
zones, with each specifying a particular AP cell fate. This
would be a hybrid paradigm between the multi-inducer
and activation-transformation models in that instead of the
presence of multiple position-specifc inducers, a universal
neural inducer works within multiple transformation zones
to specify AP neural tissues. Combinations of varying gradients of growth factor signals would defne the zones and
encode AP positional information. The dynamic expression patterns of signaling molecules and the presence
of a myriad of inhibitors for all three major caudalizing
pathways are consistent with this model. One additional
consideration is the timing of AP specifcation. Some conclusions from the classical experiments were based on the
studies using neurula-stage embryos when the AP axis had
already been well laid out. The signals at these stages are
not exactly the same as those at mid- to late gastrula stages
when AP tissues are being specifed. Dynamic temporal
expression of signaling molecules is often observed during
Xenopus embryogenesis so that neurula stage signals may
play more important roles in maintenance or elaboration of
posterior tissue development rather than inducing AP characteristics. Stage-specifc manipulations of gene functions
and dynamic temporal analyses of marker expression are
both required to address this issue.
While molecular biological studies have contributed novel
insights and proposed detailed mechanisms concerning AP
patterning, several issues remain unresolved. The patterned
expression of signaling molecules and their modulators in
distinct domains of the Spemann organizer precedes future
AP specifcation. However, little is known about molecular
circuitry controlling organizer subdivision. The competence
of neuralized animal caps or neural plate explants to respond
to caudalizing signals is lost during mid-neurulation, yet the
molecular machinery responsible for this loss is not understood. The activation-transformation model proposed by
Nieuwkoop may operate via gradient distribution of a substance with an apex level at the posterior end or different
durations that competent ectoderm is exposed to a constant
transforming agent at different AP positions. Most molecular studies so far tend to focus on the chemical gradients,
but the temporal gradient of signal exposure has not been
investigated in depth. Though multiple signals are known
to specify posterior tissues, how the signals are transduced
and converge to control downstream target gene expression has only been explored slightly. Several transcription
factors, such as Cdx members, Meis1/3, and Hox proteins,
are required for hindbrain or posterior neural development
(Dibner et al., 2001, 2004; Elkouby et al., 2010; Epstein et
al., 1997; Faas and Isaacs, 2009; Gutkovich et al., 2010;
Pillemer et al., 1998; Schyr et al., 2012). The bidirectional
interplay between these transcription factors and the caudalizing signals is not understood in detail. The eventual development of AP neural tissues requires not only instructions
from the underlying mesoderm but also further ref nement
within the neural plate via self-organization. This process
is not well appreciated at the molecular level. Most of the